Interpretable multimodal classification for age-related macular degeneration diagnosis

Journal
PLOS ONE
ISSN
1932-6203
Date Issued
2024
Author(s)
Carla Vairetti
Sebastián Maldonado
Loreto Cuitino
Type
journal-article
DOI
URL
Abstract
<jats:p>Explainable Artificial Intelligence (XAI) is an emerging machine learning field that has been successful in medical image analysis. Interpretable approaches are able to “unbox” the black-box decisions made by AI systems, aiding medical doctors to justify their diagnostics better. In this paper, we analyze the performance of three different XAI strategies for medical image analysis in ophthalmology. We consider a multimodal deep learning model that combines optical coherence tomography (OCT) and infrared reflectance (IR) imaging for the diagnosis of age-related macular degeneration (AMD). The classification model is able to achieve an accuracy of 0.94, performing better than other unimodal alternatives. We analyze the XAI methods in terms of their ability to identify retinal damage and ease of interpretation, concluding that grad-CAM and guided grad-CAM can be combined to have both a coarse visual justification and a fine-grained analysis of the retinal layers. We provide important insights and recommendations for practitioners on how to design automated and explainable screening tests based on the combination of two image sources.</jats:p>
Subjects
artificial intelligence

; 

deep learning

; 

humans

; 

image processing, computer-assisted

; 

macular degeneration

; 

multimodal imaging

; 

retina

; 

tomography, optical coherence

; 

hydroxychloroquine

; 

accuracy

; 

age related macular degeneration

; 

article

; 

artificial intelligence

; 

convolutional neural network

; 

deep learning

; 

diagnostic accuracy

; 

entropy

; 

epiretinal membrane

; 

explainable artificial intelligence

; 

human

; 

hyperpigmentation

; 

image analysis

; 

machine learning

; 

macular degeneration

; 

multimodal imaging

; 

myopia

; 

optical coherence tomography

; 

photoreceptor

; 

receiver operating characteristic

; 

reliability

; 

retina blood vessel

; 

retina injury

; 

retina maculopathy

; 

retinal nerve fiber layer thickness

; 

retinal pigment epithelium

; 

retinal thickness

; 

screening test

; 

sensitivity and specificity

; 

task performance

; 

thermography

; 

training

; 

visual field

; 

classification

; 

diagnosis

; 

diagnostic imaging

; 

image processing

; 

macular degeneration

; 

multimodal imaging

; 

optical coherence tomography

; 

pathology

; 

procedures

; 

retina