Deep Learning Based Auto-Segmentation and RECIST Evaluation After Concurrent Chemo-Radiotherapy in Locally Advanced Hepatocellular Carcinoma Patients

Background and Purpose: Hepatocellular carcinoma (HCC) is the third leading cause of cancer-related mortality, and intrahepatic progression after treatment is common. Accurate tumor evaluation is essential for treatment decisions but remains challenging due to tumor heterogeneity, cirrhotic liver background, and treatment-related artifacts. This study investigated the feasibility of a deep learning–based auto-segmentation approach for response evaluation in locally advanced HCC treated with concurrent chemoradiotherapy (CCRT). Methods: We retrospectively analyzed 83 patients with locally advanced, treatment-naïve HCC who underwent definitive CCRT between 2016 and 2021. Tumor contours were manually delineated on pre-treatment (CTpre) and first post-treatment CT (CTpost). A fully convolutional DenseNet (FCD) and an Intentional Deep Overfit Learning (IDOL) framework were trained and validated. Performance was assessed using the Dice similarity coefficient (DSC), and RECIST-based diameters were compared between manual and predicted contours. Results: In the full cohort, the FCD model achieved mean DSCs of 0.53 for CTpre and 0.33 for CTpost, while the IDOL model improved CTpost DSCs to 0.49. In the RECIST cohort (n = 63), mean DSCs were 0.61 for CTpre and 0.53 for CTpost using FCD, versus 0.63 for IDOL. Predicted RECIST diameters differed by ~6% from manual values, with concordance in 13 of 14 validation cases. Volumetric predictions showed lower correlation, with a tendency toward overestimation at tumor poles. Conclusions: The patient-specific IDOL framework improved auto-segmentation accuracy compared with conventional models and provided clinically acceptable RECIST-based response assessment. Despite limitations and lack of external validation, this study demonstrates preliminary feasibility of auto-segmentation to support response evaluation in treated HCC.