Early Diagnosis of Multiple Sclerosis Using Swept-Source Optical Coherence Tomography and Convolutional Neural Networks Trained with Data Augmentation

Abstract

Background: The aim of this paper is to implement a system to facilitate the diagnosis of multiple sclerosis (MS) in its initial stages. It does so using a convolutional neural network (CNN) to classify images captured with swept-source optical coherence tomography (SS-OCT). Methods: SS-OCT images from 48 control subjects and 48 recently diagnosed MS patients have been used. These images show the thicknesses (45 × 60 points) of the following structures: complete retina, retinal nerve fiber layer, two ganglion cell layers (GCL+, GCL++) and choroid. The Cohen distance is used to identify the structures and the regions within them with greatest discriminant capacity. The original database of OCT images is augmented by a deep convolutional generative adversarial network to expand the CNN’s training set. Results: The retinal structures with greatest discriminant capacity are the GCL++ (44.99% of image points), complete retina (26.71%) and GCL+ (22.93%). Thresholding these images and using them as inputs to a CNN comprising two convolution modules and one classification module obtains sensitivity = specificity = 1.0. Conclusions: Feature pre-selection and the use of a convolutional neural network may be a promising, nonharmful, low-cost, easy-to-perform and effective means of assisting the early diagnosis of MS based on SS-OCT thickness data.

Multiple Sclerosis Diagnosis Using Machine Learning and Deep Learning: Challenges and Opportunities

“…However, it should be noted that the studies in the literature that achieved 100% results such as [ 19 , 39 , 44 , 51 , 66 , 81 , 91 , 101 ] suffer from several limitations.…”

“…In addition to the small dataset, the samples were skewed toward MS. In the second study, Dorado et al [ 101 ] used OCT data for analyzing the retinal changes for the diagnosis of MS. A sample of 96 patients was used to train the CNN model. Data augmentation was performed as the CNN model requires a huge dataset to adequately train the model.…”

“…OCT data were only been used in a few studies and showed promising results, [ 61 , 62 , 63 , 101 , 102 ]. Palomar et al [ 61 ] proved that RNFL thickness can be used as a biomarker for MS diagnosis since it attained precision higher than 95%.…”

“…One study utilized retina features for the diagnosis of MS. López-Dorado et al [ 101 ] applied CNNs to the automatic diagnosis of MS in its early stages by analyzing images obtained using SS-OCT. The study used SS-OCT images taken from 48 MS patients and 48 control subjects.…”

Optical Coherence Tomography and Optical Coherence Tomography with Angiography in Multiple Sclerosis

“…In recent years, artificial intelligence (AI) and machine learning have proven useful in medicine. In a recent study, researchers developed a system based on a convolutional neural network that can classify the disease according to the thickness of the OCT scans, thus assisting in the early diagnosis of the disorder [ 40 ]. Machine learning has also been successfully used to predict disability progression in pwMS by analysing RNFL thickness [ 41 ].…”