MRI2Rep: Autoregressive Structured Report Generation for 3D Liver MRI

2026-06-23

AI Summary

This study addresses the challenges of time-consuming manual generation of structured reports for 3D liver MRI, compounded by the complexity of volumetric data and scarcity of paired datasets. To overcome these issues, the authors propose MRI2Rep—the first end-to-end framework for generating LI-RADS–compliant structured reports directly from 3D liver MRI scans. The method introduces a Report-to-Label Canonicalization module that converts free-text reports into structured diagnostic sequences, enabling autoregressive report generation without requiring explicit lesion annotations. It further integrates a medical vision–language model with an LLM-based evaluation proxy (LLM-Eval) for automated assessment. Experimental results demonstrate that the system achieves 76.0% case-level sensitivity, 29.4% lesion-level F1 score, and 82.4% liver-level accuracy on the test set, with 70–75% of AI-generated reports rated by radiologists as clinically acceptable.

Abstract

Manual reporting of 3D MRI studies is time-consuming, yet end-to-end structured report generation for 3D liver MRI remains underexplored due to volumetric complexity and scarce paired data. We propose MRI2Rep, an autoregressive framework for liver MRI report generation. From 3,929 real-world MRI-report pairs acquired over a 10-year single-institution cohort, a Report-to-Label Canonicalization (RLC) module converts free-text reports into structured, closed-vocabulary diagnostic sequences without lesion-level annotations. On a held-out test set, MRI2Rep achieves 76.0% case-level sensitivity, 29.4% lesion-level F1, compared with no more than 8.3% for adapted medical vision-language baselines, and 82.4% liver-level accuracy. In a blinded reader study, two radiologists rated 75% and 70% of AI-generated reports as clinically acceptable, compared with 95% and 100% for original reports. Our automated LLM-based judge, LLM-Eval, rated 61.8% of AI-generated reports as acceptable, applying a stricter standard and supporting its use as a conservative proxy. To our knowledge, this is the first end-to-end LI-RADS-structured reporting system for 3D liver MRI.

Problem

• structured report generation
• 3D liver MRI
• autoregressive modeling
• medical imaging
• LI-RADS

Innovation

• autoregressive report generation
• structured MRI reporting
• Report-to-Label Canonicalization
• LI-RADS
• 3D liver MRI