June 2023
Volume 64, Issue 9
Open Access
ARVO Imaging in the Eye Conference Abstract  |   June 2023
Detecting multiple retina diseases from fundus photographs using probabilistic deep learning model
Author Affiliations & Notes
  • Xiaoqin Huang
    Ophthalmology, The University of Tennessee Health Science Center, Memphis, Tennessee, United States
  • Jian Sun
    Integrated Data Science Section, Research Technologies Branch, National Institute of Allergy and Infectious Diseases, Bethesda, Maryland, United States
  • Ehsan Kazami
    Ophthalmology, General Hospital of Mahabad, Urmia University of Medical Sciences, Urmia, Iran (the Islamic Republic of)
  • Siamak Yousefi
    Ophthalmology, The University of Tennessee Health Science Center, Memphis, Tennessee, United States
    Department of Genetics, Genomics, and Informatics, The University of Tennessee Health Science Center, Memphis, Tennessee, United States
  • Footnotes
    Commercial Relationships   Xiaoqin Huang, None; Jian Sun, None; Ehsan Kazami, None; Siamak Yousefi, NIH EY031725 (F), NIH EY033005 (F), Research to Prevent Blindness (F)
  • Footnotes
    Support  NIH EY031725, NIH EY033005
Investigative Ophthalmology & Visual Science June 2023, Vol.64, PB0030. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Xiaoqin Huang, Jian Sun, Ehsan Kazami, Siamak Yousefi; Detecting multiple retina diseases from fundus photographs using probabilistic deep learning model. Invest. Ophthalmol. Vis. Sci. 2023;64(9):PB0030.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose : To detect multiple retina diseases, including age-related macular degeneration (AMD), diabetic retinopathy (DR), and glaucoma from fundus images using a probabilistic deep convolutional neural network (CNN) model and to evaluate the generalizability of the model.

Methods : We collected fundus photographs from patients with AMD, DR, and glaucoma, and healthy subjects from age-related eye disease studies (AREDS) dataset and multiple publicly available datasets. We developed a probabilistic deep learning model to detect four classes based on 12,418 images in the development dataset and validated the model based on two external datasets with 1305 and 600 images, respectively. The performance of the probabilistic model was compared with the conventional deep learning models.

Results : The probabilistic CNN model was developed using the ResNet152-V2 architecture. The area under the receiver operating characteristic (AUC) of the probabilistic model based on the internal testing, first, and second external datasets were 0.84 (CI: 0.82-0.86), 0.81 (0.79-0.84) and 0.81 (0.80-0.83), respectively. The AUC of the conventional CNN model based on the internal testing, first, and second external datasets were 0.94 (0.92-0.96), 0.85 (0.83-0.87), 0.86 (0.85-0.87), respectively (Fig. 1). The gradient class activation map (CAM) of the probabilistic CNN mode focused more on the macula region for AMD detection, blood vessel and hemorrhages for DR, and optic disc for glaucoma detection (Fig. 2).

Conclusions : Probabilistic CNN models can detect multiple retina diseases with a high level of generalizability with a slight compromise in performance compared with conventional CNN model. The model focused more on retinal regions corresponding to pathologies related to the underlying disease. The probabilistic model however provides both the likelihood and uncertainty of prediction that may advance the integration of CNN models in clinical settings.

This abstract was presented at the 2023 ARVO Imaging in the Eye Conference, held in New Orleans, LA, April 21-22, 2023.

 

Figure 1 ROC curve of the model result on internal testing, and two external validation datasets: left-conventional deep learning model; right-probabilistic deep learning model.

Figure 1 ROC curve of the model result on internal testing, and two external validation datasets: left-conventional deep learning model; right-probabilistic deep learning model.

 

Figure 2 Gradient CAM of the fundus images and the corresponding prediction result (top row---original images, bottom row—corresponding CAM).

Figure 2 Gradient CAM of the fundus images and the corresponding prediction result (top row---original images, bottom row—corresponding CAM).

×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×