Influence of fellow eye on the diagnosis and classification of central serous chorioretinopathy.

Abstract

To assess the influence of fellow eye information on diagnosis and classification of central serous chorioretinopathy (CSCR) using multimodal imaging-based classification. This was a retrospective, observational study including patients with CSCR with unilateral or bilateral involvement. Multimodal images of both eyes of each patient were compiled and presented to two masked retina specialists subdivided into three groups: (1) both eye data, (2) right eye data and (3) left eye data. The masked observers graded the CSCR as per the new CSCR classification into simple and complex CSCR in three different scenarios as subdivided above. Interobserver and intraobserver agreement was assessed using Cohen's kappa (95% confidence intervals). A total of 206 eyes of 103 patients with unilateral or bilateral CSCR were graded. We found a "strong" intraobserver and interobserver agreement when one eye versus both eye data was provided in terms of "simple CSCR" or "complex CSCR" (kappa value = 0.77 and 0.87, p < 0.001, and kappa value = 0.85 and 0.76, p < 0.001, respectively). Forty-three eyes (10.55%) showed discrepancy in classification between observers for individual eyes, whereas only 13 eyes (6.53%) showed discrepancy between observers when both eye data was provided. We conclude that fellow eye information was helpful in solving diagnostic dilemmas and reached 85% consensus in the diagnosis of CSCR between the observers. We found that information of fellow eyes led to a discrepancy only in 6.53% cases with 2.42% cases that had a controversial diagnosis of CSCR. Multimodal imaging-based CSCR classification provides objective approach to diagnose and classify CSCR.