Search

Purpose: The purpose of this study is to compare the ability of spectral domain optical coherence tomography angiography (SD-OCTA) and indocyanine green angiography (ICGA) to detect and measure lesion area in patients with type 1 and 2 choroidal neovascularization (CNV)., Methods: Types 1 and 2 neovascular AMD (nAMD) were included in this prospective and observational case series. ETDRS best-corrected visual acuity (BCVA), ophthalmic examination with funduscopy, OCTA (AngioVue), fluorescein angiography (FA), ICGA, and OCT (Spectralis) were performed. CNV measurements were done manually by two experienced graders using the systems' innate region selection tools., Results: Forty eyes of 39 consecutive patients with nAMD were included. Mean age was 77 ± 6.4 years, ETDRS BCVA was 67 ± 13 letters, and 11 eyes were treatment naïve. Nineteen CNV lesions were classified as type 1 and 21 as type 2. ICGA was able to identify CNV in all eyes. By contrast, OCTA detected CNV in 95% of type 1 and 86% of type 2 nAMD eyes. Mean overall CNV area (CNV-A) was 2.8 ± 2.7 mm2 in ICGA and 2.1 ± 2.7 mm2 in OCTA. Both lesion types CNV-A appeared significantly smaller in OCTA compared with ICGA (P < 0.01). Bland-Altman plot revealed a mean difference (bias) between OCTA and ICGA CNV-A of 0.76 ± 1.74 mm2. Intraclass correlation coefficient (ICC) for CNV-A was 0.91 and 0.93 for ICGA and OCTA, respectively. ICGA CNV-A in the four OCTA-negative eyes (median 4.7 mm2) was not significantly different from the 36 OCTA-positive eyes (median 1.7 mm2)., Conclusions: Type 1 and 2 CNV-A were significantly smaller in OCTA than in ICGA. OCTA was generally less successful in detecting CNV than ICGA in patients who were included into this study based on FA and OCT. However, OCTA detected all type 1 lesions except for one, indicating that the SD-OCTA signal is limited by detection limits of blood flow velocity rather than lesion type. Further efforts are needed pushing the limits of lowest detectable and fastest distinguishable flow until OCTA can deliver realistic qualitative and quantitative imaging of type 1 and 2 CNV for diagnosis and monitoring.

Purpose: To develop a data-driven interpretable predictive model of incoming drusen regression as a sign of disease activity and identify optical coherence tomography (OCT) biomarkers associated with its risk in intermediate age-related macular degeneration (AMD)., Methods: Patients with AMD were observed every 3 months, using Spectralis OCT imaging, for a minimum duration of 12 months and up to a period of 60 months. Segmentation of drusen and the overlying layers was obtained using a graph-theoretic method, and the hyperreflective foci were segmented using a voxel classification method. Automated image analysis steps were then applied to identify and characterize individual drusen at baseline, and their development was monitored at every follow-up visit. Finally, a machine learning method based on a sparse Cox proportional hazard regression was developed to estimate a risk score and predict the incoming regression of individual drusen., Results: The predictive model was trained and evaluated on a longitudinal dataset of 61 eyes from 38 patients using cross-validation. The mean follow-up time was 37.8 ± 13.8 months. A total of 944 drusen were identified at baseline, out of which 249 (26%) regressed during follow-up. The prediction performance was evaluated as area under the curve (AUC) for different time periods. Prediction within the first 2 years achieved an AUC of 0.75., Conclusions: The predictive model proposed in this study represents a promising step toward image-guided prediction of AMD progression. Machine learning is expected to accelerate and contribute to the development of new therapeutics that delay the progression of AMD.