A cohort of 380 patients who underwent curative hepatectomy for CL-HCC at our hospital between 2009 and 2015 were retrospectively studied. We randomly divided the subjects into training (n = 210) and validation (n = 170) groups. Univariate and multivariate survival analysis were used to identify prognostic factors. Visually orientated nomograms were constructed using Cox proportional hazards models. The performance of the nomogram was evaluated by the area under the ROC curve (AUC), calibration curve and compared with the conventional staging systems.
The statistical nomogram for OS built on the basis of ALBI grade, tumor number, tumor size, classification, hepatectomy methods, capsule formation and microvascular invasion (MVI) had good calibration and discriminatory abilities, with AUC of 0.746 (65-month survival). The nomogram for DFS was based on tumor number, tumor size, classification, HBV-DNA load, capsule formation and MVI, with AUC of 0.733 (65-month survival). These nomograms showed satisfactory performance in the validation cohort (AUC, 0.733 for 65-month OS; and 0.702 for 65-month DFS). The AUC of our nomograms were greater than those of conventional staging systems in the validation cohort.
The established nomograms might be useful for estimating survival for patients with CL-HCC after liver resection.