Multimodal artificial intelligence for retinal detachment diagnosis using fundus imaging and patient questionnaires
Clinical Snapshot
PICO Framework
| P — Population | Patients with suspected retinal detachment and healthy controls at Tsukazaki Hospital |
| I — Intervention | Multimodal AI system combining ultra-widefield fundus imaging and patient questionnaire data |
| C — Comparator | Single-modal AI models (image-only and questionnaire-only) and human clinician diagnosis |
| O — Outcomes | Diagnostic accuracy, sensitivity (recall), specificity, and F1 scores for retinal detachment detection |
Bottom Line
This study presents a multimodal AI system combining fundus imaging with patient questionnaires for retinal detachment diagnosis, achieving approximately 90% accuracy with high sensitivity (95% per-subject). While the multimodal approach outperformed single-modal methods, the study has significant methodological limitations including unclear reference standards, incomplete diagnostic metrics, and single-centre design. The high sensitivity suggests potential utility as a screening tool, but specificity data are critically missing. Clinical implementation would require validation in diverse populations, clear integration pathways, and cost-effectiveness analysis. The technology shows promise but needs substantial additional research before clinical adoption, particularly regarding false-positive rates and workflow integration in emergency and primary care settings where retinal detachment screening is most needed.
Key Findings
P Value: Not reported
Effect Size: Multimodal model accuracy: 89.9% (per-image), 89.3% (per-subject)
Primary Outcome: Diagnostic accuracy for retinal detachment detection
Nnt Or Sensitivity: Sensitivity (recall): 90.2% (per-image), 94.9% (per-subject)
Confidence Interval: Standard deviations: ±5.4% (per-image), ±7.1% (per-subject)
Clinical Application
Requires ultra-widefield fundus imaging capability and structured questionnaire implementation Would need validation in Australian populations; integration with existing ophthalmology services and potential Medicare item numbers for AI-assisted diagnosis would require evaluation Patients presenting with symptoms suggestive of retinal detachment in ophthalmology settings
Abstract
BACKGROUND/AIMS: This study aimed to develop a multimodal artificial intelligence (AI) system that integrates fundus imaging and patient questionnaire data to achieve clinician-level diagnostic accuracy for diagnosing retinal detachment (RD). METHODS: Ultra-widefield fundus images and comprehensive patient questionnaires were collected from patients with RD and healthy controls at Tsukazaki Hospital. A multimodal model was developed using the Contrastive Language-Image Pretraining framework to classify RD cases, alongside separate image-only and questionnaire-only models for comparison. Per-image and per-subject analyses were conducted to assess model performance. RESULTS: The multimodal model outperformed single-modal models in per-image and per-subject assessments. It achieved accuracy, recall and F1 scores of 0.899±0.054, 0.902±0.043 and 0.902±0.048 in the per-image analysis and 0.893±0.071, 0.949±0.044 and 0.873±0.074 in the per-subject analysis, respectively. The AI model's overall diagnostic accuracy was slightly lower than that of human clinicians; however, it exhibited a higher recall rate, indicating improved detection of true RD cases. CONCLUSION: Integrating fundus imaging with patient questionnaire data significantly improves AI-based RD diagnosis. Future research needs to focus on expanding the dataset and refining the questionnaire design to further improve model performance.
Related Research
Indian journal of ophthalmology
Internet of things in assistive technology for people with visual impairment: A scoping review
1 Aug 2026
BMJ open ophthalmology
Screening for diabetic retinopathy with artificial intelligence in a primary care setting: a comparative cost analysis
1 Aug 2026
Science advances
Multispectral infrared-to-full-color upconversion expanding human vision
1 Aug 2026
This content is for educational purposes for healthcare professionals only and does not constitute clinical advice. Clinical decisions should be based on individual patient assessment, current guidelines, and appropriate specialist consultation. Editorial Standards · Privacy Policy · Terms of Service