RESUMO
<p><b>OBJECTIVE</b>To validate the predictive power of the 5th and 6th editions of TNM staging system (TNM-5, TNM-6) in a Chinese patient cohort with hepatocellular carcinoma (HCC) sized > or = 5 cm after radical hepatectomy.</p><p><b>METHODS</b>Consecutive 121 patients with HCC sized > or = 5 cm undergoing radical hepatectomy between January 1995 and December 2002 were included. The impact of clinicopathological variables on prognosis was determined by univariate and multivariate analyses, after excluding 2 perioperative deaths.</p><p><b>RESULTS</b>In univariate analysis, TNM-5 stage did not show prognostic significance for overall or disease-free survival, as opposed to TNM-6 stage, Edmondson-Steiner grade, portal vein tumor thrombosis (PVTT), vascular invasion, satellite nodule, Child-Pugh grade, and hepatitis B surface antigen (HBsAg) positivity. When these significant variables were entered in multivariate analysis, Edmondson-Steiner grade was the sole independent prognosticator for both overall and disease-free survival, whereas Child-Pugh grade independently influenced disease-free survival. However, TNM-6 stage lost its predictive potential in multivariate analysis.</p><p><b>CONCLUSIONS</b>Neither TNM-5 nor TNM-6 staging system is revealed to be independently prognostic in patients with HCC sized > or = 5 cm after radical hepatectomy. Therefore, TNM-6 calls for more support in many subsets of HCC patients.</p>