Development and validation of an imaging-based nomogram: predicting Ki-67 status and prognosis in hepatocellular carcinoma

To develop and validate a contrast-enhanced computed tomography (CT)-based radiomics nomogram for the preoperative prediction of Ki-67 expression status and its prognostic significance in patients with hepatocellular carcinoma (HCC). This retrospective study enrolled 167 HCC patients who underwent curative surgical resection. Preoperative arterial-phase CT images were manually segmented, and 851 radiomics features were extracted. Features with high inter-observer agreement were retained and subsequently selected via least absolute shrinkage and selection operator (LASSO) regression to construct a radiomics signature (Rad-score). Clinical predictors were identified through univariate and multivariate logistic regression. A radiomics nomogram was developed by integrating independent clinical predictors with the Rad-score. The discriminatory performance was assessed using the area under the receiver operating characteristic curve (AUC). Calibration and clinical utility were evaluated with calibration curves and decision curve analysis (DCA), respectively. Kaplan-Meier survival analysis was performed to assess the prognostic stratification ability based on the nomogram-predicted Ki-67 status. Tumor size was independent clinical predictor of high Ki-67 expression. The Rad-score, derived from 10 selected features, was significantly higher in the high Ki-67 group. The radiomics nomogram demonstrated the best predictive performance, with AUCs of 0.869 (training cohort) and 0.798 (validation cohort), outperforming the clinical model (AUCs: 0.722 and 0.673) and the Rad-score alone (AUCs: 0.835 and 0.793). DCA confirmed its superior clinical net benefit. The nomogram effectively stratified patients into groups with significantly different overall survival, correlating with pathological Ki-67 status. The CT-based radiomics nomogram noninvasively and accurately predicts Ki-67 expression in HCC preoperatively and effectively stratifies patient prognosis. This tool integrates imaging phenotypes with clinical variable, offering a practical framework for individualized risk assessment and treatment optimization.