Background: In 2002, the American Joint Committee on Cancer and the International Union Against Cancer redefined the T-classification for hepatocellular carcinoma, shifting the cutoff value for tumor size from 2 to 5 cm and giving more emphasis to vascular invasion.
Methods: A retrospective cohort study was conducted on 223 consecutive patients with hepatocellular carcinoma observed between 1990 and 2002. One hundred twelve were resected and considered for retrospective analysis. Univariate and multivariate analyses were performed on several clinicopathologic variables. After classification according to each staging system, the long-term survival of different stages was compared. The prognostic value of each staging system was further evaluated by entering each stage, in turn, into the Cox regression model with other clinicopathologic variables. The median follow-up was 19 months.
Results: On multivariate analysis, the viral etiology of cirrhosis and the presence of multiple nodules were independent prognostic factors. When the new staging system was entered into the multivariate analysis, it was the only independent factor (P = .02). When stratified according to the old tumor-node-metastasis system, there were no significant differences in the survival between stage I and II (P = .14) or between stage IIIA and IVA (P = .33); only the survival of stage II and IIIA was different (P < .01). When stratified according to the new tumor-node-metastasis system, there were significant differences between stage I and II (71.7% vs. 54.7%; P = .02).
Conclusions: The new staging system is a more reliable and objective method for T classification. It is easy to use in clinical practice and is better at stratifying curatively resected patients with respect to prognosis.