# Towards Clinically Interpretable Ophthalmic VQA via Spatially-Grounded Lesion Evidence

arXiv: https://arxiv.org/abs/2605.22414 · HTML: https://mlsanity.com/paper/2605.22414/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 19 of its 28 references that are in the map.
- First posted on arXiv 21 May 2026.

## Builds on (most cited first, 19 of 19)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Qwen3-VL Technical Report](https://mlsanity.com/paper/2511.21631/) (2025)
- [InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Models](https://mlsanity.com/paper/2504.10479/) (2025)
- [OpenAI GPT-5 System Card](https://mlsanity.com/paper/2601.03267/) (2025)
- [SLAKE: A Semantically-Labeled Knowledge-Enhanced Dataset for Medical Visual Question Answering](https://mlsanity.com/paper/2102.09542/) (2021)
- [MedGemma Technical Report](https://mlsanity.com/paper/2507.05201/) (2025)
- [Lingshu: A Generalist Foundation Model for Unified Multimodal Medical Understanding and Reasoning](https://mlsanity.com/paper/2506.07044/) (2025)
- [REFUGE Challenge: A Unified Framework for Evaluating Automated Methods for Glaucoma Assessment from Fundus Photographs](https://mlsanity.com/paper/1910.03667/) (2019)
- [A Benchmark for Studying Diabetic Retinopathy: Segmentation, Grading, and Transferability](https://mlsanity.com/paper/2008.09772/) (2020)
- [Understanding Catastrophic Forgetting and Remembering in Continual Learning with Optimal Relevance Mapping](https://mlsanity.com/paper/2102.11343/) (2021)
- [WSI-VQA: Interpreting Whole Slide Images by Generative Visual Question Answering](https://mlsanity.com/paper/2407.05603/) (2024)
- [Learn to Segment Retinal Lesions and Beyond](https://mlsanity.com/paper/1912.11619/) (2019)
- [GEMeX: A Large-Scale, Groundable, and Explainable Medical VQA Benchmark for Chest X-ray Diagnosis](https://mlsanity.com/paper/2411.16778/) (2024)
- [ADAM Challenge: Detecting Age-related Macular Degeneration from Fundus Images](https://mlsanity.com/paper/2202.07983/) (2022)
- [EyecareGPT: Boosting Comprehensive Ophthalmology Understanding with Tailored Dataset, Benchmark and Model](https://mlsanity.com/paper/2504.13650/) (2025)
- [MIMO: A medical vision language model with visual referring multimodal input and pixel grounding multimodal output](https://mlsanity.com/paper/2510.10011/) (2025)
- [GEMeX-RMCoT: An Enhanced Med-VQA Dataset for Region-Aware Multimodal Chain-of-Thought Reasoning](https://mlsanity.com/paper/2506.17939/) (2025)
- [FunBench: Benchmarking Fundus Reading Skills of MLLMs](https://mlsanity.com/paper/2503.00901/) (2025)
- [Benchmarking Large Multimodal Models for Ophthalmic Visual Question Answering with OphthalWeChat](https://mlsanity.com/paper/2505.19624/) (2025)

## Built on by (most cited first, 1 of 1)

- [EyeVQA: Benchmarking Ophthalmic Vision-Language Models from Recognition to Spatial Grounding](https://mlsanity.com/paper/2609.32352/) (2026)

## Abstract (arXiv)

Visual Question Answering (VQA) holds great promise for clinical support, particularly in ophthalmology, where retinal fundus photography is essential for diagnosis. However, ophthalmic VQA benchmarks primarily emphasize answer accuracy, neglecting the explicit visual evidence necessary for clinical interpretability. In this work, we introduce FundusGround, a new benchmark for clinically interpretable ophthalmic VQA with spatially-grounded lesion evidence. Specifically, we propose a three-stage pipeline that collects 10,719 fundus images with 15,595 image-level meticulously annotated lesions. To ensure anatomical consistency and clinical validity, all lesions are spatially localized using the Early Treatment Diabetic Retinopathy Study (ETDRS) grid, enabling standardized mapping to nine clinically meaningful retinal regions. Built upon this structured lesion evidence, 72,706 questions are then generated spanning four formats: open-ended, closed-ended, single-choice, and multiple-choice. We further benchmark multiple general- and medical- large vision-language models using dual metrics for answer accuracy and lesion-level reasoning. The experiments demonstrate that incorporating lesion-level visual evidence consistently improves model performance and transparency, highlighting the necessity of explicit spatial grounding for reliable and explainable ophthalmic VQA.