Interobserver Agreement of Novel Classification of Central Serous Chorioretinopathy

Sahoo NK;Parameshwarappa DC;Jhingan M;Tatti F;Iovino C;Peiretti E

2022-01-01

---

Abstract

Objective To validate the newly proposed multimodal-imaging-based classification for central serous chorioretinopathy (CSCR). Methods This was a retrospective study performed in a total of 87 eyes of 44 patients with a diagnosis of CSCR. Multimodal images in the form of auto-fluorescence, fundus fluorescein angiography, indocyanine green angiography, and optical coherence tomography (OCT) images, of all the patients, were presented to two masked retina specialists. The masked observers graded each eye into simple or complex; primary, recurrent, resolved; and specific features such as foveal involvement, outer retinal atrophy, and choroidal neovascularization (CNV). Interobserver agreement was assessed using Cohen's kappa. In areas of non-consensus, a detailed discussion was carried out with a third independent grader. Results The mean age of the 44 patients (32 males and 12 females) was 49.2±9.3 years. We found a moderate-strong agreement between the two observers in all subclassifications, that included "simple or complex" (kappa value=0.91, 95% CI 0.82-0.99, p<0.001); "primary/recurrent/resolved" (kappa value=0.88, 95% CI 0.80-0.96, p<0.001) and "foveal involvement" (kappa value=0.89,95%CI 0.8-0.98, p<0.001). However, there was less agreement between the two graders with respect to classification of "outer retinal atrophy" (kappa value=0.72, 95%CI 0.57-0.87, p<0.001) and "presence/absence of CNV" (kappa value=0.75, 95% CI 0.58-0.92, p<0.001). Non-consensus in categorizing "outer retinal atrophy" was seen in eyes with sub-retinal hyper-reflective material (SHRM) and outer nuclear layer (ONL) thinning overlying subretinal fluid, and non-consensus in categorizing "CNV" was seen in eyes with inner choroidal atrophy. Conclusion Our study reports the validity and strong interobserver agreement in several aspects of the multimodal-imaging-based classification. This could support its implementation in clinical practice and pave way for appropriate treatment guidelines.