Comparison Among Methods of Retinopathy Assessment (CAMRA) study: Mydriatic, Non-Mydriatic, and Smartphone Fundus Photography

Abstract

 

Purpose

 

Fundus photography is a convenient though expensive tool for the assessment of diabetic retinopathy (DR)^1-3. A new alternative is to use a smartphone camera^4,5. The purpose of this study was to compare the effectiveness of 3 fundus photography modalities to detect DR.

 

Methods

 

Three hundred diabetic patients (600 eyes) were recruited at a tertiary care center in Chennai, India. 3-field nonmydriatic (NM) photography was performed using the Nidek [NIDEK Co., Hiroishi, Japan] Model AFC-230. Next, the patient’s eyes were dilated and photographs were taken using the iPhone 5 (Apple, Inc., Cupertino, CA), Filmic Pro software [Cinegenix LLC, Seattle, WA] and a 20D condensing lens using previously reported techniques⁴. Then 7-field photography was taken with the Zeiss FF450 Plus [Carl Zeiss Meditec, Inc. Dublin, CA]. Photos were reviewed in a masked fashion by two retina specialists for presence and severity of DR. Presence of macular edema or a grade worse than severe non-proliferative DR was considered sight threatening (STDR).

 

Results

 

Of 600 eyes, 11 (1.8%), 9 (1.5%) and 0 of the images were not graded due to poor quality for the SP, NM, and mydriatic cameras respectively. The sensitivity and specificity of NM detection of any DR compared to the mydriatic photos was 0.81 (95% CI 0.75-0.86) and 0.94 (95% CI 0.92-0.96), respectively. The sensitivity and specificity of SP detection of any DR was 0.50 (95% CI 0.43-0.56) and 0.94 (95% CI 0.92-0.97). The sensitivity and specificity of NM detection of STDR was 0.54 (95% CI 0.40-0.67) and 0.99 (95% CI 0.98-1.00), respectively. The sensitivity and specificity of SP detection of sight threatening DR was 0.59 (95% CI 0.46-0.72) and 1.00 (95% CI 0.99-1.00). Representative images are shown in Figure 1.

 

Conclusions

 

SP photography offers advantages such as connectivity, portability and low cost. The SP was less sensitive for detection of DR compared with NM imaging. However, they performed similarly for detection of STDR. We determine that this technique of SP photography is inferior to NM fundus photos for screening of DR in this population. As SP technology and advanced optics improve, we feel that future technologies have potential to become reasonable alternatives to traditional fundus cameras in telemedical screening.  

 

Representative photos from each modality. Top and bottom series show no retinopathy and macular edema respectively