HalluciGuard-eye: knowledge graph-enhanced retrieval-augmented generation for hallucination detection and mitigation in ophthalmic multimodal large language models
- 1Ophthalmic Center, The Second Affiliated Hospital, Jiangxi Medical College, Nanchang University
Abstract
Background: Diabetic retinopathy (DR) is the leading cause of preventable blindness in working-age adults, yet scalable early detection is limited by the unreliability of AI-assisted screening. Multimodal large language models can interpret fundus photographs and support DR grading, but they frequently hallucinate - fabricating lesions or misclassifying referable DR - which directly threatens timely treatment and prevention of vision loss. We developed HalluciGuard-Eye to make AI-assisted DR assessment trustworthy.
Methods: HalluciGuard-Eye combines knowledge graph-enhanced retrieval-augmented generation with hierarchical hallucination detection. An Ophthalmic Clinical Knowledge Graph containing 47,832 medical entities and 156,294 semantic relationships was constructed from clinical guidelines and published literature. A dual-stream architecture uses a Visual-Semantic Alignment Module to verify cross-modal consistency and a Clinical Factuality Grounding Network to anchor generated content to reranked evidence. The Hierarchical Factuality Score assesses anatomical, pathological, and clinical-recommendation accuracy. The framework was evaluated on four public ophthalmic imaging benchmarks.
Results: HalluciGuard-Eye reduced hallucination rates by 42.2% compared with state-of-the-art medical multimodal large language models and achieved 91.6% factual accuracy for clinical entity recognition on EyePACS. For diabetic retinopathy grading, sensitivity was 89.6% and specificity was 92.1%, representing improvements of 2.4% and 2.7%, respectively, over MedRAG. In the component ablation, removing knowledge-graph retrieval lowered EyePACS HFS by 9.4 percentage points. Additional analyses yielded 87.6% critical-evidence recall@20, 89.2% lesion-location accuracy, and 88.9% sensitivity in a patient-clustered cohort.
Conclusions: These results indicate that knowledge-grounded retrieval and hierarchical factuality assessment can reduce hallucinations while preserving diagnostic performance in ophthalmic multimodal language models.
Related
- Hallucination in Time-series Large Language Models: An empirical lnvestigation and analysis of mitigation strategies
- Artificial intelligence for ophthalmic vessel segmentation: from multimodal imaging to clinical translation
- Feasibility and Diagnostic Potential of Multimodal Neuro-Ophthalmic Biomarkers in Young-Onset Dementia and Functional Cognitive Disorder.
- Ophthalmic AI With Just Ten Examples: Comparing Multimodal Large Language Models, Vision Foundation Models, and Vision-Language Models
- IoT-enabled multimodal large language model-based framework for compositional food recognition, nutrient profiling, NOVA classification, and cancer risk mitigation
BibTeX
@article{wu2026halluciguardeye,
title = {{HalluciGuard-eye: knowledge graph-enhanced retrieval-augmented generation for hallucination detection and mitigation in ophthalmic multimodal large language models}},
author = {Ge Wu and Chong Ai and Weiqi Wu},
journal = {Frontiers in Endocrinology},
year = {2026},
month = oct,
doi = {10.3389/fendo.2026.1976190},
url = {https://www.frontiersin.org/articles/10.3389/fendo.2026.1976190/full},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - HalluciGuard-eye: knowledge graph-enhanced retrieval-augmented generation for hallucination detection and mitigation in ophthalmic multimodal large language models
AU - Ge Wu
AU - Chong Ai
AU - Weiqi Wu
PY - 2026
DA - 2026/10/20
JO - Frontiers in Endocrinology
DO - 10.3389/fendo.2026.1976190
UR - https://www.frontiersin.org/articles/10.3389/fendo.2026.1976190/full
AB - Background: Diabetic retinopathy (DR) is the leading cause of preventable blindness in working-age adults, yet scalable early detection is limited by the unreliability of AI-assisted screening. Multimodal large language models can interpret fundus photographs and support DR grading, but they frequently hallucinate - fabricating lesions or misclassifying referable DR - which directly threatens timely treatment and prevention of vision loss. We developed HalluciGuard-Eye to make AI-assisted DR assessment trustworthy. Methods: HalluciGuard-Eye combines knowledge graph-enhanced retrieval-augmented generation with hierarchical hallucination detection. An Ophthalmic Clinical Knowledge Graph containing 47,832 medical entities and 156,294 semantic relationships was constructed from clinical guidelines and published literature. A dual-stream architecture uses a Visual-Semantic Alignment Module to verify cross-modal consistency and a Clinical Factuality Grounding Network to anchor generated content to reranked evidence. The Hierarchical Factuality Score assesses anatomical, pathological, and clinical-recommendation accuracy. The framework was evaluated on four public ophthalmic imaging benchmarks. Results: HalluciGuard-Eye reduced hallucination rates by 42.2% compared with state-of-the-art medical multimodal large language models and achieved 91.6% factual accuracy for clinical entity recognition on EyePACS. For diabetic retinopathy grading, sensitivity was 89.6% and specificity was 92.1%, representing improvements of 2.4% and 2.7%, respectively, over MedRAG. In the component ablation, removing knowledge-graph retrieval lowered EyePACS HFS by 9.4 percentage points. Additional analyses yielded 87.6% critical-evidence recall@20, 89.2% lesion-location accuracy, and 88.9% sensitivity in a patient-clustered cohort. Conclusions: These results indicate that knowledge-grounded retrieval and hierarchical factuality assessment can reduce hallucinations while preserving diagnostic performance in ophthalmic multimodal language models.
SN - 1664-2392
ER -