Skip to content
DraconiticResearch
Search Sources API
190,851 studies31 sourcesUpdated 51 min ago
Peer-reviewedCC BY-NC-NDTranslational vision science & technology

FFA-SegGPT: A Two-Stage Cross-Modal AI System to Optimize the Clinical Workflow of Diabetic Retinopathy FFA Interpretation.

  • Fan Gan1,2
  • Xian Chen1,3
  • Yinan Shao4
  • Henghui Li5
  • Wanyun Chen6
  • Weiguo Qin7
  • Xin Long3
  • Huimin Fan1,2
  • Hanying Hu1,2
  • Zihan Yu1,2
  • Kaiwen Chen1,2
  • Nuo Xu1,2
4 more authors
  • Liping Chen1,2
  • Jian Cao1,2
  • Jing Li1,2
  • Zhipeng You1,2
7 affiliations
  1. 1Department of Ophthalmology, The Affiliated Eye Hospital, Jiangxi Medical College, Nanchang University, Nanchang, Jiangxi, People's Republic of China.
  2. 2Jiangxi Province Key Laboratory of Ophthalmology and Vision Sciences, Jiangxi Clinical Research Center for Ophthalmic Disease, Jiangxi Provincial Key Laboratory of Vitreoretinal Diseases for Health, Nanchang, Jiangxi, People's Republic of China.
  3. 3School of Mathematics and Computer Sciences, Nanchang University, Nanchang, Jiangxi, People's Republic of China.
  4. 4The First Affiliated Hospital of Baotou Medical College, Baotou, Inner Mongolia Autonomous Region, People's Republic of China.
  5. 5Department of Ophthalmology, Gaoxin Branch of the First Affiliated Hospital of Nanchang University, Nanchang, Jiangxi, People's Republic of China.
  6. 6Shenshan Medical Center, SunYat-sen Memorial Hospital, Sun Yat-sen University, Shanwei, Guangdong, People's Republic of China.
  7. 7Jiangxi Provincial Changzheng Hospital, Nanchang, Jiangxi, People's Republic of China.
Read full textPDFDOICite

Abstract

Purpose

The purpose of this study was to develop an interpretable, two-stage artificial intelligence (AI) system (fundus fluorescein angiography [FFA]-SegGPT) designed to optimize the clinical workflow of diabetic retinopathy (DR) FFA interpretation and reduce the risk of factual inaccuracies common in end-to-end models.Methods

We developed a cross-modal system that decouples lesion quantification from report generation. A segmentation module first explicitly segments non-perfusion areas (NPAs) and neovascularization (NV). These segmentation maps then serve as visual constraints to guide an evidence-based report generator, ensuring diagnostic text is strictly grounded in detected lesions.Results

In a simulated clinical workflow, AI-aided reporting significantly reduced mean interpretation time by 50% (from 8.13 ± 0.47 to 4.09 ± 0.21 minutes, P < 0.001) and improved median report quality (4.0 vs. 3.5, P < 0.001) compared to manual interpretation. Objectively, the model achieved high Dice Similarity Coefficients (NV = 85.35% and NPAs = 86.49%) and maintained high semantic fidelity for report generation (BERTScore = 0.91), with robust generalization demonstrated across three external cohorts.Conclusions

FFA-SegGPT enhances both the efficiency and quality of DR diagnostics. By decoupling lesion quantification from report generation, the system provides a transparent, dual-output mechanism that reduces the risk of factual inaccuracies, offering an interpretable workflow to facilitate safe clinical adoption.Translational relevance

This interpretable AI system optimizes the clinical workflow of DR by providing verifiable lesion segmentation and automated reporting, potentially reducing clinician workload and standardizing diagnostic quality.

Related

  1. Clinical setting-dependent diagnostic accuracy of artificial intelligence and store-and-forward diabetic retinopathy screening: a systematic review and meta-analysis
    Kai-Yang Chen, Hoi-Chun Chan, Chi-Ming Chan
    npj Digital MedicinePeer-reviewedCC BY11 citedMedical imaging & radiologyOpenAlex
  2. Identification of key clinical features and development of a machine learning model for screening-based detection of diabetic retinopathy
    Li Huilin, Hao Shaofeng, Li Na +4
    Scientific ReportsPeer-reviewedCC BYFull textMedical imaging & radiologyDOAJ
  3. CBAM-enhanced convolutional networks for diabetic retinopathy classification in the ODIR dataset
    Haewon Byeon
    Scientific ReportsPeer-reviewedCC BY-NC-ND 4.0Full textMedical imaging & radiologyCrossref
  4. Successful management of high-risk proliferative diabetic retinopathy during pregnancy with intravitreal faricimab
    Oxford medical case reportsPeer-reviewedCC BY-NCFull textGeneral medicineEurope PMC
  5. ISG15+ Müller glial subtype contributes to early blood–retinal barrier breakdown in diabetic retinopathy
    The Journal of clinical investigationPeer-reviewedCC BYFull textGeneral medicineEurope PMC

Cite

BibTeXRIS
BibTeX
@article{gan2026ffaseggpt,
  title = {{FFA-SegGPT: A Two-Stage Cross-Modal AI System to Optimize the Clinical Workflow of Diabetic Retinopathy FFA Interpretation.}},
  author = {Fan Gan and Xian Chen and Yinan Shao and Henghui Li and Wanyun Chen and Weiguo Qin and Xin Long and Huimin Fan and Hanying Hu and Zihan Yu and Kaiwen Chen and Nuo Xu and Liping Chen and Jian Cao and Jing Li and Zhipeng You},
  journal = {Translational vision science \& technology},
  year = {2026},
  month = oct,
  doi = {10.1167/tvst.15.10.16},
  url = {https://europepmc.org/article/MED/42872913},
  note = {CC BY-NC-ND},
}
RIS
TY  - JOUR
TI  - FFA-SegGPT: A Two-Stage Cross-Modal AI System to Optimize the Clinical Workflow of Diabetic Retinopathy FFA Interpretation.
AU  - Fan Gan
AU  - Xian Chen
AU  - Yinan Shao
AU  - Henghui Li
AU  - Wanyun Chen
AU  - Weiguo Qin
AU  - Xin Long
AU  - Huimin Fan
AU  - Hanying Hu
AU  - Zihan Yu
AU  - Kaiwen Chen
AU  - Nuo Xu
AU  - Liping Chen
AU  - Jian Cao
AU  - Jing Li
AU  - Zhipeng You
PY  - 2026
DA  - 2026/10/01
JO  - Translational vision science & technology
DO  - 10.1167/tvst.15.10.16
UR  - https://europepmc.org/article/MED/42872913
AB  - Purpose  The purpose of this study was to develop an interpretable, two-stage artificial intelligence (AI) system (fundus fluorescein angiography [FFA]-SegGPT) designed to optimize the clinical workflow of diabetic retinopathy (DR) FFA interpretation and reduce the risk of factual inaccuracies common in end-to-end models.Methods  We developed a cross-modal system that decouples lesion quantification from report generation. A segmentation module first explicitly segments non-perfusion areas (NPAs) and neovascularization (NV). These segmentation maps then serve as visual constraints to guide an evidence-based report generator, ensuring diagnostic text is strictly grounded in detected lesions.Results  In a simulated clinical workflow, AI-aided reporting significantly reduced mean interpretation time by 50% (from 8.13 ± 0.47 to 4.09 ± 0.21 minutes, P < 0.001) and improved median report quality (4.0 vs. 3.5, P < 0.001) compared to manual interpretation. Objectively, the model achieved high Dice Similarity Coefficients (NV = 85.35% and NPAs = 86.49%) and maintained high semantic fidelity for report generation (BERTScore = 0.91), with robust generalization demonstrated across three external cohorts.Conclusions  FFA-SegGPT enhances both the efficiency and quality of DR diagnostics. By decoupling lesion quantification from report generation, the system provides a transparent, dual-output mechanism that reduces the risk of factual inaccuracies, offering an interpretable workflow to facilitate safe clinical adoption.Translational relevance  This interpretable AI system optimizes the clinical workflow of DR by providing verifiable lesion segmentation and automated reporting, potentially reducing clinician workload and standardizing diagnostic quality.
SN  - 2164-2591
ER  -