Synthetic Vasculature and Pathology Enhance Vision-Language Model Reasoning

TL;DR

This paper introduces Synthetic Vasculature Reasoning (SVR), a framework for generating synthetic OCTA images and explanations to improve vision-language models' medical reasoning, especially in data-scarce domains.

Contribution

The paper presents a novel synthetic data generation method for OCTA images with pathology annotations, enabling training of VLMs with improved reasoning and explanation capabilities.

Findings

VLM trained on synthetic data achieves 89.67% accuracy on real OCTA images.

Synthetic data enhances explanation quality and pathology localization.

Model outperforms supervised baselines in zero-shot classification.

Abstract

Vision-Language Models (VLMs) offer a promising path toward interpretable medical diagnosis by allowing users to ask about clinical explanations alongside predictions and across different modalities. However, training VLMs for detailed reasoning requires large-scale image-text datasets. In many specialized domains, for example in reading Optical Coherence Tomography Angiography (OCTA) images, such precise text with grounded description of pathologies is scarce or even non-existent. To overcome this bottleneck, we introduce Synthetic Vasculature Reasoning (SVR), a framework that controllably synthesizes images and corresponding text, specifically: realistic retinal vasculature with Diabetic Retinopathy (DR) features: capillary dropout, microaneurysms, neovascularization, and tortuosity, while automatically generating granular reasoning texts. Based on this we curate OCTA-100K-SVR, an OCTA image-reasoning dataset with 100,000 pairs. Our experiments show that a general-purpose VLM (Qwen3-VL-8b) trained on the dataset achieves a zero-shot balanced classification accuracy of 89.67% on real OCTA images, outperforming supervised baselines. Through human expert evaluation we also demonstrate that it significantly enhances explanation quality and pathology localization on clinical data.