CineMR: Tool-Integrated Vision-Language Reasoning for Quantitative Cardiac MRI Assessment

2026-10-01

AI Summary

This study addresses the difficulty of existing vision-language models (VLMs) in reliably extracting quantitative metrics from multidimensional cardiac MRI by proposing a tool-augmented VLM. The method leverages an interleaved reasoning mechanism to invoke a medical image analysis toolchain, including segmentation and volumetry, for quantitative assessment. A core innovation is the introduction of a Group Relative Policy Optimization (GRPO) training strategy with a conditional tool-use reward, alongside the construction of a multi-cohort benchmark to enhance tool-calling reliability. Through joint optimization via supervised fine-tuning and GRPO, the model achieves a Pass@1 of 35.9% and a tool-calling accuracy of 99.8%, improving ventricular measurement precision by over 20% compared to direct prediction.

Abstract

Cardiovascular magnetic resonance (CMR), including cine imaging, is a reference standard for the noninvasive assessment of cardiac morphology and ventricular function. Cine CMR interpretation integrates qualitative visual assessment with quantitative measurements of ventricular volumes, ejection fraction, myocardial mass, wall thickness, and regional wall motion. Current medical vision-language models (VLMs) cannot reliably derive quantitative measurements from multidimensional cine images without analysis tools. We present CineMR, a tool-augmented VLM that invokes cardiac image-analysis tools and integrates their outputs into interleaved reasoning for quantitative CMR assessment. We also construct a multi-cohort visual question answering benchmark covering quantitative metric extraction, multiclass diagnosis, and differential diagnosis, together with tools for segmentation, phase selection, volumetry, morphometry, and regional wall motion analysis. CineMR is trained with supervised fine-tuning (SFT) on tool-interaction traces followed by Group Relative Policy Optimization (GRPO) with conditional tool-use rewards. On the multi-cohort cine CMR benchmark, CineMR achieves 35.9% pass@1 and 58.9% pass@4, compared with 1.5% pass@1 for the Qwen3-VL-8B backbone and 0.0% and 7.0% pass@1 for LLaVA-Med v1.5 and MedGemma-4B, respectively. Correct tool invocation reaches 99.8% after GRPO, up from 78.9% after SFT. Live tool outputs improve ventricular measurement accuracy by 20.4--23.7% over direct model predictions, and removing all tools reduces pass@1 from 35.9% to 27.9%. These results highlight the importance of reliable tool use for quantitative cine CMR reasoning and support CineMR as a promising approach for assistive cardiac image assessment. Code, benchmark resources, and model weights are available at https://github.com/AI-MIND-Lab/CineMR.

Problem

• Cardiovascular Magnetic Resonance
• Vision-Language Models
• Quantitative Assessment
• Cine Imaging
• Tool Integration

Innovation

• Tool-Integrated VLM
• Cardiac MRI
• Group Relative Policy Optimization (GRPO)
• Visual Question Answering
• Quantitative Reasoning