Key points are not available for this paper at this time.
Objective: To construct a radiomics-based model for predicting postoperative recurrence in hepatocellular carcinoma (HCC) patients with vessels encapsulating tumor clusters (VETC) positive based on CT scan. Methods: This retrospective study enrolled patients who underwent surgical resection between January 2016 and January 2024 at Union Hospital, with pathologic confirmation of HCC and VETC status. An external test set was drawn from Chegu Hospital, covering January 2018 to January 2022. Tumor segmentation was performed on portal venous phase CT scan, and then radiomics features were extracted. These features were further analyzed using the LASSO algorithm and combined with clinical features to construct a radiomics-clinical combination model. Model performance was evaluated using receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA). Patients were divided into high- and low-risk groups based on model scores, and Kaplan-Meier (KM) curves were compared. Results: A total of 243 patients were included (median age 56.7 years, 211 males). Nine radiomics features and two clinical features were selected to construct the combined model. The area under the ROC curve (AUC) for predicting 1-year recurrence was 0.898 (95% CI: 0.797-0.999) in the internal test set and 0.804 (95% CI: 0.641-0.967) in the external test set. Calibration curves and DCA demonstrated high net clinical benefit of the combined model. The median recurrence-free survival (RFS) of patients in the high-risk group was significantly lower than that in the low-risk group (internal test set: 13.5 vs 30.0 months, respectively. P=0.004; external test set: 13.0 vs 31.0 months, respectively. P< 0.0001). Conclusion: The radiomics-clinical combination model showed high accuracy for preoperatively predicting recurrence in patients with VETC-positive HCC receiving hepatectomy.
Lei et al. (Sat,) studied this question.