Evaluation of Semiautomated Measurement of Geographic Atrophy in Age-Related Macular Degeneration by Fundus Autofluorescence in Clinical Setting

Abstract

Region Finder is a reliable tool for the identification and quantification of geographic atrophy in patients with age-related macular degeneration, in a clinical setting even when performed by junior reader.

Prevalence and quantification of geographic atrophy associated with newly diagnosed and treatment-naïve exudative age-related macular degeneration

“…Using data from RFA and multimodal imaging, which is a more reliable approach compared with evaluation of FAF imaging only,26 GA was present in 24.3% of study eyes newly diagnosed with exudative AMD. In 2.9% of eyes (corresponding to patients with poor image quality of FAF frames or with advanced neovascular lesion with extensive subretinal fibrosis which made a grading impossible) neither multimodal reading nor reading by the third grader allowed to decide whether atrophy was present or not.…”

“…RFA, based on FAF, offers an accurate, reproducible and time-efficient tool for identification and quantification of GA23 30 and can be used in daily clinical practice 26. Other additional imaging methods, including SD-OCT with integrated microperimetry,34 or the ‘Index of Retinal Autofluorescence’ applied to FAF imaging,35 may further improve the accuracy and reproducibility of atrophy quantification.…”

“…The increased autofluorescence intensity in the perilesional zone of atrophy helps to predict future atrophy progression 23 24. Using Region Finder Analyser (RegionFinder Software (RFA), Heidelberg Engineering, Heidelberg, Germany), a novel semi-automated customised software based on FAF imaging and embedded in Spectralis (Spectralis HRA+OCT; Heidelberg Engineering, Heidelberg, Germany) atrophic patches can be quantified and the spread of the total size can be determined over time 25 26…”

Automated segmentation of geographic atrophy of the retinal epithelium via random forests in AREDS color fundus images

“…FAF and OCT segmentation methods undoubtedly result in the high quality segmentation. Segmentation methods on FAF and OCT have achieved high accuracy, reproducibility, and agreement, commonly displaying segmentation accuracy above 0.9 [12,15,23,24,30,33,35,36,39,45]. However, this disparity in segmentation performance between color fundus and FAF and OCT is largely due to differences in imaging modalities rather than faults in segmentation method [35].…”

“…Semi-automated GA segmentation of FAF images has also been pursued by using commercial packages such as RegionFinder software (Heidelberg Engineering, Heidelberg, Germany) [33,36]. …”