Diagnostic Performance of LI-RADS v2024 Radiation Treatment Response Algorithm: A Multi-Reader Study with Pathologic Correlation

Motivation: LI-RADS Treatment Response Algorithm (TRA) v2018 had limitations in assessing radiation-based therapies for hepatocellular carcinoma (HCC). The newly released v2024 includes specific criteria for radiation therapy response assessment. Goal(s): To evaluate the diagnostic performance and inter-reader agreement of LI-RADS v2024 radiation TRA using pathologic correlation as the reference standard. Approach: Multi-reader study with 4 fellowship-trained radiologists interpreting post-treatment imaging in 25 patients who underwent liver transplantation after radiation therapy. Results: The radiation TRA demonstrated high specificity (91.7%) and moderate sensitivity (46.2-61.5%) across readers, with strong inter-reader agreement (Fleiss' kappa = 0.714). Impact: This first validation study of LI-RADS v2024 radiation TRA shows excellent inter-reader agreement and diagnostic performance, supporting its clinical implementation for standardized post-radiation therapy response assessment in HCC patients.