Deep learning-based single-shot prediction of differential effects of anti-VEGF treatment in patients with diabetic macular edema

Abstract

Anti-vascular endothelial growth factor (VEGF) agents are widely regarded as the first line of therapy for diabetic macular edema (DME) but are not universally effective. An automatic method that can predict whether a patient is likely to respond to anti-VEGF therapy can avoid unnecessary trial and error treatment strategies and promote the selection of more effective first-line therapies. The objective of this study is to automatically predict the efficacy of anti-VEGF treatment of DME in individual patients based on optical coherence tomography (OCT) images. We performed a retrospective study of 127 subjects treated for DME with three consecutive injections of anti-VEGF agents. Patients' retinas were imaged using spectral-domain OCT (SD-OCT) before and after anti-VEGF therapy, and the total retinal thicknesses before and after treatment were extracted from OCT B-scans. A novel deep convolutional neural network was designed and evaluated using pre-treatment OCT scans as input and differential retinal thickness as output, with 5-fold cross-validation. The group of patients responsive to anti-VEGF treatment was defined as those with at least a 10% reduction in retinal thickness following treatment. The predictive performance of the system was evaluated by calculating the precision, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC). The algorithm achieved an average AUC of 0.866 in discriminating responsive from non-responsive patients, with an average precision, sensitivity, and specificity of 85.5%, 80.1%, and 85.0%, respectively. Classification precision was significantly higher when differentiating between very responsive and very unresponsive patients. The proposed automatic algorithm accurately predicts the response to anti-VEGF treatment in DME patients based on OCT images. This pilot study is a critical step toward using non-invasive imaging and automated analysis to select the most effective therapy for a patient's specific disease condition.

Prediction of response to anti‐vascular endothelial growth factor treatment in diabetic macular oedema using an optical coherence tomography‐based machine learning method

“…Rasti et al (2020) recently proposed a novel deep convolutional neural network to predict anti‐VEGF treatment response and reached an average AUC, sensitivity and specificity of 0.866, 0.801 and 0.850. However, only total retinal thicknesses from 127 patients was extracted to perform this task, which was hard to reflect specific pathophysiological mechanisms, and made the model less explainable.…”

“…Nevertheless, to our best knowledge, few studies integrated precursors on OCT to make predictions for DME patients who might receive anti‐VEGF therapy. Recently, Rasti et al (2020) developed a deep learning algorithm to predict the response to anti‐VEGF treatment in DME patients based on total retinal thicknesses extracted from OCT B‐scans. However, the lack of anatomic and pathologic features integrated in the study made it less explainable and hard to reflect relevant pathophysiology mechanisms.…”

Digital health during COVID-19: lessons from operationalising new models of care in ophthalmology

“…23 Moreover, new prospective AI applications have been made use of in ophthalmology for predicting the individual risk of future disease progression of common eye diseases such as diabetic retinopathy, age-related macular degeneration, and glaucoma, [24][25][26][27][28] as well as the likelihood of a successful treatment response to invasive interventions, such as intravitreal injections with antivascular endothelial growth factor (anti-VEGF) agents in retinal diseases. 29 In particular, AI support for treatment decisions for a patient has been applied for the management of age-related macular degeneration, a leading cause of blindness that is often asymptomatic in early stages. 30,31 Novel AI algorithms have been developed to detect, in ophthalmic imaging, early features or quantify subjective measurements, or both, that drive treatment decisions in a reproducible manner, such as the volume of intraretinal fluid and subretinal fluid.…”

Multi-Compartment Spatially-Derived Radiomics From Optical Coherence Tomography Predict Anti-VEGF Treatment Durability in Macular Edema Secondary to Retinal Vascular Disease: Preliminary Findings

“…Recently deep learning (DL)-based models have been evaluated for assessment of various ocular diseases, including diabetic eye disease [35] – [38] . Rasti et al utilized a novel deep convolutional neural network (CNN) using pre-treatment OCT scans as the input for predicting differential retinal thickness following three consecutive anti-VEGF injections with 5-fold cross-validation [35] .…”