A New Scoring System for Gallbladder Cancer: The First Step of a Long Walk

Abstract

In this issue of the Annals of Surgical Oncology Shukla et al. present us with a new scoring system to help in the decision process of the management of patients with gallbladder cancer. Gallbladder cancer is a silent disease, often diagnosed at an advanced stage, when the odds are against the patient for the most part. In some areas of the world such as northern India and Chile, this disease is highly prevalent, possibly due to the high prevalence of gallstones. In Chile, it represents the first cause of cancer death in women. Surgery is the only known curative treatment of gallbladder cancer, provided that a complete macroand microscopic tumor resection is achieved. The importance of a tool to predict resectability of gallbladder cancer patients derives from the fact that frequently these patients are found to have unresectable disease at the time of laparotomy. Until now, resectability of gallbladder cancer has depended greatly on the extension of the disease, ascertained both objectively and subjectively by the surgeon, and on his or her surgical ability, and is usually determined during the operation. Thus, an exploratory laparotomy could be considered a gold standard when it comes to assessing resectability, but the idea of a predictive score is to have reliable information before the operation, avoiding it in those cases that will not benefit from it. Thus, avoiding an unnecessary procedure would result in better patient management, and better administration of human and technical resources. In this regard, it would be a very useful tool. Ideally, a tool to predict the resectability of a tumor should be: (1) simple and easily performed, (2) cheap, (3) widely available, (4) precise, and (5) have a high negative predictive value. Shukla et al. propose such a scoring system. It is based on retrospective analysis of a series of 124 patients treated at Tata Memorial Hospital. With a 0–10 point score based on serum bilirubin, CA 19.9, and computed tomography (CT) scan findings, three groups of patients are identified: A, those highly likely to be resectable; B, those that may be resectable; and C, those highly likely to be unresectable. A testing sample of 335 patients was then retrospectively analyzed, seeking to correlate the prognostic score with the treatment actually offered to the patients; they found a highly significant correlation. The statistical methods used for the design of the score, however, are not described. In the 2004 abstract regarding the scoring system, a reference is made to the use of CA 19.9. The same group reports that there is a high correlation between radiological features, status of resectability, and CA 19.9. This suggests that CA 19.9 may not be an independent variable. The authors do not tell us whether a multivariate analysis was performed to ensure independence of CA 19.9. The fact is that we do not know the true contribution of this test to the overall prognostic score. Other variables were registered in the testing sample, but no further mention of them is made. Perhaps other biochemical markers besides those studied here are worthy of attention. While for the training sample 58% of the patients were considered amenable for surgery, in the testing sample this treatment was offered to 32.5% of the population. This underscores the heterogeneity of the cohorts, and, perhaps, calls for testing in a larger group of patients. Of the 109 patients operated in the testing sample, one would like to know how many Published online August 15, 2008. Address correspondence and reprint requests to: Augusto R. Leon, MD; E-mail: aleon@med.puc.cl