Deep Learning for Diagnosing and Segmenting Choroidal Neovascularization in OCT Angiography in a Large Real-World Data Set

Abstract

Purpose To diagnose and segment choroidal neovascularization (CNV) in a real-world multicenter clinical OCT angiography (OCTA) data set using deep learning. Methods A total of 105,66 OCTA scans from 3135 eyes, including 4701 with CNV and 5865 without, were collected in five eye clinics. Both 3 × 3-mm and 6 × 6-mm scans of the central and temporal macula were included. Scans with CNV were collected from multiple diseases, and scans without CNV were collected from both healthy controls and those with multiple diseases. No scans were removed during training or testing due to poor quality. The trained hybrid multitask convolutional neural network outputs a CNV diagnosis and membrane segmentation, respectively. Results The model demonstrated a highly accurate CNV diagnosis (area under receiver operating characteristic curve = 0.97), achieving a sensitivity of 95% at 95% specificity. The model also correctly segmented CNV lesions (F1 score = 0.78 ± 0.19). Additionally, model performance was comparable on both high-definition 3 × 3-mm scans and low-definition 6 × 6-mm scans. The model did not suffer large performance variations under different diseases. We also show that a subclinical lesion in a patient with neovascular age-related macular degeneration can be monitored over a multiyear time frame using our approach. Conclusions The proposed method can accurately diagnose and segment CNV in a large real-world clinical data set. Translational Relevance The algorithm could enable automated CNV screening and quantification in the clinic, which will help improve CNV diagnosis and treatment evaluation.

Optical Coherence Tomography Angiography: A 2023 Focused Update on Age-Related Macular Degeneration

“…MNV, for instance, may be underestimated as a result of signal attenuation and projection errors, or it may be misinterpreted as non-MNV pathology that is included in the segmentation result (e.g., drusen). Errors can also be associated with low contrast between neovascularization vessels and the background signal, requiring time-consuming review of images [ 96 ]. Recently, DL systems incorporated OCTA into the model training for computer-aided diagnosis (CAD) of retinal diseases, including AMD [ 96 – 100 ].…”

“…Errors can also be associated with low contrast between neovascularization vessels and the background signal, requiring time-consuming review of images [ 96 ]. Recently, DL systems incorporated OCTA into the model training for computer-aided diagnosis (CAD) of retinal diseases, including AMD [ 96 – 100 ]. Recently, Wang and colleagues proposed a CNN-based model by using OCTA real-world multicenter dataset for MNV diagnosis and segmentation in AMD and non-AMD pathologies.…”

“…On the other hand, 6 × 6 mm scans typically reported reduced signal-to-noise ratio and an increased rate of artifacts. Furthermore, this model effectively evaluated MNV progression, as a rising number of pixels in correspondence of neovascularization could predict its exudation [ 96 ]. Notably, several multimodal imaging DL models also including OCTA have been proposed.…”

A Complete Review of Automatic Detection, Segmentation, and Quantification of Neovascularization in Optical Coherence Tomography Angiography Images

“…One of the major challenges faced in this research area is the complexity of images. Most techniques struggle with low image quality, as well as noise and background patterns that resemble NV intensity, NV patterns, and scattered artifacts [23][24][25][26][27][28][29]. These issues often lead to over-segmentation and under-segmentation, resulting in low precision and recall, especially in CNV cases [30].…”

Quantitative Characterization of Retinal Features in Translated OCTA

“…In recent years, OCTA has been demonstrated to identify, detect, and predict DR, 12–19 AMD, 20–22 Glaucoma 23 and several other retinal diseases. 24–31 Despite the advantages, widespread deployment of OCTA has been limited due to the high device cost. 32,33 The additional requirements of hardware and software for an OCTA device pose a financial burden for clinics as well as patients, therefore, there are only a limited number of hospitals and retinal clinics that use OCTA on a daily basis, for routine ophthalmic check-ups.…”