No abstract

To investigate the impact of qualitatively graded and deep learning quantified imaging biomarkers on growth of geographic atrophy (GA) secondary to age-related macular degeneration. METHODS: This prospective study included 1062 visits of 181 eyes of 100 patients with GA. Spectral-domain optical coherence tomography (SD-OCT) and fundus autofluorescence (FAF) images were acquired at each visit. Hyperreflective foci (HRF) were quantitatively assessed in SD-OCT volumes using a validated deep learning algorithm. FAF images were graded for FAF patterns, subretinal drusenoid deposits (SDD), GA lesion configuration and atrophy enlargement. Linear mixed models were calculated to investigate associations between all parameters and GA progression. RESULTS: FAF patterns were significantly associated with GA progression (p < 0.001). SDD was associated with faster GA growth (p = 0.005). Eyes with higher HRF concentrations showed a trend towards faster GA progression (p = 0.072) and revealed a significant impact on GA enlargement in interaction with FAF patterns (p = 0.01). The fellow eye status had no significant effect on lesion enlargement (p > 0.05). The diffuse-trickling FAF pattern exhibited significantly higher HRF concentrations than any other pattern (p < 0.001). CONCLUSION: Among a wide range of investigated biomarkers, SDD and FAF patterns, particularly in interaction with HRF, significantly impact GA progression. Fully automated quantification of retinal imaging biomarkers such as HRF is both reliable and merited as HRF are indicators of retinal pigment epithelium dysmorphia, a central pathogenetic mechanism in GA. Identifying disease markers using the combination of FAF and SD-OCT is of high prognostic value and facilitates individualized patient management in a clinical setting.

Purpose: To compare area measurements between swept source optical coherence tomography angiography (SSOCTA), fluorescein angiography (FA), and indocyanine green angiography (ICGA) after applying a novel deep-learning-assisted algorithm for accurate image registration. Methods: We applied an algorithm for the segmentation of blood vessels in FA, ICGA, and SSOCTA images of 24 eyes with treatment-naive neovascular age-related macular degeneration. We trained a model based on U-Net and Mask R-CNN for each imaging modality using vessel annotations and junctions to estimate scaling, translation, and rotation. For fine-tuning of the registration, vessels and the elastix framework were used. Area, perimeter, and circularity measurements were performed manually using ImageJ. Results: Choroidal neovascularization lesion size, perimeter, and circularity delineations showed no significant difference between SSOCTA and ICGA (all P > 0.05). Choroidal neovascularization area showed excellent correlation between SSOCTA and ICGA (r = 0.992) and a Bland–Altman bias of −0.10 ± 0.24 mm2. There was no significant difference in foveal avascular zone size between SSOCTA and FA (P = 0.96) and an extremely small bias of 0.0004 ± 0.04 mm2 and excellent correlation (r = 0.933). Foveal avascular zone perimeter was not significantly different, but foveal avascular zone circularity was significantly different (P = 0.047), indicating that some small cavities or gaps may be missed leading to higher circularity values representing a more round-shaped foveal avascular zone in FA. Conclusion: We found no statistically significant differences between SSOCTA and FA and ICGA area measurements in patients with treatment-naive neovascular age-related macular degeneration after applying a deep-learning-assisted approach for image registration. These findings encourage a paradigm shift to using SSOCTA as a first-line diagnostic tool in neovascular age-related macular degeneration.

No abstract