Comparison of Glaucoma Probability Score and Moorfields Regression Analysis to Discriminate Glaucomatous and Healthy Eyes

Abstract

To compare sensitivity and specificity values of glaucoma probability score (GPS) and Moorfields regression analysis (MRA) of Heidelberg retina tomograph (HRT) to discriminate between glaucomatous and healthy eyes. A total of 160 eyes of 160 individuals (80 glaucoma patients and 80 healthy subjects) were enrolled in this prospective cross-sectional study. Confocal scanning laser ophthalmoscopy was performed with HRT. Performance of GPS and MRA were evaluated by considering borderline (BL) cases as within normal limits (WNL) or as outside normal limits (ONL). For further analysis glaucoma group was divided into two subsets: initial and moderate/severe damage group. There was no difference between glaucoma patients and normal subjects in terms of gender, disc area, and pachymetry (p>0.05). The GPS was higher, average visual field MD and PSD values were worse, and patients were older in glaucoma group (p=0.001). Sensitivity and specificity values were 72.5% and 93.8% respectively for MRA and 75.0% and 88.8% for GPS when BL cases considered as WNL, and when BL cases considered as ONL these values were 83.8% and 73.8% for MRA and 88.8% and 70.0% for GPS. There was no difference between sensitivity and specificity values of GPS and MRA for either situation (p>0.05). Sensitivity of GPS (76.2%) was higher than of MRA (61.9%) in initial glaucomatous eyes (p=0.317). Diagnostic performance of GPS was similar to MRA. It was found that GPS might differentiate between glaucomatous and healthy eyes with relatively better sensitivity but worse specificity and represent considerable advantage over MRA in early glaucoma cases.