ABSTRACT
Objective To compare the CLIP score, the JIS score, and the China staging system (CS) in the prediction of survival of patients with resectable hepatocellular carcinoma (HCC). Methods The Clinicopathologic and follow-up data of 224 patients who underwent hepatic resection for HCC from January 2000 to July 2005 were retrospectively studied. The patient distribution and the survival curve of each staging system were used to compare the ability to stratify and to discriminate prognosis. The likelihood ratio, chi-square test and the linear trend chi-square test were used to compare the homogeneity and the monotonicity of the relationship between stage and mortality rate of each staging system. The increase in the -2 log likelihood statistic on removal of any one staging system was in turn used as a means of ranking the individual staging systems according to their importance within the regression model. The statistical package used was SPSS version 16. 0 and Stata SE version 8.0. Results Based on the China staging system, the percentages of patients categorized as Ⅰa, Ⅰ b, Ⅱa,Ⅱb and Ⅲ a were 14. 3%, 17.4%, 21.9%, 31.7% and 14. 7% respectively, showing excellent stratification ability. However, nearly 81. 6% of the patients were classified as a CLIP score of 0-2, which showed poor stratification ability, and only 3. 1 % of the patients were classified as score 0 category of the JIS scoring system. In the follow-up period, the log-rank test and the corresponding Kaplan-Meier survival curves confirmed each staging system to be able to differentiate patient survival in the different stages. Individual pairwise comparisons revealed inconsistencies across the different staging systems. In particular, using the log-rank test, the JIS scoring system and the China staging system showed significant differences in patient survival on all pairwise comparisons. By contrast, the CLIP scoring system failed to differentiate significantly between score 2 and score 3 patients. The JIS scoring system could identify the best prognostic group who would benefit from curative and aggressive treatments, whereas the discriminatory value of the CLIP score was noted in the intermediate- and advanced-phase HCC patients. The China staging system was shown to have the best homogeneity, overall discriminatory capacity and monotonicity of gradient. The change in the -2 log likelihood statistic on removal of any staging system revealed that for this cohort of patients, the appropriate importance in the ranking of the independent contribution of each factor to the regression model was: CS> CLIP>JIS. Conclusion Among three clinical staging systems, the China staging system had the highest prognostic value, with better stratification and higher discriminatory capacity than the CLIP scoring system and the JIS scoring system for this cohort of patients who received partial hepatectomy for HCC. The CLIP scoring system performed better in identifying the worst prognostic patients.