Automated assisted clinical diagnosis of retinopathy of prematurity based on deep learning

Objective

To evaluate the application value of an intelligent fundus assisted diagnosis system for detecting retinopathy of prematurity (ROP) based on deep learning.

Methods

A total of 38 895 fundus images for premature infants screening were collected from Renmin Hospital of Wuhan University Eye Center and were labeled by 10 licensed ophthalmologists.A deep learning network model was established to acquire automatic classification of disease stages and plus disease.The accuracy, sensitivity and specificity of the algorithm were calculated to evaluate the performance of the artificial intelligence system for ROP automatic diagnosis.This study protocol was approved by Ethic Committee of Renmin Hospital of Wuhan University (No.WDRY2019-K032). Written informed consent was obtained from the guardians of the children before entering the study cohort.

Results

The intelligent system achieved an accuracy of 0.931.Specifically, the accuracies in detecting demarcation line (stage Ⅰ) was 0.876, ridge (stage Ⅱ) was 0.942, ridge with extra retinal fibrovascular (stage Ⅲ) was 0.968, subtotal retinal detachment (stage Ⅳ) was 0.998, total retinal detachment (stage Ⅴ) was 0.999, vascular tortuosity and dilatation (plus disease) was 0.896, optic disc was 0.954, macular was 0.781, and laser scars were 0.974, respectively.

Conclusions

Deep learning algorithm can detect the stages and plus disease of ROP with excellent accuracy, and it provides the feasibility of applying the algorithm for ROP automated screening in clinical.

Artificial intelligence; Deep learning; Retinopathy of prematurity; Assisted diagnosis