Detection of glaucoma progression on longitudinal series of en-face macular optical coherence tomography angiography images with a deep learning model.
Mohammadzadeh Vahid, Liang Youwei, Moghimi Sasan, Xie Pengtao, Nishida Takashi, Mahmoudinezhad Golnoush, Eslani Medi, Walker Evan, Kamalipour Alireza, Micheletti Eleonora
AI Summary
A deep learning model effectively detected glaucoma progression using longitudinal macular OCTA images, outperforming traditional methods. This could enhance early clinical detection and management.
Abstract
Background/aims: To design a deep learning (DL) model for the detection of glaucoma progression with a longitudinal series of macular optical coherence tomography angiography (OCTA) images.
Methods
202 eyes of 134 patients with open-angle glaucoma with ≥4 OCTA visits were followed for an average of 3.5 years. Glaucoma progression was defined as having a statistically significant negative 24-2 visual field (VF) mean deviation (MD) rate. The baseline and final macular OCTA images were aligned according to centre of fovea avascular zone automatically, by checking the highest value of correlation between the two images. A customised convolutional neural network (CNN) was designed for classification. A comparison of the CNN to logistic regression model for whole image vessel density (wiVD) loss on detection of glaucoma progression was performed. The performance of the model was defined based on the confusion matrix of the validation dataset and the area under receiver operating characteristics (AUC).
Results
The average (95% CI) baseline VF MD was -3.4 (-4.1 to -2.7) dB. 28 (14%) eyes demonstrated glaucoma progression. The AUC (95% CI) of the DL model for the detection of glaucoma progression was 0.81 (0.59 to 0.93). The sensitivity, specificity and accuracy (95% CI) of DL model were 67% (34% to 78%), 83% (42% to 97%) and 80% (52% to 95%), respectively. The AUC (95% CI) for the detection of glaucoma progression based on the logistic regression model was lower than the DL model (0.69 (0.50 to 0.88)).
Conclusion
The optimised DL model detected glaucoma progression based on longitudinal macular OCTA images showed good performance. With external validation, it could enhance detection of glaucoma progression.
Trial registration number: NCT00221897.
MeSH Terms
Shields Classification
Key Concepts4
The deep learning (DL) model, a customised convolutional neural network (CNN), for the detection of glaucoma progression using longitudinal series of macular optical coherence tomography angiography (OCTA) images, achieved an AUC of 0.81 (95% CI: 0.59 to 0.93) in 202 eyes of 134 patients with open-angle glaucoma followed for an average of 3.5 years.
The deep learning (DL) model, a customised convolutional neural network (CNN), for the detection of glaucoma progression using longitudinal series of macular optical coherence tomography angiography (OCTA) images, demonstrated a sensitivity of 67% (95% CI: 34% to 78%), specificity of 83% (95% CI: 42% to 97%), and accuracy of 80% (95% CI: 52% to 95%) in 202 eyes of 134 patients with open-angle glaucoma.
The AUC for the detection of glaucoma progression based on a logistic regression model was 0.69 (95% CI: 0.50 to 0.88), which was lower than the deep learning (DL) model, in 202 eyes of 134 patients with open-angle glaucoma followed for an average of 3.5 years.
In a cohort of 202 eyes of 134 patients with open-angle glaucoma followed for an average of 3.5 years, 28 (14%) eyes demonstrated glaucoma progression, defined as having a statistically significant negative 24-2 visual field (VF) mean deviation (MD) rate.
Related Articles5
Optic Disc Microvasculature Reduction and Visual Field Progression in Primary Open-Angle Glaucoma.
Case SeriesDecreased macular deep capillary plexus is associated with functional progression of normal tension glaucoma patients with unilateral visual field loss.
Longitudinal StudyLongitudinal changes in optical coherence tomography angiography characteristics in normal-tension glaucoma with or without high myopia.
Cohort StudyFrequency of Agreement Between Structural and Functional Glaucoma Testing: A Longitudinal Study of 3D OCT and Current Clinical Tests.
Cohort StudyRelationship Between Retinal Oxygen Saturation and the Severity of Visual Field Damage in Glaucoma.
Cross-Sectional StudyIs this article assigned to the wrong chapter(s)? Let us know.