We read with interest a paper by Kudo et al. published in a recent issue of HEPATOLOGY that compared the predictive power between Japan Integrated Staging (JIS) system and Cancer of the Liver Italian Program (CLIP) system in a large patient cohort with hepatocellular carcinoma (HCC). 1 The authors concluded that the JIS system is better and may be a preferred staging model. Because the CLIP system has been prospectively validated and proposed as the primary staging system, it would be interesting to examine how these commonly used staging systems were derived and to explore the potential limitations of their conclusion.
The vast majority (96%) of patients in the Kudo et al. study had undergone radical treatment (resection or loco-regional therapy), suggesting that most of them belonged to an early or intermediate stage. This distinct feature made the JIS system, which is intrinsically similar to the Barcelona-ClI `nic Liver Cancer (BCLC) system proposed by the Barcelona group, 2 more favorable. A recent study showed that the BCLC system was the best one compared with the CLIP, Okuda, and other systems in a surgically oriented referral center. 3 It should be noted that the CLIP and Okuda systems were originally derived from a large, unselected population, and the majority of patients were treated conservatively. Therefore, although the selected prognostic predictors are not mutually exclusive, certain risk factors (tumor size ฯฝ 3 or 2 cm in the BCLC or tumor-node-metastasis [TNM]/JIS system) may only be meaningful in the population that is predominantly treated with radical therapy. This may explain why the JIS system is better, because the outcome is intimately associated with baseline demographics and subsequent treatment strategy. In keeping with these findings, another study in which more than half (52%) of patients were treated conservatively because of advanced stage demonstrated that CLIP was a good predictive model. 4 Moreover, in a multicenter survey we have recently shown that the CLIP system was not superior to the JIS system in patients undergoing resection. 5 Therefore, it is not surprising that the JIS system can beat others in an appropriate study environment.
Another important factor that may alter the predictive ability of a staging system is the distribution of patients. The proportion in each category was considered fairly balanced for JIS scores 0 (12%), 1 (31%), 2 (33%), 3 (17%), and 4 or 5 (7%), whereas the distribution in the CLIP model was rather skewed. 1 There will be a substantial degree of swing in the survival curves if the denominator is too small or the predictors are analyzed in an otherwise unbalanced population.
Collectively, the JIS system contains treatment-derived parameters and is similar to the BCLC system that may work well in areas where HCC is diagnosed at a relatively early stage, whereas the CLIP system would only prevail when patients predominantly belonged to an intermediate or late stage, a condition in which aggressive therapy is less likely. It is necessary to consider all risk factors from early to advanced stages when building an ideal staging system for HCC.