A clinical and contrast-enhanced CT-based model for preoperative prediction of CK19- and GPC3-positive dual-phenotype hepatocellular carcinoma.

Abstract

Objective: To explore the predictive value of preoperative clinical and multiphase dynamic contrast-enhanced CT (CECT)-based qualitative and quantitative features for identifying Cytokeratin 19 (CK19)- and Glypican 3 (GPC3)-positive dual-phenotype hepatocellular carcinoma (DPHCC_CG+).

Materials and methods: A total of 363 HCC patients who received preoperative CECT and surgical resection from two medical centers were retrospectively included. Patients were divided into DPHCC_CG+ (CK19- and GPC3-positive) or Non-DPHCC_CG+ (CK19- and/or GPC3-negative) groups based on histopathology. Qualitative and quantitative CECT features, along with clinical variables, were compared between the two groups. Univariate and multivariate logistic regression analyses were performed to identify independent predictors of DPHCC_CG+. Three predictive models (CT qualitative-clinical, CT quantitative-clinical and combined model) were developed in the training cohort and externally validated. Model performance was assessed and compared using receiver operating characteristic (ROC) analysis and the DeLong test.

Results: Both the CT qualitative-clinical and CT quantitative-clinical models demonstrated good predictive performance for identifying DPHCC_CG + in both the training and external validation cohorts (AUCs > 0.79). Independent predictors of DPHCC_CG+ included serum AFP ≥ 400 ng/mL, presence of mosaic appearance, arterial phase tumor-to-aorta attenuation ratio (A-TAR ≤ 0.262), and delayed phase normalized washout ratio (D-NWR ≤ -0.141). The combined model integrating these predictors achieved optimal diagnostic performance, with AUCs of 0.878 (95% CI: 0.822-0.935) in the training cohort and 0.806 (95% CI: 0.723-0.889) in the external validation cohort.

Conclusion: The combined model integrating preoperative serum AFP level with CECT qualitative and quantitative parameters enables effective, noninvasive identification of DPHCC_CG+ before treatment, providing potential clinical value for risk stratification and personalized management of HCC patients.