Detecting retinal lesions with automated perimetry

Abstract

This paper considers the clinical problems associated with the detection of localized retinal lesions using automated perimeters. We present four case studies of patients who had documented macular lesions, some about the size of the papilla, but who had normal visual field results when tested with automated perimeters using central threshold programs. Three factors were identified that limit an automated perimeter in detecting a lesion. These are: 1. misalignment between the test grid and lesion, 2. subject threshold fluctuations and, 3. inappropriate use of grey scales. Several strategies such as the use of sectional cuts or fine grids, the use of a simple statistical analysis or the use o..