Improved predictive diagnosis of diabetic macular edema based on hybrid models: an observational study

Abstract

Diabetic Macular Edema (DME) is the most common sight-threatening complication of type 2 diabetes. Our goal was to develop an alternative method to optical coherence tomography (OCT) for DME diagnosis by introducing spectral information derived from spontaneous electroretinogram (ERG) signals as a single input or combined with eye fundus. To this end, an observational study was completed (n = 233 participants). Basal ERGs were used to generate scalograms and spectrograms via Wavelet and Fourier transforms, respectively. Using transfer learning, distinct Convolutional Neural Networks (CNN) were trained as classifiers for DME using OCT, scalogram, spectrogram, and fundus images. Input data were randomly split into training and test sets with a proportion of 80% to 20%, respectively. The top performers for each input type were selected, OpticNet-71 for OCT and DenseNet-201 for fundus and non-evoked ERG-derived scalograms, to generate a combined model by assigning different weights for each of the selected models. Model validation was performed using a dataset alien to the training phase of the models. None of the models powered by non-evoked ERG-derived input performed well. Metrics of the best hybrid models were all above 0.81 for fundus combined with non-evoked ERG-derived information; and above 0.85 for OCT combined with non-evoked ERG-derived scalogram images. These data show that the spontaneous ERG-based model improves all the performance metrics of the fundus and OCT-based models, with the exception of sensitivity for the OCT model, to predict DME. Combining non-evoked ERG with OCT represents an improvement to the existing OCT-based models, and combining non-evoked ERG with fundus is a reliable and economical alternative for the diagnosis of DME in underserved areas where OCT is unavailable.