Premium
Agreement between ophthalmoloscopy and ultrawide field image analysis in an outpatient clinic setting
Author(s) -
Bordet J.,
Caliot J.,
Vardi K.,
Barbe C.,
Ducasse A.,
Angioi K.,
Arndt C.
Publication year - 2015
Publication title -
acta ophthalmologica
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.534
H-Index - 87
eISSN - 1755-3768
pISSN - 1755-375X
DOI - 10.1111/j.1755-3768.2015.0390
Subject(s) - medicine , concordance , outpatient clinic , ophthalmology , optic disc , cohen's kappa , kappa , visual field , optometry , retinal , linguistics , philosophy , machine learning , computer science
Purpose Ultrawide field imaging of the retina is a rapidly developing technology. The purpose of this study was to evaluate the concordance between fundoscopy and ultrawide field image reading both performed by an ophthalmologist. Methods A prospective two center study was performed at two university hospitals (Nancy, Reims) from January 2011 to November 2014. The first eligible patient presenting at the outpatient every working day was included. A wide field imaging optomap ® (optos) was performed by a nurse or an orthoptist before the medical examination. Indirect slit lamp fundoscopy was performed, with or without dilated pupils according to the clinical situation. Then the image lecture was performed by the ophthalmologist. In both cases, the characteristics of the optic disc, the vessel and the retina were noted. The agreement between both examination methods was considered excellent when kappa coefficient (k) was >0.8. Results 901 patients were included in the study. A highly substantial agreement between both examination methods was found with the overall results of k = 0.88 [0.85–0.90]. When comparing the agreement with retinal findings, (k = 0.83 [0.80–0.86]), optic disc findings (k = 0.89 [0.86–0.92]) and vessel analysis (k = 0.84 [0.79–0.89]), it was highly substantial. An abnormal finding was detected on ultrawide field image analysis only and was missed with fundoscopy in 42 eyes. In 12 eyes, an abnormal finding was seen with fundoscopy and was not detected on ultrawide field imaging. Conclusions There appears to be an excellent agreement between ultrawide field image analysis and fundoscopy. However the false negative rate appears to be lower with fundoscopy and ultrawide field image analysis associated.