Do patient- and tumor-related factors predict the peritoneal spread of pancreatic adenocarcinoma? Ingmar Ko¨nigsrainer • Derek Zieker • Stephan Symons • Katharina Horlacher Alfred Ko¨nigsrainer • Stefan Beckert
Received: 6 December 2011 / Accepted: 11 December 2012 Ó Springer Japan 2013
Abstract Purpose In pancreatic cancer, the presence of peritoneal carcinomatosis (PC) precludes the possibility of a surgical cure, irrespective of the resectability of the primary tumor. However, peritoneal spread cannot be reliably detected radiographically during preoperative tumor staging. Methods The pancreatic adenocarcinoma database of the Tu¨bingen Comprehensive Cancer Center included 29 patients in whom PC was incidentally detected during the surgery. These patients were retrospectively compared for patient- and tumor-related factors with 29 randomly selected patients without PC who underwent curative resection. Results Clinical jaundice and diarrhea were more frequently present in patients without PC. The CA 19-9 levels were significantly higher in patients with PC compared to those in patients without PC. No other differences were observed in the patient- or tumor-related factors between the two groups. Conclusion In pancreatic cancer patients, markedly elevated CA 19-9 levels may serve as surrogate marker for peritoneal dissemination, irrespective of the local resectability of the tumor. In such patients, laparoscopy should
I. Ko¨nigsrainer (&) D. Zieker K. Horlacher A. Ko¨nigsrainer S. Beckert Department of General, Visceral and Transplant Surgery, University of Tu¨bingen Comprehensive Cancer Center, Hoppe-Seyler-Strasse 3, 72076 Tu¨bingen, Germany e-mail: [email protected] S. Symons Center for Bioinformatics Tu¨bingen, Sand 14, 72076 Tu¨bingen, Germany
be considered as an additional staging tool to rule out peritoneal carcinomatosis. Keywords Pancreatic cancer Peritoneal carcinomatosis Risk factors CA 19-9
Introduction Pancreatic cancer is a very aggressive tumor entity with a generally poor prognosis. Surgical resection is the only curative treatment strategy, with five-year survival rates of up to 25 % [1–4]. In unresectable pancreatic cancer, palliative chemotherapy or best supportive care is indicated. The presence of peritoneal carcinomatosis (PC) precludes the possibility of a surgical cure, regardless of whether the primary tumor is resectable. However, since PC cannot be reliably detectable in the course of preoperative staging, it is often incidentally found during the surgery. In this case, the primary tumor is not resected and the procedure is terminated unless there is evidence of biliary or intestinal obstruction as the rationale for biliary or gastrointestinal bypass surgery. Since surgical trauma may induce rapid tumor progression, it is crucial to identify PC patients prior to surgery to avoid unnecessary laparotomy [5–8]. Although there are established risk factors for surgical unresectability, no parameters are available so far to distinguish between patients with and without PC, since the presence of PC is not necessarily associated with unresectability of the primary tumor. The CA 19-9 tumor marker is one of the factors predicting unresectability, but its role in predicting PC remains unclear [9, 10]. We retrospectively compared the tumor- and patient-related variables in patients with and without intraoperatively confirmed PC who were operated on with the intention of curative resection.
Patients and methods Between 1984 and 2005, 523 patients with pancreatic adenocarcinoma were treated at the Comprehensive Cancer Center of Tu¨bingen University Hospital, Germany. Information on all patients was collected in a database. The patient data included the sex, age, history of clinical jaundice, abdominal pain, weight loss, smoking, alcohol abuse, diabetes, family pre-disposition, diarrhea and duration of clinical symptoms as described by patients and documented in their records. From this database, 29 patients with peritoneal carcinomatosis that was incidentally detected intraoperatively (PC group) were identified. These patients were originally operated on with curative intent after showing no signs of PC in their CT scan. However, because PC was surgically detected, the primary tumor was not resected, and biliary or gastric bypass was performed only if there was biliary or intestinal obstruction. Out of the remaining 494 patients in the database, a random numeric patient list was created. Control patients (non-PC group) were selected in ascending order by skipping patients that did not undergo curative resection. This particular selection protocol was continued until the control group (non-PC group) consisted of 29 patients. The patientand tumor-related factors were retrospectively analyzed and compared in the PC and non-PC groups. Patient-related factors The patients’ weight loss, age, serum CA 19-9 levels, presence of type II diabetes mellitus, clinical jaundice, smoking, evidence of alcohol abuse, familial pre-disposition, diarrhea, duration of clinical symptoms and abdominal pain were retrospectively evaluated by an analysis of the patient records. Patients who reported consuming more than two alcoholic beverages per day on a regular basis were considered to have abused alcohol. The presence of scleral icterus was considered to indicate clinical jaundice. Family pre-disposition was defined as one or more relatives who had died from pancreatic carcinoma. Weight loss was calculated as the weight loss per week as described by patients.
compared using the Chi square test, and quantitative differences were compared using the Mann–Whitney U test. A value of p \ 0.05 was considered to be significant. The SPSS version 13.0 software program (SPSS, Chicago, Illinois, USA) was used for all statistical analyses.
Results Both groups were comparable with respect to the baseline demographic data (Table 1) and risk factors (Table 3). The serum CA 19-9 level was significantly higher in patients with PC than in patients without PC (2,330 U/ml [16–41,000] versus 387 U/ml [18–27,950]; p = 0.041). Clinical jaundice and diarrhea were more frequently detected in the non-PC group (24 versus 59 %; p = 0.008 and 21 versus 3 %; p = 0.044) (Table 2). The tumor size did not differ significantly between the two groups (p = 0.188), with a statistical trend toward there being more pancreatic head tumors in the non-PC group (p = 0.127). No further differences were observed with respect to the patient- or tumor-related factors (Table 1).
Discussion The clinical symptoms of pancreatic cancer include jaundice, pain, weight loss and diarrhea. If PC is evident, the prognosis is even worse, and palliative chemotherapy or Table 1 Patient demographic data Patients with PC
Patients without PC
CA 19-9 (U/ml)
Table 2 Clinical symptoms of the patients in both groups
Tumor-related factors The tumor size and location were determined by CT scans. The tumor location was categorized as pancreatic cancer of the head or body/tail. The largest tumor diameter seen on the CT scan was taken as the tumor size. Statistical analyses The data are presented as the medians [min–max] or n (%), unless stated otherwise. Qualitative differences were
Patients with PC
Patients without PC
Weight loss in kg/week
Duration of symptoms prior to the diagnosis of pancreatic cancer
Pancreatic head cancer
Surg Today Table 3 Risk factors for pancreatic cancer
Patients with PC
Patients without PC
Family history of pancreatic cancer
best supportive care is indicated. PC is often found incidentally in patients who are operated on with curative intent. CT scanning is not sufficient to detect small-volume PC, and staging laparoscopy is necessary to identify patients with PC, helping to avoid an unnecessary laparotomy . However, laparoscopy is currently not performed very frequently. This investigation compared two groups of patients (with or without PC), and both of whom were operated on with a curative intent. The accepted risk factors for pancreatic cancer, such as nicotine or alcohol abuse, family history of pancreatic cancer or chronic pancreatitis did not differ between the two groups. Moreover, we observed no preoperative differences in patients with regard to weight loss or abdominal pain. However, the patients in the non-PC group had a significantly higher rate of biliary obstruction, clinically designated as scleral icterus. One explanation for this could be that the nonPC group had a tendency to have more pancreatic head tumors, causing stenosis of the bile duct. Because these tumors are more likely to cause early clinical symptoms, pancreas head tumors might be associated with PC less frequently than are cancers of the pancreatic corpus or tail. In fact, icterus is a very early sign of a pancreatic head malignancy. Another sign was the higher incidence of diarrhea in these patients. However, again, pancreatic head tumors or tumors of the processus uncinatus seem to be associated with a higher incidence of diarrhea, so it may be unrelated to PC. Even though an elevated CA 19-9 level has been described as a significant risk factor for unresectability and poor survival [9, 10], its predictive value for anticipating peritoneal carcinomatosis remains unclear. In all our patients with a resectable primary tumor, the level of CA 19-9 was significantly higher in patients with PC as opposed to those without PC. Hyperbilirubinemia may be associated with falsely high levels of CA 19-9. However, since clinical jaundice was more frequently detected in the non-PC group, it is not likely that the elevated CA 19-9 levels in our investigation resulted from increased serum bilirubin concentrations. Nevertheless, hyperbilirubinemia might explain, at least in part, the almost ten times higher (in median) CA 19-9 levels in the non-PC group.
Since PC can already be present even though the primary tumor is still resectable, high CA 19-9 levels might not only be predictive of local unresectability but also serve as a surrogate marker for peritoneal dissemination. In clinically asymptomatic patients, laparoscopy should be considered as an additional staging tool to rule out peritoneal spread and avoid unnecessary explorative laparotomy, which might itself provoke further tumor growth [5–8]. Moreover, in patients with high CA 19-9 levels, other authors have already shown staging laparoscopy to be beneficial with respect to the local tumor resectability . However, our present study is limited by the low number of patients, which does not allow a statistically solid CA 19-9 cut-off level to be calculated. Future studies with a larger number of patients will be needed to confirm our results and define a cut-off level for CA 19-9 that predicts the presence of PC. In conclusion, in patients with high CA 19-9 levels and no radiographic signs of local unresectability or peritoneal spread, laparoscopy should be considered as an additional staging tool prior to laparotomy.
References 1. Warshaw AL, Ferna´ndez-del Castillo C. Pancreatic carcinoma. N Engl J Med. 1992;326:455–65. 2. Friess H, Kleeff J, Fischer L, Mu¨ller M, Bu¨chler MW. Surgical standard therapy for cancer of the pancreas. Chirurg. 2003;74: 183–90. 3. Sener SF, Fremgen A, Menck HR, Winchester DP. Pancreatic cancer: a report of treatment and survival trends for 100,313 patients diagnosed from 1985–1995, using the National Cancer Database. J Am Coll Surg. 1999;189:1–7. 4. Malleo G, Marchegiani G, Salvia R, Butturini G, Pederzoli P, Bassi C. Pancreaticoduodenectomy for pancreatic cancer: the Verona experience. Surg Today. 2011;41:463–70. 5. Van den Tol MP, ten Raa S, van Grevenstein WM, van Rossen ME, Jeekel J, van Eijck CH. The post-surgical inflammatory response provokes enhanced tumour recurrence: a crucial role for neutrophils. Dig Surg. 2007;24:388–94. 6. van Rossen ME, Hofland LJ, van den Tol MP, van Koetsveld PM, Jeekel J, Marquet RL, van Eijck CH. Effect of inflammatory cytokines and growth factors on tumour cell adhesion to the peritoneum. J Pathol. 2001;193:530–7. 7. Ko¨nigsrainer I, Zieker D, Beckert S, von Weyhern C, Lo¨b S, Falch C, et al. Local peritonectomy highly attracts free floating intraperitoneal colorectal tumour cells in a rat model. Cell Physiol Biochem. 2009;23:371–8. 8. Raa ST, Oosterling SJ, van der Kaaij NP, van den Tol MP, Beelen RH, Meijer S, et al. Surgery promotes implantation of disseminated tumor cells, but does not increase growth of tumor cell clusters. J Surg Oncol. 2005;92:124–9. 9. Satoi S, Yanagimoto H, Toyokawa H, Inoue K, Wada K, Yamamoto T, et al. Selective use of staging laparoscopy based on carbohydrate antigen 19–9 level and tumor size in patients with radiographically defined potentially or borderline resectable pancreatic cancer. Pancreas. 2011;40:426–32.
Surg Today 10. Mehta J, Prabhu R, Eshpuniyani P, Kantharia C, Supe A. Evaluating the efficacy of tumor markers CA 19-9 and CEA to predict operability and survival in pancreatic malignancies. Trop Gastroenterol. 2010;31:190–4.
11. Contreras CM, Stanelle EJ, Mansour J, Hinshaw JL, Rikkers LF, Rettammel R, et al. Staging laparoscopy enhances the detection of occult metastases in patients with pancreatic adenocarcinoma. J Surg Oncol. 2009;100:663–9.