Abstract
Purpose: :
To investigate if patients experiencing a stroke differ from matched normals in static retinal vessel analysis.
Methods: :
A group of 30 patients after recent ischemic cerebral stroke were imaged by a retinal digital fundus camera. Patients were matched by age and gender to 30 normal control subjects. Retinal photographs were analyzed by a semi-automatic vessel tracking software (Talia Technology, Lod, Israel). The central retinal artery equivalent (CRAE), central retinal vein equivalent (CRVE), as well as the retinal arterio-venous ratio (AVR) were determined. In addition a new index combining several other retinal vessel features such as irregularity was calculated.
Results: :
Mean patient age was 60 years in stroke patients and 61 years in normal subjects (n.s.), 63% were male in both groups. Mean CRAE was 101.9+/-7.7 versus 106.5+/-8.3 microns (p=0.006), CRVE was 122.4+/-11.6 versus 127.6+/-11.8 microns (n.s.) and AVR 0.838+/-0.08 versus 0.838+/-0.06 (n.s.) for stroke patients versus normal subjects, respectively. By considering those conventional vessel features 60% sensitivity at 63% specificity were obtained. Group prediction was improved to a perfect fit (100% sensitivity and specificity) by considering additional vessel features in this limited patient group.
Conclusions: :
Retinal vessel changes are present in patients with stroke versus matched normal subjects. AVR as well as vessel diameters do not distinguish groups sufficiently, with CRAE being most predictive. However, additional vessel features can be extracted by static retinal vessel analysis, which should allow a considerably better prediction.
Keywords: retina • imaging/image analysis: clinical • clinical (human) or epidemiologic studies: risk factor assessment