Purchase this article with an account.
Sheila C Nemeth, Ian J.C. MacCormack, Simon P. Harding, Susan Lewallen, Terrie E. Taylor, Peter Soliz, Vinayak S Joshi; Comparing the effectiveness of single posterior pole retinal images against binocular indirect ophthalmoscopic exam in detecting retinal hemorrhages in pediatric cerebral malaria. Invest. Ophthalmol. Vis. Sci. 2015;56(7 ):5912.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Detecting malarial retinopathy (MR) is important for accurate case classification and clinical diagnosis of cerebral malaria (CM). The use of portable retinal cameras has potential for MR detection in low-resource clinical environments. However, they may have limitations in their field of view (FOV). We investigated the ability of single FOV retinal images to detect one of the characteristic lesions of MR, retinal hemorrhages.<br />
We analyzed images the posterior pole from 42 consecutive pediatric admissions with suspected CM (Research Ward, Blantyre, Malawi, 2014). Images were acquired with a Topcon 50-EX fundus camera, and were graded for the presence/absence of retinal hemorrhages in the macula, defined as a circle centered on the center of the fovea with a radius of 3.8mm, or, 2.5 ETDRS disc diameters. The clinical binocular indirect ophthalmoscopy examination (BIO) of the entire retina was then compared to these findings. Specificity and sensitivity of the detection of hemorrhages were calculated in relation to the BIO findings.<br />
The manual grading of a posterior pole image for presence/absence of hemorrhages against BIO as the reference standard achieved the following performance metrics: sensitivity = 0.68 (95% CI, 0.48-0.83); specificity = 1.0 (95% CI, 0.73-1); Positive predictive value = 1.0 (95% CI, 0.79-1); Negative predictive value = 0.61 (95% CI, 0.39-0.80).<br /> <br />
Single field posterior pole photography has only moderate sensitivity at 68% for the detection of hemorrhages associated with MR. As expected, single field posterior pole imaging is highly specific in determining the absence of hemorrhages characteristic of CM. Our preliminary findings suggest that peripheral detection of the retina using portable cameras for MR detection will need to be included to add the ability to detect hemorrhages outside the posterior pole. Including other signs of CM (such as retinal whitening) may also increase the sensitivity.
This PDF is available to Subscribers Only