Abstract
Abstract: :
Purpose: To determine how reliably clinicians estimate the proportion of retinal involvement of cytomegalovirus retinitis (CMVR) from fundus photographs. Methods: The proportion of post-equatorial retinal surface area (%PERSA) of 5 small CMVR lesions was measured from Kodachrome transparencies using a quantitative, computerized mapping system. High quality duplicates of the photographs were sent to 75 clinicians with a clinical interest in CMVR. They were asked to estimate the %PERSA of each of the 5 lesions. Clinicians were also sent a questionnaire regarding their specialty, type of practice, and experience with CMVR. Results: There were 40 responses out of 75 solicited (53%). Clinicians overestimated lesion size 3 to 4-fold for each of the 5 lesions (summarized in the table). There were no differences in lesion size estimates based on subspecialty, years of practice, or experience with CMVR. Conclusions: Fundus photographs are used by clinicians to assist in assessment and management of clinical disease. In addition, fundus photographs have been valuable in clinical trial design. CMV retinitis is a prototype disease for which fundus photographs have been useful in the clinical and research arenas. Based on photographic documentation, clinicians overestimate the relative size of small CMV lesions. The area of retinal involvement with CMVR has clinical importance. Area of CMV involvement correlates with visual field loss and retinal detachment risk. If clinicians systematically misjudge this area, they will be unable to meaningfully generalize the results of clinical research into clinical practice. These data likely have broader relevance for other disorders assessed by fundus photographs.
Keywords: cytomegalovirus • imaging/image analysis: clinical