Tumor Biol. (2011) 32 (Suppl 1):S5–S56 DOI 10.1007/s13277-011-0238-6

ABSTRACTS

Oral Presentations

O1 GI CANCERS (GASTRIC, COLORECTAL, LIVER, PANCREATIC) O1.1 - Epithelial-mesenchymal transition in hepatocellular carcinoma H. Shiraha, S. Nishina, S. Horiguchi, S. Tanaka, M Iwamuro, J. Kataoka, M. Matsubara, A. Takaki, H. Onishi, S. Nakamura, K. Nouso, K. Yamamoto Department of Gastroenterology and Hepatology, Okayama University Medical School, Okayama, Japan Background: The epithelial-mesenchymal transition (EMT) is considered to be a critical event in the malignant transformation of cancer including hepatocellular carcinoma (HCC). As EMT is characterized by the loss of cell-cell adhesion and leads to cell individualization, the downregulation of E-cadherin is a major hallmark of EMT. The progression of EMT is also correlated with N-cadherin and vimentin expressions. It is reported that twist expression strongly induce EMT in HCC. Although loss or decreased expression of runt-related transcription factor 3 (RUNX3) has been frequently observed in human HCC, the role of decreased RUNX3 expression has not yet been fully elucidated. Aims: To elucidate the regulatory mechanism of the EMT in HCC progression. Methods: HCC cell lines, Hep3B, Huh7, HLE, and SKHep1 were used. EMT markers, jagged-1 (JAG1), twist, and RUNX3 expressions were analyzed by immunoblotting. Ectopic RUNX3 and twist expressions were carried out by introducing cDNA. Immunohistochemistry was performed for RUNX3 and EMT markers in 30 HCC tissues. Correlations between EMT markers and RUNX3

mRNA expression were also analyzed using publicly available OncomineTM datasets (www.oncomine.org). Results: Ectopic RUNX3 expression had anti-EMT effect in twist-negative human HCC cells: increased Ecadherin expression, decreased N-cadherin and vimentin expression, and decreased JAG1 expression. The immunohistochemical analysis revealed that E-cadherin expression was positively correlated, and N-cadherin expression was negatively correlated with RUNX3 expression in twist-negative human HCC tissues. Similarly, mRNA expression for E-cadherin was significantly correlated with that for RUNX3 in twist-negative HCC, while mRNA expression for JAG1 was negatively correlated with that for RUNX3. N-cadherin mRNA and vimentin mRNA expression were negatively correlated with RUNX3 mRNA expression. Conclusions: We have uncovered a novel function of RUNX3 in regulating EMT in HCC. The loss or decreased expression of RUNX3 induced EMT via JAG1. As RUNX3 regulates EMT in twist-negative HCC, RUNX3 could be the regulator in the early stage of progression.

O1.2 - Chemometrics of malignancy using IR spectroscopy: potential for clinical applications R. K. Sahu1,2 and S. Mordechai1,2 1Department of Physics, Ben-Gurion University of the Negev, Beer-heva, Israel Background: Spectroscopic analysis using FTIR absorbance or reflectance spectra of biological samples is becoming popular due to the ease with which it can be performed and the information that can be obtained about the molecular compositions, using simpler instrumentations. FTIR spectros-

S6

copy being reagent free requires minimum preparation of the sample excluding the need of widely used reagents such as antibodies and different expensive biological reagents. The technique is widely perceived as a potential diagnostic tool for detection of abnormalities in cells and tissues with an implication for cancer diagnosis where a unique spectral fingerprint is attributed to a normal cell or tissue that is altered during carcinogenesis. Aims: The utility of the different instrumentations currently available and their advantages, disadvantages and their relevance to various aspects of clinical and non clinical cancer research will be highlighted. Methods: The basic aspects of data acquisition and analysis using different modes of FTIR spectroscopy and the various factors that influence the parameters would be discussed with an emphasis on ways to minimize variations within similar samples while maximizing the separation between classes that are different from each other in a minimum number of recordable parameters. Results: While the FTIR data acquisition is simpler, it often requires complex computational analyses of the large amount of spectral data, which necessitates the inclusion of suitable software with each type of application. Conclusions: The lecture will focus on markers for colon cancer diagnosis using the crypt model developed by our group [1,2]. 1. R. K. Sahu and S. Mordechai, “Spectral signatures of colonic malignancies in mid IR region: from basic research to clinical applicability” Future Oncology, Vol. 6, No. 10, 1653–1667 2010. 2. Sahu RK, Argov S, Walfisch S, Bogomolny E, Moreh R and Mordechai S, “Prediction potential of IR-micro spectroscopy for colon cancer relapse”. Analyst. 135 (3):538–44 2010.

O1.3 - The Immunologic faecal occult blood test for the detection of significant colorectal neoplasia Auge JM, Sasot M, Escudero JM, Filella X, Molina R, Buron A, Grau J, Macia F, Castells T, Andreu M, Pellise M, Piraces M, Serradesanferm A, Hernandez C. Hospital Clinic and Hospital del Mar (Barcelona) Background: In Barcelona, including a target population of 203139 inhabitants, a biennial immunologic faecal occult blood test (iFOBT) is used with the offer of colonoscopy for those who test positively.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

Aims: To incorporate in routine and evaluate prospectively an iFOBT for the early detection of significant colorectal neoplasia. Methods: The OC-SENSOR DIANA (Eiken Chemical Co., Ltd.) analyzer was used to measures faecal human haemoglobin concentration by quantitative latex agglutination. The preanalytical phase was adapted for the peculiarities of the screening programme. The iFOBT was performed on 26106 participants without previous history of CRC, without known gastrointestinal symptoms and digestive diseases. Results: In 1015 citizens the iFOBT was positive (≥ 100 ng/mL). In 261 (26%) the colonoscopy were normal-negative, The pathologic findings were as follows: eight (1%) intestinal inflammatory disease, 93 (9%) with poliposi, inflamatory or hyperplastic polips, 182 (18%) low risk adenomas and in 376 (37%) high risk adenomas, in 95 (9%) the final result was CRC. With significant highest faecal haemoglobin levels (median: 639 ng/mL) in patients with significant colorectal neoplasia (3 or more adenomas or 1 villous adenoma or 1 high grade dysplasia adenoma or 1 adenoma ≥1 cm or colorectal carcinoma) when compared with other findings (median: 221 ng/mL). Conclusions: The faecal occult immunologic test shows consistent clinical results, with a strong relation between haemoglobin faecal levels and colonoscopy findings (significant colorectal neoplasia). The lab participation in a CRC screening programme needs a strong lab implication

O1.4 - Tumor suppressive role of HOPX gene in pancreatic carcinogenesis M. Waraya, K. Yamashita, M. Watanabe Department of Surgery, Kitasato University Background: DNA methylation of tumor suppressor gene is cancer-specific event, and it could be a useful biomarker and even identification of therapeutic point. We previously identified HOPX (homeodomain only protein X) as a tumour suppressor gene which is silenced by CpG islands methylation in human cancers by pharmacological unmasking microarray. HOPX is frequently methylated in esophageal, gastric, and colorectal cancer, and inhibits their aggressive phenotypes, however its clinical relevance and functional role remain elusive. Aims: In this current study, we examined methylation level of HOPX and tested the functional relevance in pancreatic cancer. Methods: Clinical features of HOPX promoter hypermethylation was investigated in 89 formalin-fixed pancreatic cancer tissues, and immunohistochemistry was added. We also

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

examined its functional relevance in phenotype assays such as soft ager, proliferation, invasion, and cell cycle analysis in cancer cells. Results: Pancreatic cancer tissues showed HOPX gene hypermethylation (73/89; 82.0%) as compared to the corresponding normal pancreas tissues (20/84; 23.8%), and its uniqueness was robust to discriminate tumor from normal tissues (AUC=0.85, Ppression is reduced, especially in pancreatic cancer tissues with promoter hypermethylation. Stable transfectants of HOPX gene promote G1 arrest with subG1 accumulation, and inversely affect tumor forming and invasive ability. Conclusions: Defective expression of HOPX which is consistent with CpG islands promoter DNA hypermethylation may explain aggressive phenotype of pancreatic cancer, and intense expression of HOPX in the Langerhans cells might in turn uniquely contribute to pancreatic carcinogenesis.

S7

stream EGFR, such as NRAS, HRAS, BRAF and PIK3CA mutations. With regard to BRAF mutation, not only many evidences underline its negative predictive impact, but attention is now focused on its extremely negative prognostic power and on adequate strategies able to counteract BRAF mutant tumours’ aggressiveness. Results: At present, more and more efforts are ongoing to increase the availability of molecular tools able to drive clinical choices, with particular regard to the search of still lacking molecular predictors of benefit from the antiangiogenic bevacizumab. At the same time, the need for deepening molecular characterization of mCRC acquires relevance with the aim to identify new potential targets for the development of new drugs. Conclusions: An even more intriguing challenge consists in the attempt to catch mCRC’s dynamism, by analyzing in vivo mechanisms of early and acquired resistance. This pharmacodynamic approach might help to put to use appropriate strategies to prevent or revert resistance, as well as to adequately plan subsequent lines of treatment.

O1.5 - Predictive biomarkers of response in colorectal cancer C. Cremolini, F. Loupakis, A. Falcone

O1.6 - Integration of biologic agents with cytotoxic chemotherapy in metastatic colorectal cancer

Polo Oncologico-University of Pisa

Gianluca Masi

Background: The growing amount of therapeutic options and the challenge of treatments’ personalization has disclosed the need for a deeper molecular characterization of colorectal cancer (CRC). KRAS story has established a sort of revolution in metastatic CRC (mCRC), by making for the first time a molecular determinant essential in clinical decision making. The role of KRAS mutations as predictors of resistance to anti-EGFR monoclonal antibodies was firstly suggested by small retrospective experiences, than corroborated by larger retrospective series and from post-hoc analyses of phase III randomized trials. Aims: At the same time, the increasing number of candidate markers, potentially useful in predicting treatments’ efficacy, has disclosed the need for a rigorous methodology in the validation of molecular markers before their admission to clinical practice. The necessity of prospective or at least meta-analytical confirmations is now well-established. Methods: In order to further select patients to receive antiEGFR monoclonal antibodies, an amount of data has been collected with regard to the predictive role of other components of intracellular signalling pathways down-

U.O. Oncologia Medica Universitaria, Azienda OsledalieroUniversitaria Pisana, Pisa, Italy Background: Colorectal cancer (CRC) remains a leading cause of cancer death in the developed world. Metastatic disease eventually develops in nearly 50% of patients with CRC. Chemotherapy is the mainstay of treatment in metastatic CRC (mCRC); however the majority of patients remain incurable with current therapeutic options. Progress made in the field of surgery, locoregional treatment for low-volume metastatic disease, systemic chemotherapy and targeted agents has created new treatment paradigms and improved survival in mCRC. Aims: To present recent developements in the tratmemnt of metastic colorectal cancer. Methods: Literature review. Results: Development of new cytotoxic drugs and the advent of targeted agents over the past decade have demonstrated the median overall survival (OS) for mCRC increase from 9 months to >2 years. Recent results with the biologic agents bevacizumab, a chimeric human-mouse monoclonal antibody against VEGF, cetuximab, a chimeric human-mouse monoclonal antibody against EGF receptor and panitumumab a fully- human monoclonal antibody against

S8

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

EGF receptor have demonstarated that they improve clinical outcomes when added to current regimens in patients with metastatic colorectal cancer. Dual biologic therapy in combination with chemotherapy has, however, yielded disappointing results. Identification of biological markers is expected to help determine which patients are most likely to respond to these newer agents and thus improve targeted therapy. Conclusions: Optimizing treatment for metastatic colorectal cancer patients has become increasingly complex, requiring a multidisciplinary approach not only to identify those who are curable with resectable disease but also to determine when it is best to incorporate targeted drugs, with which chemotherapy, and in which patients.

TPS, CEA, and CA 19-9 as well as by low BV1-3 of TPA, CEA, and CA 19-9. In adjuvant therapies, a favourable 3-year OS was indicated by low BV1 of M30, CEA, and CA 19-9 as well as by low BV1-3 of all cytokeratins. Similar to CEA and CA 19-9, CYFRA 21-1 was elevated in CRC patients, compared to benign diseases and healthy controls, and showed stage-dependency, while M30 was not found stage dependent. Conclusions: Cytokeratins may be useful as additive markers in predicting therapy response and assessing prognosis with colorectal cancer patients, and they should be included in future multi-marker analyses. Supported by the research project VZMSM 0021620819, and the grant IGA MZ CR 10230.

O1.7 - Cytokeratins in predicting therapy response and prognosis of colorectal cancer patients

O1.8 - HER2 in gastric cancer: a new prognostic factor and a novel therapeutic target

L. Holubec, J. Finek, V. Liska, V. Treska, O. Topolcan, P. Stieber, S. Holdenrieder

F. Di Costanzo, E. Giommoni, L. Antonuzzo

Faculty Hospital and Charles University Pilsen, Czech Republic; Institute of Clinical Chemistry, University Hospital Munich,Institute of Clinical Chemistry and Clinical Pharmacology, University Hospital Bonn, Germany. Background: Cytokeratins are valuable markers in various cancer diseases. Aims: The study Aims to investigate their role in predicting therapy response and assessing prognosis with colorectal cancer (CRC) patients undergoing chemotherapy. Methods: The levels of CYFRA 21-1, TPA, TPS, M30, CEA, and CA 19-9 were analyzed in prospectively collected sera of 30 patients with colorectal cancer during adjuvant (N=15) and primary (N=15) chemotherapy, and the marker levels were correlated with therapy response after two months and overall survival (OS). In addition to that, pre-therapeutic serum levels of CYFRA 21-1, M30, CEA and CA 19-9 were determined in 42 CRC patients, 45 benign colorectal disease patients and 51 healthy controls. Results: Median values of cytokeratins, CEA, and CA 19-9 were not different in the adjuvant and primary setting. Concerning response of primary therapy, pre-therapeutic values (BV1) and values before cycle 2 (BV2) of CYFRA 21-1, TPA, and TPS were lower in responding patients, the same with M30 (BV1). Further on, kinetics from cycle 1 to 2 (BV1-2) or 1 to 3 (BV1-3) were different in the response groups. Favourable 3year OS was indicated by low BV1 of CYFRA 21-1, TPA,

Medical Oncology Unit, A Background: In gastric carcinomas (GC) HER2 is more frequently overexpressed in cancers of the gastroesophageal junction than in stomach cancers (33.2 vs. 20.9%) and in intestinal type rather than in those with diffuse pathology (33 vs. 6%). Globally, about 22% of cases of AGC had overexpression of HER2. Aims: To determine HER2 status, the analyses must be performed on gastric tumor samples using both IHC and FISH as recommended by the validation study for HER2 testing in GC, as described by Hoffman et al. Methods: In GC, HER1 and HER2 overexpression are prognostic factors and targets of novel biologic agents. However, the clinical significance of such overexpression is not yet fully clear, and not all studies have demonstrated an association between HER2-overexpression and poor prognosis. Although the effects of HER3 or HER4 expression in GC not have been clarified, HER3 expression is frequently observed in advanced GC with poor prognosis, and HER4 gene expression seems to be higher in tumor tissue in comparison with adjacent gastric mucosa. Results: Prognostic significance of HER3-HER4 coexpression remains unclear. Trastuzumab is a humanized mouse monoclonal antibody directed against the HER2 growth factor receptor. Some clinical trials have explored the combination of CT plus trastuzumab in HER2-positive gastric and gastroesophageal adenocarcinoma. A large international clinical study, ToGA, screened 3,807 patients and randomized 594 (22.1%) patients with advanced HER2-positive GC to

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

receive, in first-line setting, chemotherapy alone versus the same combinations plus trastuzumab. Conclusions: Median overall survival (OS) was significantly improved in the arm with trastuzumab compared to CT alone: 13.5 vs. 11.1 months (HR. 0.74; 95% CI: 0.600.91; p=0.0048) respectively. The ORR was 47.3% in the trastuzumab arm and 34.5% in the CT arm (p=0.0017). Preventing the formation of HER2-HER3 using antibodies that inhibit dimerization has already demonstrated clinical benefits in HER2-positive breast cancer patients, and this treatment might be particularly sensitive to HER2-positive GCs, which reveal resistance to monoclonal antibodies targeting HER2. The development of new predictive markers analyzing dimers (HER2-HER3) rather than isolated markers could allow us to develop new drugs against GC and other solid tumors.

O1.9 - Profiling microRNA expression in hepatocellular adenoma associated with glycogen storage disease type Ia identifies miR-130b as a potential biomarker for malignant transformation Li, Ling-Hui (1, 3), Chiu, Liya (1), Kishnani, Priya (2), Chuang, Tzu-Po (1), Bali, Deeksha (2), Koeberl, Dwight (2), Austin, Stephanie (2) Boyette, Keri (2), Chen, Yuan-Tsong (2, 3) (1) National Genotyping Center, Academia Sinica, Taipei, Taiwan; (2) Division of Medical Genetics, Duke University Medical Center, Durham, North Carolina, USA; (3) Institute of Biomedical Sciences, Academia Sinica, Taipei, Taiwan Background: Hepatocellular adenoma (HCA) is a frequent long-term complication of type I glycogen storage disease (GSD I) and malignant transformation to hepatocellular carcinoma (HCC) is known to occur in some cases. MiRNAs, a group of non-coding single-stranded RNAs of ~22 nucleotides, negatively regulate expression of gene targets and have been implicated in cancer formation. Deregulation of miRNAs was observed in HCC, suggesting these miRNAs may be involved in HCC pathogenesis. However, there has been no systematic study of expression profiling of miRNA in HCA associated with GSD Ia. Aims: In this study, we aimed to identify miRNAs that were deregulated in GSD Ia HCA and HCC cell lines, and evaluate the potential of selected miRNAs as biomarkers for risk assessment of malignant transformation of GSD Ia HCA. Methods: MicroRNA profiling for adenoma and paired normal liver tissues from seven GSD Ia patients was performed using massively parallel sequencing. Selected differentially

S9

expressed miRNAs were validated in GSD Ia HCAs, general population HCAs, and HCC cell lines using quantitative RTPCR. Serum levels of candidate miRNAs in GSD Ia patients with or without HCA, general patients with HCC and healthy individuals were determined using quantitative RT-PCR. Results: A total of 47 mature miRNA were significantly deregulated in GSD Ia HCA. Among the up-regulated miRNAs, slight over-expression of miR-130b was validated in two out of ten GSD Ia HCAs, while high over-expression was found in six out of seven HCC cell lines. In addition, over-expression of miR-130b was detected in two individual serum samples from GSD Ia patients with HCA and four independent pooled serum samples from general patients with HCC, but none from GSD Ia patients without HCA. Conclusions: These results strongly suggest that miR-130b could serve as a circulating biomarker for risk assessment of malignant transformation of GSD Ia HCA.

O1.10 - Individualized multimodal therapy of pancreatic cancer-review R. Klapdor Internal Medicine, ZeTdT GmbH, Hamburg, Germany Background: During the past decade most of the randomized prospective clinical treatment studies investigated the survival of Pancreatic Cancer patients in relation to one treatment regimen. Because of their very limited effects pancreatic cancer is generally accepted as a tumor disease with a very poor prognosis. Aims: Can an efficacy orientated sequential polychemotherapy approach improve treatment of pancraatic cancer? Methods: The presentation is a review of the literature. Results: However, during recent years, adjuvant treatment studies could demonstrate a significant prolongation of PFST and survival, a new drug with different activity compared to cytostatics could be clinically introduced (Erlotinib), a new Gemcitabine-free drug combination could be shown to improve PFST and survival to a significant extend (FOLFIRI), and prospectively randomized clinical second line trials confirmed, that a second line treatment may add a second PFST to the effects of a firstline treatment. These results seem to support our concept of an efficacy orientated sequential poly-chemotherapy published since 2000 with up to 3–5 treatment sequences in the individual patients. Our results seen during the past decade

S10

suggest a prolongation of survival in relation to the number of effective treatment sequences. In view of recent data the offer to these patients should also include modern concepts of radiotherapy (like stereotactic or locoregional (SIRT) radiotherapy), modern palliative surgical endoscopy as well as actual concepts of best supportive care like pain therapy, nutritional support , pancreatic enzyme application, psychological and social advice and cooperation with self-help groups. Conclusions: Nowadays, Pancreatic Cancer therapy should therefore no longer be planned as 1-line treatment trials, but more and more as individualized efficacy orientated sequential multimodal strategies. A relevant precondition for this concept is already given: tumor marker determinations every 2–4 weeks in combination with imaging methods every 6– 8 weeks allow a rapid and valid detection of efficacy/ nonefficacy and/or new progress response in order to treat a patient with 3 and 4 different regimens, even in the case of aggressive growing Pancreatic Tumors, if diagnosed in time.

O1.11 - Use of specific antibodies capable of controlling pancreatic Ca

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

Methods: Animal models demonstrated that human colon and pancreas cancer placed into the hind thigh of a nude mouse would grow to about 5 cm, within a week. If two intraperitoneal injections of the specific tumor mAb, defined by IHC and demonstrating the presence of the specific tumor antigen were then delivered with effector cells into the peritoneal cavity, marked reduction in the size of the tumors would occur over the ensuing 2–4 wks. Results: When time of expression of the tumor antigens were investigated, it could be shown that many normal cells, several months prior to appearance of tumor, would already be producing antigen. While the pathologist is only able to define phenotypically changed cells, immunohistochemistry could, in our studies, reveal those normal cells expressing tumor protein on inception of genotypic changes. Conclusions: Neogenix monoclonal antibodies appear unique in that they serve a dual role in both diagnosis and therapy. They appear capable of controlling tumor by ADCC more effectively than other antibodies presently being defined for commercialization, ie those that appear to target growth factors.

O1.12 - CA 19.9 in pancreatic cancer: retrospective evaluation of patients with suspicioun of pancreatic cancer

Arlen M., Arlen P., Wang X., Luca J., Saric O. North Shore Univ. Hosp. and Neogenix Oncology Background: We have defined and produced several monoclonal antibodies (mAb’s) specific for colon and pancreatic cancer. These mAbs define a specific immunogenic marker (TAA) found only in the malignancy and which can act as a target for apoptosis. The antibodies, first developed in hybridomas and later sequenced and transfected into CHO, when employed in their chimeric form, are highly specific and effective in defining and attacking colon and many pancreatic cancers. These mAbs demonstrate a high degree of antibody dependent cell cytotoxicity (ADCC) in-vitro in the 50%–90% range. Aims: In reviewing the results of a previous Phase II and Phase III vaccine trial for colon cancer using the Hollinshead pooled allogeneic membrane vaccine from which the monoclonals were derived, a high degree of improved survival was shown to be specifically related to antibody production (IgG1’s) secondary to the delivery of the vaccines. The mAbs resulting from antigen (tumor vaccine) administration were found to be virtually identical to the mAbs developed at Neogenix from the original TAA’s. A cell mediated response did not appear important Such findings have been reported by others.

V. Molina, L.Visa, C.Conill, S. Navarro, JM. Escudero, JM Auge, X. Filella5, L.Fernandez-Cruz, A. Lopez-Boado, J. Ferrer, R Molina. Hospital Clinic. School of Medicine. Barcelona, Spain Background: CA 19.9 is the tumor marker of choice in pancreatic cancer Aims: To evaluate the clinical utility of CA 19.9 in patients suspicious of having pancreatic cancer Methods: CA 19.9 serum levels were prospectively determined in 573 patients admitted in the Hospital with suspicious signs of pancreatic cancer. Final diagnosis was: 77 patients with no malignancy, 389 patients with pancreatic cancer, 37 neuroendocrine pancreatic cancer, 28 cholangiomas, 4 gallbladder, 27 ampullar carcinomas and 11 periampular carcinomas. CA 19.9 was determined using a commercial available assay from Roche diagnostics, and 37 U/ml was considered as the upper limit of normality. Results: Abnormal CA 19.9 serum levels were found in 27%, 81, 5%, 85, 7%, 59.3%, 63.6% and 18.9% of patients with benign diseases, pancreatic cancer, and cholangiomas, and ampullar, peri-ampular or neuroendocrine tumors. Significantly higher concentrations of CA 19.9 were found in patients with malignancy than in those without it or with

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

neuroendocrine tumors. CA 19.9 were higher in biliary tract tumors (mainly in pancreatic cancer) than in other malignancies. CA 19.9 was useful as aid in the diagnosis of pancreatic cancer (sensitivity 77.9%, specificity 95.9%. It was also helpful with respect to tumor resectability. Conclusions: CA 19.9 is the tumor marker of choice in pancreatic adenocarcinomas, with a clear relationship to tumor location, stage and resectability.

S11

survival 0.710, p=0.007 (and 0.725, p=0.004 for 200 days survival). AUC for albumin for 100 days survival was not significant (0.639, p = 0.073) but was significant for 200 days survival (0.756, p=0.001). Conclusions: This study points out ferritin as an independent mortality predictor in patients with pancreas cancer. High serum levels of ferritin at the time of diagnosis of pancreas cancer indicate bad prognosis for patients. Acknowledgement: The study was supported by research projects MSM0021620807 and MZO0000VFN2005.

O1.13 - Ferritin as an independent mortality predictor in patients with pancreas cancer. Results of a pilot study M. Kalousova1, T. Krechler2, M. Jachymova1, A. Zak2, T. Zima1 1 Institute of Clinical Biochemistry and Laboratory Diagnostics, 2 4th Department of Medicine - Department of Gastroenterology and Hepatology, First Faculty of Medicine, Charles University in Prague and General University Hospital in Prague, Czech Republic Background: Prognosis of patients with pancreas cancer is very poor. Aims: To test the significance of laboratory parameters in evaluating the prognosis of patients with pancreas cancer. Methods: The studied group included 57 patients (31 men and 26 women, mean age 65±9 years). Blood was collected at the time of diagnosis of pancreas cancer and basic biochemical and haematological parameters, nutritional and inflammatory markers and tumour markers were measured. Patients were followed up until death (median survival 147 days). Results: Ferritin, iron, albumin, prealbumin, cholinesterase, haemoglobin, C-reactive protein, alkaline phosphatase and carcinoembryonic antigen were of significance for patients´ prognosis in univariate analysis, while CA 19-9, bilirubin, liver, pancreas and kidney tests and lipids were not. Multivariate Cox regression (forward and backward tests) demonstrated ferritin, iron and albumin as independent mortality predictors (RR(95%CI), per unit: ferritin 1.001 (1.000-1.002), p=0.002, albumin: 0.943(0.892-0.996), p= 0.035, iron: 0.936(0.890-0.985), p=0.010). Iron correlated significantly with albumin (r=0.397, p=0.002) but neither iron nor albumin correlated with ferritin. Patients who survived 100 days had significantly lower ferritin (median 246 (IQR 112–647) ug/l vs. non-survivors 435 (IQR 336– 1042) ug/l, p=0.0137), significantly higher albumin (41.7± 5.3 g/l vs. 38.4±5.3 g/l, p=0.029) but no difference in serum iron (10.8±5.3 umol/l vs. 10.0±4.7 umol/l, p= 0.064). ROC analysis for ferritin revealed AUC for 100 days

O1.14 - Existence of distinct two subgroups of local recurrence after radiofrequency ablation for hepatocellular carcinoma N. Matsumoto, S. Ozawa, M. Hatsugai, D. Hasegawa, M. Kobayashi, H. Ikeda, H. Takahashi, K. Matsunaga, T. Ishii, C. Okuse, M. Suzuki, F. Itoh Division of Gastroenterology and Hepatology, St. Marianna University School of Medicine, Kanagwa, Japan Background: Radiofrequency ablation (RFA) is widely used as a local therapy for hepatocellular carcinoma (HCC). However, it is known that some patients show unfavorable prognosis. And we have already reported that in vitro heat treatment can induce additional transformation of HCC cell line, suggesting the insufficient RFA may induce further malignant transformation of HCC in clinical settings. Aims: The aim of this study was to identify HCC patients with unfavorable prognosis after RFA. Methods: Patients and Methods: Between April 2005 and March 2009, RFA was performed on 112 patients with a total of 132 HCC nodules. Among them, we retrospectively analyzed 68 patients with fewer than three HCCs each less than 3 cm in maximum diameter, with liver function estimated as Child-Pugh Class A or B, and with followup for >1 year from the last RFA. Results: Among the 63 patients, 21 patients were recurrencefree, 21 patients had ectopic recurrence (ER), and 21 patients displayed local recurrence (LR). Patient background, HCC stage, AFP, sizes of tumors and survival did not differ significantly between patients with LR and ER, and the number of initial HCCs was the same. The Kaplan-Meier survival curve showed no significant difference in survival between LR and ER (p=0.8651). However, we realized that few deaths were observed among LR patients with longer survival, suggesting two distinct sub-groups. LR patients were further divided into rapidly emerging LR (rLR) and slowly

S12

emerging LR (sLR), based on the median recurrence-free period (287.0 days). Among patients with ER, rLR, and sLR, those with rLR showed the poorest prognosis (sLR vs. rLR, p =0.0002; rLR vs. ER, p=0.0019), while sLR showed the best prognosis (sLR vs. ER, p=0.0064, log-rank test). Patients with rLR showed significantly higher levels of tumor markers. Conclusions: Our data showed that patients with LR could be classified into 2 subgroups, and rLR is related to poor prognosis. High levels of tumor markers are characteristics of type of aggressive LR.

O1.15 - Prognosis of HCC-patients treated with TACE: Validation and improvement of staging systems

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

Conclusions: Although initially not developed for this subgroup of HCC-patients, CLIP was the top-ranked original SS for patients treated with TACE. It could be further optimized by rearranging its strata. TNM was the least suitable score for this group of patients. Combination of SS with newly identified prognostic factors like CRP led to an additional improvement of AIC. Before we can start prospective and external evaluation, further internal validation to identify the best score is mandatory.

O1.16 - HOPX tumor suppressive pathways may be epigenetically controlled in colorectal cancer Yamashita K, Ooki A, Katoh H, Waraya M, and Watanabe M

M. op den Winkel1, D. Nagel2, P. op den Winkel1, J. Sappl1, G. Straub1, C. Zech3, P. Stieber2, F.T. Kolligs1 1Medical Department 2, 2Department of Clinical Chemistry, 3Department of Radiology, LMU University Hospital Munich, Germany Background: Several Staging Systems (SS), developed in sometimes inhomogeneous patient-populations, exist for HCC. Aims: By applying the established SS (BCLC, CLIP, GETCH, JIS, Okuda, TNM and Child-Pugh), we analyzed survival rates of patients treated with TACE in order to determine which of the SS is most suitable for predicting survival in this well-defined patient population. We aimed at improving the established and developing a superior SS. Methods: HCC-Patients treated with TACE as the primary mode of therapy were identified from an electronic medical record database (2000–2009). Overall survival was the only endpoint. All parameters of the SS and additional laboratory parameters were tested in uni- and multivariate analysis. The original SS and modified versions (with rearrangement of subcategories) were ranked according to their ability to predict survival by using the Akaike Information Criterion (AIC). Results: At the time of censorship, 72% of the 167 identified patients had died. 14 negative prognostic clinical, laboratory and tumor factors were identified in univariate analysis. AFP, bilirubin, CRP, albumine, platelets and the BCLC-tumor-features remained significant in multivariate analysis. In Cox regression analysis all of the scores except for TNM remained significant. Modification of the SS resulted in an improved stratification. With regard to AIC CLIP proved to be the top-ranked original SS surpassed only by its modified version. Construction of a new score consisting of the 5 significant parameters in multivariate analysis resulted in an AIC of 969. Adding these parameters to Okuda, CLIP, Child-Pugh and BCLC, respectively yielded a further AIC-improvement.

Department of Surgery, Kitasato University School of Medicine Background: Homeobox only protein (HOPX) is a small protein with homeobox domain, and identified as a tumor suppressor gene candidate by epigenetic unmasking microarray. We have shown epigenetic change of HOPX gene, especially for DNA cytosine methylation of the CpG islands of the promoter region, in primary colorectal (CRC) cancers. Materials and methods: We examined methylation level of HOPX by TaqMan MSP, and examined its relation to clinicopahological factors in CRC. HOPX transfection was made and compared with mock transfection. Phenotypic and gene expression changes were assessed by cDNA microarray. Result: (1) Promoter DNA methylation of HOPX was frequently found in half (50%) of primary CRC. As a whole, defective expression of HOPX was consistent with DNA methylation status, and reactivated by demethylating agents. (2) HOPX methylation level was significantly associated with poor prognosis in stage III CRC. (3) HOPX transfection commonly suppresses anchorage independent growth in soft agar and nude mice and Matrigel invasion. (4) HOPX-affected transcripts were identified by cDNA microarray, and the well-known tumor suppressor genes such as PTEN, Syk, IL20R, and macroH2A with epigenetically altered were listed and validated in CRC transfectants and primary CRC tissues. Conclusion: HOPX is a definitive tumor suppressor gene in CRC cancers, and it is frequently silenced by promoter DNA methylation. This pathway may be deactivated by DNA methylation even for other downstream tumor suppressor genes in CRC. To reveal the clinical significance of HOPX methylation in primary CRC, and functional relevance of defective expression of HOPX.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

O2 BREAST CANCER O2.1 - Role of the tumor suppressor gene PAR-4 on cell proliferation, survival and morphogenesis of mammary epithelial cells MCF10A and breast cancer cells MCF-7

S13

O2.2 - Accuracy of an individual reference limit (IRL) during intensive post-operative monitoring of diseasefree breast cancer patients with serum CEA-TPACA15.3 tumor marker (TM) panel A. Nicolini, S Fancelli, P Ferrari, L Anselmi, M Conte, P Miccoli

de Bessa-Garcia, SA, Pereira, MC, Nagai MA Disciplina de Oncologia, Departamento de Radiologia e Oncologia da Faculdade de Medicina da Universidade de Sao Paulo, Av. Dr. Arnaldo 455, 4 andar, sala 4112, 01246903; [email protected] Background: Experimental evidence indicates that PAWR (PKC apoptosis WT1 regulator; also named PAR-4, prostate apoptosis response-4) is one of the central players in cancer cell survival and could be a target for cancerselective targeted therapeutics. Recently, we have demonstrated that reduced nuclear expression of Par-4 is associated with poor breast cancer prognosis (Nagai et al., Int J Oncol, 37: 41, 2010): however, little is currently known regarding the role of Par-4 in breast cancer tumorigenesis. Aims: Our main goal is to investigate the effects of overexpression and suppression of PAR-4 in cell proliferation, differentiation and survival in mammary epithelial cells. Methods: MCF10A and MCF-7 cells were transfected with expression vectors for PAR-4 over-expression (pCMV6PAR-4) or suppression (pRS-shRNA-PAR-4). Proliferation assays were carried out using MTT. Using a system of three-dimensional (3D) cell culture we also evaluated the expression pattern of Par-4 by immunofluorescence and quantitative real time PCR (qPCR) during the morphogenesis of MCF10A cells. PAR-4 over-expression decreased the proliferation rates in both MCF10A and MCF-7 cells compared to control cells. On the other hand, PAR-4 knockdown leads to increased proliferation in MCF7 cells, indicating that PAR-4 has inhibitory effects on mammary epithelial cells proliferation and survival. PAR-4 was highly expressed in the MCF10A cells inside the acini structure, suggesting PAR-4 might have a role in the lumen acini formation. During the morphogenesis of MCF10A cells in 3D cell culture, the cells within the lumen showed caspase-3 activation, indicating apoptotic activity. PAR-4 was only partially co-expressed with activated caspase-3 on these cells. Conclusions: Our findings suggest, for the first time, that PAR-4 has inhibitory effects on mammary epithelial cells proliferation and survival. In addition, our results provide evidence that PAR-4 is differentially expressed during the process of MCF10A acinar morphogenesis. Supported by FAPESP, CNPq.

Department of Internal Medicine, Department of Surgery, University of Pisa, Italy. Background: Intensive post-operative monitoring with serum TMs for an early detection of breast cancer relapses still is controversial. Aims: To improve cost-effectiveness ratio of intensive monitoring, specificity and sensitivity of CEA-TPACA15.3 TM panel have been evaluated using an individual reference limit (IRL). Methods: Previously we used the commercial kit cut-off value and dynamic evaluation (constant elevation (CE) or progressive increase (PI) in one or more of TM value) as significant increase to suspect a relapse. In this study, in 131 disease-free breast cancer patients an IRL was calculated with 5 consecutive serum values regularly obtained during 6 (mean; 4–9 range) months at the beginning of the follow-up. IRL was calculated considering mean value of the 5 consecutive determinations plus 2SD for CEA and CA15.3, and 3SD for TPA. Two consecutive values higher than IRL were considered a significant increase to suspect a relapse. In this instance, immediately a complete instrumental study was carried out to find out and define the site of recurrences. Results: Using the commercial kit cut-off value, we have reported 31% CE and/or PI for CEA-TPA-CA15.3 TM panel. In this study, total determinations were 678, during a total follow-up of 1425 months. Two consecutive values higher than IRL occurred 21 times (3%) in 21 (16%) of the 126 non relapsed patients with IRL specificity of 84%. In 4 of the 5 patients who relapsed during the follow-up, IRL increase was the first suspicion of recurrence, with 2.5 (m + sd; 2 to 3 range) months lead time. In the last patient, IRL was falsely negative with IRL sensitivity of 80%. Accuracy in the early detection of recurrences was 83% and negative predictive value was 99%. Conclusions: IRL strongly increased specificity of serum CEA-TPA-CA15.3 TM panel compared to that using a conventional dynamic cut-off. These preliminary data point out that an appropriate use of IRL permits to carry out in these patients an intensive monitoring likely with a favourable cost-effectiveness ratio.

S14

O2.3 - Sensitivity, specificity and predictive values of serum HER-2/ECD for detecting occult metastases in tissue HER-2 positive breast cancer patients

Patricia Diana Sorensen1,2, Jonna Skov Madsen1,2, Rikke Fredslund Andersen1, Erik Hugger Jakobsen3, Ivan Brandslund1,2 1Department of Clinical Biochemistry Vejle, Lillebaelt Hospital, Denmark 2IRS, University of Southern, Denmark 3Department of Oncology, Lillebaelt Hospital, Denmark Background: HER-2 gene amplification and protein overexpression are associated with poor prognosis in breast cancer patients. HER-2 status is assessed in tissue, usually from primary tumor, using immunohistochemistry (IHC)/fluorencence in situ hybridization (FISH). The extracellular domain of the HER-2 protein (HER-2/ ECD) may be cleaved and shed from the surface of breast cancer cells and released into circulation where it can be measured in serum. Aims: The current study aim is to determine sensitivity, specificity and predictive values of serum HER-2/ECD for detecting occult metastases in tissue HER-2 positive breast cancer patients. Methods: One hundred eighty nine tissue HER-2 positive breast cancer patients were monitored with serum HER-2/ECD at clinically determined time intervals (typically 3–12 months) according to routine practice at the Oncology Department. Serum HER-2/ ECD was measured using the ADVIA Centaur HER-2 assay, values more than 15 ng/mL were considered positive. Tissue HER-2 was determined routinely using IHC/FISH where IHC +3 or IHC +2 and FISH >2 is considered positive. Disease recurrence was diagnosed according to routine clinical practice. Results: Forty five patients had recurrence in the form of metastases, 33 were serum HER-2/ECD positive (>15 ng/ mL) before the event, giving a sensitivity of 73% (95% CI (58% – 85%)). One hundred thirteen patients did not have metastases in the period, 82 were serum HER-2/ECD negative giving a specificity of 72% (95% CI, 64% – 80%) Thus the positive predictive value is 51% (95% CI (39% – 64%)) and negative predictive value 87% (95% CI (79%-93%)). Thirty one patients out of 189 were excluded: 29 because of missing blood test before recurrence and 2 because of local relapse.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

Conclusions: Serum HER-2/ECD is a useful marker for detecting occult metastases in tissue HER-2 positive breast cancer patients. It should be investigated in future studies whether detecting occult metastases at an earlier time increases overall clinical outcome.

O2.4 - Tumor markers as prognostic and predictive factors in primary locoregional breast cancer R Molina, J M. Auge, B. Farrus, J. Pahisa, M. Muñoz, A. Torne, X. Filella, JM Escudero, P. Fernandez, M.Velasco Hospital Cl nic. Medical School, Barcelona, Spain Background: Tumor marker utility as prognostic factors in primary breast cancer is unclear. Aims: To evaluate the tumor marker utility as predictive and prognostic factors in breast cancer. Methods: CEA and CA 15.3, were prospectively studied in the sera of 2062 patients (HER-2/neu in 1332 patients) with untreated primary breast cancer diagnosed between 1984 to 2008. All three tumor markers were also determined during neoadjuvant treatment in 124 patients (103 chemotherapy, 15 combinations of chemotherapy and hormonotherapy, 6 other treatments). Results: Increased CEA (>5 ng/mL), CA 15.3 (>30 mU/ mL) and HER-2/neu (>15 ng/ml) concentrations were found in 12.7%, 19.6% and 10.6% of the patients, respectively, and one or both tumor markers were increased in 34%. Increases in each tumor marker were correlated with larger tumor sizes and nodal involvement (HER-2/neu only in those with tissue overexpression). Tumor size, ER, and CEA were independent prognostic factors by multivariate analysis in the total group (disease free survival (DFS) and overall survival (OS)) as well as in node-positive (NP) and node-negative (NN) patients. Nodal involvement and histological grade were independent prognostic factors in the total group as well as in NP patients. By contrast, neoadjuvant treatment, adjuvant treatment, histology and CA 15.3 were independent prognostic factors only in NN patients (DFS and OS). All patients with CEA >7.5 -1/4g/L, had recurrence during follow-up. Use of both tumor markers allowed discrimination of the groups of risk in T1 NN patients: 56.3% of recurrences were seen when one or both tumor markers were increased, whereas only 9.4% of recurrences were seen in T1 NN patients without increases of either marker. Patients with abnormal HER-2/ neu serum levels before treatment showed a better tumor response (p=0.018).

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

S15

Conclusions: CEA and CA 15.3 are useful prognostic factors in NP and NN breast cancer patients. CEA >7.5 −1/4g/L is associated with a high probability of subclinical metastases.

ity, followed by TPS and CA 15-3 in the prognosis estimation of Breast Cancer pts and should be included in clinical trials.

O2.5 - Prognostic significance of cytokeratin markers in breast cancer “a meta analysis”

O2.6 - The sum of serological biomarkers for the early detection of metastatic breast cancer

V Barak (1), AC Aronsson (2), Novikov I (3), T Peretz (1), R Einarsson (2)

Petra Stieber1, D Di Gioia2, I Bauerfeind3, I Ertl1, M Untch4, D Nagel1, V Heinemann2

(1) Immunology Laboratory for Tumor Diagnosis, Oncology Department, Hadassah- Hebrew University Medical Center, Jerusalem, (2) Biostatistical Unit, Gertner Institute for Epidemiology and Health Policy Research, Israel (3) IDL Biotech AB, Stockholm

1Institute of Clinical Chemistry, 2Department of Medical Oncology, 3Gynaecological Department, University of Munich “ Campus Grosshadern, Germany; 4Gynecological Department, Helios Klinikum, Berlin-Buch, Germany

Background: Cytokeratin fragments 16, 18, 19 serum levels, have been shown to be sensitive biomarkers with prognostic significance in a variety of cancers. Aims: A Meta-analysis evaluating the potential of Cytokeratin Markers (TPS, TPA) compared to established Tumor Markers (CA 15-3, CEA, CA125, CA 27-29) to improve clinical status evaluation of advanced Breast Cancer patients (pts), to assess their response to therapy, and calculate the Lead Time to Metastasis development by those markers. Methods: A Meta-analysis combining info from 49 small studies, from 1985 up to now, summarizing 3379 advanced Breast Cancer pts.It included descriptive/graphical methods (L'Abbe plot, Forest plot, Funnel plot), estimated an overall common effect and between-studies heterogeneity, mainly using a random effect model and determined ability to draw Conclusions. Results: Sensitivity (%) of markers from all studies was recalculated (95% CI) and found; CA 15-3- 67.9 (67.8 / 68.0), CEA “ 53.0 (52.8 / 53.1), TPS “ 71.1 (71.0 / 71.2), TPA “ 70.6 (70.3 / 70.9), the combination CA 15-3 + TPS “ 81.7 (81.5/ 81.9) or CA 15-3 + CEA + TPA “ 89.0 (88.7 / 89.3). Clinical correlations to marker levels, overall estimates (95 % CI) revealed: TPS “ 0.795 (0.594, 0.903), TPA “ 0.638 (0.047, 0.898), CA 15-3 “ 0.554 (0.340, 0.713), CEA “ 0.614 (0.405, 0.761). Lead times (months) for local recurrences or Metastasis (months, 95% CI) were: TPS “ 5.9 (3.11, 6.82), TPA “ 3.5 (1.76, 3.13), CA 15-3 “ 4.4 (2.93, 5.93), or the panel CEA + TPA+ CA 15-3 “ 11.9 (2.93, 20.89). Conclusions: Cytokeratin markers demonstrated the best indications of response to therapy and provided the most important prognosis in Breast Cancer pts. The combination of CEA, CA 15-3, TPA demonstrated the highest sensitiv-

Background: Whether the early detection of asymptomatic metastatic breast cancer can lead to more effective and specific treatment with better or longer survival is not known nowadays. But it is evident that the first signs of metastatic spread can be recognized in blood. Aims: Early and specific detection of metastatic disease (=asymptomatic patients) in breast cancer patients in a non randomized prospective trial using kinetics of CEA, CA153 and CA 125 in 6-week intervals during follow-up after completion of perioperative treatment. Methods: Among the 681 patients we analysed the clinical data of those who developed distant metastases (n=91) were investigated. A reproducible increase of CEA (Abbott, AxSYM) or CA15-3 or CA 125 (both Roche, Elecsys) ‰ ¥100%, corresponding to a predefined specificity of 98%, was the indicator for metastases. Results: 91 patients developed metastases. The medians of the individual baseline values of these patients when they entered the trial with no evidence of disease after primary treatment were: CEA: 1.1 ng/ml; CA 15-3 16.8 U/ml; CA 125 12.8 U/ml. 66 patients (73%) showed the previously defined increase of CEA or CA15-3 or CA 125 at the time of first metastases (true-positive, TP). In 29 patients (32%) CA15-3 alone, in 16 patients (18%) CEA alone, in 10 patients CA 125 alone (11%), in 8 patients (9%) 2 of the 3 markers increased in parallel and in 2 patients (2%) all 3 markers increased. 24 (36%) of the TP patients suffered from oligometastases. All patients with liver-only metastases were TP (n=12), 8 of them suffered from oligometastases. 25 patients (27%) did not have a significant increase of tumor markers at the time of first metastases (falsenegative, FN), of these 10 (40%) had oligometastases (4 bone, 3 lung, 2 brain, 1 skin).

S16

The median disease-free interval until detection of metastases did not differ between TP (46 months) and FN patients (57 months) (p=0.5). Conclusions: Kinetics of CA 15-3, CA 125, and CEA during follow-up of breast cancer patients allow early detection of asymptomatic and oligometastatic disease.

O2.7 - uPA and PAI-1: clinically validated prognosis factors for individualized therapy decisions in early breast cancer C. Thomssen, (1); Schmitt, Manfred (2); Vetter, Martina (1); Kantelhardt, Eva (1); J ₪nicke, Fritz (3); Harbeck, Nadia (4)

Departments of Gynaecology: (1) Martin-Luther-University, Halle (Saale); (2) Technische Universit ₪t Munich; (3) University Hospital Hamburg; (4) University Hospital Cologne; Germany Background: In node-negative breast cancer, more than 70% of all patients are cured by loco-regional therapy alone. However, current consensus guidelines recommend adjuvant therapy for nearly all node-negative breast cancer patients. Thus, in node-negative breast cancer, appropriate identification of low-risk patients who do not need potentially toxic adjuvant chemotherapy is not resolved. Currently, gene expression profiling and molecular typing are widely discussed as appropriate tools for decision making on adjuvant therapy. However, for none of these tools data from prospective studies are available. Aims: The invasion markers urokinase-type plasminogen activator (uPA) and its inhibitor PAI-1 were studied for more than twenty years, and level-one evidence has accumulated that these markers are of significant relevance for prognostic information in node-negative breast cancer. In addition, there is also data that tumours with high uPA and/or PAI-1 levels may be more sensitive to adjuvant chemotherapy. Methods: Standardized determination of uPA and PAI-1 is performed by ELISA in extracts of fresh frozen tumours. Results: In a unicentre, prospective observational study, it was demonstrated that in node-negative breast cancer, patients with tumours that express low levels of both, uPA and PAI-1, have a very low risk of relapse comprising about 50% of all node-negative patients. The data of a Europeanwide pooled analysis with more than 8000 patients confirmed the prognostic value these markers. In the prospective multicentre trial CHEMO-N0, we confirmed the significant and independent prognostic impact of uPA/ PAI-1. With ten-year follow up, also benefit from adjuvant chemotherapy was shown in patients with high uPA/PAI-1

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

tumour concentrations. Data from the NNBC-3 Europe trial suggests that in luminal-type cancers, differentiation between high and low risk may also be feasible. Conclusions: Routine use of uPA and PAI-1 for risk estimation in node-negative breast cancer patients is well supported and can avoid adjuvant chemotherapy in about half of all node-negative patients.

O2.8 - Characterization of Human Breast Tumor Cell Line BT-20 Adapted to Long-Term High Nitric Oxide Exposure

H De Vitto1-4, BS Mendonça1, KM Elseth2-4, A Onul2-4, A Noweder2-4, J Xue2-4,BJ Vesper2-4, CVM Gallo5, FD Rumjanek1, JA Radosevich2-4 1Universidade Federal do Rio de Janeiro, IBqM, Rio de Janeiro, Brazil; 2Center Molecular Biology of Oral Diseases, College of Dentistry, University of Illinois at Chicago, Chicago, IL, USA; 3Jesse Brown VAMC, Chicago, IL,USA; 4Dept. of Oral Medicine and D Background: Background: Reactive oxygen species (ROS) and nutrient deprivation play an important role in breast cancer mitochondrial adaptation. This adaptation allows cells to survive in the stressful microenvironment of the tumor. Aims: This study is directed at understanding the consequensces of High Nitric Oxide (HNO) exposure to mitochondria in human breast cancer cells. Methods: Methods and materials The breast cancer cell line BT-20 (parent) was adapted to HNO, resulting in the BT-20-HNO cell line. Both cell lines were analyzed using MTT, FACS, LDH leakage assay, DNA sequencing, and Western blot analysis. MTT growth assays and FACS analysis revealed that BT20-HNO grew faster and had more cells in S-phase than the parent cell line. The LDH assay revealed that BT-20-HNO was more prone to use the glycolytic pathway than the parent cell line. The BT-20-HNO cells were also more resistant to the apoptotic inducing agent Salinomycin, suggesting that p53 may be mutated in these cells. PCR followed by DNA sequencing of the p53 gene showed that it was in fact mutated in the DNA binding site (L194F). Western blot analysis showed that p53, iNOS, and eNOS were significantly up-regulated in these cells. Conclusions: Conclusions: Our results suggest that free radicals, such as NO, pressure human breast tumor cells to acquire an aggressive phenotype with resistance to apoptosis. These findings provide a mechanism by which the dysregulation of ROS in the mitochondria of breast cancer cells results in DNA damage.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

O3 OVARIAN CANCER O3.1 - Serological tumor-associated markers CA 125 and HE4 in monitoring of ovarian cancer patients N.S. Sergeeva, N.V. Marshutina, I.A. Korneeva, E.G.Novikova Moscow Herzen Research Oncological Institute, Moscow, Russia Background: Relations of absolute marker value to its discriminative level for all measuring points of markers during dynamic patient observation were calculated. Aims: To compare the dynamic of CA 125 and HE4 in monitoring of serous ovarian cancer (OC) patients. Methods: HE4 concentration was evaluated in blood serum of serous OC patients with “Architect HE4 assay” sets (Abbott Diagnostics, USA). CA 125 concentration was evaluated using “CA 125 AxSym” sets (Abbott Diagnostics, USA). Discriminative levels for CA 125 was 35 U/ml, and for HE4 - 70 pmol/l. Markers levels were determined in dynamic of treatment and observation, certainly including periods of remission and relapse. Results: For comparing of degree of two markers reaction for treatment and relapse, the following indicators were considering: 1) a primary level (relating to upper limit), 2) decreasing rate (during treatment), 3) rate and degree of increasing during development of relapse. In most cases a similar markers changes treatment and development of relapses were found. CA 125 dynamic was more expressed (comparing to HE4) in 8 cases of 12. At the same time HE4 was more ”active” than CA 125 in other three cases. The changes degree of two markers was similar in one case. Group of 3 patients (HE4 preference) differs from group of 8 patients (CA 125 preference) by relapse beginning in long time (3–5 years) after primary treatment. Conclusions: The addition of new tumor marker – HE4 to CA 125 monitoring of OC patients are justified at least for part of patients. A following study make possible to define HE4 place in treatment and observation stages in OC patients.

O3.2 - Therapeutic intervention opportunities for ovarian cancer patients diagnosed with late stage disease WC Hitt1, AF Burnett1, RA Dennis2, MJ Cannon3, K O’Brien1, K Shigemasa4, JB Beard1, TJ O’Brien1 1University of Arkansas for Medical Sciences, Department of Obstetrics and Gynecology 2Central Arkansas Veterans Healthcare System 3University of Arkansas for Medical Sciences, Department of Microbiology and Immunology 4National Hospital Organization Fuku

S17

Background: Ovarian cancer remains a major challenge to caregivers in that it is the number one cause of death in women with gynecologic cancer and its diagnosis at early stage still remains extraordinarily low (about 20–25%). Unfortunately, women diagnosed with late stage disease almost all have recurrent disease and eventually die of metastatic carcinoma. Aims: To develop new markers which could be utilized to complement early diagnosis for ovarian cancer patients and which would offer new targets for therapeutic intervention in patients with persistent disease. Methods: To address the selection of new biomarkers for early diagnosis and therapeutic intervention we have examined multiple families of genes which could contribute to tumor growth and spread. Members of selected families were amplified by PCR and sequenced to determine which family members were overexpressed. In some cases individual genes were cloned to confirm family membership. Follow-up PCR on a panel of ovarian tumors confirmed these genes were overexpressed in tumors and may be a candidate panel for early diagnosis and therapeutic intervention. Results: After developing ELISA assays to the selected markers, we have now confirmed that these individual markers are present in the serum of patients with early stage disease. Following patients with known disease further confirms the presence of these markers in patient serum with recurrent and/or progressive disease. Because these enzymes are constitutive mediators of tumor growth and spread they offer an opportunity for inhibition utilizing protease inhibitors thereby down regulating tumor growth and spread. Conclusions: Profiling patients after diagnosis can allow identification of patients which express individual functional biomarkers and can therefore offer a window of therapeutic opportunity for treatment with protease inhibitors when no alternative therapy is available for patients who have become resistant to standard chemotherapy.

O3.3 - HE-4, CA 125 and ROMA in the differential diagnosis of patients with gynecological diseases R Molina, JM. Escudero, JM. Auge, X Filella ,L. Ferro, A Torne, J. Pahisa Hospital Clinic. Barcelona, Spain Background: CA 125 is the marker of choice in ovarian cancer.

S18

Aims: To evaluate a new tumor marker, HE4, in comparison with CA 125 and the Risk of Ovarian Malignancy Algorithm (ROMA) in the differential diagnosis of gynaecological diseases. Methods: CA125 and HE4 serum levels were determined in 166 healthy women, 285 patients with benign gynaecological diseases (68 endometriosis, 56 myomas, 137 ovarian cysts and 24 with other diseases), 33 patients with non-active gynaecological cancer and 143 with active gynaecological cancer (121 ovarian cancers). CA 125 and HE4 cut-offs were 35 U/mL and 130 pmol/ L, respectively. ROMA algorithm cut-off was 13.1 and 27.7 for premenopausal or postmenopausal women, respectively. Results: HE4, CA125 and ROMA results were abnormal in 1.1%, 12.4% and 11.7% of healthy women and in 1.1%, 30.2% and 12.3% of patients with benign diseases, respectively. Among patients with cancer, HE4 had significantly higher concentrations in ovarian cancer than in other malignancies (p > 100 U/ml in premenopausal and >35 U/ml in postmenopausal women). Sensitivity was 94% and 83% in postmenopausal and premenopausal (women with a positive predictive value of 87% and 56%, respectively. Conclusions: HE4 has a higher specificity and similar sensitivity than CA 125. Combination of both markers increases the utility in the diagnosis of ovarian masses.

O3.4 - TP53-related microRNA in primary ovarian cancer A. Swiercz(1), M. Zajdel(1), M. Kulinczak(1), M. Chechlinska(1), A. Dansonka-Mieszkowska(2), A. Rembiszewska(2), J. Kupryjanczyk(2), J. K. Siwicki(1) (1)Immunology and (2)Molecular Pathology Departments, Cancer Centre and Institute of Oncology, Warsaw, Poland Background: Changes in TP53 and TP53-dependent pathways are critical for the pathogenesis and biology of ovarian cancer (OvCa). Some microRNAs (miRs) are important regulators of TP53-dependent pathways. The role of TP53-related miRs in cancer biology is unclear, and

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

in OvCa, the available data come mostly from cell line studies. Aims: We aimed to study the expression of miR-34a,b,c, miR-145 and miR-125b known to be related to TP53dependent pathways, and miR-205 overexpressed in OvCa, in the clinical samples of primary epithelial OvCa (PEOvCa) of different TP53 status and tumour grade. Methods: miR expression was measured by RT-qPCR in PEOvCa with >75% cancer cells, and in A2780, IGROV-1, TOV-112D and SKOV3 lines. Normal human ovarian surface epithelium (HOSE) served as a calibrator. Results: 1. PEOvCa, compared to HOSE, presented significant alterations in miR-34a, -34b, -34c, -145, 125b and −205 expression, independent of tumour grade. 2. PEOvCa with the wild-type TP53, expressed decreased miR-205 and inconsistently changed miR-34b levels, while cancers with TP53 mutations frequently overexpressed miR-34b and −205. 3. Regardless of the TP53 status, miR-34a was downregulated, miR-145 and -125b were upregulated and miR34c expression varied in PEOvCa. 4. In OvCa cell lines, miR-34a and -34b were downregulated, -34c was downregulated in A2780 and SKOV3 and not expressed in IGROV-1 and TOV-112D, miR-145 was upregulated in IGROV1 only, but miR-125b in none of the lines, and miR-205 was downregulated or not expressed. Conclusions: 1. Tumour grade does not influence the expression of the studied miRs. 2. In PEOvCa, the expression of miR-34b and miR-205 relates to the TP53 status. miR-205 emerges for the first time, to be associated with TP53. 3. miR-34b level seems to be regulated by an additional way to a direct TP53 transactivation, thus miR-34 family members appear to be under different regulatory mechanisms. 4. miR expression differs between clinical samples and OvCa cell lines. Functional studies with ectopic TP53/miR expression/ silencing are ongoing. The work was partly supported by the grant NN401050138 from the Polish Ministry of Science and Higher Education.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

O3.5 - The ROMA score and the influencing factors upon the diagnostic accuracy Petra Stieber1, A Kirschenhofer2, S F ¼rst2, D Nagel1, L Hertlein2, M Lenhard2, D Mayr3, A Burges2

S19

O3.6 - Gynecological Cancer Inter Group (GCIG) criteria to interpret serial CA 125 concentrations: their monitoring performance evaluated in a computer simulation model G. Soletormos1, F. Lund1, P. Hyltoft Petersen2, M Frejstrup Pedersen1, T. Hillig1, MK Tuxen3

Institute of Clinical Chemistry1, Department of Gynecology and Obstetrics2, Institute of Pathology3, University of Munich “ Campus Großhadern, Munich, Germany

1 Dpt. Clinical Biochemistry, Hillerod Hospital, Denmark. 2 NOKLUS, Bergen University, Norway

Background: Early diagnosis of ovarian cancer needs significant improvement. Aims: CA 125 and HE4 are described to possess additive diagnostic accuracy in the differential diagnosis of ovarian tumors. Both biomarkers are neither tumor- nor organspecific. We investigated their release in benign and malignant diseases. Methods: The sera (stored at -80 °C) of 206 healthy individuals, 671 patients with various benign disorders and 720 patients suffering from cancer obtained at time of diagnosis before first treatment were analyzed retrospectively measuring CA 125 and HE4 on the Architect system (Abbott Diagnostics, US). Results: The median (95th perc) of HE4 in healthy men is 26 pM (47), in women 40 pM (74), of CA 125 10 U/ml (20) in men and 14 U/ml (31) in women. The medians of benign diseases are for HE4 between 30 pM and 51 pM (except benign urol diseases with 660 pM), for CA 125 between 13 and 20 U/ml, the 95th perc. reaches HE4-values of 400 pM in gastrointest. diseases and 3500 pM in urol. dis, for CA 125 940 U/ml in benign gastroint. dis. For HE4 and CA 125 the highest values are reached in ovarian cancer with median values of 242 for HE4 and 391 for CA125, followed by lung cancer, oesophagus cancer and cervix uteri cancer for HE4 and cervix uteri, stomach, hepatocelluar and lung cancer for CA 125. Highest sensitivity shows HE4 (68%) and CA 125 (69%) at 95% specificity in ovarian cancer and Borderline Tumor (control group: benign gynaecological disorders) with AUC 88.9% for HE4 and 91,9% for CA 125. Conclusions: HE4 and CA 125 show the highest sensitivity in ovarian cancer and borderline tumors, but are both panmarkers and released in various benign and malignant diseases. In patients with a reduced renal function HE4 levels as well as combinations of HE4 with other biomarkers must be interpreted with caution and should not be the single indicator for invasive diagnostic or treatment procedures.

Background: In the management of ovarian cancer several algorithms for interpretation of serial CA 125 concentrations during monitoring have been proposed. The GCIG has proposed criteria to interpret increments starting from below cutoff and above cutoff, respectively. Criterion for CA 125 increments starting below cutoff (Criterion A): CA 125 concentrations ‰ twice the upper limit of normal on two occasions. Criterion for CA 125 increments starting above cutoff (Criterion B): CA 125 concentrations ‰ the nadir value on 2 occasions. Aims: To investigate the monitoring performance of the two GCIG criteria in a standardized computer simulated model. Methods: A computer simulation model was developed using parameters for the tumor biomarker CA 125 regarding biological variation and different rates of increases during progressive disease. The two GCIG algorithms were applied to the simulated data corresponding to 1000 surrogate patients. The steady-state biological variation (CV-withinsubjects) was set to 35.5%, and based on Gaussian random numbers. CA 125 increase indicating clinical progressive disease was simulated by using three different exponential increases ( » = 0.009, 0.017, and 0.07 kU/L/day). Results: The main outcome of the simulations was that Criterion A performed well with a very low frequency of false positive signals of tumor growth, whereas Criterion B provided an unacceptable high frequency of false positive signals of tumor growth. Conclusions: A recent publication at ASCO 2009 by the OV05 and 55955 Collaborators found that œWomen can be reassured that there is no benefit from early detection of relapse by routine CA125 measurements even if CA125 rises. Chemotherapy can be delayed until signs or symptoms of tumor recurrence. However, this study applied a modified version of the GCIG Criterion A, using CA 125 sampling intervals of 3 months which may have invalidated the lead time potential of the biomarker.

S20

O4 THYROID CANCER

O4.1 - Medullary thyroid cancer: from genetic to therapy R. Elise., V. Bottici, V. Cappagli, D. Viola, L. Valerio and A. Pinchera Department of Endocrinology and Metabolism, University of Pisa, Italy Background: Medullary thyroid carcinoma (MTC) is a rare tumor arising from parafollicular thyroid C cells and can occur sporadically (70-80%) or as part of the multiple endocrine neoplasia (MEN) type 2 syndromes (20-30%). In these cases, other endocrine organs (i.e., parathyroid and adrenal glands) may be involved, and three different syndromes, MEN 2A, MEN 2B, and familial medullary thyroid carcinoma (FMTC), can be distinguished. According to the literature, the clinical prevalence of the 3 hereditary syndromes is about 60% for MEN 2A, 10% for MEN 2B and 30% for FMTC. Conversely, the sporadic form of MTC is defined by the absence of a familial history of MTC and of the other MEN 2-related tumors. Aims: RET is a protooncogene coding for a tyrosine kinase transmembrane receptor involved in cell growth and proliferation. In 1993, germline RET mutations were recognized as the causative molecular alterations in MEN 2 syndromes. As recommended in the recently published American Thyroid Association (ATA) guidelines for the management of patients with MTC, all patients with clinical suspicion of hereditary MTC should be screened for the presence of a germline RET mutation and, if positive, all first degree family members should be tested to identify the gene carriers. Moreover, a RET somatic mutation is also present in 40% of sporadic MTC. The RET mutated sporadic cases have a worse prognosis than those with RET wild type. Methods: Literature data Results: Recently, several new drugs have been developed which are able to block the activated tyrosine kinase receptors. Several of them such as AMG706,

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

ZD6474 and XL184 have been tested first in vitro and then in vivo for their putative anti-RET activity. Promising results have been obtained in patients with advanced MTC and treated with these drugs in phase II/ III clinical trials in terms of prolonged recurrence free time. Conclusions: These data are expected to be confirmed by further prospective multicenter randomized trials. O4.2 - Treatment of differentiated thyroid cancer F. Pacini Section of Endocrinology and Metabolism, Department of Internal Medicine, Endocrinology & Metabolism and Biochemistry, University of Siena, Italy Background: Initial treatment for thyroid cancer is total thyroidectomy and dissection of involved lymph node compartements. After total thyroidectomy, patients with differentiated thyroid cancer (DTC) are treated with 131I. This procedure decreases the risk of locoregional recurrence and facilitates the long term follow-up. In addition the high activity of 131I allows obtaining a highly sensitive post-therapeutic whole body scintigraphy (WBS). Radioiodine ablation is recommended in high risk and in low risk patients, while there is no indication in very low risk patients. Effective thyroid ablation requires stimulation by TSH, while the patient is on levo-thyroxine (LT4) therapy. Aims: After thyroid ablation, aim of follow-up is the early discovery and treatment of persistent or recurrent disease. Methods: Literature data Results: In our Center, two to three months after initial treatment thyroid function tests (FT3, FT4, TSH) are obtained to check the adequacy of LT4 suppressive therapy. At 6 to 12 months the follow-up is aimed to ascertain whether the patient is free of disease. At this time most (nearly 80%) of the patients belong to the low risk categories and disclose normal neck ultrasound and undetectable (n of surgery and radioiodine therapy. External beam radiotherapy also may be indicated. Distant metastases are more successfully cured if they take up radioiodine, are of small size located in the lungs (not visible at X-rays). Bone metastases have the worst prognosis.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

Conclusions: When appropriately treated, 2/3 of DTC patients with local disease and 1/3 of those with distant disease may achieve complete remission. Whenever radioiodine therapy is not effective and the disease progress, enrollment of the patients in experimental trials with tyrosine kinase inhibitor should be the treatment of choice.

O4.3 - New targeted therapies in advanced thyroid cancer Lisa Licitra Medical Oncology, Istituto Nazionale Tumori, Milan Background: Biological agents are rapidly developing for the treatment of metastatic RAI resistant thyroid cancer. Aims: Agents that target BRAF and VEGFR have shown the most promising results. Methods: Literature data. Results: BRAF V600E mutation seems to be positively associated with tumour response by using BRAF targeting agents. With these agents, impressive clinical responses and prolonged disease stabilisation were observed. This activity compares favourably with that of chemotherapy with less prominent toxicity, although typically associated drug sideeffects should be promptly recognised and managed. To date no drug has proved to prolong survival but at least a one placebo randomized controlled study in medullary cancer showing a statistical improvement of progression free survival with vandetanib. Conclusions: Other randomized studies are ongoing. The therapeutic challenge is to better correlate tumor biological profile and treatment activity, to look for drugs that are tolerable in the long term, to discover new drug able targets and to start refining treatment sequences based on the development of secondary resistances.

O4.4 - Molecular hallmarks of differentiated thyroid cancers Fuvio Basolo Section of Pathology, Department of Surgery, University of Pisa, Italy.

S21

Background: Emerging evidence indicates that interactions between chemokine receptors and their ligands may play a critical role in several steps of tumor development, including tumor growth, progression, and metastasis. The chemokine receptor CXCR4 (CXC chemokine receptor 4, also called fusin) interacts specifically with the chemokine CXCL12 (ckemokine C-X-C motif - ligand 12) to exert proliferative and chemotactic effects. CXCL12 is constitutively expressed in the lymph nodes, bone marrow, and liver tissue. The CXCL12/CXCR4 axis is involved in migration and metastatic processes in several human neoplasms, including carcinoma of the breast, prostate, lung, colon, pancreas and melanoma. Aims: We investigated the relationship between the clinicopathological features of the tumors and mutations in the BRAF gene to verify whether over-expression of CXCR4 is linked to more aggressive behavior in well differentiated thyroid tumors. Methods: CXCR4 protein expression was evaluated by immunohistochemical staining in a consecutive series of 200 papillary thyroid carcinomas. A final staining score was calculated by adding the score representing the percentage of positive cells to the intensity score. The CXCR4 expression of each papillary thyroid carcinoma sample was normalized by calculating the z score for each final staining score. BRAF mutation was investigated by direct genomic sequencing analysis. Results: Univariate analysis was used to correlate CXCR4 expression with the papillary thyroid carcinoma variant, the degree of neoplastic infiltration, AJCC stage, the presence of lymphocytic thyroiditis and the mutation status of the BRAF gene. Multiple regression analysis confirmed a strong association between CXCR4, BRAF mutation and the degree of neoplastic infiltration. These data clearly indicate that the chemokine receptor expression induced by oncogenic activation could be the major determinant of the local aggressiveness of neoplastic cells. Conclusions: Our data indicate that CXCR4 expression and BRAF mutation status could cooperatively induce and promote a more aggressive phenotype in papillary thyroid carcinoma via several pathways and specifically increase the tumors™ spread outside of the thyroid gland.

S22

O5 PROTEOMICS & EPIGENETICS

O5.1 - Epigenetic changes in repetitive elements are closely associated with genetic and clinical features in multiple myeloma M. Nojima1, Y. Aoki2, H. Yasui2, R. Maruyama3, M. Ashida3, H. Suzuki3, T. Ishida2, and Y. Shinomura2 1Department of Public Health, 21st Department of Internal Medicine, and 3Department of Molecular Biology, Sapporo Medical University School of Medicine Background: Multiple myeloma (MM) is still a lethal malignancy despite the development of new therapies. One of the prognostic factors in MM is various chromosomal aberrations including hyperdiploidy, monosomy (e.g. chromosome 13), and translocations. From the aspect of epigenetics, it has been reported that hypomethylation of long interspersed nuclear elements 1 (LINE-1), a family of repetitive elements, is related to chromosomal instability. However, it has not been clear how it affects such chromosomal aberrations. Aims: To assess an association of repetitive elements hypomethylation with the chromosomal and clinical features in MM Methods: To assess the methylation levels of repetitive elements including LINE-1, Alu and Satellite-alpha (Sat-a), pyrosequencing was performed. Chromosomal aberrations were detected by array-based comparative genomic hybridization (aCGH). Bone marrow plasma cells (CD138+) collected from 74 MM and 7 MGUS cases and 4 normal plasma cells (NPC) were used for this study. Results: Significant linear reductions in methylation levels of repetitive elements were observed according to the degree of malignancy (NPC, MGUS, and MM). After hierarchical clustering based on aCGH, MM samples were categorized into approximately 5 groups. Remarkably, hypomethylation of repetitive elements were closely associated with the groups with frequent chromosomal deletions. For example, the mean methylation levels of LINE-1 was 36.1% with chromosome 13 deletion, and 44.0% without its deletion (P=0.010). Also, the number of deleted probes was significantly correlated with LINE-1 methylation levels (R=−0.531). Finally, we observed significantly poorer prognosis in the lower LINE-1 methylation group (P=0.015, compared to the higher methylation group).

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

Conclusions: DNA methylation levels of repetitive elements, especially of LINE-1, are associated with the frequency of chromosomal deletions and prognosis in MM. We believe our findings may contribute to the clinical practice in MM. This presentation is dedicated to Dr. Minoru Toyota.

O5.2 - The origin of colorectal cancer with CpG island methylator phenotype H. Suzuki1,2, E.Yamamoto1,2, M. Nojima3, T.Shimizu1, Y. Shinomura2, K. Imai4, M. Toyota1 1Department of Molecular Biology, Sapporo Medical University, Japan Background: We have previously reported that a subset of CRCs show genome wide DNA methylation defects, which we termed CpG island methylator phenotype (CIMP). Subsequently, we and others classified CRCs into three subgroups, CIMP1, CIMP2 and CIMP-negative (CIMP-N). The CIMP1 tumors are characterized by frequent BRAF mutations and infrequent p53 mutations, whereas CIMP2 and CIMP-N tumors show frequent KRAS and p53 mutations, respectively. Aims: We aimed to understand the molecular mechanism of the CIMP cancer development. Methods: We analyzed 252 precancerous lesions and 78 advanced CRC specimens. Genome-wide analysis was carried out using Methylated CpG island Amplification (MCA)-microarray and array-based comparative genomic hybridization (array CGH). Gene methylation and mutation were analyzed by pyrosequencing. Results: MCA-microarray revealed that precancerous lesions could be categorized into 3 clusters according to their methylation status (M-cluster-1, -2 and -3, which are corresponding to CIMP1, CIMP2, and CIMP-N). Lesions in M-cluster-1 exhibited frequent BRAF or KRAS mutations, and sessile serrated adenomas are significantly enriched in this group. The M-cluster-2 was highly enriched with tubulovillous adenomas with frequent KRAS mutations, whereas the M-cluster-3 consisted of tubular adenomas and hyperplastic polyps in which BRAF or KRAS mutations were less prevalent. Array CGH analysis suggested that chromosomal aberrations are acquired during the development of CIMP2 and CIMP-N cancers. Conclusions: Our data suggested that early precancerous lesions exhibit similar epigenetic alterations as advanced

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

CRCs, and that CRCs develop through at least three pathways which exhibit distinct genetic and epigenetic signatures.

O5.3 - How to handle proteome complexity in search for biomarkers in biological fluids

S23

blindfolded, searching for a treasure without any prior knowledge of what to look for (call it a treasure map). Guidelines will be given on the proper protocols that could lead to a successful biomarker discovery.

O5.4 - Systematic quantification of peptides/proteins in urine using a novel mass spectrometry based approach

P.G. Righetti B. Domon Politecnico di Milano, Department of Chemistry, Materials and Chemical REVIEW The large proteome complexity (10 to 12 orders of magnitude in human sera) calls for pre-fractionation in order to have access to low-abundance species (LAP), where one might find those precious biomarkers of diseases. There are at least twomain approaches to this problem: a) immuno-subtraction (depletion) of the 6 to 20 highestabundance proteins (in sera and other biological fluids); b) enrichment approaches such as combinatorial peptide ligand libraries (CPLL). I will focus on enrichment with CPLLs, since the first approach has been recently heavily criticized on the grounds that it will only give access to medium to high abundance proteins. On the contrary, CPLLs have given access to genuine LAPs, having concentrations below 1 μg/L, a level rarely seen with common tools. For instance, in exploring the cytoplasmic proteome of human red blood cells, 1578 unique gene products could be identified, vs. barley 252 in the best report not using CPLLs. In cerebrospinal fluid, >1200 unique species could be listed, with a very high score, via prior CPLL treatment. In the urinary proteome, via a novel CPLL elution protocol, ca. 3200 protein spots could be separated and visualized in a silver-stained 2D map. In human sera, via a single CPLL sweep, >3800 unique gene products could be detected at once. Even in human saliva, which could also be a reservoir of biomarkers, >2000 unique species could be identified via CPLLs. These examples clearly indicate that the CPLL methodology has the unique power of bringing to the limelight those very rare proteins in biological fluids which could be potential markers of just about any possible disease. Yet, before singing the paean of this methodology, one should be aware of the fact that the last ten years of research on biomarkers have been a failure, simply because just about all labs went to dig into sera

Luxembourg Clinical Proteomics Center, CRP-Sant ©, Strassen, Luxembourg REVIEW Proteomics aims at generating comprehensive and quantitative data sets that will help modeling and understanding complex biological processes. Even with today’s technology, it remains very challenging as two major hurdles have to be overcome: i) the vast sample complexity due to a large number of proteins and their structural diversity, and ii) the wide dynamic range of protein concentrations spanning up to ten orders of magnitude. In spite of the recent technology developments, alternative strategies are still needed to provide sufficient penetration into a proteome. The benefits and pitfalls of current strategies will be presented. LC-MS based proteomics approaches are routinely used to detect, quantify, and identify proteins in complex samples. It is very effective in identifying biomarker candidates isolated from tissues, bodily fluids. Selected reaction monitoring performed on triple quadrupole instruments is a reference method to precisely analyze biomarkers in complex biological samples such urine. The technique, offering unique selectivity and sensitivity, can leverage candidates from discovery proteomics experiments as well as knowledge from derived from transcriptomics data sets or literature mining. This approach can be used in conjunction with isotope dilution (addition of isotopically labeled peptides) to the samples to ensure precise quantification, and use chromatographic (elution time) and mass spectrometric (ion intensities) properties to corroborate the identity of the analytes. Alternatively, a new technique leveraging the highresolution / accurate mass capabilities of a quadrupole / orbitrap instrument will be presented and discussed. This high resolution and the fast acquisition capabilities of this novel instrument offers unique selectivity to reliably

S24

quantify and confirm the identity large sets of biomarkers in clinical samples such as urine.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

O5.6 - Next generation methods to study DNAmethylation T. De Meyer, G. Trooskens, W. Van Criekinge

O5.5 - Proteomic analyses of cancer cells and body fluids in bladder cancer M. Sánchez- Carbayo Tumor Markers Group, Spanish National Cancer Research Center (CNIO) Proteomic approaches are accelerating the discovery of the molecular alterations involved in tumorigenesis and bladder cancer progression by the combination of analyses characterizing cultured cancer cells in vitro and also human clinical material both using tissue and body fluids. Such alterations may represent critical molecular events that could be developed into potential targets for therapeutic interventions and also be translated into cancer biomarkers. Aims: To develop potential targets for therapeutic interventions and also to translate them into cancer biomarkers. Methods: The presentations is based on review of the literature Quantitative proteomics is strongly contributing to the discovery of candidate disease-specific target and biomarkers. While protein and antibody arrays permit differential quantification of known proteins, mass spectrometry techniques are leading for protein identification. The potential for expanding our understanding of bladder carcinogenesis can be explored in cancer cells in vitro and compared to the molecular alterations found in tumor specimens. Furthermore, the application and optimization of high-throughput proteomic technologies in non-invasive specimens such as blood and urine samples would contribute to translate proteomic alterations characteristic of bladder tumors into potential biomarkers that could be applied to the clinics. Several proteomic strategies to identify critical pathways and candidates involved in bladder cancer initiation, progression and metastasis will be discussed such as iTRAQ and SILAC approaches. With the aim of diagnosing and predicting clinical behavior for patients with bladder cancer, proteomic technologies will be described using clinical material. Discussed tissue strategies will include tissue and reverse phase arrays, serum approaches will refer to antibody and protein arrays, while DIGE and MS technologies applications will be described for urine samples.

Ghent University; Belgium Background: Over the last decade, DNA-methylation research has shifted from a gene-based approach to genome-wide analyses, and a novel generation of methodologies enables researchers to pursue this in a highthroughput manner. Aims: The pros and cons of the most important methodologies will be discussed. Methods: This comparison will be based on the general data characteristics and real data examples. Results: The conversion of cytosines by a bisulfite chemical treatment results in a methylation status dependent DNA-sequence. In a methylation specific PCR (MSP), methylation status specific primers are used to assess the methylation degree of a very specific locus. However, genome-scale methodologies need a higher throughput and can for example be obtained by second generation whole genome sequencing of bisulfite treated samples. This approach is currently the gold standard, but in most cases still too expensive for a biomarker discovery approach. A typical solution is to restrict the analysis to a subset of genomic regions of interest, or to avoid sequencing at all by opting for a microarray based approach (e.g. Illumina Infinium BeadChips). While all of these "genome-scale" methods cannot be considered as truly genome-wide, the huge coverage of the BeadChip and low cost make this one of the best alternatives to genome-wide bisulfite sequencing. A completely different approach captures methylated DNAfragments in a pool of sonicated DNA, either by antibody or by a methyl-binding domain (MBD), followed by second-generation sequencing (or also the use of microarrays). In contrast to bisulfite sequencing, the enrichment based sequencing methods do not offer single-nucleotide resolution, although it should be noted that DNAmethylation does not function at this resolution itself. However, one major disadvantage of these methods remains the fact that the sequencing coverage depends on sequence features, although this is not necessarily a problem for biomarker discovery. These methods are the best alternative to whole genome bisulfite sequencing if one aims at a cost-efficient, truly genome-wide approac

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

Conclusions: Of the different next generation methods, DNA-methylation specific microarrays and MBD-based enrichment sequencing are good and alternative alternatives to whole genome bisulfite sequencing, at least until the latter can be performed on many samples at a reasonable cost.

O5.7 - Cancer epigenetics: tumor biology and clinical applications

S25

combination of genetic and epigenetic inactivation. While it is clear that both genetics and epigenetics combine to altering gene expression and function, a debate persists as to which type of alteration is dominant. The reversible nature of these alterations and the possibility of detection in cancer cells provide opportunities for use of this information in the clinic. Here I will provide an overview of epigenetic alterations in cancer and discuss how they will contribute improving control of CRC in the future.

M. van Engeland Dept. of Pathology GROW-School for Oncology & Developmental Biology Maastricht University Medical Center Maastricht The Netherlands Background: Review Results: Cancer is characterized by accumulating genetic alterations. This is complemented by the recent understanding of epigenetic alterations (heritable changes in gene expression that are not accompanied by changes in DNA sequence) in CRC. It is increasingly clear that epigenetic alterations in cancer affect every component of gene regulation, including DNA methylation, loss of imprinting (loss of monoallelic gene regulation conferred by parent-of-origin-specific DNA methylation), posttranslational histone modifications, nucleosome positioning, chromatin looping and small non-coding RNAs. The most extensively characterized epigenetic alteration in CRC is promoter hypermethylation, which occurs at CpG dinucleotide-dense regions, called CpG islands, present at the 5™ region of approximately 60% of genes. Hypermethylation of promoter CpG islands has been observed for numerous tumor suppressor- and DNA repair genes and functions equivalently to coding-region mutations or deletion, therefore considered the ™third hit™™ in the Knudson model for inactivation of tumor suppressor genes. Promoter CpG island methylation for some genes is related to aging, but cancer-specific hypermethylation events are frequent as well and affect WNT-, RTK-, NOTCH-, TP53-, PI3K-, retinoic acid-, and IGF signaling as well as other pathways regulating cell cycle regulation, transcription regulation, DNA repair/ stability, apoptosis, adhesion angiogenesis, invasion and metastasis and chromatin organization. Thus, epigenetic alterations target many of the same pathways targeted for mutational events. Some genes are exclusively altered by epigenetic inactivation, others targeted by genetic events and a subgroup of genes have a

O5.8 - Emergent biomolecular markers for improving preoperative diagnosis of thyroid tumors 1A. Carpi, 2A. Nicolini 1 Department of Reproduction and Ageing, University of Pisa, Pisa, Italy Background: Microarray analysis can generate gene expression profiles and multiple protein expression analyses of thyroid tumors. The latter includes proteomics which describes the global analysis of the proteome. Microarrays were used to charcterize thyroid tumor develoment and to improve diagnostics including preoperative evaluations. Many biomolecular markers have been evaluated on preoperative thyroid nodule tissue. However, important obstacles have limited clinical use of these emerging biotechnologies. Aims: To develop a classification system of these technologies, to analyze the above mentioned obstacles, to show that the use of a new preoperative tissue substrate derived from the percutaneous large needle aspiration biopsy (LNAB) can eliminate an important number of the obstacles. Methods: The presentation will be a review of the literature Results: Biomolecular markers can be classified according to their biological-biochemical types, according to differentiation or dedifferentiation status, and according to the molecular or biological substrates involved or according to the method of identification. The obstacles to the use of emerging biomolecular markers are related to the pathology of the thyroid nodule or to the specific techniques used. Conclusions: The tissue substrate derived from LNAB is very suitable for biomolecular marker determination, in particular for galectin-3 which importantly improves preoperative diagnostic accuracy of thyroid nodules.

S26

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

O6 MELANOMA (Skin and Uveal)

O6.2 - Prediction and early detection of metastasis in uveal melanoma by biological markers

O6.1 - Melanoma Immunotherapy “ from localized immune modulation to melanoma vaccines

V. Barak1, S. Frenkel2, I. Kalickman1 and J. Pe’er2

S. Merims, Lotem M, Ospovat I, Frankenburg S, Machlenkin A, Hamburger T, Peretz T Sharett Institute of Oncology, Hadassah Medical Center, Jerusalem, Israel Background: The year 2010 marks a new hallmark in the history of cancer vaccines: for the first time, the American FDA recognized the benefit and approved a vaccine against cancer (Dendreons Provenge vaccine against metastatic prostate cancer). An additional breakthrough in cancer immunotherapy was further introduced with the approval of anti-CTLA4, Ipilimumab (Yervoy, Bristol-Myers Squibb) for the treatment of advanced melanoma patients. Immunotherapy is an important modality in the therapy of patients with malignant melanoma and melanoma vaccines aim to increase specific cellular and humoral responses (T and B-cell responses), enhance dendritic cells’ capacity to present antigens to T cells, thus promoting immune response, and promote resistance to local immunosuppressive factors secreted in melanoma. Aims: This overview will discuss the current status of immunotherapy for melanoma, with a focus on our institutions experience in cellular vaccines. Methods: Since 1996 our group has been involved in the development of melanoma cell vaccines. We have conducted multiple clinical trials both in the adjuvant and metastatic setting including immune-monitoring of more than 100 melanoma patients. Results: We have shown that an autologous vaccine, composed of self-melanoma tumor cells, elicits both CD4 and CD8 T cell responses that correlate with improved patient survival. In the absence of self-tumor cell lines, a group of patients was vaccinated with allogeneic (non-self) tumor cells expressing at least one MHC allele identical to the patients own. Intriguingly, patients with an HLA-B35 phenotype, representing approximately one third of the patients, had significantly better overall and disease free survival than the rest of the patients. Conclusions: Development of novel immunotherapeutic approaches, along with optimization of existing melanoma vaccines, continues to hold a great promise in the field of melanoma immune-therapy and new breakthroughs will evolve from our understanding of the complexity of immune responses in this disease.

1Immunol Lab for Tumor Diagnosis, Oncology and 2Ophthalmology Departments, 1Hadassah -Hebrew University Medical Center, Jerusalem, Israel Background: We have previously shown that serum biomarkers as Osteopontin (OPN), S-100, MelanomaInhibitory Activity (MIA), and Tissue Polypeptide-Specific Antigen (TPS) are significantly elevated in patients (pts) with metastatic Uveal Melanoma, as compared to disease-free (DF) pts. Aims: 1. To examine changes in biomarkers for assessing treatment’s utility of enucleation or brachy-therapy in Uveal Melanoma pts. 2. To evaluate the prediction of Liver Metastasis in those patients, by significant increases in biomarker levels. Methods: Levels of serum biomarkers were analyzed for 93 Uveal Melanoma pts. and 95 Controls.A matchedpairs analysis was used to compare baseline (pretreatment) marker levels with those measured at 1, 4, and 10 months, post treatment. Differences in biomarker levels were analyzed for the entire group, or for each treatment group separately and correlated to metastasis vs DF status of pts. Results: Of the 93 pts, 80 underwent brachy-therapy and 13 were enucleated. One month following treatment S-100, MIA, and TPS levels did not change, while OPN still increased (11.51 to 13.11 ng/ml, p=0.0430). Four months after treatment there were no changes in marker levels from baseline, except for a significant decrease in S-100 levels in the enucleated pts (0.09 to 0.06 ¼g/l, p=0.0492). S-100 levels decreased in the brachy-therapy pts group (0.11 to 0.08 ¼g/l, p=0.0251) only after 10 months. OPN, MIA TPS and S −100 significant (p=0.002) increasing levels, were predictive of metastases development. VEGF was found elevated in Metastasis, with a wide heterogeneity, performing less than other markers. Conclusions: Successful local treatment (enucleation) of primary Uveal Melanoma and brachy-therapy resulted in significant decreases, mostly in S-100 and in OPN levels, providing a good parameter for assessing treatment utility. Significant increases (5–25 fold) in all markers, predicted and detected earlier than CT or liver enzymes, the development of Liver Metastases in Uveal Melanoma pts-enabling earlier treatment and a better survival of pts, as demonstrated recently.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

O6.3 - Novel biomarkers for metastatic melanoma 1,2Ortenberg R, 1Sapoznik S, 3Faranesh S, 1 Ben-Ami E, 1,2Besser MJ, 3Peretz T, 3Lotem M, 1Schachter J, 1,2Markel G 1Department of Clinical Microbiology and Immunology, Sackler Faculty of Medicine, Tel Aviv University, Israel; 2Ella Institute of Melanoma, Sheba Medical Center, Ramat Gan, Israel; 3Sharet Institute of Oncology, Hadassah Medical Organization, Jerusalem, I Background: One of the main challenges in management of melanoma patients is the lack of specific biomarkers. The biomarkers used today are LDH and S100, which are non-specific or related to the mechanism of disease. We have previously shown that CEACAM1 protects melanoma cells from immune-mediated elimination. Here we show that serum CEACAM1 (sCEACAM1) can serve as a novel melanoma biomarker Aims: To test the potential of sCEACAM1 as new biomarker for melanoma monitoring, prognosis and as a medical decision making tool Methods: sCEACAM1 is quantified with a custom developed ELISA. Production of sCEACAM1 by melanoma cells was studied in vitro by using various melanoma cell lines and primary melanoma cultures. The potential of sCEACAM1 as a marker for tumor burden was studied in the serum of mice implanted with human melanoma cells, following surgical removal or sham surgery. sCEACAM1 in humans was tested in 40 healthy donors and >100 melanoma patients. In addition, sCEACAM1 was tested in melanoma patients before and after treatment either with adoptive cell therapy (30 patients) or with autologous whole cell vaccination (30 patients). Results: In vitro, we show that soluble CEACAM1 is produced by melanoma cells. This depended on de novo protein synthesis and intra-cellular transport, and was independent of metalloproteinase-mediated shedding. In the melanoma xenograft model, serum sCEACAM1 levels increased with tumor growth, dropped sharply following surgical removal and re-emerged only in cases of relapse. In melanoma patients, serum sCEACAM1 levels were high only in metastatic patients with evidence of disease. Furthermore, sCEACAM1 levels inversely correlated with prognosis in stage IV patients, enabled monitoring of response to immunotherapy and predicted response to TIL-mediated immunotherapy. Conclusions: sCEACAM1 reflects disease burden and provides prognostic and predictive value, which could guide decision making in melanoma. More clinical studies are required to validate the role of CEACAM1 as a new biomarker for melanoma.

S27

O7 PATHOLOGY & IMAGING O7.1 - The role of the c-KIT receptor in malignant transformation of thyroid epithelium: clinical implications C. Mazzanti2, S. Tomei2, I. Marchetti, L.Rossi 4, K. Zagaglia 2, F. Lessi 2, A. Apollo2, P. Aretini2, G. Di Coscio3, G.Bevilacqua1 Division of Surgical, Molecular, and Ultrastructural Pathology1, Section of Molecular Pathology 2, Section of Cytopathology 3, Department of Human Morphology and Applied Biology 4, University of Pisa and Pisa University Hospital, Pisa Background: Thyroid carcinomas represent 90% of all endocrine malignancies, typically as papillary thyroid carcinomas(PTC).Fine needle aspiration is the best way to evaluate a thyroid nodule, however the diagnosis is not always simple. Aims: Molecular pathology has been proven to be a useful diagnostic tool in different malignacies, and recently several investigations have analyzed cytological and histological thyroid samples for the BRAF gene V600E mutation, which has been demonstrated to be a novel diagnostic and prognostic biomarker in thyroid cancer.However by using only the BRAF V600E analysis there will still be a certain percentage of malignant tumors that would remain unclassified, so it is crucial to find other specific tumour markers.The present study evaluates c-KIT mRNA expression level and gene status in FNA smears, with the aim to verify its potential role as a diagnostic marker in order to improve the actual cytological diagnostics of thyroid nodules. Methods: In this study we extracted RNA from 82 FNA smears, 46 malignant and 36 benign according to histology in order to evaluate by Real Time PCR the expression levels of the c-KIT gene. Moreover to explore the diagnostic utility of c-KIT expression, we divided the gene expression values in four arbitrarily defined classes of c-KIT expression values and we calculated a malignacy and benignancy index. Results: We have found an obvious decrease rather than increase in transcripts of c-KIT in malignant thyroid lesions compared to the benign ones. The molecular analysis was proven by ROC analysis to be highly specific and sensitive, improving the cytological diagnostic accuracy. Conclusions: In Conclusions c-KIT expression could be useful as an additional diagnostic marker on FNA thyroid samples. The use of the c-KIT expression value

S28

could be used as an indicator of malignancy/benignancy in those samples with an undetermined cytological diagnosis, in which the BRAF V600E is absent and not informative.

O7.2 - Proposed algorithm for effective receptor assessment on core needle biopsy S.C. Seferina,1 MD., M. Nap, MD.2, PhD, F. van den Berkmortel, MD, PhD 3, J. Wals, MD.3, 1Division of Medical Oncology, Department of Internal Medicine, GROW “ School for Oncology and Developmental Biology, Maastricht University Medical Centre, Maastricht; 2 Department of Pathology, Atrium Medical Centre Parkstad, Heerlen; 3Department of Inte Background: Neo-adjuvant treatment in breast cancer patients brought a new challenge in the assessment of estrogen (ER), progesterone (PR) and human epidermal growth factor receptor 2 (HER2).Hormonal therapy and HER2 targeted therapy given in the neo-adjuvant setting fully depends on the results of these parameters on core needle biopsy (CNB). When CNB is compared to surgical specimen (SS), a number of false positive(FP) and false negative(FN) test results are observed.In order to gain more insight in the risk and consequences following these discrepancies,receptor status on both CNB and SS were compared Aims: Assess and compare the CNB and the SS on ER,PR and HER2,and develop a cost-effective algorithm for assessment of receptor status based on CNB Methods: We retrospectively collected data from a database containing records of patients with a new diagnosis of breast cancer in Atrium Medical Centre Parkstad,a large teaching hospital within the Netherlands.We assessed the hormone and HER2 receptor status in both the CNB and SS resulting in a total of 513 breast cancer patients.The SS was taken as the gold standard. Results: The IHC ER assessment had a FN result in 27.7%; 95% CI (19.2-38.0)of the cases and a FP result in 5.9%; 95% CI(3.9-8.7);PR had a FN result in 29.8%;95% CI (23.6-37.0)of the cases and a FP result in 9.9%;95% CI(6.913.9).In the HER2 IHC results, we found a concordance in

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

54.3% in scores 0-1+,47.9% in the 2+ group and,a concordance of 75.3% in the 3+ group.Analyses regarding in situ hybridization-data are ongoing. Conclusions: In patients with a negative ER/PR on CNB, a re-assessment on the SS may be considered if available, in relation to the relatively large amount of FN test results. The opposite is seen in the IHC HER2 assessment; where more FP are observed.In a cost-effective approach, IHC assessment on CNB could be limited to those patients eligible for neo-adjuvant treatment whereas in all other cases CNB and SS can be pooled.

O7.3 - Hypothesis of a viral etiology of Breast Cancer G. Bevilacqua Division of Surgical, Molecular, and Ultrastructural Pathology, University of Pisa Background: Notwithstanding the fact that breast carcinoma represents the most frequent cancer in women and that it has been largely studied all over the world for many decades, its etiology is largely unknown. In mouse the etiological role of the Murine Mammary Tumor Virus (MMTV) in the development of tumors of the mammary gland has been demonstrated for a long time. The strong similarities between the human and the murine disease led to a halfcentury quest for a possible viral etiology of breast carcinoma in women. MMTV viral antigens were found in human breast tumors, MMTV particles were described in human cells and milk and MMTV sequences were found in humans. These data were not convincing for a long time but today a possible viral etiology is receiving a new consideration, mainly because the demonstration of the causative role of HPV for carcinoma of the cervix has opened this new frontier. Aims: If MMTVels is involved in human breast carcinogenesis, it could play an early, an intermediate or a late role. In other words, it could behave as a transforming agent, an etiological cofactor, and/or a pathogenetic agent. Establishing when exogenous viral sequences appear during the natural history of the disease would represent an interesting

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

contribution to answer the previous question. This paper investigates for the first time the presence of MMTV envelope gene-like exogenous sequences in preinvasive lesions, as well as in infiltrating carcinomas and in normal tissues. Methods: A Leica ASLMD automatic laser microdissector was used to select the epithelial cell population to be studied. Stromal and inflammatory cells were carefully excluded. DNA extraction and PCR was performed on microdissected cells. CISH esperiments were performed to confirm the presence of the viral sequence in the nuclei of breast cancer cells. Results: Results demonstrate a very early appearance of MMTV env-like sequences in human breast carcinoma progression. For the first time we report the presence of the MMTV env-like viral sequence by CISH experiments, which detected viral sequence hybridization signals in the nuclei breast cancer cells. Conclusions: In conclusion, exogenous MMTV sequences are easily identifiable in the early steps of breast cancer progression, suggesting their possible role in the transformation process, at least as an etiological cofactor.

O7.4 - F-18 choline imaging in relation to other imaging procedures and in vitro PSA levels in prostate cancer P. Oehr Dpt. Nuclear Medicine, University Bonn, Germany Background: Choline is a compound which is incorporated into cellular membranes. Because cancer cells duplicate very quickly, the biosynthesis of membranes is also very rapid and associated with increased uptake of choline. These observations have led to the use of choline PET or PET/CT for imaging of malignancies, especially for detection of prostate cancer. Aims: To describe current value of choline imaging in relation to other imaging procedures and the in vitro determinations of PSA levels in prostate cancer. Methods: This is a review of the current literature. Results: F-18 FDG glucose uptake has very little accuracy for diagnosis or staging of this malignancy. Increased uptake of acetate or choline is an alternative

S29

approach, and evidence is coming from imaging studies with choline. C-11 choline and F-18 choline have been successfully applied to prostate cancer for staging of both primary and recurrent disease. Choline PET or choline PET/CT combined with MRT is an indication to improve the diagnostic specificity for intra- or extraprostatic localizations of primary prostate cancer in a multimodal approach, giving additional information regarding the localization of the main primary prostate cancer lesions. Localization of prostate cancer in selected patients with increased risk of having the disease in view of having persistently elevated levels of PSA and negative prostate needle biopsy may become another indication for choline PET. The most common sites of metastases from prostate cancer are lymph nodes and bone. Concerning initial N-staging, choline PET shows better results than the presently recommended conventional procedure TRUS. For detection of bone metastases, both choline PET/CT and F-18 PET/CT have a better sensitivity and specificity than bone scan and can detect metastatic disease in patients with PSA levels of >2 ¼g/l, whereas the conventional procedure bone scan may reliably detect bone metastases at PSA levels of >10 ¼g/l. This has an impact on PSA follow-up and treatment decisions. Concerning other imaging approaches in prostate cancer, 11C-methionine, 18FFDHT and anti-18F-FACBC are promising but remain to be elucidated further. Conclusions: The data support that not all established standard methods may still be appropriate and the introduction of choline-PET-CT is recommended, combined with MRI.

O7.5 - Data vizualisation linking information across different hospital departments Nap Marius and Van der Linden Eric-Jan Atrium MC Heerlen, NL and Magnaview, Eindhoven, NL Background: Patient contacts generate hugh amounts of data, often in such dimensions that a rapid overview and close examination of relations is impossible.

S30

Aims: To investigate the possibility to combine data from the department of Pathology and Clinical Chemistry and to generate visual representations carrying meaningfull information. Methods: 900 cases of colon and rectum carcinoma operated from 2005 to 2009 were retrieved from the department of pathology. Sex, age, date of surgery and PIDrecord were collected. Based on the PID all available Clinical Chemistry data from 2000 to 2009 were retrieved. Each record consisted of a test name, test result, date of analysis and PID. All together 1 million records were combined with the PID as the linking key. Methods: Using Magnaview, visual representations were generated which were inspected and improved by repeated clustering of testnames in relation to the related organ status. For comparisons of abberations on a background of multidimensional results, we related every single test result to the mean value of its peers and allowed filtering of data based on their magnitude of difference from 2SD of their peers. Results: Integrating the data from two different departments allowed addition of classifications and enabled display in relation to pre- and postoperative periods. The interactive display of a broad variety of tests allows the visual manipulation of a million datapoints with one mouse click and supports the generation of new hypothesis, quickly testing their plausibility before going to energy absorbing statistical work. Conclusions: Further improvement of this approach will unlock very large datasets for inspection and finally improve our understanding of health care consumption.

O7.6 - Analysis of free beta subunit of hCG secretion by common epithelial cancer cell lines B. Burczynska1,2, S. Butler1, X. Wen1, A. Shiled1, R. K. Iles1,3. 1Center for Investigative and Diagnostic Oncology, Middlesex University, UK 2Dept of Cell Biology, Poznan University of Medical Sciences, Poland 3ELK Foundation for Health Research, Oxford University, UK

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

Background: Ectopic production of hCGβ is a known feature of many common epithelial cancers although its use diagnostically is married by poor consistency in detection. It has been suggested that growth hormone-like action of hCGβ in cancer results in the inhibition of apoptosis and is related to the ability of the protein to homodimerise and subsequently cross-talk with known cystine knot growth factor receptors. Our previous study showed the presence of ‘trophoblastic’ CGB3,5,7,8 mRNA in all selected non-trophoblastic cancer cell lines but only those cells which express CGB1/2 showed the ability to secrete hCGβ into the culture media. Aims: To examine if cancer cell lines actively transcribing genes CGB1/2 produce hCGβ. To investigate the possible structure/function of secreted protein products of CGB genes. Methods: A variety of cancer cell lines were immunofluorescently stained with anti-free beta hCG antibody and co-localized with calnexin and giantin. The cell culture media were analysed on MALDI-TOF and compared with non-cancer cells (TCL-1), recombinant hCGβ (Sigma) and purified hCGβ from pregnant women urine (Abcam). Results: Many cancer cell lines which failed to secrete hCGβ into their culture media were positive in immunostaining; these were found previously to express CGB 3,5,7,8 but not CGB1/2 genes. Confocal microscopy found that hCGβ was located in the Endoplasmatic R. Conversely cell lines like ScaBER, that secrete hCGβ into the media, were less immunohistochemically positive and expressed high level of CGB1/2. hCGβ secreted into culture cell medium was detected by MALDI-TOF MS. The predominant protein form of rhCGβ and TCL-1 cell line culture media was a glycosylated monomer (m/z ~23 kDa); in contrast cancer cells secreting hCGβ into the media produce predominantly homodimeric form of hCGβ (m/z ~46 kDa). Conclusions: Transcription of CGB genes may not necessarily lead to the translation of mature active protein and neither may it be secreted from the cells. The impact of this diagnostically means that the method selection of hCGβ detection is critical to provide the best prognostic information regarding patient status.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

O8 LUNG Cancer O8.1 - Autoantibodies in lung cancer and their use in early cancer diagnosis C Chapman, G Healey, A Murray, L Peek, J Robertson The University of Nottingham, UK; Oncimmune Ltd, UK Background: The EarlyCDT-Lung test, which measures circulating autoantibodies to tumor-associated autoantibodies (AAbs), has been reported to identify approximately 40% of lung cancers (LCa) at all stages of disease. Aims: Improvements in the specificity of an AAb test for LCa would improve its cost effectiveness, and decrease the number of individuals who would suffer the anxiety associated with a false positive test result. Methods: Samples from 235 patients with newly diagnosed LCa and gender, age and smoking matched individuals with no history of cancer (Optimization Set) were measured for the presence of AAbs to 6 established antigens (p53, NY-ESO-1, CAGE, GBU4-5, Annexin I and SOX2) as well as 2 other lung cancer associated antigens MAGE A4 and HuD. The sensitivity and specificity of the original panel of 6 AAb assays was compared with a new panel of 7 AAb assays (p53, NY-ESO-1, CAGE, GBU4-5, SOX2, HuD and MAGE A4). The sensitivity and specificity of the 6 and the 7 AAb panels were then compared in an independent clinical sample set. This comprising 2 consecutive series of 774 and 689 individuals, deemed by their physician to be at increased risk of developing LCa, and who had also been sent for a CT scan due to these concerns. Results: The panel of 6 AAb assays gave a sensitivity and specificity of 39%/89%, while the modified panel of 7 AAb assays resulted in a sensitivity and specificity of 41%/91% in the optimization set which, once adjusted for occult cancers in the high risk population, resulted in an overall panel specificity of 93%. For a LCa incidence in a high risk population of 2.4% this improvement would result in a positive

S31

predictive value of 1 in 8 and an overall accuracy of the test of 91.6%. Analysis of the 2 different panels in the clinical sample set confirmed that the 7 antigen panel maintained its sensitivity, and resulted in a significant increase in the specificity of the test from 82% to 91%, in this high risk population. Conclusions: The EarlyCDT-Lung test can be used as part of the armamentarium of tests, available to the clinician, to aid the diagnosis of early stage lung cancer. The change from a 6 to a 7 AAb assay improved the overall specificity of the test, increasing its clinical usefulness.

O8.2 - Nitric oxide modulates HIF-1α expression in human lung cancer E Gouvas1,2,3, C Amegashie4, KM Elseth1,2,3, BJ Vesper1,2,3, GK Haines III5, G Tarjan6,7, JA Radosevich1,2,3 1Center for Molecular Biology of Oral Diseases, College of Dentistry, University of Illinois at Chicago, Chicago, IL, USA; 2Dept. of Oral Medicine and Diagnostic Sciences, College of Dentistry, University of Illinois at Chicago, Chicago, IL, USA; 3Jesse B Background: The HIF-1α gene plays an important role in regulating oxygen homeostasis. Tumors overexpressing HIF-1α typically respond poorly to radiation. Nitric oxide (NO) plays an important role in tumor progression. Patients with tumors expressing elevated NO levels portend a worse clinical outcome compared to those expressing low NO levels. We have created a model system of tumor cells adapted to high levels of NO (HNO). Aims: We predict that tumors in HNO microenvironments will grow better under hypoxic conditions than tumors in low level NO microenvironments and expect that HNOadapted cells will exhibit high expression of HIF-1α in hypoxic conditions. Methods: A549, a human lung adenocarcinoma cell line was used. A549-HNO and A549-Parent cells were tested under hypoxic conditions by exposing the cells

S32

to varying concentrations of cobalt chloride. CoCl2 treatment mimics hypoxic conditions. MTT and migration assays were performed at 24-hour intervals. HIF-1α expression was measured using immunohistochemistry on cytospins. Gene chip analysis was conducted for both cell lines. Results: A549-HNO cells treated with CoCl2 grew and migrated faster than parent cells at a range of concentrations (0.05-1 mM) in a time-dependent manner (24 and 48 hrs). Immunocytospins confirmed higher HIF-1α expression in the A549-HNO cells. Gene chip analysis indicated a significant increase in HIF-1α expression in A549-HNO relative to A549-Parent. Conclusions: Conclusions: These results suggest tumors growing in microenvironments containing HNO levels will thrive better in hypoxic conditions than tumors growing in low NO environments. The HIF-1α gene is a potential biomarker for aggressiveness in tumors.

O8.3 - Serum biomarkers in the therapy monitoring of lung cancer S. Holdenrieder, D Nagel, J v Pawel, P Stieber University Hospital Munich; Asklepios Lung Hospital Gauting; University Hospital Bonn, Ge Background: Although cancer biomarkers are widely used in therapy monitoring of lung cancer patients, there are only rare systematic evaluations of their clinical significance. Aims: We analyzed lung biomarker kinetics in patients with non-small cell (NSCLC) and small cell lung cancer (SCLC) on their power to specifically identify patients with poor and good response to systemic chemotherapy. Methods: Lung biomarkers CYFRA 21-1, CEA, NSE and ProGRP were evaluated in prospectively collected serum samples of 311 patients with advanced NSCLC with 1st-line and 161 with 2nd-line chemotherapy, and 128 patients with SCLC (1st-line chemotherapy). Absolute marker levels and relative changes at time of staging after two treatment cycles were correlated with concurrent radiological response to therapy.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

Results: Absolute levels and kinetics of CYFRA 21-1 and CEA at staging discriminated significantly between non-progressive and progressive patients in NSCLC (1st-/2nd-line therapies). Best ROC curves for identification of progression were obtained for CYFRA 21-1 levels (AUC 81.3% and 79.6%). Because several patients with tumor response had increasing values, sensitivity for detection of progression was only 31% and 22% (CYFRA 21-1; 1st/2nd-line), and 17% and 12% (CEA; 1st-/2nd-line) at 90%-specificity for nonprogression. For detection of remission, sensitivity was 29% and 55% (CYFRA 21-1; 1st/2nd-line), and 31% and 21% (CEA; 1st-/2nd-line) at 90%-specificity for non-remission. In SCLC, ProGRP, NSE and CYFRA 21-1 levels at staging discriminated well between response groups while kinetics were less meaningful. Best ROC curves for non-response were obtained for the combinations of NSE and CYFRA 21-1 (AUC 91.3%) and ProGRP and CYFRA 21-1 (88.3%). Concerning kinetics, sensitivity for detection of non-remission at 90%specificity was 20%, 40%, 10%, and 30% for ProGRP, NSE, CYFRA 21-1 and CEA kinetics; for detection of remission, sensitivity was 49%, 13%, 19% and 39% for ProGRP, NSE, CYFRA 21-1 and CEA kinetics, respectively. Conclusions: Lung cancer biomarker levels are well suited to estimate the therapy response in NSCLC and SCLC patients. Kinetic interpretation is particularly useful for response detection while increases in some well-responding patients limit the specific detection of progression.

O8.4 - Improved lung cancer diagnosis by the combination of tumor markers Molina R(1), Leon A(2), Trape J(3), Garrido A(4), Auge JM(1), Filella X(1), de la Torre P(4), Escudero JM(1), Blavia R(3), Perich D(4). (1)Hospital Clinic, Barcelona, Spain; (2) Hospital Virgen del Rocio, Sevilla, Spain; (3) Hospital Althaia, Barcelona, Spain; (4) Hospital Donostia, San Sebastian, Spain Background: In previous studies, our group reported a tumor marker algorithm for the differential diagnosis and to aid in the histological typing of lung cancer.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

Aims: The aims of this study were: 1) To apply this algorithm to other populations of patients with the suspicious signs of lung cancer and to clarify the clinical applications. 2) To elaborate a new algorithm to subdivide NSCLC 3) To create an informatics program to calculate cancer risk and the histological subtype. Methods: CEA, CYFRA 21-1, SCC, CA15.3, ProGRP and NSE were prospectively studied in 1858 patients with suspicious signs of lung cancer from 4 different Spanish hospitals. We considered as suspicious for malignancy when tumor markers were higher than 5 ng/mL, 3.3 ng/mL, 2 ng/mL, 30 UI/mL, 50 pg/mL and 25 ng/mL for CEA, CYFRA 21-1, SCC, CA15.3, ProGRP and NSE, respectively. Results: Tumor marker sensitivity for lung cancer in the diagnosis of symptomatic patients was similar to that previously reported: 82% in NSCLC and 85% in SCLC and the specificity 86%. Tumor markers suggested a histological diagnosis in 83.3% of patients with lung cancer. The positive predictive value of tumor markers for the histological type was 97.3% in NSCLC and 96.9% in SCLC. The use of tumor markers in this population improves the histological diagnosis in patients with unclear histology obtained by cytology (15% of patients), and in patients with unclear imaging findings (8%). Use of markers may also decrease the time to diagnosis (18% of patients). Tumor markers in NSCLC also suggest the diagnosis of NSCLC histological subtype (adenocarcinoma vs. squamous in 81.8% of patients and 86% in the subgroup with unclear histology. Our informatics program allow to obtain similar results Conclusions: We confirm that a tumor marker algorithm is useful in symptomatic patients to predict the risk of cancer, to suggest the histological type and to improve the diagnostic techniques in patients with suspicious signs of lung cancer.

O8.5 - Serological biomarkers in lung cancer P. Stieber Institute of Clinical Chemistry, Campus Grosshadern, University of Munich, Germany

S33

Background: Detecting and treating lung cancer effectively and efficiently is still a major health care challenge. Aims: During lung cancerogenesis, diverse biochemical changes are observed influencing tumor cell growth, proliferation and tumor spread that are associated with the release of characteristic markers into the blood stream. Methods: Recent guidelines of the National Academy of Clinical Biochemistry (NACB) emphasize the high relevance of the markers carcinoembryonic antigen (CEA), cytokeratin-19 fragments (CYFRA 21-1), squamous cancer cell antigen (SCCA), neuron-specific enolase (NSE) and progastrin-releasing peptide (ProGRP) for their usage in lung cancer. Results: These 5 biomarkers have been investigated for the clinical indications: primary diagnosis and differential diagnosis, prognosis, post-operative surveillance, monitoring therapy in advanced disease and detection of recurrent disease in different histological subtypes. Due to their lack in tumor “ and organspecificity and despite their high frequency of increased release in lung cancer, single determinations of most of these biomarkers at present are not helpful for lung cancer screening of asymptomatic patients “ with the exception of ProGRP which reveals a highly specific and sensitive sensitivity for small cell lung cancer. The currently most important field of indication for ProGRP, CEA, CYFRA 21-1, SCCA and NSE is their determination in context of the primary diagnosis of lung cancer to support the histological diagnosis or to review the histology or, at an inoperable stage, to establish a diagnosis. Conclusions: The relatively poor prognosis of lung cancer patients in general, and the lack of satisfactory treatment modalities for recurrent disease, at least in the past, have led to a restrictive attitude for the application of tumor marker determinations, especially in follow-up care. Generally, follow-up investigations of any kind are of limited use in the absence of alternative therapy. But besides these limitations tumor markers as non invasive, reproducible, fast and not expensive method allow very early detection of recurrent disease and treatment control.

S34

O9 LYMPHOMA and LEUKEMIA

O9.1 - Epigenetic modulation of telomerase gene transcription in human T cell leukemia Toshifumi Hara and Mastaka Nakamura Human Gene Sciences Center, Tokyo Medical and Dental University Background: In our previous studies on function of the human T cell leukemia virus type I (HTLV-1) Tax in leukemogenesis of adult T cell leukemia (ATL), we found a novel element in the hTERT (human telomerase reverse transcriptase) gene promoter that critically regulates hTERT transcription in human T cells. The element has been demonstrated to be involved in repression of hTERT transcription in resting normal human T cells. Aims: We wish to understand how the element in the promoter regulates hTERT transcription in human T cell leukemia, because hTERT expression is constitutive in leukemia cells and is closely associated with cellular immortalization. Methods: Peripheral normal T cells and human T leukemia cell lines Jurkat and Kit 225 were used. A nuclear factor to bind the element was cloned by the yeast one-hybrid method. Factor binding was examined by gel mobility shit assay and chromation immunoprecipitation (ChIP) assay. Promoter activity was monitored by reporter assay and semi-quantative PCR for hTERT mRNA. DNA methylation was analyzed by the bisulphate DNA sequencing method. Chromatin modification was examined by ChIP assay with antibodies to modified histones. Results: The transcription factor Krüppel-like factor 2 (KLF2) was found to directly bind the hTERT element. The binding was observed only in resting, but not growing, normal T cells, in which hTERT transcription is repressed. The element in the hTERT promoter was methylated in Jurkat and Kit 225 leukemia cell lines, in which the hTERT gene is constitutively transcribed. Kit 225 cells expressed KLF2. However no association of KLF2 with the promoter element was seen. Demethylation of the element induced KLF2 binding and repression of hTERT transcription. In Jurkat negative for KLF2, there was no effect of demethylation on hTERT

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

expression. KLF2 binding to the element seemed to parallel histone H3 modification toward transcriptional repression. Conclusions: KLF2 binds the hTERT promoter in human T cells, resulting in repression of hTERT transcription. DNA methylation of the element in T leukemia cell lines inhibits KLF2 binding, contributing constitutive expression of hTERT.

O9.2 - Multiple myeloma: update in risk consideration, diagnosis, therapy, with trends and perspectives R. Lamerz Medical Department II, LMU-Klinikum Campus-Grosshadern, Munich, Germany Background: During the last 5–10 years, increasing improvements in the detection, risk estimation, diagnosis and treatment of multiple myeloma have occured. Aims: The aim of this presentation is to review and update the progress reached in risk consideration, diagnosis, and therapy of this still incurable disease and to mention some trends in future perspectives. Methods: In order to do this, the main recent and relevant literature was reviewed. Results: Multiple myeloma accounts for 1% of neoplastic diseases and 13% of haematological cancers, with a median incidence in Western countries of 5.6 cases/100,000 persons and a median age of about 70 years. This review presents a concise update on the oncogenomics/genomics, precursor disease MGUS=monoclonal gammopathy of undetermined significance, SMM=smoldering or asymptomatic myeloma) and risk stratification, diagnosis, and actual treatment of multiple myeloma in naïve patients and those with relapsed or refractory disease. Since the introduction of melphalan and prednisone in the 1960s and of high-dose melphalan with autologous stem cell transplant and bisphosphonates in the 1990s, a steady improvement of therapy has been reached by the involvement of immunomodulatory agents thalidomide and lenalidomide and the proteasome inhibitor bortezomib in the 2000s. Further progress is ongoing by the development of additional small molecules, multidrug combinations, monoclonal antibodies and molecular imaging in the 2010s.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

Conclusions: The recent treatment of multiple myeloma has steadily extended overall survival in patients below the age of 60 years to a 10-year survival rate of about 30% which represents a remarkable ongoing success for a still incurable disease

O9.3 - Drug monitoring of imatinib plasma levels in patients affected by chronic myeloid leukemia S. Galimberti1, A. Di Paolo2, M. Lastella2, M. Polillo2, M. Rousseau1, E. Sordi1, C. Barat1, R. Danesi2, M. Petrini1 1-Department of Oncology, Transplant, New Advances in Medicine, Section of Hematology, University of Pisa, Italy Background: In Western Countries, the incidence of chronic myeloid leukemia (CML) is 1–2 cases/ 100000/year. Leukemia cells are characterized by the presence of the Philadelphia (Ph™) chromosome, and by the BCR-ABL rearrangement. The chronic indolent phase of CML may last for 3–4 years until when an accelerated/blastic phase occurs. Recently, the introduction into clinical practice of imatinib, able to inhibit the tyrosine-kinase activity of BCR-ABL, has led to a significant improvement of survival. The highest therapeutic benefit is expected when minimal plasma concentrations of imatinib are equal or greater than 1000 ng/mL. However, imatinib pharmacokinetics is influenced by several physiological, pathological and pharmacogenetic factors. Aims: The application of a therapeutic monitoring protocol for imatinib in CML patients. Methods: Twenty-one patients, 11 men and 10 women, median age 58 years, affected by CML and treated with imatinib were enrolled. Drug daily dose was 400 mg. Plasma levels of imatinib were measured by a high performance liquid chromatography method with UV detection. Results: Pharmacokinetic analysis revealed a wide interpatient variability for all of the parameters investigated. In particular, clearance (CL/F) and volume of distribution (Vd/F) of imatinib were 14.2 ±7.1 L/h (CV, 50.2%) and 284 ±48 L (CV, 17.0%), respectively, while terminal half-life accounted for 16.4 ±6.7 h (CV, 41.1%). The Vd/F was significantly greater in men (312.5 ±39.3 L) with respect to that calculated in women (252.5 ±36.9 L).

S35

Mean maximal plasma concentration of imatinib was 2380 ± 540 ng/mL, while trough concentration accounted for 802 ± 465 ng/mL. It is worth noting that only 6 and 13 patients had Cmin values higher than 1000 and 500 ng/mL, respectively. Conclusions: This pharmacokinetic model could be useful to check patient compliance to imatinib and to early detect possible onset of resistance, with immediate shift to a IIgeneration TKIs.

O9.4 - Treatment of marginal zone lymphoma O. E, Buda G, Galimberti S, Pelosini M, Cecconi N, Cervetti G, Petrini M Azienda Ospedaliera-Universitaria Pisana, Hematology Unit, University of Pisa, Pisa, Italy Background: MZL accounts between 5% and 17% of all non-Hodgkin’s lymphomas. MZL are most often very indolent malignancies with 10 years survival approximately 75-80%. Standard treatment option for stage I-II disease has not been yet established. Patients with systemic stage III-IV disease should be considered for a more aggressive treatment. However a standard chemotherapic approach for MZl is still missing. Aims: We are reporting a retrospective study on the use of 2Cda and rituximab in the treatment of advanced stage MZL patients. Methods: From January 2000, patients referring to our Hospital because of MZL and needing for treatment underwent to chemotherapy with 2Cda 5 mg/sqm weekly, for 6 consecutive weeks, +/− rituximab, or received 100 mg/die orally CTX for 15 consecutive days a month and rituximab on day 8 of CTX therapy, monthly, for a 6 months treatment program. Results: With 2Cda ORR was 89.3%, with 53.6% CR and 35.7% PR. The remaining 10.7% patients obtained a stable disease or progression disease at the end of treatment. With CTX, the ORR was 90.3%, with 51.6% CR and 38.7% PR. Conclusions: To conclude, 2Cda therapy and CTX therapy are effective in the treatment of any subtype of MZL. Adding rituximab allowed increasing the response rate and the quality of response, significantly prolonging the TTF. Prospective randomized studies are necessary to validate our results and to define a standard approach in advanced stage MZL patients

S36

O9.5 - Thymidine kinase 1 as biomarker in patients with lymphoma and leukemia

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

O9.6 - Superior performance of isothermal Loop-mediated AMPlification (LAMP) in molecular diagnosis of leukemia

S Eriksson Departement of Anatomy, Physiology & Biochemistry, The Swedish University of Agricultural Science, BMC, Uppsala and AroCell AB, Uppsala, Sweden Background: Thymidine kinase 1 (TK1) is a DNA precursor enzyme and a marker for cell proliferation. Increased levels of TK1 activity in blood have been found and the TK REA and TK Liaison assays are clinically established for monitoring and prognostics of hematological malignancies. Aims: To review recent results with TK1 determinations, particularly TK1 protein measurements in patients with lymphoma and leukemia. Methods: Assay methods that will be reviewed are; a) serum TK activities using the Profiligen TK-Rea kit (Diasorin) (Mathews C, et al. Eur J Hematol, 77, 309–317, 2006; Letestu R et al., Blood, 116, 4588–90, 2010), TK1 protein levels determined by immuno assays based on peptide antibodies against TK1 either by b) a chemoluminecent dot blot assay (Pan, Z-L et al., J Cancer Res Clin Oncol, 136, 1193–9, 2010) or c) the TK 210 ELISA (Ruby C et al., Tumor Biol 31, supl 1, s82-83,2010). Patients in a) were 91 chronic lymphocytic leukemia (CLL) and 339 CLL pateints diagnosed with Binet stage A followed until progression, respectively; in b) 37 B-CLL patients with information on treatment response and 5-years survival and in c) 40 acute leukemia sera and 26 plasma samples from CLL patients. Results: Higher serum TK levels were found in IgVH unmutated, compared with IgVH mutated patients. Serum TK was an independent predictor of progression free survival and could be combined with 3 other markers (beta-2-microglubulin, lymphocytosis, CD38) to obtain cost-effective strategy for reliable prognostication. When the mean serum TK1 protein levels were determined with the dot blot method, values before treatment and at 1 and 28 days after start of the treatment could be used to predict outcome of patients after therapy, including 5-year survival. A TK 210 ELISA was recently established which determines concentrations of TK1 protein in serum and plasma. It shows excellent correlation with the TK activity levels. Conclusions: Serum TK activity and protein measurements can serve as valuable biomarkers for monitoring and prognostication of patients with lymphoma and leukemia

G. Minnucci, O. Spinelli*, A. Bussini*, S. Salmoiraghi*, A.Rambaldi*, E. D'Agostini, R. Mesturini, F. Colotta and G. Amicarelli DiaSorin Spa, Saluggia (VC), Italy; *Ospedali Riuniti di Bergamo, Bergamo, Italy Background: Several molecular markers associated to leukemias are actually diagnosed by PCR-based techniques. The main limitations encountered are timeconsuming and multi-steps procedures and the requirement of specialized equipment and expertise. Aims: To develop a real-time non-PCR improved method, based on the isothermal LAMP (Loop mediated AMPlification) reaction, applicable to the detection of point mutations and chromosome translocations in leukemia cells. Methods: Have developed a novel LAMP assay to selectively detect the JAK2 V617F mutated sequence. The sensitivity has been established on serial dilutions of mutant DNA from the UKE1 cell line into the wild type DNA from the BJAB cell line. The specificity has been determined on 240 DNA replicates from several wild type cell lines and healthy donors. Finally the performances have been validated on 40 clinical samples from patients affected by myeloproliferative neoplasms and compared with the results obtained by the conventional Allele Specific PCR. Two different LAMP assays have been specifically designed for the detection of BCR-ABL and PML-RARalpha fusion tra! nscripts and are in advanced phase of development. Results: In JAK2 mutated cell lines and clinical samples the modified LAMP assay consistently demonstrated superior sensitivity, specificity and simplicity of assay set up in respect to conventional PCR. Preliminary data confirm and extend the superior performance of LAMP to the detection of BCR-ABL and PMLRARalpha translocations, in which nucleic acid amplification is performed directly on RNA, thus decreasing the time-to-result and the risk of contaminations from classical multiple-steps procedures. Since this technology doesn’t need thermocycling, it requires a simple, not-expensive instrument. Conclusion: The modified LAMP technology holds superior performance versus conventional PCR methods in detecting point mutations and chromosome translocations in leukemia cells.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

O10 TUMOR MARKERS IN CLINICAL TRIALS O10.1 - Bevacizumab pharmacogenetics: still looking for the right pieces of the puzzle Guido Bocci Division of Pharmacology, Department of Internal Medicine, University of Pisa, Pisa, Italy Background: Bevacizumab has been the first moleculartargeted antiangiogenic therapy approved by FDA and EMEA and, despite the numerous clinical successes, many challenges remain.Indeed, no genetic markers to predict or monitor the efficacy of - or the resistance to bevacizumab have been found yet. Aims: Thus, any efforts to identify these markers represent a challenging field of clinical oncology research. The important role of both tumor cells and tumor microenvironment in determining the complex process of angiogenesis clearly supports a pharmacogenetic investigation of tumor and germline single nucleotide polymorphisms (SNPs) for antiangiogenic drugs such as bevacizumab. Methods: An obvious strategy to pursue this aim is to look at the VEGF-A SNPs because the inhibition of the biologically active vascular endothelial growth factor (VEGF) represent the best known mechanism of action of the antibody. Results: In particular, it has been reported the association between different VEGF-A genotypes and median overall survival (OS) when using bevacizumab (with paclitaxel chemotherapy) in a phase III clinical trial of metastatic breast cancer. Indeed, it has been shown, in a retrospective record of cases, that VEGF-A -2578AA genotype was associated with a superior OS in the arm treated with paclitaxel and bevacizumab vs. paclitaxel alone. Moreover, patients carrying the VEGF-A -1154A allele also demonstrated a superior OS in the combination arm. These important experiences have greatly stimulated the bevacizumab pharmacogenetic research on VEGF-A SNPs. In particular, with regard to metastatic colorectal cancer (mCRC), a recent retrospective experience has shown a significant correlation of VEGF-A -1498TT genotype with worse PFS in a population of mCRC patients treated with FOLFIRI plus bevacizumab as first-line regimen.

S37

Conclusions: Are these results ready for the clinical routine? One of the major key-points in pharmacogenetics seems to be the extremely low-rate of “successful” determinants that enter the clinics compared to the huge amount of enthusiastic preliminary data. This central point has certainly a close relation to the hot issue of methodology. Initial data should be tortured with the most accurate statistical correction to avoid the risk of false positive associations.

O10.2 - Bone alkaline phosphatase - a drug target and a protein therapeutic J. L. Millan Sanford-Burnham Medical Research Institute, La Jolla, California 92037, USA Background: Tissue-nonspecific alkaline phosphatase (TNAP), the isozyme expressed in bone, liver and kidney in humans, is the primary enzyme controlling skeletal mineralization. Studies in this laboratory using transgenic and knockout mouse models indicate that the main function of this enzyme is to hydrolyze extracellular inorganic pyrophosphate (ePPi), a potent mineralization inhibitor. Transgenic upregulation of TNAP in bone leads to systemically lower levels of ePPi and increased bone mineral density, while ablation of TNAP in the skeleton leads to rickets/osteomalacia, a condition known as hypophosphatasia. Furthermore, the ectopic upregulation of TNAP in vascular smooth muscle cells is necessary and sufficient to cause medial vascular calcification, a condition prevalent in end-stage renal disease, diabetes, obesity and aging, as well as in rare diseases such as idiopathic infantile arterial calcification. Aims: 1) To develop potent specific inhibitors of TNAP to be used for the treatment of medial vascular calcification. These same compounds could be useful to reduce the bone sclerosis caused by osteoblastic prostate cancer metastasis. 2) To develop a bonetargeted TNAP treatment for the prevention and treatment of hypophosphatasia in mice and humans.

S38

Methods: 1) Small molecule screening and medicinal chemistry optimization via structure-activity-relationships; 2) fusion protein construction and purification; 3) generation and characterization of animal models; 4) hypophosphatasia mice and patients Results: Using the exceptional resource of the Conrad Prebys Center for Chemical Genomics (CPCCG) at the Sanford-Burnham Medical Research Institute (SBMRI) we have developed potent TNAP inhibitors able to reduce calcium deposition in cultures of vascular smooth muscle cells and in aortas ex vivo. We have developed a robust mouse model of idiopathic infantile arterial calcification useful for the in vivo validation of this therapeutic approach. These same TNAP inhibitors are currently being investigated as a means of ameliorating the bone sclerosis associated with osteoblastic bone metastasis by prostate cancer cells. The use of bone-targeted TNAP (ENB-0040) for the prevention and treatment of hypophosphatasia has been validated. To-date we have documented a remarkable improvement of the skeletal and dental abnormalities of infants and children with life-threatening hypophosphatasia by weekly administration of bonetargeted TNAP. Conclusions: TNAP inhibitors show great promise for the treatment of soft-tissue calcification and bone sclerosis. Bone-targeted TNAP is likely to become an approved drug for the treatment of hypophosphatasia. Supported by grants DE12889, AR53102, AR47908 and HL101899 from the National Institutes of Health, USA.

O10.3 - Pitfalls in the evaluation of tumor markers Ricardo Moro BioCurex Inc. 7080 River Rd. Richmond, B.C., Canada V6X 1X5, Background:Tumor marker evaluation (sensitivity and specificity in discriminating cancer from non-cancer samples) is affected by sources of error that are generally overlooked:

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

1) ROC analysis: This statistical tool is simple to use and offers the perceived advantage of being distribution independent. However: 1.a) Sensitivity and specificity are undetermined within vertical or horizontal segments of the curve. Selecting the best sensitivity and specificity combination is incorrect; the best fit line for the data set passes somewhere in the middle of those straight segments. 1.b) Each point in a ROC curve is independent from all others and therefore the cutoff value depends on one point, which if incorrect will conditioning a wrong cutoff value. A better approach is to normalize the distribution if necessary (Log, Box-Cox transformation) and then analyze the data using Gaussian functions. Aims: The "golden standard": The pathology report is usually considered as "absolute truth". However, when two or more pathologists look at the same slides, there is a certain disagreement, depending on the lesion. This source of error is commonly overlooked and yet significant when evaluating screening tests. Methods: Pathology vs. "course of action": The pathology might classify a case as benign (i.e. breast “atypical hyperplasia”) but the lesion might be still removed. The sample is classified as benign, but the treatment corresponds to a malignancy. If the tumor marker is elevated, the case is considered as a false positive whereas in actuality the test agreed with the course of action taken. Thus, the course of action should be considered along with the pathology. Results: Coefficient of variation (CV) error near the cutoff value: The performance of a tumor marker depends on the performance of the device used to measure it. When evaluating new tumor markers, the latter should be made irrelevant; otherwise a good marker could be discarded because of a bad immunoassay. Samples with marker value close to the cutoff can fall on the positive or negative side just because of the CV error. These points should be included when evaluating the overall performance of a tumor marker assay, but excluded when the marker itself is being evaluated. Conclusions: The accumulation of these errors may reduce a marker’s specificity and wrongly render it useless for screening.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

O11 UROLOGICAL CANCERS O11.1 - Down-regulation of estrogen receptor-β associates with transcriptional coregulator PATZ1 delocalization in human testicular seminomas

S39

associated with ERβ down regulation in human testicular seminomas could be ascribed to an impaired cAMP mediated signalling to generate the testicular neoplasia.

O11.2 - Hyperparathyroidism during Sunitinib administration in mRCC patients

Boscia F., Esposito F., Chieffi P. 1Dipartimento di Medicina Sperimentale, II Universita di Napoli, Italy Background: Estrogen exposure has been linked to a risk for the development of testicular germ cell cancer. The effects of estrogen are now known to be mediated by estrogen receptor α (ERα) and ERβ subtypes, but only ERβ has been found in human germ cells of normal testis. However, its expression was markedly diminished in seminomas, embryonal cell carcinomas, and in mixed germ cell tumours, but remains high in teratomas. PATZ1 is a recently discovered zinc finger protein that, due to the presence of the POZ domain, acts as a transcriptional repressor affecting the basal activity of different promoters. PATZ1 is able to bind proteins involved in chromatin remodeling such as HMGA1, that we have shown to be over-expressed in human testicular seminomas, and RNF4. Aims: We have previously described that PATZ1 plays a crucial role in normal male gametogenesis and that its upregulation and mis-localization could be associated to the development of testicular germ cell tumours. Methods: Here, we show by using different techniques (IHC,Immunofluorescence, immunoprecipitation, Western blot, and others) that ERβ interacts with PATZ1, and PKA (Regulatory subunit) in normal germ cells, while down regulation of ERβ is associated with transcriptional coregulator PATZ1 delocalization in human testicular seminomas. Results: Interestingly, we found that the translocation of PATZ1 from the cytoplasm into the nucleus is regulated by cAMP that induces also an ERβ increased expression and nuclear localization, while this effect is counteracted by using the anti-estrogen ICI 182–780 Conclusions: Taken together, our results demonstrate for the first time the interaction between PATZ1 and ERb in normal germ cells. Consistently, the PATZ1 delocalization

Baldazzi V, Tassi R, Caruso S, Garofoli E, Molinara E, Mori E, Pellegrini E, Lapini A*, Carini M*, Mazzanti R. ONCOLOGIA MEDICA 2, * UROLOGIA 1; AOU CAREGGI, Firenze Background: Sunitinib malate is an orally bioavailable tyrosine kinase inhibitor, with known activity against many tyrosine kinase receptors involved in tumour growth and angiogenesis. As its use in clinical practice is quite recent, many of its possible side effects are still unknown. Aims: Here we report the occurrence of newly onset of hyperparathyroidism in a cohort of sunitinib receiving patients. The aim of our study is to evaluate parathyroid and mineral metabolism parameters among patients treated with first line sunitinib, to assess a possible relationship between the drug and phosphor-calcicium homeostasis Methods: We prospectively evaluated parathyroid and calcium metabolism function tests in patients receiving 50 mg Sunitinib daily for metastatic renal cell carcinoma according to a classical 6-weeks schedule. 17 patients were eligible for a complete assessment of intact parathyroid hormone plasma levels, serum phosphate and calcium, serum 25-hydroxyvitaminD and 1,25-dihydrovitaminD, urinary calcium and phosphate. Those biochemical evaluations were performed before the beginning of treatment and at the end of each Sunitinib ON period as routine laboratory evaluations in every patient. Results: Before the beginning of treatment PTH, serum and urinary calcium and phosphate were within the normal range in every patient. During Sunitinib administration, 11 patients (64,75%) developed elevated parathyroid hormone levels, with low-to-normal serum calcium and phosphate. Serum calcium levels, phosphate and vitamin D metabolites always were within a normal range. Patients presenting elevated PTH presented low or undetectable urinary calcium levels.

S40

Conclusions: Newly onset of hyperparathyroidism may develop in patients undergoing Sunitinib treatment. A similar observation has been reported among patients on Imatinib as a consequence of PDGFR inhibition both on osteoclasts and osteoblasts. These initials events result in a decreased bone re-absorption and temporary decrease of serum ionized calcium that induces a compensatory hyperparathyroidism with a decreased renal calcium excretion and in the end a correction of serum calcium levels.

O11.3 - Prostate cancer: the Prostate Health Index (PHI) may guide diagnostic and treatment M. Lazzeri Dept. of Urology, San Raffaele Turro, Vita-Salute San Raffaele University, Milan, Italy Background: Isoform [2′]proPSA (p2PSA), %p2PSA and the Beckman Coulter Prostate Health Index (PHI) improve the accuracy in predicting the presence of PCa at prostate biopsy and are related to PCa aggressiveness Aims: To translate statistical findings into a clinical practice cut-off and test the hypothesis which accurately predict the PCa characteristics with final pathology after radical prostatectomy. Methods: The study is an observational prospective analysis of sensibility, specificity and accuracy of serum p2PSA, % p2PSA {([−2]proPSA/10)/fPSA)} and PHI {[[−2]proPSA/ fPSA] x squart PSA}, in Caucasian men referred for prostatic biopsy for suspected PCa. All patients underwent ambulatory TRUS-guided prostate biopsies.Their pathological predictive accuracy was evaluated in men who underwent open retropubic and robot-assisted laparoscopic radical prostatectomy and pelvic lymphadenectomy- according to the surgeon™s judgement. Results: In 664 patients, PCa cancer was found in 266 (40%), p2PSA, %p2PSA and PHI values were significantly higher (p 73.4%) and PHI (AUC: 72.1%) were the most accurate predictors and significantly outperformed tPSA

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

(AUC: 51%) and %fPSA (AUC: 62.9%) in the prediction of PCa at biopsy (p‰₪0.001). At a %p2PSA cut-off of 1.26 a total of 123 (18.5%) biopsies could be avoided, but 7 (5.4%) cancers with a Gleason score of 7 or greater would have been missed. At a PHI cut-off of 30, a total of 125 (18.9%) biopsies could be avoided and 5 (3.8%) cancers with a Gleason score of 7 or greater would have been missed. The %p2PSA and PHI levels were significantly higher in patients with pT3 disease, pathologic Gleason sum ‰¥7, Gleason sum upgrading, and lymph node invasion (all p values ‰₪0.004). Conversely, %p2PSA and PHI levels were significantly lower in patients with tumour volume Conclusions: Our findings show % p2PSA and PHI could be used in real-life situation to improve decisions whether to perform a prostatic biopsy in a set of contemporary Caucasian men missing only a few aggressive cancers. Finally, they are powerful predictors of several PCa characteristics at final pathology after radical prostatectomy.

O11.4 - Tumor-associated trypsin inhibitor/SPINK1 in prostate cancer U. H. Stenman Department of Clinical Chemistry, Biomedicum, Helsinki University, Helsinki, Finland REVIEW Tumor-associated trypsin inhibitor (TATI), also called SPINK1, is an independent serum marker for adverse prognosis in ovarian, colorectal and renal cancer but in ventricular and bladder cancer, loss of TATI expression is associated with adverse prognosis (1). Recently, we showed that the serum concentrations of TATI are elevated in nearly half of the patients with prostate cancer. Tissue expression of TATI is increased in 15% of the tumors and this is associated with adverse outcome. Suppression of TATI expression in prostate cancer cells reduces their invasiveness

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

indicating that TATI acts as an oncogene (2). Interestingly, TATI is as an autocrine growth factor in many tumors, e.g., hepatocellular, breast and colorectal cancer. This has been ascribed both to its ability to stimulate the epidermal growth factor receptor (EGFR) and to inhibit a so far unknown serine protease that mediates apoptosis (reviewed in (1)). Both mechanisms could explain why TATI it is such a strong prognostic factor for many cancers. A recent study showed that TATI/SPINK1 activates (EGFR) and stimulates invasion and growth of prostate cancer cells in culture, while administration of TATI antibody inhibits growth of prostate cancer xenografts in mice. Humanized antibodies to TATI for treatment of prostate cancer are now under development (3) and measurement of TATI in serum is of potential value for selection of patients for antibody therapy. References: 1. Stenman UH. Role of the tumor-associated trypsin inhibitor SPINK1 in cancer development. Asian J Androl 2011;13:628–9. 2. Tomlins SA, Rhodes DR, Yu J, Varambally S, Mehra R, Perner S, et al. The role of SPINK1 in ETS rearrangementnegative prostate cancers. Cancer Cell 2008;13:519–28. 3. Ateeq B, Tomlins SA, Laxman B, Asangani IA, Cao Q, Cao X, et al. Therapeutic targeting of SPINK1-positive prostate cancer. Sci Transl Med 2011;3:72ra17.

O11.5 - Specific Plasminogen Isoforms in Patients with Developing Prostate Cancer

D.I. Vodolazhsky1,E.F. Shin1, A.Y. Golikov1, T.N. Belova1, D.V. Zimakov1, Y.N. Cherkasova1, N.V. Boyko1, M.I. Kogan2, M.B. Chibichian2, S.A. Moshkovsky3, D.G. Matishov1

1: Institute of Arid Zones of the Southern Scientific Centre RAS, Institution of the Russian Academy of Sciences, Rostov-on-Don, Russia. 2: Rostov State Medical University, Rostov-on-Don, Russia. 3: V.N. Orekhovich Institute of Biomedical Chemistry of Rus

S41

Background: Plasminogen (Pg) is currently of interest largely due to its function as a basis for protein products in cells, such as plasmin and angiostatin, which are important in development of malignant tumors. Although Pg is known to exist in a few isoforms, there is no data on their link to malignant tumor development. Aims: The aim is to find a range of Pg isoforms characteristic of different stages of tumor development in prostate. Methods: Blood samples from male patients, 61 to 70 years of age, were divided as follows: control (healthy donors, n=10), group I (T1-T2 stages of tumor, n=11) and group II (T3-T4 stages of tumor, n = 10). The analysis of Pg isoforms included the following stages: affinity chromatography, 2D-PAGE and mass spectrometry. An χ2 test and discriminant analysis were used to assess statistically significant variations in frequency of occurrence of specific Pg isoforms and their range in the control and pathology samples. Results: The recorded protein products vary in molecular weight (MW) and isoelectric point (pI); average ranges of Pg isoforms were obtained for each study group. A total of 16 Pg isoforms were found that can be pooled into 3 groups according to MW. Within those groups, variations in pI were observed. The MW= 116 and 114 kDa isoforms were found in all samples. Dissimilarities were observed in MW = 112 kDa group, which include 5 isoforms different in pI. The control contained two combinations (ranges) of those isoforms, in group I the combination of isoforms grew to 7 and in group II it grew to 6. Thus, 5 and 4 Pg ranges were found to be characteristic of groups I and II, respectively, and using them within the pool helps identify T1-T2 and T3-T4 stages of cancer with high confidence in 55% and 70% of cases, respectively. Conclusions: The results suggest a possibility that the said ranges of Pg isoforms may serve as a marker for identifying stages of tumor development in human prostate.

S42

O12 VALIDATION O12.1 - International multi-validation of the Prostate Cancer Prevention Trial Risk Calculator (PCPTRC): Results from the Prostate Biopsy Collaborative Group (PBCG) Ankerst D. P., A. Boeck, I. M. Thompson, A. M. Cronin, M. J. Roobol, J.Hugosson, J. St. Jones, M. W. Kattan, E. A. Klein, F. Hamdy, D. Neal, J. Donovan, D. J. Parekh, G.Bartsch, H. Klocker, W. Horninger, A. Benchikh, G. Salama, A. Villers, S. J. Freedland Texas University, Health Science Center, San Antonio, San Antonio, Texas, USA; Technische Universitaet Muenchen, Garching, Germany; and other institutions of the PBCG Background: The PCPTRC is currently recommended in US screening guidelines, but multiple validations of it have revealed conflicting results concerning its applicability to all populations. Aims: To evaluate the discrimination, calibration and net benefit performance of the PCPTRC across 10 European and U.S. biopsy cohorts. Methods: PCPTRC risks were calculated for 25,733 biopsies using prostate-specific antigen (PSA), digital rectal examination, family history and history of prior biopsy, and single imputation for missing covariates. Risks were evaluated using the areas underneath the receiver operating characteristic curves (AUC), discrimination slopes, chi-square tests of goodness of fit, and net benefit decision curves. Results: AUCs of the PCPTRC ranged from a low of 56.2% in the Goeteborg Rounds 2–6 cohort to a high of 72.0% in the Goeteborg Round 1 cohort, and were statistically significantly higher than that of PSA in 6 out of the 10 cohorts. Calibration of the PCPTRC failed in all cohorts except for the San Antonio SABOR cohort, a healthy screening cohort similar to the PCPTRC cohort. There was limited to no net benefit to using the PCPTRC risks for referral to biopsy in all five European Randomized study of Screening for Prostate Cancer cohorts but benefit at limited ranges in the other cohorts. Conclusions: External validation of the PCPTRC across ten cohorts revealed varying degree of success highly dependent on the cohort, most likely due to different criteria for and work-up before biopsy. Future validation studies of new calculators for prostate cancer should acknowledge the potential impact of the specific cohort studied when reporting successful versus failed validation.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

O12.2 - The development and validation of a prostate cancer risk calculator including information on prostate volume from digital rectal examination: ERSPC Rotterdam M. J.Roobol, H.A. van Vugt, S. Loeb(1), Yefang Zhu, Meelan Bul, Chris H.Bangma, Arno G.L.J.H. van Leenders, Ewout W.Steyerberg, Fritz H.Schröder Erasmus University Medical Centre, Rotterdam, The Netherlands and (1) New York University School of Medicine, New York, NY, US Background: The ERSPC riskcalculators (RCs) are validated useful tools for prostate cancer (PC) risk assessment, but include prostate volume (PV) data from transrectal ultrasound (TRUS). Aims: To develop and validate a digital rectal examination (DRE) based RC which circumvents the need for TRUS, but still includes information on PV. Methods: For development of the DRE based RC we studied the original ERSPC Rotterdam RC population (www.uroweb. org) including 3,624 men (885 PC cases) detected at initial screening and 2,896 men (547 PC cases) detected at repeat screening 4 years later. The validation cohort consisted of 253 men, previously screened participants of ERSPC Rotterdam. Data on TRUS-assessed PV in the development cohorts were reclassified in 3 categories (25, 40 and 60 cc). New RCs including PSA, DRE and PV categories were developed for men with and without a previous negative biopsy to predict overall and clinically significant PC (HG PC) defined as Tstage > T2B and/or Gleason score >= 7. Predictive accuracy was quantified using ROC analysis and AUC of models based on PSA alone, PSA and DRE, the DRE based RC and the original RC were compared. In the validation study, comparison with the Prostate Cancer Prevention Trial (PCPT) RC was performed. Results: AUC of PSA alone, PSA and DRE, the DRE based RC and the original ERSPC RC to predict PC at initial biopsy were 0.69, 0.73, 0.77 and 0.79 respectively. The respective AUCs for predicting HG PC were 0.74, 0.82, 0.85 and 0.86. Similar results were seen in men previously biopsied and in the validation cohort. The DRE based RC outperformed the PCPT RC and a model based on PSA and DRE only. Conclusions: A prediction tool containing PV estimates based on DRE is more accurate in risk prediction as compared to PSA or PSA and DRE alone and is easily implemented into daily practice by urologists or general practitioners.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

S43

O13 TD WORKSHOP

O14 IMMUNOLOGY

O13.1 - Characterization of antibodies to the C-terminal peptide of hCG

O14.1 - Immunogenic cell death serum biomarkers in the prognosis and the therapy prediction of cancer patients

U. H. Stenman S. Holdenrieder Dept of Clinical Chemistry, Biomedicum, Helsinki University, Helsinki, Finland Background: Monoclonal antibodies (MAbs) to the Cterminal peptide (CTP) of the β subunit of human chorionic gonadotropin (hCGβ) are completely specific for hCG and do not recognize luteinizing hormone (LH), which lacks a corresponding peptide. They are therefore used in many commercial hCG assays. Three epitopes have been identified on CTP. Epitope b8 comprises the terminal amino acids 140–145 and b9 amino acids 115– 120. A unique antibody, B152 recognizes a third epitope comprising the core-2 O-glycan on Ser132 and surrounding peptides. Aims: To characterize combinations of antibodies to the CTP submitted to the hCG antibody workshop. Methods: Antibodies recognizing epitopes to CTP were combined in sandwich assays and tested for reactivity with various forms of hCG. Results: We showed that combinations of the three epitopes on CTP gave functional assays. However, when combined with MAbs recognizing other epitopes on the subunits, only part of them gave functional assays Conclusions: Knowledge on the epitope specificity of hCG antibodies facilitates development of assays recognizing specifically the different forms of hCG occurring in biological fluids individually or together.

Institute of Clinical Chemistry and Clinical Pharmacology, University Hospital Bonn, Germany Background: Induction of immunogenic cell death has been shown to be associated with the efficiency of cytotoxic therapies. Aims: In cancer patients undergoing systemic and localized therapies, serum biomarkers of immunogenical cell death high mobility group box 1(HMGB1) and receptor of advanced glycation end products (RAGE) were investigated on their relevance for predicting or early indicating therapy response and prognosis. Methods: Several prospective trials on 49 patients with colorectal cancer with liver metastases undergoing selective internal radiation therapy (SIRT), including 50 patients with primary liver cancer undergoing transarterial chemoembolization (TACE), and 44 patients with localized breast cancer receiving neoadjuvant systemic chemotherapy were included. Blood samples were taken before as well as during the course of the treatments. In local SIRT and TACE trials, blood was taken daily after application to see the immediate effect of therapy on marker levels. HMGB1 and RAGE were quantified by ELISA. Therapy response was quantified by radiology, outcome was objectified as progression-free and overall survival. Results: In cancer patients receiving local SIRT or TACE therapies, HMGB1 levels rapidly increased already 24 hours after application while RAGE levels decreased. In

S44

CRC patients undergoing SIRT, elevated HMGB1 concentrations were associated with poor therapy response (24 h value) and poor overall survival (0 h and 24 h values). In liver cancer patients undergoing TACE, low 0 h and 24 h RAGE levels were associated with insufficient therapy response and unfavourable overall survival. In breast cancer patients, pretherapeutic high HMGB1 and low RAGE concentrations were associated with poor response to neoadjuvant chemotherapy – with borderline significance, while CEA and CA 15-3 were not. Conclusions: Proimmunogenical cell death serum marker HMGB1 and antiimmunogenical soluble RAGE are new and valuable biomarkers for the prediction of therapy response and prognosis in cancer patients.

O14.2 - Human Th17 cells: main features and possible role in cancer Romagnani S., Annunziato F. Department of Internal Medicine, University of Florence, Italy Background: CD4+ T-helper (Th) cells that selectively produce interleukin (IL)-17A (Th17) are thought to be critical for both host defense and autoimmunity. Three major dogmas were established, based on initial studies performed in murine models, and initially extrapolated by many researchers to human pathophysiology. First, Th17 cells represent a fixed CD4+ T-cell effector phenotype. Secondly, Th17 cells are responsible for pathogenicity in several chronic inflammatory disorders, Th1 cell being instead protective. Finally, Th17 cells originate from na ¯ve Th cells in response to transforming growth factor (TGF β and IL-6. Aims: To demonstrate possible species-specific differences between murine and human Th17 cells. Methods: Examination of phenotypic and functional features of Th17 cells in human T-cell clones and in freshly-derived circulating and tissue-infiltrating T cells by using different techniques. Results: Human Th17 cells expressing the Th1“related transcription factor T-bet, showed IL-12-inducible plasticity

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

to Th1 cells, and TGFβ signalling was dispensable for their development, which simply required the presence of IL1β¢ and IL-23. We also showed that human Th17 cells and Th17-derived Th1 cells express CD161, the human equivalent of the murine NK1.1 molecule. We have more recently demonstrated the existence of a RORC-dependent self regulation of human Th17 cell expansion, probably aiming to reduce their potentially high dangerousness, and we have identified the molecular mechanisms responsible for such regulation. Conclusions: More recent studies in mice have led to reassessment of the three above mentioned dogmas regarding the Th17 phenotype, suggesting that murine Th17 cells are not so different compared to man after all. Indeed, in different models of both murine and human pathology Th17 cells seem to exert their pathogenic role when they shift in the inflammatory microenvironment into Th1 cells. A few studies demonstrated an important role for Th17 cells in the protection against cancer in both mice and humans. However, whether cancer protection is exerted by Th17 or Th17-derived Th1 cells is still unclear. An intriguing question is the rarity of Th17 cells in the inflamed or neoplastic sites despite their important potential role in both immunopathology and protection

O14.3 - Chronic inflammation-induced immunosuppression: Biomarkers, underlying mechanisms and clinical implication for cancer therapy M. Baniyash, M. Sade-Feldman, J. Kanterman and E. Ish-Shalom The Lautenberg Center for General and Tumor Immunology. Hebrew University, Medical School, Jerusalem, Israel Background: It is well established that chronic inflammation and cancer are associated; chronic inflammation can predispose an individual to cancer and developing tumors induce micro- and macro-chronic inflammatory environments. Under both circumstances, an immunosuppressive milieu ensues, enabling tumor escape from the host™s immune system and may disrupt success of immune interventions.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

Aims: To extend our understanding of the mechanisms underlying chronic inflammation induced immunosuppression and its impact on tumor development and spreading, seeking to discover drugs neutralizing such an environment and to increase efficacies of given immune-based therapies. We are also establishing a high-fidelity detection system using unique biomarkers for monitoring the hosts immune status modified by chronic inflammation that could predict success of given therapies and disease regression. Methods: Mouse models for chronic inflammation and growing transplantable or spontaneously occurring tumors are used. Currently, a retrospective study analyzing blood samples from cancer patients is initiated, testing expression of biomarkers that provide indication for the hosts immune status prior to and following chemo- and immune-based therapies. Results: We proved chronic inflammation and associated myeloid derived suppressor cells as the causative link for the induced immunosuppression described in various tumors and explored the underlying mechanisms. We demonstrate that such an environment suppresses not only the hosts immune system but also newly administered immune cells and vaccination-based therapies, limiting success of cancer immunotherapies. We also show the diverged effect of chemotherapeutic drugs on the immunosuppressive environment by using novel biomarkers. Conclusions: Our studies advance the understanding of tumor-associated immunosuppression and could facilitate design of innovative strategies for cancer therapy that may be combined with monitoring the patient’s immune status prior to and following a given treatment, enabling establishment of an optimal and successful personalized cancer therapy.

O14.4 - The plasma soluble HLA peptidome as a source of cancer biomarkers Bassani-Sternberg M, Barnea M, Beer E, Admon I, Admon A. Department of Biology, Technion - Israel Institute of Technology and IBM Haifa Research Laboratory, Haifa, Israel

S45

Background: The human leukocyte antigen (HLA) molecules are membrane-bound peptide transporters that carry peptides to the cell surface for surveillance by circulating T lymphocytes. These peptides are produced by degradation of cellular proteins and their display at the cell surface serves to alert the immune system to the disease state of the cells. Many types of cancer cells release into the plasma large amounts of soluble HLA molecules (sHLA) carrying their authentic peptide repertoires, presumably to escape immune surveillance. Since healthy cells shed only small amounts of sHLA molecules relative to the cancer cells, the sHLA molecules present in the plasma of cancer patient are mostly derived from the diseased cells. We have recently developed these findings into a new methodology for cancer diagnosis and identification of potential cancer immunotherapeutics. Aims: To develop our new approach of diagnosing cancer and identifying targets for immunotherapy based on analysis of the plasma soluble HLA peptidome into a clinically viable approach. Methods: The methodology is based on a single-step immunoaffinity purification of the soluble HLA molecules from human plasma, obtained from a 10 ml test tube of fresh blood. The bound peptides are extracted, identified and quantified by capillary chromatography combined with Orbitrap mass spectrometry. Results: We compared the plasma soluble HLA peptidomes of different cancer patients, including leukemia and melanoma, before and after treatment. In general, the sHLA peptidomes change only slightly when prepared from the different blood samples taken from the same person, yet they are very different between people who differ in their HLA haplotypes. Most importantly, as the tumor loads is reduced, following treatment, some sHLA peptides are also reduced or even totally disappear. Thus, following the sHLA peptidomes, during the course of the disease, helps to identify the more significant peptides with potential for clinical use. So far, tens of thousands of sHLA peptides were identified this way from the plasma of a few dozen cancer patients and healthy controls. Conclusions: This approach facilitates the identification of large biomarkers dataset, potentially even useful for personalized treatment and immunotherapy.

S46

O15 GUIDELINES

O15.1 - The Italian Guide 2010; A comparative assessment allowing for flexibility within an evidence based framework Gion M.1, Trevisiol C.2, Pregno S. 3, Rosin C.1, Fabricio A.S.C.4,1 1Regional Center for Biomarkers, Department of Clinical Pathology, Azienda ULSS 12, Venice, Italy; 2Istituto Oncologico Veneto IOV - IRCCS, Padua, Italy; 3Cochrane Italian Center, Milan, Italy, 4ABO Foundation, Venice, Italy Background: After decades from the discovery of cancer biomarkers, their use in clinical practice is still widely inappropriate. In addition, a plethora of novel biomarkers has been discovered in the last years thanks to the development of knowledge in cell biology and of new technologies. The concurrent shrinkage of economic resources requires strategies to translate the knowledge on clinical utility of cancer biomarkers for their end-users Aims: The goal of the project was to present in synoptic tables a systematic comparison of recommendations for clinical use of circulating cancer biomarkers and critical appraisal of the guidelines addressing similar topics, facilitating the end-user to their easier utilization Methods: The "Guide 2010" was a two-years project (2009–2010). A transparent, explicit and systematic EBM approach was adopted by a multidisciplinary working group consisting of 60 experts from different fields. The project was carried out under the behalf of 7 Italian Scientific Societies representing the main healthcare specialties involved in the tumor markers clinical use. Recommendations regarding biomarkers of 17 different solid tumors with the corresponding strength of evidence and rating scheme were faithfully reported into comparative tables, and then synthesized into synoptic tables. The guideline critical appraisal was performed by at least 3 independent experts with the AGREE-I tool was also recorded Results: Fifty-two EBM guidelines were selected from over 3500 documents retrieved (published in the last 7 years). The synoptic tables and AGREE evaluations were published in the Guide 2010 and presented during several national congresses organized by the participating Scientific Societies Conclusions: The "Guide 2010" synthesizes the available recommendations for each clinical scenario and at the same time supplies a synthetic appraisal of the guidelines. This

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

approach allows the users to make critical and informed decisions on biomarkers use in clinical practice, permitting a certain degree of flexibility although always within the EBM framework

O15.2 - How to choose recommendations from different guidelines “ The available methodological instruments Pregno S.1, Trevisiol C.2 1Italian Cochrane Centre, Milan, Italy, 2Istituto Oncologico Veneto IOV - IRCCS, Padua, Italy Background: High quality, evidence-based clinical practice guidelines are helpful tools in clinical decisions. Busy health professionals need clear guidance to retrieve, evaluate and choose sound recommendations. Aims: To present a comparative analysis of the primary areas of agreement and difference between the guidelines addressing similar topics on biomarkers in oncology, evaluating the guidelines quality with the AGREE I tool through a multidisciplinary approach. Methods: The guideline quality appraisal has been done by a multidisciplinary panel representative from relevant both clinical and non-clinical areas interested in cancer biomarkers, using AGREE I tool. They were splitted in groups with different tasks, but all the group operated in a coordinated and structured system. In particular, members of the scientific and technical committee defined appropriate information to be included, and the structure of data selection and presentations. They consulted each other during several face-to face meetings and numerous conference calls. Results were discussed and approved in a oneday plenary meeting. Results: We chose to present a comparison of the main guidelines available in tabular form containing a synthesis of the recommendations, associated with the AGREE scores, as the guideline quality appraisal is the necessary completion for a critical comparison of recommendations. We produced also a in-depth assessment of methodological quality of the guidelines presented in a tabular form. Using this approach we described what are the recommendations currently available for each scenario considered and what are the relative levels of evidence and strength of recommendations; we evaluated the quality of selected guidelines; we assessed whether guidelines provide proper indicators that will allow verification in practice. Conclusions: Using the synoptic tables of recommendations, associated with the AGREE we provide clinicians the opportunity to acquire the critical skills necessary to choose sound and valid recommendations.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

S47

O15.3 - How recommendations on biomarkers are presented in guidelines produced in a clinical oriented setting

O15.4 - How recommendations on biomarkers are presented in guidelines produced in a clinical pathology oriented setting

P. Zola, V. Zanfagnin, E. Tripodi

C. M. Sturgeon

Gynaecologic Oncology Unit University of Turin Italy

Department of Clinical Biochemistry, Edinburgh Royal Infirmary, Edinburgh EH16 4SA, UK

Background: Among biomarkers, we consider the CA 125 as a good example of a marker used in diagnosis, treatment and follow-up for patients with cancer, although its utility is still debated. Aims: To present a comparison of recommendations for clinical use of CA 125 facilitating the end-user. Methods: Guidelines published in the last 10 years were reviewed. Results: The NICE™s guideline (2011) about the management in primary care recommends to measure CA125 in symptomatic women; if CA125 ‰35 IU/ml, an ultrasound scan of the abdomen/pelvis is suggested; if positive for ovarian cancer, further investigations are mandatory. If CA125 is normal, or elevated but with a normal ultrasound: asses for other causes of symptoms; if no other cause is found, advise to return if symptoms become frequent or persistent. In Cancer Unit in women with suspected ovarian cancer the dosage of CA125 is indicated, if not already done; in case of women under 40, it is strongly suggested to measure also levels of AFP and beta-hCG, to identify non epithelial ovarian tumors. The Cancer Care Ontario™s document (2011) illustrates that as a stand-alone modality, CA125 is not recommended for differential diagnosis in adnexal masses. The Italian guidelines (2004) recommend to measure the CA125 in association with ultrasound but these tools speed up the diagnosis of only eight months and this does not involve a decrease of mortality. The SIGN guidelines (2003) suggest to refer to gynaecologist patients with a pelvic mass irrespectively of the CA125 value; in secondary care the RMI scoring system is the method for predicting malignancy. During the follow up the ESMO and AIOM guidelines (2008) the CA125 measurement was recommend at each visit. These indications have to be updated taking into account the results of the published RCT (Rustin 2010) showing that the early start of chemotherapy due to a rising of CA125 value does not improve OS compared with delayed treatment until signs/symptoms for women with complete remission after first-line chemotherapy for ovarian cancer. Conclusions: The different guidelines analyzed suggest the use of CA 125 in the management of the patients affected by ovarian cancer. These recommendations have to be updated with the recent RCT Conclusions.

Background: Clinical guidelines for tumour markers provide a sound basis for the use of these tests in routine practice and should be developed as a result of systematic and critical appraisal of the medical and scientific literature. Since the contribution of tumour marker measurements cannot be considered in isolation, recommendations for their use should ideally be contained within broader guidelines considering all aspects of clinical care of the cancer patient. Aims: A multidisciplinary approach to guideline development is always desirable, but inevitably some guidelines focus primarily on clinical aspects (e.g. surgical or medical procedures) while others are more laboratory-oriented. Methods: These differences will reflect not only the membership of the guideline development group but also its remit and the methodological approach adopted. It is often necessary to restrict the number of topics considered for practical reasons (e.g. time constraints). Guideline development groups established by the National Institute of Health and Clinical Excellence (NICE) in the United Kingdom endeavour to do so by using the PICO (patient, intervention, comparison and outcome) framework with between fifteen and twenty associated questions considered a reasonable number. Identifying the most pertinent questions and achieving appropriate balance across what are quite often very wide ranging issues is clearly critical to the success and relevance of the final guidelines. Results: Taking appropriate account of laboratory issues (e.g. method-related differences in results) has been achieved to varying degrees in different guidelines. The National Academy of Clinical Biochemistry (NACB) Laboratory Medicine Practice Guidelines for use of tumour markers in the clinic have been developed following careful review of available literature and other guidelines by sub-committees of both clinical and laboratory-based experts on particular cancers. The NACB recommendations consequently focus somewhat more on relevant laboratory aspects than other guidelines, many of which appear to take little account of methodological considerations even when recommending numerical decision limits (cut-offs) for action. Conclusions: Involvement of laboratory professionals in guideline development groups is highly desirable.

S48

O16 EGTM

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

Conclusions: The identification of pathological conditions and other factors that result in false-positive TM findings when analyzing serous effusions yields high sensitivity and maximum specificity for these determinations.

O16.1 - Diagnostic accuracy of tumor markers in serous effusion J. Trape, R. Molina, F. Sant, J. Montesinos, A. Arnau, J. Franquesa, R. Blavia, E. Martin, E. Marquilles, D. Perich, C. Perez, J. Roca, M. Domenech, J. Lopez, J.M. Badal. Hospital de Sant Joan de Deu. Althaia Xarxa Assistencial de Manresa. Barcelona, Spain Background: Elucidation of the etiology of serous effusions is a challenge for the clinical diagnosis. Several studies have demonstrated the clinical utility of tumor markers (TM) for diagnosing neoplastic effusions, but the results are highly disparate. The reported sensitivity varies from 20% to 80%, depending on the combination of markers, and the specificity from 70% to 100%. The main conditions resulting in falsepositive TM results, tuberculous, complicated parapneumonic effusions and empyema, can be identified by analysis of several biological magnitudes, such as adenosine deaminase (ADA), C-reactive protein (CRP) and % of polymorphonuclear (%PN). Aims: The aim of this study is to evaluate the diagnosis performance of TM in serous effusions and to assess whether ADA, CRP and %PN allow identify false positives. Methods: We determinated in 537 serous effusions CEA, CA19-9, CA15-3, ADA, CRP and leukocyte differential count Results: We classified in two groups: those effusions with ADA 90%PN group B (identify most frequent diseases with no malignant elevations of TM). In group A (n = 438) the effusions were considered to have a malignant etiology when TM values were over the serum cut-off for CEA, CA15-3, and CA19-9; and when fluid/serum ratio was >1.2 for CEA, CA15-3 and CA19-9. Based on these criteria, sensitivity for the combination of all TMs was 77.6% at 100% specificity. In Group B (n = 109) cut-off established for CEA, CA15-3 and CA 19-9 was 50 ng/mL, 70 U/ mL and 7700 U/mL respectively. We obtained a sensitivity of 68.4% at 100% specificity. With a sensitivity for while effusion of 74% at a maximum specificity

O16.2 - Risk assessment evaluation in screening for colorectal cancer Hans J. Nielsen Department of Surgical Gastroenterology 360, Hvidovre Hospital, University of Copenhagen, DK-2650 Hvidovre, Denmark Background: Emerging results indicate that screening improves survival of patients with colorectal cancer. Therefore, screening programs are already implemented or are considered for implementation in Asia, Europe and North America. Aims: How efficient is screening for colorectal cancer? Methods: At present, a great variety of screening methods are available including colono- and sigmoidoscopy, CT- and MR-colonography, capsule endoscopy, DNA and occult blood in faeces, etc. Pros and cons of the various tests, including economical issues, are debated. Although a plethora of evaluated and validated tests even with high specificities and reasonable sensitivities are available. An international consensus on screening procedures is still not established. The rather limited compliance in present screening procedures is a significant drawback. Furthermore, some of the procedures are costly and therefore selection methods for these procedures are needed. Results: Current research into improvements of screening for colorectal cancer includes blood based biological markers, such as proteins, DNA and RNA in combination with various demographically and clinical parameters into a “risk assessment evaluation” (RAE) test. It is assumed that such a test may lead to higher acceptance among the screening populations, and thereby improve the compliances. Furthermore, an addition of media campaigns including social medias may add even more individuals to the screening programs. Conclusions: Implementation of validated RAE and progressively improved screening methods may reform the cost/benefit of screening procedures for colorectal cancer. Therefore results of present research, validating RAE tests, are awaited with interest.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

S49

O16.3 - Tumor markers in personalized medicine

O16.4 - Use of Biomarkers in Screening for Cancer

R Molina

M. J.Duffy

Barcelona, Spain.

Department of Pathology and Laboratory Medicine, St Vincent’s University Hospital, Dublin, UCD School of Medicine and Medical Science, University College Dublin, Dublin, Ireland.

Background: Tumor markers are known for more than 40 years, but the clinical application of most are not clear in the majority of guidelines. Aims: Tumor markers are commonly non-specific with increases in some benign diseases. Methods: The majority are not tissue specific and not clearly related to the tumor origin. Results: Finally their sensitivity is related to tumor stage, with lower sensitivities and specificities in patients with early tumors, invalidating their use in the diagnosis or early diagnosis, in the majority of malignancies. However, tumor marker have been used in the selection of risk patients or in the diagnosis, prognosis and follow-up of several malignancies. Data previously indicated shows why guidelines do not suggest the majority of markers in the majority of malignancies. However, tumor marker characteristcs adapted to the patient situation, with different cut-offs has a real clinical value. PSA, calcitonin, beta HGC, have shown their utility in the diagnosis of high risk patients. They are not specific to make the diagnosis, but allow the selection of patients with high risk of malignancy, that will be demonstrate by other diagnostic methods. Others studies with different tumor markers are being evaluated in colorectal cancer, ovarian cancer and lung cancer. The use of tumor markers in helping in the diagnosis had similar problems: lack of specificity and specificity. Experience shows that false positive results invalidate the use of a cut-off or suggestive value for all population. It is not possible to use CA 19.9 in patients with jaundice, or PSA in patients with prostatitis in a similar way that in a patient without these problems. Tumor marker evaluation may be adapted to the patient situation, using different values to suggest malignancy according them. It is not the same use the tumor markers in early diagnosis than in patients with advance disease, but around 15% of patients with cancer arrived and are diagnosed with loco-regional advance diseases or metastases. Serum concentrations of the majority of tumor markers in advance stage are so high, that there are not problems in the differential diagnosis of other benign diseases. Conclusions: Tumor markers may be useful in the diagnosis, early diagnosis of recurrence or therapy monitoring, if used correctly,ie, by adapting the results to the patient situation and using serial determinations in those cases with doubts.

Background: Screening asymptomatic subjects is a possible approach for reducing both the incidence and mortality of cancer. Aims: The Aims of this presentations are to review the use of biomarkers as screening test for early maligancy Methods: Review of the literature Results: Because of their ease of measurement and relatively low costs, several biomarkers have been evaluated or are currently undergoing evaluation as screening tests for early malignancy. These include the use of vanillymandelic acid and homovanillic acid in screening for neuroblastoma in newborn infants, pepsinogen in screening for stomach cancer, AFP in screening for hepatocellular cancer in high-risk subjects, CA 125 in combination with transvaginal ultrasound in screening for ovarian cancer, PSA in screening for prostate cancer and fecal occult blood testing (FOBT) in screening for CRC. Of these biomarkers, only the use of FOBT has been unequivocally shown to reduce mortality from cancer, i.e., from colorectal cancer. Indeed, 4 large randomised prospective trials have now shown that screening with FOBT significantly reduces mortality from CRC. In contrast to FOBT in CRC, recent results form a large prospective trials evaluating CA 125 with transvaginal ultrasound in screening for ovarian cancer showed no reduction in mortality. The most widely used screening biomarker is PSA for detecting early prostate cancer. It is important to stress however, that the use of PSA in this setting is controversial. It is controversial because of the lack of consistent data as to whether screening reduces mortality from prostate cancer, the possibility of overdiagnosis and thus overtreatment of indolent prostate cancer, and lack of clarity as to what is the best treatment for patients with early prostate cancer. Hopefully, updated results from the ongoing prospective trials will provide answers to these questions. Conclusions: In general, lack of sensitivity for early malignancy and lack of specificity, when combined with the low prevalence of cancer in the community, limit the use of existing biomarkers in screening for cancer.

S50

O17 NEW TREATMENTS, TARGETTING AND NEW APPROACHES

O17.1 - Cyclic beta-interferon-interleukin-2 sequence in addition to antiestrogens significantly prolongs survival of endocrine dependent metastatic breast cancer patients A. Nicolini, P. Ferrari, G. Rossi, A. Carpi, A. Aqhababyan, M. Conte, P. Miccoli Department of Internal Medicine, University of Pisa, Department of Biostatistics, CNR Pisa, Department of Reproduction and Ageing, Department of Surgery, University of Pisa, Italy. Background: In endocrine-responsive metastatic breast cancer patients tumor growth becomes resistant to antiestrogens and many mechanisms have been considered to explain this resistance. We hypothesized that in these patients during clinical benefit (CB) beta-interferon and interleukin-2 can synergize with antiestrogens and delay or overcome the occurrence of this resistance. Aims: To confirm this hypothesis, in 1992 we started an exploratory study. Methods: Among 61 non selected metastatic breast cancer patients who from April 1992 to April 2003 were treated in our Center, 26 (43%) with stable disease (SD) or responsive to first line antiestrogens were consecutively recruited for the addition of immunotherapy with beta-interferon-interleukin-2 sequence cyclically given. Results: In 2005 we reported for the first time on these 26 patients; they showed prolongation of CB and overall survival (OS) more than 3 times than that observed in 30 consecutive historical controls treated with antiestrogen therapy alone. Successively in these 26 and in 6 further recruited patients we showed that cellular immunity is significantly stimulated by interleukin-2 administration only during clinical benefit and that different changes of pro-inflammatory cytokines, CRP and inhibiting factors occur consistent with associated clinical benefit or with disease progression, respectively. In 27 of these 32 patients who progressed during 1st line salvage hormone- plus immunotherapy the lead time at the progression of metastatic disease during therapy was significantly longer than in controls who did not received immunotherapy. Also it was found that CB and OS are related to the type of response and receptor status.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

Conclusions: In the overall, data from this exploratory trial point out that the proposed schedule of immunotherapy in addition to concomitant antiestrogens significantly delays progression of disease and consistently prolongs clinical benefit and overall survival. Therefore, a prospective multicenter randomized trial is mandatory to confirm these findings.

O17.2 - Antifibronectin and antitenascin radioimmunotherapy with small immunoprotein (SIP) in metastatic cancer Paola A. Erba Nuclear Medicine Department, University of Pisa Medical School Background: 131I-L19SIP (Radretumab) is an antibodybased agent that selectively targets newly formed blood vessels within tumor tissues, and is radiolabeled with iodine-131 as the source of beta particles for anticancer therapy. This fully humanized immuno-reagent, binds specifically and with very high affinity to the extradomain B splice variant of fibronectin, an extracellular matrix protein that is abundantly expressed in tissues undergoing rapid physiologic remodeling as well as in newly-formed blood vessels of cancers. It combines the unique feature of pan-tumor-targeting (by binding to an easily accessible and stably expressed neoangiogenesis marker) with a favorable pharmocokinetic profile. The latter is characterized by selective tumor uptake, longlasting tumor residence times (exceeding 3 days), and rapid plasma clearance with an average value of the effective plasma half-life of 15.3 hours (interval range for the 95% C.L.: 12.91-18.79 hours) Aims: To evaluate the efficiency of the antibody-based 131I-L19SIP (Radretumab) on patients in vivo. Methods: When we evaluate patients, we initiate the study with the administration of a diagnostic-level activity (185 MBq) of 131I-L19SIP. Over the subsequent 4–7 days blood samples are taken for estimating the radiodosimetry burden to the bone marrow, and sequential whole-body and SPECT/CT scans were acquired for dosimetric estimates to normal organs and to target lesions. Such data serve to ascertain whether patients met a-priori defined eligibility criteria for subsequent R-RIT. These criteria included either a dosimetric ratio between target lesion and red bone marrow (the critical organ for toxicity) higher than 10:1 or a target-to-background (muscle) ratio higher than 4:1 at 24 or 48 hours post injection. Patients showing selective tumor uptake as specified were offered to

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

receive R-RIT. Post-therapy whole-body and SPECT/CT scans were acquired only at one time-point, between 7– 10 days post-administration. Results: Up to know we have concluded a phase-one study and a phase I-II dose escalation study. Currently the agent is under investigation in two clinical trials in combination with EBRT and chemotherapy for advanced lung cancer and with WBI for brain mult Conclusions: The efficiency of this targeting system is presently under evaluation.

O17.3 - Peritoneal surface malignancy treated by cytoreductive surgery and hyperthermic intraperitoneal chemotherapy: from the bench to the bedside M. Deraco, D. Baratti, S. Kusamura Fondazione IRCCS Istituto Nazionale dei Tumori, Milano Italy Background: Peritoneal Surface Malignancies (PSM),have been considered in the past as terminal conditions. During the last two decades they were approached as a curable disease. Aims: To verify the efficacy of Cytoreductive Surgery (CRS) associated with Hyperthermic Intra Peritoneal Chemotherapy (HIPEC). Methods: 430 patients treated at the Milan NCI and literature data. Results: 1)Pseudomyxoma Peritonei is a rare PSM originated from a mucinous appendiceal tumor. Ten years survival overcome 80% when treated with CRS and HIPEC compared with 3 years median survival when treated with sequential incomplete surgical debulking that is the traditional approach. 2)Diffuse Malignant Peritoneal Mesothelioma: survival grew from 12 months with a sCT treatment to 53 months with CRS + HIPEC. Significant prognostic factors resulted: the epithelial subtype compared with the biphasic or sarcomatoid, the absence of lymph node metastasis, the completeness of cytoreduction scores and the use of HIPEC. 3)Peritoneal Carcinomatosis from Colorectal Cancer : results allow to consider this disease very responsive to treatment; median survival of patients treated with CRS and HIPEC is 30–62 months according to the disease extension.. 4)Peritoneal Carcinomatosis from Gastric Cancer is very aggressive and only patients with very limited disease can benefit of this approach; median survival of patients treated with CRS and HIPEC is 9 months. 5) Peritoneal Carcinomatosis from Epithelial Ovarian Cancer: CRS and HIPEC is very effective in all phases of the

S51

natural history of this disease. Only platinum resistant patients should be excluded to the treatment. Effort should be focused on the upfront treatment where overall survival is 60% at 5 years with 30 months median progression free survival. Conclusions: CRS with HIPEC showed efficacy on outcomes, on Peritoneal Mesothelioma and Pseudomyxoma Peritonei patients as well as in a selected group of patients with Peritoneal Carcinomatosis. Biological aggressiveness, extension of disease and completeness of surgery resulted as the most relevant factors correlated to prognosis. Patient selection according to the prognostic factors is mandatory to ensure appropriate treatment.

O17.4 - A multicenter phase III randomized study comparing two dosing schedules for hypofractionated image-guided radiotherapy in patients with oligometastatic cancer L Forte, M Coppola, A Grandinetti, M Panichi, F Matteucci, F Pasqualetti, C Greco University of Pisa Background: Local aggressive therapy can potentially prolong survival or may even have curative intent in some patients with disease very limited in number and site (oligometastatic disease). As highlighted in several retrospective studies and clinical trials stereotactic body radiation therapy (SBRT) has emerged as one of the local therapy options for oligometastases in various body sites, most commonly in the lung and liver. With the development of new technologies and the introduction of Image Guided Radiotherapy (IGRT) in the clinical practice SBRT has demonstrated promising results in terms of local control and safety and can be used as a real alternative to surgery, even in combination with CT.. With the progress of SBRT in cancer treatment a second issue could be addressed about which schedule of radiotherapy could allow the better outcome. Laboratory evidence and clinical trials suggest that high-dose RT given in a single fraction (24–26 Gy) could achieve tumor eradication like the same biological dose given in few fractions with the same safety. Aims: Our study will compare the local tumor control outcomes, pattern of failure and toxicities of a single dose of 24 Gy with 27 Gy given in three fractions in oligometastatic patients. Methods: This trial is an ongoing, phase III, multicenter (MSKCC and University of Pisa), randomized

S52

study. In the first arm a single fraction of 24 Gy is delivered, in the second 9 Gy in 3 fractions. Our intent is to recruit 200 lesions to be randomized. These lesions must be located outside the brain, lung and liver and more than 3 mm from the spine. We consider oligometastatic a patient with five or less lesions. All radiotherapy treatments are delivered with IGRT. We assess the response to treatment using modified RECIST criteria. Results: Since the opening of this study at the University of Pisa we have randomized 22 lesions, 50% for each arm, 5 patients with prostate cancer, 3 lung, 2 pancreas, 2 kidney, 1 cervical, breast, esophagus, biliary tract, adrenal. Median follow-up is 7 months. Conclusions: To evaluate the toxicities and the responses to different schedules of radiotherapy and their length we need a longer follow-up and an increased enrollment.

O17.5 - Panitumumab in therapy of patients with advanced metastasizing colorectal carcinoma - personal clinical experience L. Holubec, M. Votavova, O. Fiala, VM. Matejka, V. Arnetova, P. Jungmannova, O. Topolcan, J. Finek

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

(31%); PD-12pts (33%). Median TTP was 22.4 weeks. In case of bio-chemical response rate assessment, the CEA, CA 19-9, and TPS tumor markers showed statistically significant decline of values (CEA p-value 0.0266; CA 19-9 p-value 0.0075; TPS p-value 0.0002). Regarding the TK, no significant decline was recorded (p-value 0.1671). The main undesirable effect appeared to be dermal toxicity (grade I-III with 98% of the patients), no serious infusion response occurred within our group of patients. Conclusions: Panitumumab can be used as a highly effective alternative therapy of mCRC in patients with wild type K-RAS. Assessment of the tumor markers: CEA, CA 19-9, and TPS is a suitable way to check effectiveness of the targeted biological therapy. Supported by the research project VZMSM 0021620819, and by the grant IGA MZCR 10230.

O17.6 - Maintenance of the cytostatic effect with additional 5-FU-LV significantly increases survival in gastrointestinal cancer at high risk of relapse A. Nicolini, P. Ferrari, G. Rossi, A. Carpi, S Fancelli, A. Aqhababyan, C.E. Ambrosini, M. Conte, P. Miccoli.

Dept. of Oncology, Univ. Hospital and Medical Faculty Pilsen, Czech Rep.

Department of Internal Medicine, University of Pisa, Department of Biostatistics, CNR Pisa, Department of Reproduction and Ageing, Department of Surgery, University of Pisa, Italy.

Background: Targeted biological therapy as a part of personalized medicine becomes standard therapy for oncology patients. In case of patients with wild-type of K-RAS gene, Panitumumab appears to be an effective possibility of the targeted therapy. Aims: The authors present results of a study which analyzes a group of 41 patients with advanced metastasizing colorectal carcinoma (mCRC) who underwent palliative bio-chemotherapy and were given Panitumumab in the second and further lines of the therapy. Methods: A data file of 41 advanced mCRC patients (30 men, 11 women; median age of 61 years) was evaluated. The primary aim of this study was to assess overall response rate (ORR) and time to progression (TTP), the secondary objective was to assess biochemical response rate through tumor markers (CEA, CA 19-9, TPS, and TK) assessment, to evaluate life quality of the patients and to assess toxicity of the therapy. Results: The therapy response rate was assessed with 36 patients. The ORR was achieved with 36% of the patients /CR-1 pts (3%); PR-12 pts (33%)/; SD-11pts

Background: Prognosis of metastatic colorectal and other locally advanced gastrointestinal (GI) cancers treated with potentially curative resection is poor and median 5-year disease-free (DFS) and overall survivals (OS) range from 10% to 50%. By an extension of a model applied to endocrinedependent breast metastatic breast cancer, we have hypothesized that in these GI cancer patients at high risk of relapse long-term maintenance of a cytostatic effect can interrupt or downregulate the activities of residual cancer cells and halt tumor progression and immune evasion. Aims: In 21 patients we started an exploratory prospective clinical trial with further 5-fluorouracil (5-FU)-leucovorin (LV) cycles regularly given up to the fifth year after the completion of conventional adjuvant chemotherapy and the potentially curative resection. Methods: An active prospective cohort including 21 patients (study group) was matched in a 1:1 ratio with a retrospective parallel control group of 21 patients. The study group received long-term repeated post-adjuvant administration of 5-FU-LV cycles, while the matched control group was conventionally treated. Statistical analysis was performed by Kaplan-Meier method and Cox’s proportional hazard regression model.

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

S53

Results: The five-year disease-free survival (DFS) was 77.0+/−10.1% and 31.7+/−10.6% (p=0.001; hazard ratio (HR) 5.3, 95% C.I.: 1.7–16.1, p=0.003), while the fiveyear overall survival (OS) was 88.0+/−8.1% and 37.0+/− 10.7% (p=0.001; HR 8.9, 95% C.I.: 2.0–39.9, p=0.004) in the study group and in matched controls respectively. Conclusions: These findings suggest a relevant improvement in the outcome of this population by an intermittent and prolonged cytostatic effect with 5-FU-LV.

All parameters were investigated in a Multivariate Analysis: the cytokeratins TPS and Cyfra 21-1 were found to be independent prognostic factors. Conclusions: We demonstrated SCC ( in Squamos Cancers), CEA, TPS and Cyfra21-1 as sensitive and useful markers in H&N Ca. pts. Decreases in marker levels were associated with a longer survival and a favorable response to therapy. Increases in TPS were the most sensitive predictors of advanced disease and poor prognosis, correlating also to stage of disease and nodes.

O17.7 - Diagnostic and prognostic value of tumor markers (SCC, TPS, CYFRA 21-1, CEA) in head & neck cancer patients

O17.8 - Vitamin D status and per-oral Vitamin D substitution in patients suffering R. Klapdor1, S. Klapdor 1,2, E. Richter 2

V Barak, M Wigoda, K Rachmut, B Nisman, T Peretz, J Elidan, M Gross Departments of Oncology and Otolaryngology-Head&Neck Surgery, Hadassah Hebrew -University Medical Center, Jerusalem, Israel. Background: Head & Neck Cancer is a heterogenous disease and patients prognosis is very poor, due to dependence on many factors Aims: To investigate the clinical role and prognostic significance of Tumor Markers (SCC, Cyfra 21-1, TPS, CEA) in Head and Neck Cancer patients (H&N pts). Methods: Serum levels of SCC, Cyfra 21-1, TPS, CEA were evaluated in H&N pts: Larynx −88, Nasopharynx −14, Oral cavity −37, Parotid and other salivary gland malignancies −18, controls-46. Forty eight H&N pts, prior and post therapy (surgery, radiation, or chemotherapy) were investigated and related to clinical status, lymph node status, stage, grade, and response to treatment. Results: In Laryngeal Carcinomas, levels of all markers were high before and decreased significantly after therapy [SCC: 1.46+0.3 to 0.7+0.08 (−p=0.02), TPS: 77.5+8.4 to 53.7+6.1 (p=0.03), Cyfra 21-1: 1.54+0.3 to 1.11+0.17 (p =.0.02), CEA: 23.9+16.2 to 7.9 +1.6 (p=0.02). Therapy of oral cavity SCC reduced significantly the SCC levels from 1.67+0.2 to 0.99+0.19 (p=0.05). Advanced H&N pts (T3/ T4) demonstrated higher levels of all markers compared to early stage disease pts (T1/ T2 ), [SCC: 1.93+0.9 vs 0.88+ 0.17(p= 0.07), TPS: 77.3+ 9.9 vs 61.7+ 6.7,(p =0.012), Cyfra 21-1: 1.63+0.2 vs 0.89+0.2 (p=0.02), CEA:17.9+ 7.8 vs 9.1+2.5]. Significantly higher levels of Cyfra 21-1 were seen in Larynx pts vs Oral Cavity pts (p=0.03). Node positive pts had higher TPS levels than node negative pts −76 +0.91 vs 44 +1.17, (p=0.02) . Significant correlations were demonstrated between response to therapy and markers levels (P=0.0001) Cyfra 21-1 r=0.685, TPS r=0.614, SCC r=0.505., CEA r=0.724.

1Internal Medicine, ZeTDT GmbH, Hamburg Background: Exocrine pancreatic insufficiency (EP) due to PaCa, CP or surgery of PaCa or Stomach Cancer is often associated with maldigestion of fatty soluble vitamins (vit.). Consecutively, metabolic bone and skin diseases, visual problems and coagulation problems may occur. In addition, an important role of Vit. D is discussed during recent years increasingly, e.g. in prevention oncology and cardiology. Recent own pilot studies showed that generally recommended monthly injections of ADEK preparations (10,000 IU) were not able to raise the serum Vit. D in these patients (pat.) to a normal range. Aims: Therefore, we measured prospectively the 25(OH)Dserum levels in pat. with Pancreatic Diseases (PD) and investigated whether it will be possible to supplement Vit. D orally in combination with adequate supplementation of pancreatic enzymes and nutrition advice. Methods: The study was performed in 133 pat. with PaCa or CP. Vit. D was supplemented under sequential serum control (doses between 1,000 IU up to 20,000 IU daily). For comparison, similar studies were done in 114 pat. without known PD (Co). Results: 91% of the pat. with PD showed Vit. D levels below 30 ng/ml (Median: 14 ng/ml), 74% below 20 ng/ml and 32% and 9% below 10 ng/ml and ke above 30 ng/ml in case of individual adaption of the Vit. D uptake, during summer as well as during winter periods “ even in those pat. in which i.m. application of ADEK every 2 to 4 weeks failed to raise the Vit. D levels into the normal range. Surprisingly, the results in the Co-group was similar (87% of the pat. without PD showed vit. D levels below 30 ng/ml, Median: 14,6 ng/ml) and were also easy to normalize by oral intake of vit. D preparations. Conclusions: The majority of pat. with PD showed Vit. D serum levels below 30 ng/ml. According to our studies, however, Vit. D deficiency in pat. with PD can easily be

S54

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

supplemented by oral Vit. D intake. Parenteral application does not seem to be required. This procedure should also be successful in the rare cases of Vit. A deficiency “ a complex in which we are currently interested in.

showing the biological effect of cediranib in glioblastoma cells. Together with the encouraging results obtained in the clinical trials with cediranib, our results suggested cediranib as a good candidate for glioblastoma patients’ treatment.

O17.9 - Effectiveness of receptor tyrosine kinases (RTKs) targeted therapies in glioblastoma

O17.10 - Decreased expression of carbonyl reductase is a significant predictor of outcome in uterine endometrial cancer and induce epithelial mesenchymal transition

O. Martinho1,2, R. Oliveira3, V. Miranda-Gonçalves1,2, C. Clara3, J. Almeida3, RM. Reis1,2,3 1Life and Health Sciences Research Institute (ICVS), Health Sciences School, University of Minho, Braga, Portugal; 2ICVS/3B’s - PT Government Associate Laboratory, Braga/Guimaraes, Portugal; 3Molecular Oncology Research Center, Barretos Cancer Hospital, B Background: The standard treatment for glioblastoma consists of concomitant temozolomide-based chemotherapy and radiotherapy, but virtually all patients recur, with a fatal outcome. Glioblastomas are characterized by higher frequency RTKs alterations, being RTKs targeted therapies under testing for the development of new therapies for these untreatable tumors. The most important RTKs that are targets for the development of molecular therapy in glioblastomas have been PDGFRA, KIT and VEGFR2. Aims: In the present work we aimed to compare the efficacy of three inhibitors of KIT, PDGFRA and VEGFR2-imatinib, sunitinib and cediranib-for glioblastoma treatment. Methods: We screen the effect of cediranib, sunitinib and imatinib in a panel of 10 glioblastoma cell lines in vitro by viability, cell cycle, apoptosis, migration and invasion assays, or in vivo by CAM assay. Intracellular signaling was assessed by western blot. The mutated and amplified status of KIT, PDGFRA and VEGFR2 gene was assessed by sequencing and qPCR, respectively. Results: In the panel of 10 glioblastoma cell lines, cediranib was the most potent, with the lowest mean of IC50 values, followed by sunitinib and finally imatinib. The great efficacy of cediranib in vitro was proven to be due to its higher cytostatic and cytotoxic effects, by inhibiting cellular survival, proliferation, migration, invasion and inducing apoptosis of the cells. In addition, cediranib showed to be a potent inhibitor of MAPK and AKT pathway. It’s anti-tumoral and angiogenic activity was further confirmed in vivo. Finally, we observed that the molecular status of KIT, PDGFRA and VEGFR2 were not predictors of therapy response. Conclusions: This is the first study comparing the effectiveness of those drugs in glioblastoma cells. Importantly, we provide the first glioblastoma preclinical model

A. Murakami K. Yoshidomi and N. Sugino Department of Obstetrics and Gynecology, Yamaguchi University Graduate School of Medicine Background: Carbonyl reductase (CR) is an NADPHdependent, mostly monomeric, cytosolic enzyme with broad substrate specificity for carbonyl compounds. CR appears to be involved in the regulation of tumor progression, although the molecular mechanisms of CR in cancer progression and the clinical significance of CR status were not fully understood. Aims: We investigated the molecular mechanisms by which CR affects cancer cell behavior in vitro and the clinical significance of CR using immunohistochemical analyses of human uterine endometrial adenocarcinoma tissues. Methods: Sense and antisense CR cDNAs were transfected into a human uterine endometrial adenocarcinoma cell line (SNG M) to investigate the role of CR in cancer cell invasion, metastasis, and its molecular mechanisms. Relationship between CR expression analyzed by immunohistochemistry and prognosis such as progression free survival (PFS) and overall survival (OS) were examined in uterine endometrial adenocarcinoma tissues from FIGO stage I-IV (n=109). Results: In the in vitro experiment, suppression of CR increased cancer cell invasion and diminished E-cadherin expression and cellular aggregation. On the other hand, over-expression of CR increased E-cadherin expression and cellular aggregation and decreased cancer cell invasion and migration. The transcriptional suppressors of E-cadherin, Snail and ZEB1, were increased by suppression of CR, and decreased by over-expression of CR. Immunohistochemical analyses showed that the decreased CR expression is significantly related with poor PFS and OS compared with strong CR expression. In the multivariate analyses, the decreased CR expression was the independent prognostic factor for PFS and OS. Conclusions: CR regulates cancer cell invasion via Ecadherin and its repressors in uterine endometrial cancer. The decreased CR expression can be a useful marker to

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

S55

predict an unfavorable clinical outcome in patients with endometrial cancer.

O18 STEM CELLS

O17.11 - New functional biomarkers offer opportunities for early diagnosis and targets for therapy in patients with progressive cervical disease

O18.1 - Cancer cell reprogramming: stem cells differentiation stage factors (SCDSF) and their regulatory influence on tumor growth

WC Hitt1, AF Burnett1, RA Dennis2, MJ Cannon3, K OBrien1, D Rose1, K Shigemasa4, D Bard1, JB Beard1, TJ OBrien1

P. M. Biava, Alberto Frosi

1University of Arkansas for Medical Sciences, Department of OB/GYN Background: Cervical cancer continues to be a major problem in women with gynecologic cancer. Approximately 12,300 women were diagnosed in 2010 and approximately 4,100 died of their disease. To complement pap smears and to provide a relatively easy recognition of disease a blood test with biomarkers with potential for early diagnosis and as targets for treatment of persistent disease could offer patients in general an opportunity for early diagnosis and offer the caregiver a therapeutic window for treatment of persistent disease. Aims: To identify such functional biomarkers our group put together a list of criteria for the identification of functional biomarkers which would qualify as candidates for the role of early diagnosis and therapeutic targets which include the following criteria: overexpressed in tumor cells, exported or secreted from tumor cells, relatively small molecular weight. Methods: We developed redundant primers to gene family signature sequences which allow comparison of families of genes expressed in tumors and normal tissues. By defining which family members were overexpressed in cervical cancer and validating their expression by PCR amplification and Immunohistochemistry, we selected eight candidate biomarkers with potential for use as diagnostic candidates and as therapeutic targets in patients with cervical cancer. Results: For each marker we developed an ELISA assay and examined patients with cervical disease and also followed patients with disease through their treatment protocol. Many patients showed elevated levels (>95% of controls) of one or more markers and proteases commonly peaked during treatment follow-up providing an opportunity for mitigation of tumor growth and spread through inhibition of these serine proteases. Conclusions: New Biomarkers present an opportunity for both early diagnosis and for therapeutic intervention in patients with persistent disease.

Istituto di Ricovero e Cura a Carattere Scientifico (IRCCS) Multimedica-Milano, Italy Background: Experiments on different human tumor cell lines treated with stem cell differentiation stage factors (SCDSF) taken from zebrafish embryos during precise stages of cell differentiation demonstrated a significant slowdown in tumor growth. In addition, a significant decrease in Lewis lung carcinoma injected in C57BL/ Aims: To ascertain if SCDSF could synergistically/additively interact with 5-Fluorouracil (5-FU). Methods: Whole cell-count, flow-cytometry analysis and apoptotic parameters were recorded in human colon cancer cells (Caco2). Two clinical trials were conducted in 179 consecutive patients with intermediate-advanced hepatocellular carcinoma and in 49 consecutive patients with HCC in advanced-terminal stage respectively. All the patients received a product prepared for human therapy containing SCDSF. Results: Cell proliferation was significantly reduced by SCDSF, meanwhile SCDSF+5-FU leads to an almost complete growth-inhibition. SCDSF produces a significant apoptotic effect, meanwhile the association with 5-FU leads to an enhanced additive apoptotic rate at both 24 and 72 hrs. SCDSF alone and in association with 5-FU triggers both the extrinsic and the intrinsic apoptotic pathways, activating caspase 8, 3 and 7. SCDSF and 5-FU alone exerted opposite effects on Bax and Bcl-xL proteins, meanwhile SCDSF+5-FU induced an almost complete suppression of Bcl-xL release and a dramatic increase in the Bax/Bcl-xL ratio. These data suggest that zebrafish embryo factors could improve chemotherapy efficacy. Finally, 19.8% regression, 16% stable disease and a significant difference in survival between the group of patients who responded to treatment versus the group with progression of disease was observed in an open randomized clinical trial on 179 consecutive patients with intermediate-advanced hepatocellular carcinoma. In the other clinical trial on 49 patients with HCC in advanced-terminal stage consecutively recruited and retrospectively evaluated, 38 had “advanced” stage and 11 “terminal” stage. In 5 patients with “advanced” stage a sustained CR was reported (13.1%). No side effects occurred.

S56

Conclusions: SCDSF treatment confirmed its efficacy in patients with “advanced” HCC, in terms of PS and tumoral response

Tumor Biol. (2011) 32 (Suppl 1):S5–S56

O18.3 - The targeted therapy age: potential applications for circulating tumor cells (CTC) investigation Rita Zamarchi

O18.2 - The case for a new tumor stem cell theory: “tumor cell plasticity” JA Radosevich1,2,3, KM Elseth1,2,3, BJ Vesper1,2,3 1Center for Molecular Biology of Oral Diseases, College of Dentistry, University of Illinois at Chicago, Chicago, IL, USA; 2Dept. of Oral Medicine and Diagnostic Sciences, College of Dentistry, University of Illinois at Chicago, Chicago, IL, USA; 3Jesse B Background: The tumor cell theory, as set forth, proposes that there are rare “tumor stem cells” (TSCs) within a tumor mass or tissue culture cell line. It is these rare TSCs that give rise to other TSCs as well as “tumor cells” (TCs). It is proposed that the bulk of a tumor is comprised of TCs and that current treatment modalities work by killing TCs. It is hypothesized that rare remaining TSCs do survive and are the cause of recurrent disease. Within the current theory, TCs produce only daughter TCs; that is, they are not a source of TSCs themselves. FACS sorting using Hoechst 33342 dye (H33342) is the only accepted method to isolate TSCs: TCs are positive, and TSCs are negative in this isolation procedure. Aims: We propose to set forth both experimental and theoretical considerations supporting the need to adapt a new “tumor cell plasticity theory” in place of the current “stem cell theory.” Adaptation of this new theory would fundamentally change TSC research and improve the chances of finding more effective ways to treat cancer. Methods: Using the human lung adenocarcinoma cell line A549 it was first determined that A549 cells could grow in the presence of H33342 at the concentration used to sort cells by FACS. Cells were then treated with H33342 and FACS sorted 5–7 times consecutively on one day. Only a very conservatively selected population of tumor cells which were H33342 positive were collected during each sort; thus, each sorted sample could not contain any dye negative TSCs. The cells were then cultured for 3–5 weeks under various conditions (but without H33342), and then resorted. Results: In each case the populations of TCs that were cultured resulted in a TSC population reappearing in the second sort. Conclusions: The results of these experiments support the idea that either the H33342 FACS sorting method is not reliable or that TCs have the plasticity to become TSCs. In either case, the TSC research community needs to redefine the functional definition of what a TSC really is, if it exists at all.

IOV-IRCCS Background: The capability to block the growth and spread of cancer by interfering with specific molecules involved in tumor growth and progression represents the new deal of oncologic therapy, hopefully leading (at least for carriers sensitizing mutations) to substantial changes in the dismal prognosis and shooting down adverse side effects. To evaluate molecularly targeted therapies early clinical trials should incorporate validated pharmacodynamic biomarkers for giving either proof of mechanism (drug hits target) and proof of concept (tumor responds to drug). Moreover, the development of companion diagnostic test will be mandatory for approving new targeted drugs. Aims: Being the "liquid phase" of solid tumors, the circulating tumor cells (CTC) are generally considered a good source of tumor samples, reflecting "just in time" the disease condition. Notwithstanding the small number of published reports, often devoted to investigation of little pilot groups of patients, achieved data indicate that automated assessment of CTC extending phenotypic analysis to critical targets, allow the effective serial monitoring of patients throughout the continuum of care. Especially, the characterization of CTC unveils modification of the tumor biology after second and/or third line of therapy, potentially disclosing the emergence of additional sensitizing or drug-resistance mutations during therapy. Methods: Recently we developed an M30-integrated CTC assay for monitoring changes in the amount of apoptotic CTC (in addition to CTC count) during therapy. By exploiting this quantitative immunofluorescence, tumor burden in peripheral blood can be accurately quantified showing that live CTC are associated with disease recurrence, consistent with enhanced cell migration and invasion. Results: Here live/apoptotic CTC changes were measured in metastatic renal cancer (mRCC) patients receiving firstline Sunitinib. We find that the presence of EpCam-positive, live CTC predict disease recurrence in individual mRCC patient, being linked with remote relapses under first-line Sunitinib. Conclusions: Original data obtained in mRCC shown how characterizing CTC may help in enabling clinicians to more accurately predict for treatment benefit.