Developing an image-based grading scale for peripheral drusen to investigate associations of peripheral drusen type with age-related macular degeneration

Paripoorna Sharma,Fritz Gerald P. Kalaw,Andrew Lin,Evan H. Walker,Shyamanga Borooah
DOI: https://doi.org/10.1038/s41598-024-70352-3
IF: 4.6
2024-09-02
Scientific Reports
Abstract:Age-related macular degeneration (AMD) is a leading cause of blindness. It is associated with peripheral drusen which has not been categorized. We investigated peripheral drusen to validate an image grading system and to understand possible associations between peripheral drusen and AMD. We collated clinical data, ultra-widefield (UWF) pseudocolor fundus images and Spectral-Domain Optical Coherence Tomography (SD-OCT) scans from consecutive retinal patients. SD-OCT scans were used to determine AMD stage. A masked retinal specialist recorded the types of peripheral drusen observed in UWF images. Eyes whose UWF images did not pass quality screening and those without AMD and peripheral drusen were excluded from the study. Statistical tests were utilized to determine the validity of our grading system and associations of peripheral drusen with AMD. A total of 481 eyes (283 subjects) were included in the study (mean age 73.1 ± 1.2years, 64.3% female). Interobserver and test–retest statistical analyses to evaluate the UWF image grading system resulted in Cohen's Kappa 0.649 ( p < 0.001) and 0.922 ( p < 0.001) respectively. A total of 284 (59.0%), 28 (5.8%), 15 (3.1%), 22 (4.6%), 4 (0.8%), 39 (8.1%), and 32 (6.7%) eyes had hard, soft, reticular, cuticular, atrophic, mixed drusen, and mixed drusen and atrophy respectively in at least one peripheral retinal quadrant. Hard peripheral drusen was significantly associated with the presence of AMD ( p = 0.010). Peripheral drusen types were variably seen in retinal patients with and without AMD. We validated a peripheral drusen grading system and provided an image library to assist in the identification of peripheral drusen. Our study found an association between peripheral hard drusen and an AMD diagnosis but did not find a link between peripheral drusen and severity of AMD.
multidisciplinary sciences
What problem does this paper attempt to address?