Development and nomogram prediction of early postoperative recurrence in hepatocellular carcinoma based on preoperative CT imaging radiomic features and serum features related to microvascular infiltration.

Abstract

Background: Hepatocellular carcinoma (HCC) is characterized by high postoperative recurrence rates, and predicting early recurrence is crucial for improving clinical outcomes, yet remains challenging. Both preoperative computed tomography (CT) imaging radiomic features and serum biomarkers related to microvascular infiltration are important indicators of HCC prognosis. This study aimed to develop a nomogram model incorporating both preoperative CT radiomic features and serum biomarkers associated with microvascular infiltration to predict early postoperative recurrence in HCC patients.

Methods: The study included 156 HCC patients who underwent radical surgery at the Tumor Hospital Affiliated to Nantong University between January 2021 and January 2022. Preoperative CT imaging data were obtained for each patient, and radiomic features were extracted using the 3D Slicer software. Preoperative serum biomarkers related to microvascular invasion were collected, including alpha-fetoprotein (AFP), vascular endothelial growth factor A (VEGF-A), Speckled Protein 100 (SP100), and the Fibrosis-4 (FIB-4) index levels. Postoperative follow-up was conducted for 2 years, during which recurrence data were collected. The radiomics score was generated through dimensionality reduction and least absolute shrinkage and selection operator (LASSO) regression analysis. Univariate and logistic regression analyses were used to identify independent risk factors for early postoperative recurrence of HCC. The nomogram model was constructed using R language, and its predictive performance was evaluated using receiver operating characteristic curves, calibration curves, and decision curve analysis curves.

Results: Among the 156 patients, 60 experienced early recurrence, while 96 did not. Feature reduction through LASSO regression identified 10 optimal features from the venous phase and 4 optimal features from the arterial phase, leading to the development of a radiomics score formula. The early recurrence group had significantly higher radiomics scores than the non-early recurrence group [-1.35 (-2.29, 1.21) vs. 0.94 (-0.40, 1.87), P<0.001]. Logistic multivariate regression analysis identified lesion number, Edmondson grade, AFP and VEGF-A levels, and radiomics score as independent risk factors for early postoperative recurrence of HCC (P<0.05). The nomogram model demonstrated high predictive performance with area under the curve (AUC) values of 0.9265 and 0.9255 in the training and internal test sets, respectively. The model demonstrated good net benefit across a threshold range of 0.01-75%, effectively identifying high-risk patients for early postoperative recurrence.

Conclusions: The nomogram model based on preoperative serum biomarkers related to microvascular infiltration and CT radiomic features demonstrated high predictive performance for early postoperative recurrence of HCC. However, further studies, including external validation, are needed to establish the model's generalizability and clinical applicability.