Cluster analysis of indicators of liver functional and preoperative low branched-chain amino acid tyrosine ration indicate a high risk of early recurrence in analysis of 165 hepatocellular carcinoma patients after initial hepatectomy.

Abstract

Cluster analysis is used for dividing many prognostic indicators, including liver function, tumor progression, and operative variables, into specific clusters. The albumin (ALB), hepatocyte growth factor (HGF), and branched chain amino-acid to tyrosine ratio (BTR) may represent the severity of liver disease and function of the hepatic reserve. We developed the ALB-BTR and HGF-BTR classifications depending on each level to find specific unique subgroups. Our aim was to identify specific subgroups destined for favorable and poor prognoses after initial hepatectomy. Between 2002 and 2008, 165 patients were analyzed retrospectively. Liver function indicators, including BTR, tumor-related factors, and operative variables, were evaluated by cluster analysis with Ward's criterion. The ALB-BTR classification was divided into 4 groups depending on ALB (cutoff value, 4.0 g/dL) and BTR (cutoff value, 6.0). The HGF-BTR classification was also divided into 4 groups depending on HGF (cutoff value, 0.35 ng/mL) and BTR (cutoff value, 6.0). The prognoses of the subgroups were compared by the log-rank test. Cluster analysis divided multiple indicators into 5 different clusters. In each cluster, we further analyzed subgroups using the ALB-BTR and HGF-BTR classification. Mean recurrence-free survival times in ALB-GI (19.1 ± 2.4 months) and HGF-GIII (29.4 ± 3.8 months) were less than their mean overall survival times. Cluster analysis is useful to find similar and different indicators. Even though liver function was well preserved, low BTR could identify early recurrence in hepatocellular carcinoma patients after resection.