Purpose: To compare deep learning-reconstructed T1-weighted volumetric interpolated breath-hold examination (DL-VIBE) with conventional VIBE (Conv-VIBE) for gadoxetic acid-enhanced hepatobiliary phase (HBP) MRI in biliary visualization, lesion characterization, and artifact reduction.

Methods: This retrospective study analyzed 94 patients (145 lesions) undergoing 3T MRI (March-May 2025). Paired HBP acquisitions used DL-VIBE (1.5 mm slices, CAIPIRINHA acceleration factor=6, inline DL reconstruction) and Conv-VIBE (3.0 mm). Two blinded radiologists evaluated biliary anatomy, lesion margins, and artifacts via quantitative metrics (edge rise distance [ERD], CNR) and qualitative 5-point scales. Statistical analysis included paired t-tests and Cohen’s kappa.

Results: DL-VIBE significantly improved tertiary bile duct detection (2.51 vs. 0.95 branches, p<0.001) and duct continuity scores (3.04 vs. 2.66, p<0.001). Lesion margin sharpness increased (ERD: 0.95±0.51mm vs. 1.54±0.10mm; p<0.001), with higher overall lesion detection (97.2% vs. 91.0%, p=0.005) and superior subcentimeter lesion detection (96.3% vs. 87.6%, p=0.004). DL-VIBE reduced image noise (3.91±0.31 vs. 2.32±0.78, p<0.001) and ringing artifacts (3.62±0.53 vs. 2.37±0.72, p<0.001) while maintaining bile duct CNR (235.75±15.08 vs. 138.23±12.22, p<0.001). Texture naturalness was marginally lower (2.91±0.65 vs. 3.52±0.54, p<0.001).

Conclusion: DL-VIBE enhances biliary microanatomy delineation, lesion margin resolution, and small lesion detection within standard breath-holds, offering a clinically superior alternative to Conv-VIBE for precision hepatobiliary imaging.