LI-RADS CT/MRI Nonradiation Treatment Response Assessment Version 2024 for Detecting Local Recurrence of Surgically...

BACKGROUND. LI-RADS CT/MRI Nonradiation Treatment Response Assessment (TRA) version 2024 (v2024) specifies that the updated algorithm may be applied for evaluating the surgical margin after resection of hepatocellular carcinoma (HCC). OBJECTIVE. The purpose of this study was to compare detection of local recurrences after surgical resection of HCC between LI-RADS CT/MRI Nonradiation TRA v2024 and the LI-RADS CT/MRI Core version 2018 (v2018) diagnostic algorithm. METHODS. This retrospective study included patients with surgically resected HCC who underwent at least one liver CT or MRI examination performed at least 6 months postoperatively. Two radiologists independently reviewed all postoperative CT and MRI examinations of included patients in separate sessions, assessing the surgical margin by use of LI-RADS CT/MRI Nonradiation TRA v2024 and LI-RADS CT/MRI Core v2018, respectively; a third radiologist resolved discrepancies. Local recurrence was defined by v2024 as LR-TR Viable (i.e., masslike enhancement at the margin, noted in any phase and of any size) and by v2018 as LR-5, LR-M, or LR-TIV. RESULTS. The study included 200 patients (142 men and 58 women; median age, 67 years) with a median follow-up of 30.5 months. Interreader agreement for v2024 categorization was substantial (κ = 0.71). By v2024, LR-TR Viable at the surgical margin was assigned in 56 patients (28.0%) (on the initial postoperative examination in 21 patients, directly after an LR-Nonviable assignment without an intervening LR-TR Equivocal assignment in 31 patients, and after an LR-TR Equivocal assignment in the remaining four patients); the assignments were made after a median of 5.1 months and at a median observation size of 1.5 cm. By v2018, LR-5, LR-M, or LR-TIV was assigned at the surgical margin in 52 patients (26.0%) after a median of 12.0 months (p < .001) and at a median size of 2.1 cm (p < .001). At the time of the 56 initial LR-TR Viable assignments, the categories assigned for v2018 were LR-3, LR-4, LR-5, LR-M, and LR-TIV in 21, 10, 18, three, and four patients, respectively. Upgrading of LR-3 and LR-4 observations to LR-5, LR-M, or LR-TIV occurred after a median of 5.1 and 3.0 months, respectively. CONCLUSION. v2024 detected local recurrence in 28% of patients at an earlier time and at a smaller size than did v2018. CLINICAL IMPACT. The findings support application of v2024 for surgical margin evaluation after HCC resection.