Multidisease Deep Learning Neural Network for the Diagnosis of Corneal Diseases

Am J Ophthalmol. 2021 Jun:226:252-261. doi: 10.1016/j.ajo.2021.01.018. Epub 2021 Jan 30.

Abstract

Purpose: To report a multidisease deep learning diagnostic network (MDDN) of common corneal diseases: dry eye syndrome (DES), Fuchs endothelial dystrophy (FED), and keratoconus (KCN) using anterior segment optical coherence tomography (AS-OCT) images.

Study design: Development of a deep learning neural network diagnosis algorithm.

Methods: A total of 158,220 AS-OCT images from 879 eyes of 478 subjects were used to develop and validate a classification deep network. After a quality check, the network was trained and validated using 134,460 images. We tested the network using a test set of consecutive patients involving 23,760 AS-OCT images of 132 eyes of 69 patients. The area under receiver operating characteristic curve (AUROC), area under precision-recall curve (AUPRC), and F1 score and 95% confidence intervals (CIs) were computed.

Results: The MDDN achieved eye-level AUROCs >0.99 (95% CI: 0.90, 1.0), AUPRCs > 0.96 (95% CI: 0.90, 1.0), and F1 scores > 0.90 (95% CI: 0.81, 1.0) for DES, FED, and KCN, respectively.

Conclusions: MDDN is a novel diagnostic tool for corneal diseases that can be used to automatically diagnose KCN, FED, and DES using only AS-OCT images.

Publication types

  • Research Support, N.I.H., Extramural
  • Research Support, Non-U.S. Gov't
  • Validation Study

MeSH terms

  • Area Under Curve
  • Corneal Diseases / diagnosis
  • Deep Learning*
  • Diagnosis, Computer-Assisted*
  • Dry Eye Syndromes / diagnosis*
  • Female
  • Fuchs' Endothelial Dystrophy / diagnosis*
  • Humans
  • Keratoconus / diagnosis*
  • Male
  • Middle Aged
  • Neural Networks, Computer*
  • Prospective Studies
  • ROC Curve
  • Tomography, Optical Coherence