LI-RADS Treatment Response Algorithm v2024 for Post-Treatment Assessment of Hepatocellular Carcinoma: A Systematic Review Across CEUS, CT/MRI, and Locoregional Therapies

Background: Accurate post-treatment imaging assessment of hepatocellular carcinoma (HCC) is essential for guiding retreatment, transplantation eligibility, and prognosis. The Liver Imaging Reporting and Data System Treatment Response Algorithm (LI-RADS TRA) was updated in 2024, introducing refinements across contrast-enhanced ultrasound (CEUS), CT/MRI, and radiation-based therapies, including the TR-Nonprogressing category. We aimed to systematically review current evidence on the diagnostic performance, reproducibility, and clinical implications of LI-RADS TRA v2024 for post-treatment assessment of HCC across imaging modalities and locoregional therapies. Methods: A systematic search of PubMed, Embase, Scopus, Web of Science, and CENTRAL was conducted according to PRISMA 2020 guidelines. Original studies applying LI-RADS TRA v2024 (or comparisons with earlier versions or mRECIST) after locoregional therapies were included. Outcomes of interest comprised diagnostic accuracy metrics, inter-reader and inter-modality agreement, temporal behavior of TRA categories, and prognostic associations. Given substantial heterogeneity, results were synthesized narratively. Results: Six studies met inclusion criteria. After thermal ablation, CEUS applying the non-radiation TRA v2024 showed very high specificity (approximately 88–100%) and excellent negative predictive value (about 94–98%), with substantial-to-almost-perfect inter-reader agreement (κ up to 0.92) and excellent inter-modality agreement with CT/MRI (ICC ≈ 0.90). After transarterial chemoembolization, CEUS performance was time-dependent, with higher sensitivity at early follow-up (~15 days) and higher specificity at later assessment (~30 days). On MRI, non-radiation TRA v2024 combined with ancillary features improved sensitivity and, in some analyses, accuracy compared with v2017 and v2024 without ancillary features, while remaining more specific than mRECIST. In radiation-based therapies, the radiation TRA v2024 captured delayed-response patterns through the TR-Nonprogressing category, which demonstrated meaningful temporal evolution and prognostic separation from TR-Viable. Surgical validation confirmed preserved diagnostic performance against histology. Conclusions: Despite heterogeneous data, available evidence supports LI-RADS TRA v2024 as a clinically useful framework that improves clarity, reproducibility, and actionability of post-treatment HCC imaging. CEUS is particularly effective after ablation; timing is critical after TACE, ancillary features should be routinely applied on MRI, and TR-Nonprogressing appropriately reflects radiation biology while preserving prognostic value.