Answers are grounded in this paper and cite their sources.
Background & Aims Microvascular invasion (MVI) is a key determinant of recurrence and poor outcomes in hepatocellular carcinoma (HCC), yet accurate preoperative detection remains challenging. A deep-learning model integrating multiphase MRI and clinical features was developed and externally validated to noninvasively predict microvascular invasion and stratify postoperative recurrence risk. Methods Clinico-radiological data from 924 resected HCC patients (2014–2023, five tertiary centers in China) were retrospectively assembled. A deep-learning model (DL-TriFusion) integrating multiphase MRI and clinical variables for MVI prediction was trained (n=361), internally validated (n=155), and externally validated across three centers (n=408; 188/136/84); whether recurrence risk stratification using imaging-based surrogates is superior to pathology-confirmed MVI alone was also evaluated. Statistical analysis included classification and detection metrics. Results DL-TriFusion achieved AUCs ranging from 0.957 to 0.959, significantly outperforming all unimodal and bimodal models (p < 0.001). Ablation studies confirmed incremental value from combining clinical variables with imaging features. Prognostically, DL-TriFusion outperformed pathology-based MVI, with C-indices 0.837/0.755 vs 0.447/0.520 (both p <0.001); in the external cohort, AUCs were 0.846 vs 0.485 at 2 years and 0.938 vs 0.501 at 5 years ( p <0.001). Performance was consistent across HBV, histological, and center-based subgroups. Conclusions DL-TriFusion enables robust preoperative prediction of MVI. It also improves postoperative stratification of early and late recurrence. Impact and implications DL-TriFusion represents a reliable, noninvasive biomarker with strong potential for clinical translation. By leveraging routinely available MRI and clinical data, it enables accurate preoperative assessment of MVI and stratification of recurrence risk. This capability could optimize surgical planning, identify candidates for neoadjuvant or adjuvant therapy, refine surveillance strategies, and reduce unnecessary overtreatment—ultimately supporting personalized management of HCC.
No takes yet. Share an insight, caveat, or question.
Fan et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: