EUVIMEDOpdag EUVIMED
ValidierungsstudieOpen access · CC BY 4.0

Construction and Validation of a Long‐Term Visual Prognosis Prediction Model for Proliferative Diabetic Retinopathy After Vitrectomy Based on Machine Learning

Entwicklung und Validierung eines Modells zur Vorhersage der langfristigen Sehprognose bei proliferativer diabetischer Retinopathie nach Vitrektomie auf Basis von Machine Learning

Xiaolu Wang iD, Jinghua Shi iD, Yan Gao iD, Tiantian Li iD, Xuehong Han iD, Lixia Guo iD, Tingting Jia iD, Yan Wang iD

Endocrinology, Diabetes & Metabolism · 2026-09 · 9 (5)

DOI: 10.1002/edm2.70314 ↗

Auf einen Blick

Bei 609 PDR-Patient:innen nach PPV wurde ein Machine-Learning-Modell zur Vorhersage der langfristigen Sehprognose entwickelt. LightGBM schnitt am besten ab und zeigte in der Validierung gute Kalibrierung.

Fragestellung

Lässt sich die langfristige Sehprognose bei PDR nach PPV mit einem Machine-Learning-Modell vorhersagen?

Methode

609 PDR-Patient:innen (609 Augen) nach PPV; zufällige Aufteilung 8:2 in Trainings- und Validierungsset; LASSO, multivariate logistische Regression und Machine Learning; Bewertung mit AUC-ROC, Accuracy, Precision, Recall, F1, Kalibrierung und DCA.

Ergebnisse

Als unabhängige Prädiktoren wurden Niereninsuffizienz (OR 6,932; 95 %-KI 3,394–14,158), präoperative Irisneovaskularisation (OR 7,674; 95 %-KI 3,699–15,920), Silikonöl-Tamponade (OR 2,799; 95 %-KI 1,641–4,707) und IBIL (OR 0,902; 95 %-KI 0,829–0,981) identifiziert. LightGBM war das beste Modell.

Einschränkungen

Externe Validierung und prospektive Evaluation stehen noch aus.

Schlussfolgerung der Autor:innen

Die Autor:innen berichten, dass das Modell die frühe Risik पहचानung unterstützen und die personalisierte Behandlungsplanung fördern kann, vorbehaltlich externer Validierung und prospektiver Prüfung.

Kernaussagen

  • 609 PDR-Patient:innen nach PPV
  • LightGBM war das beste Vorhersagemodell
  • Externe Validierung und prospektive Prüfung fehlen

Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.

At a glance

Validierungsstudie in Endocrinology, Diabetes & Metabolism, published online on 2026-09-07. Written by Xiaolu Wang and 7 co-authors. Institution: School of Nursing, Shanxi Medical University, People's Republic of China. Major MEDLINE topics: Diabetic Retinopathy, Vitrectomy, Machine Learning. 41 references. The full text is freely available (open access, CC BY 4.0). Funded by: Research Fund Project of Shanxi Eye Hospital.

Auf einen BlickAbstractSubjectsBibliografiske oplysningerCiteRelated studies

Abstract (original language)

Introduction

This study aims to develop a machine‐learning‐based risk prediction model for a long‐term visual prognosis in patients with proliferative diabetic retinopathy (PDR) following pars plana vitrectomy (PPV).

Methods

We analysed 609 PDR patients (609 eyes) who underwent PPV at Shanxi Eye Hospital from January 1, 2022, to January 1, 2025. The dataset was randomly split into training and validation sets at an 8:2 ratio. Candidate risk factors were identified using LASSO regression, followed by multivariate logistic regression to determine independent predictors. Machine learning algorithms were then trained to construct the prediction model, with performance evaluated by AUC‐ROC, accuracy, precision, recall, and F1 score. Calibration was assessed via calibration curves, and clinical utility by decision curve analysis (DCA).

Results

LASSO regression and multivariate logistic regression indicated that several factors influenced postoperative long‐term visual prognosis. These included Renal Insufficiency (OR = 6.932, 95% CI: 3.394–14.158), Preoperative iris neovascularization (OR = 7.674, 95% CI: 3.699–15.920), Silicone oil tamponade (OR = 2.799, 95% CI: 1.641–4.707), Indirect bilirubin (IBIL) (OR = 0.902, 95% CI: 0.829–0.981) ( p < 0.05). LightGBM was found to be the best for predicting postoperative long‐term visual prognosis risk in PDR patients. The LightGBM model demonstrated good calibration in both training and validation sets. DCA of the validation set showed clinical net benefit at low‐to‐moderate risk thresholds, outperforming both ‘treat‐all’ and ‘treat‐none’ strategies.

Conclusions

In conclusion, this study developed a machine learning‐based prognostic model for long‐term visual prognosis in PDR patients after PPV, visualized as a proof‐of‐concept web calculator to assist clinical staff in early risk identification and support personalised treatment planning, pending external validation and prospective evaluation.

Subjects

MeSH terms (MEDLINE)

Diabetic RetinopathyVitrectomyMachine LearningHumansFemalePrognosisPredictive Learning ModelsMiddle AgedMaleRisk FactorsPrediction AlgorithmsVisual AcuityAgedRetrospective Studies

Author keywords

diabetic retinopathypredictive modelvitrectomy

Authors and institutions

  • School of Nursing, Shanxi Medical University, People's Republic of China
  • Department of Nursing, Shanxi Eye Hospital, People's Republic of China
  • Department of Vitreoretinopathy, Shanxi Eye Hospital, People's Republic of China

ORCID profiles: Xiaolu Wang, Jinghua Shi, Yan Gao, Tiantian Li, Xuehong Han, Lixia Guo, Tingting Jia, Yan Wang

Funding

  • Research Fund Project of Shanxi Eye Hospital

Cite

APA

Wang, X., Shi, J., Gao, Y., Li, T., Han, X., Guo, L., Jia, T., & Wang, Y. (2026). Construction and Validation of a Long‐Term Visual Prognosis Prediction Model for Proliferative Diabetic Retinopathy After Vitrectomy Based on Machine Learning. Endocrinology, Diabetes & Metabolism, 9(5). https://doi.org/10.1002/edm2.70314

Vancouver

Wang X, Shi J, Gao Y, Li T, Han X, Guo L, et al. Construction and Validation of a Long‐Term Visual Prognosis Prediction Model for Proliferative Diabetic Retinopathy After Vitrectomy Based on Machine Learning. Endocrinology, Diabetes & Metabolism. 2026;9(5). doi:10.1002/edm2.70314

BibTeX
@article{Wang2026,
  author = {Wang, Xiaolu and Shi, Jinghua and Gao, Yan and Li, Tiantian and Han, Xuehong and Guo, Lixia and Jia, Tingting and Wang, Yan},
  title = {Construction and Validation of a Long‐Term Visual Prognosis Prediction Model for Proliferative Diabetic Retinopathy After Vitrectomy Based on Machine Learning},
  journal = {Endocrinology, Diabetes & Metabolism},
  year = {2026},
  volume = {9},
  number = {5},
  doi = {10.1002/edm2.70314}
}
Start for free →