Analysis of Risk Factors and Construction of Risk Prediction Model for Patients With First Recurrence After Initial Curative Resection for Hepatocellular Carcinoma.

Abstract

Objective: To investigate treatment strategies and relevant clinical indicators for patients with first recurrence after initial curative resection for hepatocellular carcinoma (HCC), and to construct a nomogram model for predicting the risk of survival and disease progression, as well as to validate its predictive performance.

Methods: A retrospective analysis was conducted on 200 patients who experienced their first recurrence after initial curative resection for HCC in our hospital between January 2018 and July 2025. Patients were randomly divided into a training cohort and a validation cohort at a ratio of 7:3. Univariate and multivariate Cox regression analyses were performed to identify independent prognostic factors affecting survival and disease status. On the basis of these factors, nomogram prediction models were established respectively for overall survival (OS) and progression-free survival (PFS). The prognostic capabilities were assessed by calibration curves and receiver operating characteristic (ROC) curves. The prediction models derived from the training cohort were validated in the validation cohort to assess their accuracy and feasibility.

Results: Multivariate analysis revealed that independent factors influencing OS included alpha-fetoprotein (AFP) level, maximum tumor area, lymph node metastasis, histologic grade of the tumor at initial resection, and microvascular invasion (MVI). Independent factors affecting PFS included age, AFP level, and histologic grade of the tumor at initial resection. On the basis of these factors, individualized risk prediction models were constructed respectively for OS and PFS in patients with first postoperative recurrence. In the training cohort, the area under the curve (AUC) values for OS at 1-year, 2-year, and 3-year were 0.872, 0.919, and 0.831, respectively; for PFS, the corresponding AUC values were 0.740, 0.746, and 0.748. Calibration curves and ROC curves demonstrated good consistency between model predictions and actual observations. Different treatment strategies did not significantly affect the OS or PFS.

Conclusion: AFP level, maximum tumor area, lymph node metastasis, histologic grade of the tumor at initial resection, and MVI are independent risk factors for OS in patients with first recurrence after initial curative resection for HCC. Age, AFP level, and histologic grade of the tumor at initial resection are independent risk factors for PFS. The risk prediction models developed in this study effectively predict 1-year, 2-year, and 3-year risks of OS and PFS, demonstrating strong predictive performance and providing valuable insights for clinical medical professionals in patients with first recurrence after initial curative resection for HCC.