Nomogram for predicting early cancer-related death due to recurrence after liver resection in hepatocellular carcinoma patients with Barcelona Clinic Liver Cancer (BCLC) stage B/C: a multicenter study.

Background: Early identification of the risk of early cancer-related death (within one year, ECRD) due to recurrence after liver resection for hepatocellular carcinoma (HCC) patients with Barcelona Clinic Liver Cancer (BCLC) stage B/C is important for surgeons to make clinical decisions. Our study aimed to establish a nomogram to predict the ECRD due to recurrence for HCC patients with BCLC stage B/C.

Methods: A total of 672 HCC patients with BCLC stages B/C from four medical centers between January 2012 and December 2018 were included in our study. The patients were randomly divided into a training cohort (n = 404) and a validation cohort (n = 268) at a ratio of 6:4. The least absolute shrinkage and selection operator (LASSO) logistic regression model was used to establish a nomogram model.

Results: In our LASSO-logistic regression model, three variables were independently associated with the ECRD due to recurrence: the alpha-fetoprotein-tumor burden score (ATS score, Odd Ratio [OR]: 1.12, p = 0.001), BCLC stage (OR: 4.39, p < 0.001) and the aspartate transaminase (AST) to alanine transaminase (ALT) ratio (AAR, OR: 1.49, p = 0.027) and we established the nomogram model based on these three variables. The nomogram model showed superior predictive ability in the training cohort (Area under the curve [AUC]: 0.754, 95%CI: 0.703-0.804) and the validation cohort (AUC: 0.741, 95%CI: 0.660-0.823). Compared with the ATS score, BCLC stage and AAR, the nomogram both had better predictive ability in both the training cohort (ATS score, AUC: 0.699, 95%CI: 0.646-0.752, p = 0.010; BCLC stage, AUC: AUC: 0.684, 95%CI: 0.637-0.732, p < 0.001; AAR, AUC: 0.593, 95%CI: 0.522-0.663, p < 0.001) and the validation cohort (ATS score, AUC: 0.659, 95%CI: 0.577-0.740, p = 0.002; BCLC stage, AUC: 0.688, 95%CI: 0.622-0.753, p = 0.009; AAR, AUC: 0.540, 95%CI: 0.436-0.645, p < 0.001).

Conclusions: We established a nomogram that had excellent predictive power for predicting the ECRD due to recurrence in HCC patients with BCLC stage B/C, which might help surgeons to avoid futile liver resection.