Annals of Surgical Oncology 15(11):3132–3137

DOI: 10.1245/s10434-008-9917-y

A New Scoring System for Gallbladder Cancer (Aiding Treatment Algorithm): An Analysis of 335 Patients Parul J. Shukla, MS, FRCS,1 Rakesh Neve, MS,1 Savio G. Barreto, MS,1 Rohini Hawaldar, BSc, DCM,2 Mandar S. Nadkarni, MS, DNB, MNAMS,1 K. M. Mohandas, MD, DNB,3 and Shailesh V. Shrikhande, MS, MD1

1

Department of Gastrointestinal Surgical Oncology, Tata Memorial Hospital, Parel, Mumbai 400 012, India 2 Clinical Research Secretariat, Tata Memorial Hospital, Parel, Mumbai 400 012, India 3 Department of Digestive Diseases and Clinical Nutrition, Tata Memorial Hospital, Parel, Mumbai 400 012, India

Background: There is currently no preoperative staging/scoring system available for gallbladder cancer. Unfortunately, in gallbladder cancer, patients manifest advanced stages of the disease. There is need for a methodology that can aid accurate preoperative staging and the subsequent treatment algorithm. We thus sought to validate a new scoring system, the Tata Memorial Hospital Staging System (TMHSS), for gallbladder cancer. Methods: TMHSS is based on the cumulative impact of specific features of computed tomographic scan, presence or absence of jaundice, and serum cancer antigen 19–9 levels. This scoring system was first proposed in 2004. Patients with gallbladder cancer were enrolled onto the testing sample for TMHSS to ascertain its validity. A total of 335 consecutive patients with gallbladder cancer who sought care at the Tata Memorial Hospital between May 1, 2005, and December 31, 2006, were studied. Treatment was suggested on the basis of current existing protocols. Each patient was assigned a TMHSS score, and the treatment decision taken was compared with the algorithm generated for each individual score. Concurrence of the decision taken with the score generated algorithm was tested by the Kendall tau-b test. Results: Ordinal-by-ordinal analysis of the value of the test was .75, which showed excellent concurrence and a statistically significant P value (P \ .0001). Conclusion: TMHSS provides an excellent correlative treatment plan for patients with gallbladder cancer. It has the potential to reduce unnecessary surgical explorations and to direct patients to the ideal treatment strategy, thereby offering a degree of prognostication. Key Words: Gallbladder—Scoring—Outcomes—Management—Cancer.

and modest benefits.3,4 In many patients who are offered radical surgery, the disease is found to be too extensive to permit adequate resection, which results in patients undergoing an unnecessary surgical exploration.5 It would be a great help if a preoperative guide existed that facilitated clinical decision making for patients with gallbladder cancer so such exploration would be unnecessary. The currently available staging systems for gallbladder cancer include the tumor, node, metastasis staging system,6 the modified Nevin stating system,7

Gallbladder cancer remains a difficult malignancy to treat, primarily because of the poor prognosis associated with it.1 The only modality of treatment offering a potential of long-term survival remains radical surgery (except for T1a tumors).1,2 Chemotherapy and radiotherapy have limited applications Published online May 6, 2008. Address correspondence and reprint requests to: Parul J. Shukla, MS, FRCS; E-mail: [email protected] Published by Springer Science+Business Media, LLC Ó 2008 The Society of Surgical Oncology, Inc.

3132

3133

NEW SCORING SYSTEM FOR GALLBLADDER CANCER

the Japanese Biliary Surgical Society System,8 and the Bartlett9 staging system. They are all based on histopathological examination of a cholecystectomy specimen and do not have the potential to provide useful preoperative information that would permit management. At the World Congress of the International Hepato-Pancreato-Biliary Association in 2004,10 we proposed a new scoring system, the Tata Memorial Hospital Scoring System (TMHSS), for gallbladder cancer, which is based on radiological, clinical, and biochemical features. The rationale behind this system was to amalgamate the key features in the investigative algorithm to streamline treatment strategies; the objective is to predict resectability and offer prognostication. The inspiration for this scoring system was the Child criteria,11 used widely in evaluating patients with liver disease that is being considered for surgical resection. The aim of this study was to apply TMHSS in a large cohort of patients with gallbladder cancer to validate it. PATIENTS AND METHODS Training Sample Between July 1, 2001, and December 31, 2004, the data of all 124 patients with gallbladder cancer who sought care at the Department of Gastrointestinal Surgical Oncology, Tata Memorial Hospital, were retrospectively analyzed. Careful attention was paid to the group of patients presenting with gallbladder cancer, either for the first time or after undergoing a cholecystectomy elsewhere and then being referred for a revision radical surgery. The investigations performed, especially serum cancer antigen (CA) 19–9, serum bilirubin, and computed tomographic (CT) scan, were carefully assessed for correlation to outcomes of these patients, such as radical surgery, palliative chemotherapy, endoscopic retrograde cholangiopancreatography, and stenting, or simply palliative care. These tests were noted to correlate with outcome. Of the 124 patients, only 72 (58.06%) had disease amenable to resection. CA 19-9 Levels When the serum CA 19-9 values are [ 90 U/mL, 94% of patients had unresectable disease, and when the level rose to [ 450 U/mL, 100% of those patients had unresectable disease. The normal range of CA 19-9 is 0 to 30 U/mL.

TABLE 1. Tata Memorial Hospital scoring system for gallbladder cancer Characteristic

Score

CA 19-9 levels (U/mL) 0–30 30–90 90–450 [450 Serum bilirubin levels (mg/dL) \3 [3 Computed tomographic scan featuresa Normal Gallbladder mass Liver infiltration Medially placed mass/intrahepatic biliary radicle dilatation Metastatic disease

1–4 1 2 3 4 0–2 0 2 0–4 0 1 2 3 4

a

In the presence of more than one finding, the score remains that of the finding with the highest value. Score calculated as A + B + C (maximum score = 10).

Jaundice A total of 14 patients (11.2%) had serum bilirubin levels of [ 3 mg/dL, and all of them had unresectable disease. The median CA 19-9 levels in these patients was 632.4 U/mL. On the basis of these observations, TMHSS was devised incorporating the CT findings, serum CA 199 levels, and serum bilirubin. The total score for TMHSS was intended to be up to a maximum of 10. The CT findings were classified into five groups, from 0 for normal CT to 4 for obvious metastasis; intervening scores indicate gradual increments in invasive pattern of disease on imaging. Any patient with liver metastasis would automatically receive a score of 4. Serum CA 19-9 levels were classified into four groups on the basis of the study conducted above. Because we had found that patients with a serum bilirubin value [ 3 mg/dL were most likely to have inoperable disease, we assigned a score of 2 to a serum bilirubin value of [ 3 mg/dL (Table 1). Testing Sample Between May 1, 2005, and December 31, 2006, all patients with gallbladder cancer who sought care at the Department of Gastrointestinal Surgical Oncology, Tata Memorial Hospital, were included in the testing sample in this study. Each of the patients was thoroughly examined and investigated for confirmation of the diagnosis of the disease and estimation of the extent of the disease for further management. Liver function tests, complete blood counts, CA 19-9 levels, and CT scan were performed in all patients. The diagnosis was confirmed by histopathological Ann. Surg. Oncol. Vol. 15, No. 11, 2008

3134

P. J. SHUKLA ET AL.

TABLE 2. Tata Memorial Hospital gallbladder cancer scoring system and its relationship to interpretation and managementa Group

Score

Interpretation

Management strategy

A B C

0–3 4–6 7–10

Highly likely to be resectable Maybe resectable Highly likely to be unresectable

Surgery (staging laparoscopy—resection) Neoadjuvant options/staging laparoscopy Palliative options (palliative chemotherapy/stenting/symptomatic care)

a

If evidence of metastatic disease appears on computed tomography (score = 4), it is treated as group C disease.

examination of the gallbladder or by guided fineneedle aspiration cytology in patients with metastatic or advanced disease. Patients who had undergone a simple cholecystectomy at another institution and who were referred to our institution for radical surgery were also included after we confirmed the diagnosis of the primary tumor to be in the gallbladder. Variables such as age, sex, clinical examination findings, levels of serum bilirubin, serum transaminases, serum alkaline phosphatase, CA 19-9, and hemoglobin, total and differential white cell counts, prothrombin and activated partial thromboplastin times, chest X-rays, and CT scan findings were recorded in all patients. The follow-up protocol was tailored according to stage of disease, treatment offered, and expected outcome. Establishing a New Prognostic Score We sought to construct a new prognostic model based on the following principles: It is preferable to have break points for continuous variables such as serum bilirubin or CA 19-9 because their distribution is wide, and a single break point may not be optimal. Variables must be those commonly assessed in practice to enable comparison between different institutions. The model should not include established classifications because they may be modified in the future. A total of 354 patients sought care at our department from May 1, 2005, to December 31, 2006. Of these patients, complete records were unavailable of 19, so the analysis was performed with data for 335 patients. There were 227 women (67.8%) and 108 men (32.2%) with a mean age of 51.2 ± 11.1 years (range, 18–81 years). All of the patients had received a diagnosis of gallbladder cancer confirmed by either histopathological examination of the resected specimens (radical/revision radical surgery), review of specimen slides from patients operated and not considered for surgical resection at the Tata Hospital, or by fine-needle aspiration cytology of the gallbladder mass under CT or ultrasound. Treatment decisions were made on the basis of patients’ complete clinical and imaging profile. The Ann. Surg. Oncol. Vol. 15, No. 11, 2008

patients were offered various treatments ranging from simple cholecystectomy for T1a lesions (guided by frozen section) and radical cholecystectomy for tumors [ T1b, revision radical cholecystectomy for patients with incidental gallbladder cancer who had undergone surgery elsewhere, endoscopic retrograde cholangiopancreatography and stenting for patients with obstructive jaundice, palliative chemotherapy for advanced disease, and symptomatic care for advanced malignancy with a poor general condition. Patients were considered to have unresectable disease if they had evidence of liver metastases, peritoneal metastases, noncontiguous organ involvement, positive lymph node station involvement beyond N1 (N1 disease includes nodes on the cystic duct, portal vein in the hepatoduodenal ligament, and the hilum of the liver), and/or involvement of the hepatic artery or the portal vein. To analyze the outcomes of the score, we grouped these patients into two categories. The first group comprised patients to whom curative or potentially curative treatment options, such as surgery, were offered. The second group comprised patients to whom palliative care was provided or to whom palliative treatment options, such as stenting or palliative chemotherapy, were provided. On the basis of our current existing protocols of management, 109 (32.5%) patients underwent surgical exploration; 226 patients (67.5%) were underwent palliative strategies of management. We then scored each patient by TMHSS scheme (Table 2). On the basis of the scoring system, 106 patients, 98 patients, and 131 patients were classified into the groups A, B, and C, respectively. Statistical analysis was performed by SPSS version 14.0 (SPSS, Chicago, IL), and the Kendall tau-b test was used to confirm the concurrence of the score with the actual treatment given.

RESULTS When we cross-tabulated the data, we compared the distribution of the possible treatment modality according to the scoring system with the actual

3135

NEW SCORING SYSTEM FOR GALLBLADDER CANCER

TABLE 3. Data distribution of scores versus actual treatment offered Scorewise stage

Treatment provided Curative No. of patients Percentage within total score Palliative No. of patients Percentage within total score Total

Total number Group A Group B Group C of patients 99 93.4%

10 10.2%

0 0.0%

109 32.5%

7 6.6%

88 89.8%

131 100%

226 67.5%

106

98

131

335

FIG. 2. Univariate linear correlation analysis between cancer antigen 19-9 and the proposed scoring system.

to be .75, which showed excellent concurrence and a significant P value (.0001).

DISCUSSION

FIG. 1. Univariate linear correlation analysis between computed tomographic scan and the proposed scoring system.

treatment given (Table 3). By Pearson correlation, we determined the correlation of each of the factors in the scoring system to the treatment group to be .827 for CT scan (Fig. 1), .821 for CA 19-9 (Fig. 2), and .691 for serum bilirubin. Detailed examination of the data revealed that serum bilirubin, which was high in only 80 patients, had a 100% correlation with the patient being offered palliative care (none of the patients with serum bilirubin [ 3 mg/dL could be offered surgery). By regression analysis with 95% confidence intervals, we assessed the significance of each of these scoring parameters. The CT score, serum bilirubin scores, and the CA 19-9 scores were compared separately with the treatment provided. It was found that all three components of the scoring system attained statistical significance (P \ .0001), indicating that none was more significant than the other. Ordinal-by-ordinal analysis revealed the value of the test

Hawkins et al.5 have cited the importance of preoperatively detecting patients who are unlikely to benefit from a exploratory laparotomy. One of the reasons for this is the increased postoperative morbidity observed in patients with advanced disease. Avoiding unnecessary surgeries reduces the unnecessary cost of investigations and hospitalization, and it also allows patients and physicians to focus on palliation and improving the patient’s quality of life.5 Jaundice has been noted by Oertli et al.12 to be associated with advanced disease. The main causes for hyperbilirubinemia in gallbladder cancer patients are a medially placed tumor infiltrating into the porta hepatis and a lymph nodal mass infiltrating or encircling the common hepatic duct or the common bile duct and causing obstruction. Biliary tree invasion indicates aggressive tumor biology.5 As we found in our scoring system analysis, all 80 patients with bilirubin levels of [ 3 mg/dL had disease that was ultimately suitable only for palliative treatment. CA 19-9 has been routinely used a serum tumor marker in gallbladder cancer as an adjunct to ambiguous or indeterminate radiologic imaging.13 Ritts et al.14 noted 79.4% sensitivity and 79.2% specificity when serum levels were [ 20 U/mL. CT scanning has been widely used in the diagnosis of gallbladder cancer to visualize the appearance of Ann. Surg. Oncol. Vol. 15, No. 11, 2008

3136

P. J. SHUKLA ET AL.

the primary tumor (mass replacing the gallbladder, wall thickening, intraluminal polyp), to study the tumor’s extension into surrounding tissues, and to stage the tumor.15–18 Some researchers, while studying the sensitivity of conventional CT in gallbladder cancer, found that despite the low-moderate sensitivity in the detection of gallbladder cancer extension, CT had a high positive predictive value in determining resectability and thus assisting treatment planning, especially in advanced disease.19,20 Multidetector CT scans allow a faster examination with lower collimation thickness and more reliable volumetric reconstructions. This allows better detection of perivesicular tumor infiltration while minimizing partial volume artifacts, thereby improving T staging of the tumor.18,21,22 The existing protocols relating to gallbladder cancer focus on radical surgery for operable disease and chemotherapy/chemoradiotherapy options for patients with inoperable disease. Although it is true that in some cases the choice of therapy is straightforward, in many cases, this is not the case. Very often, in patients found by imaging to have borderline operable disease, the clinician is faced with the dilemma of whether surgery or some other modality of treatment should be offered. Here, we propose a new scoring system that takes into account the important features of serum bilirubin, CA 19-9, and imaging characteristics useful in prognostic assessment of patients with gallbladder cancer. The score is easy to calculate and is based on variables that are routinely assessed. We envisage that this system—the first of its kind in gallbladder cancer—will be useful in the clinical decision-making process, thereby helping the treating surgeon in deciding the patient’s chances for a curative resection. At the present time, such a scoring system does not exist. The existing staging systems for gallbladder cancer rely primarily on the histopathological examination of the resected specimen. The proposed system has a good discriminant ability, revealing patient subgroups that are good candidates for surgery and are likely to benefit from a radical attempt at resection. It also allows identification of a subgroup of patients who seem to have a borderline chance of resectability according to clinical investigation, but who may benefit from exploratory laparotomy with intent to cure. The final group of patients, with clearly unresectable disease, are unlikely to be candidates for a curative resection. We are hopeful that the proposed TMHSS will be used and evaluated by other similar large-volume centers and surgeons treating gallbladder cancer. The Ann. Surg. Oncol. Vol. 15, No. 11, 2008

true test of this scoring system would be the confirmation of its validity in aiding treatment algorithms. TMHSS complements the existing tumor, node, metastasis and Nevin staging systems and provides a practical, clinically based system. We believe that this will be a valuable tool to guide surgeons in managing gallbladder cancers.

REFERENCES 1. Balachandran P, Agarwal S, Krishnanin N, et al. Predictors of long-tern survival in patients with gallbladder cancer. J Gastrointest Surg 2006; 10:848–54. 2. Foster JM, Hoshi H, Gibbs JF, et al. Gallbladder cancer: defining the indications for primary radical resection and radical re-resection. Ann Surg Oncol 2007; 14:833–40. 3. Todoroki T, Iwasaki K, Irii K, et al. Resection combined with intraoperative radiotherapy (IORT) for stage IV (TNM) gallbladder carcinoma. World J Surg 1991; 15:357–62. 4. Pradeep R, Kaushik SP, Sikora SS, et al. Predictors of survival in carcinoma of the gallbladder. Cancer 1995; 76:1145–9. 5. Hawkins WG, DeMatteo RP, Jarnagin WR, et al. Jaundice predicts advanced disease and early mortality in patients with Gallbladder Cancer. Ann Surg Oncol 2004; 11:310–5. 6. Wittekind C, Greene FL, Mutter RVP, et al. (2004) TNM staging of gallbladder cancers. In: TNM Atlas. 5th edition. New York: Springer. pp 124–34. 7. Donohue JH, Nagorney DM, Grant CS, et al. Carcinoma of the gallbladder. Does radical resection improve outcome? Arch Surg 1990; 125:237–41. 8. Onoyama H, Yamamoto M, Tseng A, et al. Extended cholecystectomy for carcinoma of the gallbladder. World J Surg 1995; 19:758–63. 9. Bartlett DL, Fong Y, Fortner JG, et al. Long-term results after resection for gallbladder cancer. Ann Surg 1996; 224:639–46. 10. Shukla PJ, Nadkarni MS, Shrikhande SV. Proposal for a new clinico-biochemical-radiological staging system for gallbladder cancer (abstract). HPB 2004; 6(Suppl 1):97. 11. Child CG III, Turcotte JG. (1964) Surgery and portal hypertension. In: Child CG III (ed). Major Problems in Clinical Surgery. WB Saunders, Philadelphia. pp 11–85. 12. Oertli D, Herzog U, Tondelli P. Primary carcinoma of the gallbladder: operative experience during a 16-year period. Eur J Surg 1993; 159:415–20. 13. Bartlett DL, Fong Y. (2000) Tumour of the gallbladder. In: Blumgart LH, Fong Y (eds). Surgery of the liver and biliary tract. 3rd edition.WB Saunders, Philadelphia. pp 993–1016. 14. Ritts RE, Nagorney DM, Jacobson DA, et al. Comparison of preoperative serum CA 19-9 levels with results of diagnostic imaging modalities in patients undergoing laparotomy for suspected pancreatic or gallbladder disease. Pancreas 1994; 9:707–16. 15. Franquet T, Montes M, Ruiz de Azua Y, et al. Primary gallbladder carcinoma: imaging findings in 50 patients with pathologic correlation. Gastrointest Radiol 1991; 16:143–8. 16. Tsuchiya Y. Early carcinoma of the gallbladder: macroscopic features and US findings. Radiology 1991; 179:171–5. 17. Kumar A, Aggarwal S. Carcinoma of the gallbladder: CT findings in 50 cases. Abdom Imaging 1994; 19:304–8. 18. Rodriguez-Fernandez A, Gomez-Rio M, Medina-Benitez A, et al. Application of modern imaging methods in diagnosis of gallbladder cancer. J Surg Oncol 2006; 93:650–64. 19. Ohtani T, Shirai Y, Tsukada K, et al. Spread of gallbladder carcinoma: CT evaluation with pathologic correlation. Abdom Imaging 1996; 21:195–220.

NEW SCORING SYSTEM FOR GALLBLADDER CANCER

20. Misra S, Chaturvedi A, Misra NC, et al. Carcinoma of the gallbladder. Lancet Oncol 2003; 4:167–76. 21. Yoshimitsu K, Honda H, Shinozaki K, et al. Helical CT of the local spread of carcinoma of the gallbladder: evaluation

3137

according to the TNM system in patients who underwent surgical resection. Am J Roentgenol 2002; 179:423–8. 22. Kim BS, Ha HK, Lee D, et al. Accuracy of CT in local staging of gallbladder carcinoma. Acta Radiol 2002; 43:71–6.

Ann. Surg. Oncol. Vol. 15, No. 11, 2008