Nomogram for prediction of hepatocellular carcinoma recurrence after liver resection

Background: Hepatocellular carcinoma (HCC) is a common malignancy with high mortality. Liver resection (LR) is a curative treatment for early-stage HCC, but the prognosis of HCC patients after LR is unsatisfactory because of tumor recurrence. Prognostic prediction models with great performance are urgently needed. The present study aimed to establish a novel prognostic nomogram to predict tumor recurrence in HCC patients after LR. Methods: We retrospectively analyzed 726 HCC patients who underwent LR between October 2011 and December 2016. Patients were randomly divided into the training cohort (n = 508) and the testing cohort (n = 218). The protein expression of 14 biomarkers in tumor tissues was assessed by immunohistochemistry. The nomogram predicting recurrence-free survival (RFS) was established by a multivariate Cox regression analysis model and was evaluated by calibration curves, Kaplan-Meier survival curves, time-dependent areas under the receiver operating characteristic (ROC) curves (AUCs), and decision curve analyses in both the training and testing cohorts. Results: Alpha-fetoprotein [hazard ratio (HR) = 1.013, P = 0.002], portal vein tumor thrombosis (HR = 1.833, P < 0.001), ascites (HR = 2.024, P = 0.014), tumor diameter (HR = 1.075, P < 0.001), E-cadherin (HR = 0.859, P = 0.011), EMA (HR = 1.196, P = 0.022), and PCNA (HR = 1.174, P = 0.031) immunohistochemistry scores were found to be independent factors for RFS. The 1-year and 3-year AUCs of the nomogram for RFS were 0.813 and 0.739, respectively. The patients were divided into the high-risk group and the low-risk group by median value which was generated from the nomogram, and Kaplan-Meier analysis revealed that the high-risk group had a shorter RFS than the low-risk group in both the training (P < 0.001) and testing cohorts (P < 0.001). Conclusions: Our newly developed nomogram integrated clinicopathological data and key gene expression data, and was verified to have high accuracy in predicting the RFS of HCC patients after LR. This model could be used for early identification of patients at high-risk of postoperative recurrence. (c) 2024 First Affiliated Hospital, Zhejiang University School of Medicine in China. Published by Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.