Visual Field Prediction using Recurrent Neural Network

Keunheung Park,Jinmi Kim,Jiwoong Lee

doi:10.1038/s41598-019-44852-6

Keunheung Park, Jinmi Kim + Show 1 more

Open Access

https://doi.org/10.1038/s41598-019-44852-6

Copy DOI

Abstract

Artificial intelligence capabilities have, recently, greatly improved. In the past few years, one of the deep learning algorithms, the recurrent neural network (RNN), has shown an outstanding ability in sequence labeling and prediction tasks for sequential data. We built a reliable visual field prediction algorithm using RNN and evaluated its performance in comparison with the conventional pointwise ordinary linear regression (OLR) method. A total of 1,408 eyes were used as a training dataset and another dataset, comprising 281 eyes, was used as a test dataset. Five consecutive visual field tests were provided to the constructed RNN as input and a 6th visual field test was compared with the output of the RNN. The performance of the RNN was compared with that of OLR by predicting the 6th visual field in the test dataset. The overall prediction performance of RNN was significantly better than OLR. The pointwise prediction error of the RNN was significantly smaller than that of the OLR in most areas known to be vulnerable to glaucomatous damage. The RNN was also more robust and reliable regarding worsening in the visual field examination. In clinical practice, the RNN model can therefore assist in decision-making for further treatment of glaucoma.

Highlights

Artificial intelligence capabilities have, recently, greatly improved
We evaluated the performance of the recurrent neural network (RNN) by comparison with the conventional ordinary linear regression method (OLR)
The study was approved by the institutional review board (IRB) of Pusan National University Hospital; the requirement for patient consent was waived by the IRB because of the retrospective nature of the study

Summary

Introduction

In the past few years, one of the deep learning algorithms, the recurrent neural network (RNN), has shown an outstanding ability in sequence labeling and prediction tasks for sequential data. We built a reliable visual field prediction algorithm using RNN and evaluated its performance in comparison with the conventional pointwise ordinary linear regression (OLR) method. The performance of the RNN was compared with that of OLR by predicting the 6th visual field in the test dataset. Murata et al.[10] used a type of machine learning algorithm, variational Bayes linear regression (VBLR); they reported that it demonstrated superior performance, compared with pointwise linear regression. Visual field progression is a prototypical complicated problem with many unpredictable errors and large variations among patients. In the past few years, 1 deep learning algorithm, the recurrent neural network (RNN), has shown outstanding achievement. A series of visual field examinations www.nature.com/scientificreports/

Objectives

Methods

Results

Conclusion