The hepatoma arterial-embolization prognostic (HAP) score and its several modifications predict survival outcomes in patients with hepatocellular carcinoma (HCC) treated with trans-arterial chemoembolization (TACE). We investigated whether HAP-based risk score is applicable in patients treated with TACE due to recurred HCC after curative resection.Methods
A total of 448 patients with HCC who underwent curative resection between 2003 and 2015 were enrolled. Cox regression analyses and area under the curves (AUC) were used to identify risk factors and to calculate the predictive performance of risk scores, respectively.Results
The median age of the study population (378 men, 70 female) was 59.4 years. The median time from resection to recurrence was 17.7 (interquartile range, 7.3–37.1) months. Multivariate analysis indicated that alpha-fetoprotein >400 ng/mL (hazard ratio [HR]=2.367; 95% confidence interval [CI] 1.603–3.495), and serum albumin <3.6 g/dL (HR=2.072; 95% CI 1.449–2.964), tumour number ≥2 (HR=1.813; 95% CI 1.362–2.415), tumour size >7 cm (HR=0.971; 95% CI 0.416–2.269), segmental portal vein invasion (HR=2.695, 95% CI, 1.620–4.485), and time from resection to recurrence <2 years (HR=1.630, 95% CI 1.287–2.066) were the independent predictors for survival (all p<0.05). The AUC to predict survival at 3 and 5 years was 0.713, and 0.649, respectively, for modified HAP-II, which were higher than those of HAP (0.602 and 0.584) and mHAP (0.606 and 0.589). When HAPpostop was established according to multivariate analysis, the AUC to predict survival at 3 and 5 years were 0.799 and 0.735, respectively, which were significantly higher than those of other HAP-based models (all p<0.05).Conclusions
The HAP-based risk models significantly predicted survival in patients treated with TACE due to recurred HCC after curative resection. However, HAPpostop showed superior performance in this cohort.