Magdalena Baratsits

Drusen volume development over time and its relevance to the course of age-related macular degeneration

Aims To quantify the change in drusen volume over time and identify its prognostic value for individual risk assessment. Methods A prospective observational study over a minimum of 3 years and maximum of 5 years and follow-up examination every 3 months was conducted at the ophthalmology department of the Medical University of Vienna. 109 patients presenting early and intermediate age-related macular degeneration (AMD) were included, of which 30 patients concluded a regular follow-up for at least 3 years. 50 eyes of 30 patients were imaged every 3 months using spectral-domain and polarisation-sensitive optical coherence tomography (OCT). Drusen volume was measured using an automated algorithm. Data of a 6-month follow-up were segmented manually by expert graders. Results Gradings from 24 000 individual B-scans showed solid correlation between manual and automated segmentation with an initial mean drusen volume of 0.17 mm3. The increase in drusen volume was shown to be comparable among all eyes, and a model for long-term drusen volume development could be fitted as a cubic polynomial function and an R2=0.955. Spontaneous drusen regression was observed in 22 of 50 eyes. In this group, four eyes developed choroidal neovascularisation and three geographic atrophy. Conclusions Drusen volume increase over time can be described by a cubic function. Spontaneous regression appears to precede conversion to advanced AMD. OCT might be a promising tool for predicting the individual risk of progression of AMD.

Machine Learning of the Progression of Intermediate Age-Related Macular Degeneration Based on OCT Imaging

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.

Detection and Differentiation of Intraretinal Hemorrhage in Spectral Domain Optical Coherence Tomography

ABSTRACT Purpose: The purpose of this study was to classify and detect intraretinal hemorrhage (IRH) in spectral domain optical coherence tomography (SD-OCT). Methods: Initially the presentation of IRH in BRVO-patients in SD-OCT was described by one reader comparing color-fundus (CF) and SD-OCT using dedicated software. Based on these established characteristics, the presence and the severity of IRH in SD-OCT and CF were assessed by two other masked readers and the inter-device and the inter-observer agreement were evaluated. Further the area of IRH was compared. Results: About 895 single B-scans of 24 eyes were analyzed. About 61% of SD-OCT scans and 46% of the CF-images were graded for the presence of IRH (concordance: 73%, inter-device agreement: k = 0.5). However, subdivided into previously established severity levels of dense (CF: 21.3% versus SD-OCT: 34.7%, k = 0.2), flame-like (CF: 15.5% versus SD-OCT: 45.5%, k = 0.3), and dot-like (CF: 32% versus SD-OCT: 24.4%, k = 0.2) IRH, the inter-device agreement was weak. The inter-observer agreement was strong with k = 0.9 for SD-OCT and k = 0.8 for CF. The mean area of IRH detected on SD-OCT was significantly greater than on CF (SD-OCT: 11.5 ± 4.3 mm2 versus CF: 8.1 ± 5.5 mm2, p = 0.008). Conclusions: IRH seems to be detectable on SD-OCT; however, the previously established severity grading agreed weakly with that assessed by CF.

Multimodal Imaging of Cotton Wool Spots in Branch Retinal Vein Occlusion

Purpose: To describe and follow cotton wool spots (CWS) in branch retinal vein occlusion (BRVO) using multimodal imaging. Methods: In this prospective cohort study including 24 patients with new-onset BRVO, CWS were described and analyzed in color fundus photography (CF), spectral domain optical coherence tomography (SD-OCT), infrared (IR) and fluorescein angiography (FA) every 3 months for 3 years. The CWS area on SD-OCT and CF was evaluated using OCT-Tool-Kit software: CWS were marked in each single OCT B-scan and the software calculated the area by interpolation. Results: 29 central CWS lesions were found. 100% of these CWS were visible on SD-OCT, 100% on FA and 86.2% on IR imaging, but only 65.5% on CF imaging. CWS were visible for 12.4 ± 7.5 months on SD-OCT, for 4.4 ± 3 months and 4.3 ± 3.4 months on CF and on IR, respectively, and for 17.5 ± 7.1 months on FA. The evaluated CWS area on SD-OCT was larger than on CF (0.26 ± 0.17 mm2 vs. 0.13 ± 0.1 mm2, p < 0.0001). The CWS area on SD-OCT and surrounding pathology such as intraretinal cysts, avascular zones and intraretinal hemorrhage were predictive for how long CWS remained visible (r2 = 0.497, p < 0.002). Conclusions: The lifetime and presentation of CWS in BRVO seem comparable to other diseases. SD-OCT shows a higher sensitivity for detecting CWS compared to CF. The duration of visibility of CWS varies among different image modalities and depends on the surrounding pathology and the CWS size.