MRI-based clinicoradiologic model for identifying MT-HCC with distinct postoperative prognosis and potential association with postoperative TKI-ICI therapy.

Background: This study aims to develop a non-invasive MRI-based clinicoradiologic model for predicting microvascular invasion-positive (MVI+) and tertiary lymphoid structures-positive (TLSs+) hepatocellular carcinoma (HCC) and to explore whether this subgroup is associated with different postoperative outcomes and a potential differential association with postoperative TKI-ICI combination therapy. Methods: A retrospective analysis was conducted on 338 HCC patients across two centers, stratified by MVI and TLSs status, and divided into training (n=237) and validation (n=101) cohorts. MRI features such as tumor size, hemorrhage, and capsule, were assessed. Model performance was assessed using the area under the receiver operating characteristic curve (AUC). Kaplan-Meier curves were used to compare recurrence-free survival (RFS). Results: The MRI-based clinicoradiologic model for predicting MVI+/TLSs+ HCC (MT-HCC) integrated AFP levels and MRI features, including absence of intratumoral hemorrhage, incomplete capsule, and mosaic structure, achieved an AUC of 0.831 in the validation cohort. In exploratory analyses, the model-predicted MT-HCC group showed a shorter RFS than the model-predicted non-MT-HCC group among patients without postoperative TKI-ICI exposure (p = 0.002), whereas a different RFS pattern was observed among patients with postoperative TKI-ICI exposure (p = 0.017). Conclusion: The MRI-based clinicoradiologic model may help identify MT-HCC preoperatively and may help identify a subgroup with distinct postoperative prognosis and a possible differential outcome pattern according to postoperative TKI-ICI exposure.