CAS-IQA: Teaching Vision-Language Models for Synthetic Angiography Quality Assessment

TL;DR

This paper introduces CAS-IQA, a vision-language framework for assessing the quality of synthetic angiographies, leveraging auxiliary images and task-specific metrics to improve clinical relevance and outperform existing methods.

Contribution

The paper proposes a novel vision-language model for fine-grained quality assessment of synthetic angiographies, incorporating auxiliary images and a new multi-path feature fusion module.

Findings

CAS-IQA outperforms state-of-the-art IQA methods on CAS-3K dataset.

Constructed CAS-3K dataset with 3,565 annotated synthetic angiographies.

Defined three task-specific evaluation metrics for clinical relevance.

Abstract

Synthetic X-ray angiographies generated by modern generative models hold great potential to reduce the use of contrast agents in vascular interventional procedures. However, low-quality synthetic angiographies can significantly increase procedural risk, underscoring the need for reliable image quality assessment (IQA) methods. Existing IQA models, however, fail to leverage auxiliary images as references during evaluation and lack fine-grained, task-specific metrics necessary for clinical relevance. To address these limitations, this paper proposes CAS-IQA, a vision-language model (VLM)-based framework that predicts fine-grained quality scores by effectively incorporating auxiliary information from related images. In the absence of angiography datasets, CAS-3K is constructed, comprising 3,565 synthetic angiographies along with score annotations. To ensure clinically meaningful assessment, three task-specific evaluation metrics are defined. Furthermore, a Multi-path featUre fuSion and rouTing (MUST) module is designed to enhance image representations by adaptively fusing and routing visual tokens to metric-specific branches. Extensive experiments on the CAS-3K dataset demonstrate that CAS-IQA significantly outperforms state-of-the-art IQA methods by a considerable margin.