Nomogram model based on dual-energy CT quantitative parameters and radiomics for preoperative prediction of microvascular invasion in hepatocellular carcinoma

Abstract Objective To explore the potential value of a nomogram model based on dual-energy CT (DECT) quantitative parameters and radiomics for preoperative prediction of microvascular invasion (MVI) in hepatocellular carcinoma. Methods A retrospective analysis was conducted on 145 patients with hepatocellular carcinoma who underwent surgical resection. The cohort was randomly divided into a training cohort (101 cases) and a validation cohort (44 cases) at a ratio of 7:3. Clinical and imaging data, including laboratory tests, tumor imaging features, and quantitative parameters of DECT, were analyzed. Intraclass correlation coefficient, Pearson correlation analysis, and least absolute shrinkage and selection operator regression were employed to select radiomics features from portal venous phase (PV) and virtual monoenergetic image (VMI, 40 keV), which were then used to calculate the radiomics score (Rad-score). Logistic regression was used to establish clinical and radiomics models for predicting MVI, and a combined clinical-radiomics model was further established. The performance of the models was assessed in terms of discrimination, calibration, and clinical application. Results Gender, tumor size, normalized iodine concentration, PV Rad-score and VMI Rad-score are independent predictors of MVI. The Area under the receiver operating characteristic curve of the clinical model, radiomics model, and combined model for predicting MVI in the training cohort were 0.683, 0.886, and 0.898, respectively, and in the validation cohort were 0.702, 0.829, and 0.844, respectively. Calibration and clinical decision curves indicated that the combined model had a higher fit and clinical benefit. Conclusion Nomogram model based on DECT quantitative parameters and radiomics can effectively predict MVI, which may be of value for clinical decision-making in hepatocellular carcinoma patients.