Discrepancy in current central serous chorioretinopathy classification

Abstract

Aim: To report the discordance in central serous chorioretinopathy (CSCR) classification among practising retina specialists. 

Methods: The study conducted was a multicentre survey. Multimodal retinal images along with relevant clinical details of 100 cases diagnosed as CSCR (from six centres) were circulated among six retina specialists across the globe. The image sets included colour fundus photographs, fundus autofluorescence images, optical coherence tomography b-scans, fluorescein and indocyanine green angiography of the study and fellow eyes. The graders were asked to classify the disease of study eye, according to their own criteria. The graders were masked to the responses of other graders. The final analysis of the pooled response data was done based on the diagnosis of study eye only. The main outcome measure was degree of agreement between six independent observers using Fleiss Kappa statistics. 

Results: Grading for 100 eyes of 100 patients (men, 93%) was included in the analysis. 20 patients had a history of steroid use. The graders provided 36 different terms to classify the disease, with poor agreement among graders (Fleiss Kappa=0.134). The consistency in diagnosing acute CSCR was statistically higher than for either chronic (p=0.012) or recurrent CSCR (p<0.0001). When collapsing descriptors into six main terms, agreement remained poor (Fleiss Kappa=0.218). 

Conclusion: The high discordance among experienced retina specialists in describing CSCR clinical subtypes is highlighted. The current work demonstrates the limitations of current empirical CSCR classification methods and the need for a more objective and refined system to bring uniformity in diagnosis and prognostication of the disease.

More information

Identifiers

Catalogue record

Export record

ASCII CitationAtomBibTeXData Cite XMLDublin CoreDublin CoreEP3 XMLEndNoteHTML CitationHTML CitationHTML ListJSONMETSMODSMPEG-21 DIDLOpenURL ContextObjectOpenURL ContextObject in SpanRDF+N-TriplesRDF+N3RDF+XMLRIOXX2 XMLReferReference ManagerSimple Metadata

Altmetrics