Abstract
Purpose: :
To determine and compare the sensitivity and specificity of Glaucoma Probability Score (GPS) and Moorfields Regression Analysis (MRA) of the Heidelberg Retina Tomograph (HRT) in differentiating healthy from glaucomatous eyes.
Methods: :
Cross sectional study on 40 POAG patients (40 eyes) 14 males, 26 females (mean age 68.27±13.44 years) and 40 normal subjects (40 eyes) 12 males, 28 females (mean age 57.25±14.85 years). Complete examination, SAP (Humphrey, 30-2 SITA standard), HRT II (software version 3.0) imaging of the optic disc were performed. The sensitivity and specificity of optic nerve head examinations using the GPS (an automated, contour line independent method of optic disc analysis) and the MRA were evaluated by two criteria: by including borderline classifications either as test negatives (criteria 1, most specific) or as test positives (criteria 2, least specific). According to EGS guidelines (Hodapp classification), POAG patients were also divided into 3 groups: early (MD>-6 dB), moderate (-12 dB<MD<-6 dB) and advanced glaucoma (MD<-12 dB). The agreement between global results of both analyses was evaluated with tables of contingency.
Results: :
When borderline cases were considered as test negatives (criteria 1) sensitivity and specificity resulted 67.5% and 85% with GPS and 72.5% and 95% with MRA respectively. Considering borderline cases as test positives (criteria 2) sensitivity and specificity resulted 87.5% and 62.5% with GPS and 92.5% and 85% with MRA respectively. The sensitivity of GPS in eyes with different stages of glaucoma was 60% for early, 73% for moderate and 100% for advanced glaucoma with criteria 1; 88%, 82% and 100% for criteria 2 respectively. The MRA sensitivity was 68% for early, 73% for moderate and 100% for advanced glaucoma with criteria 1; 92%, 91% and 100% respectively with criteria 2. A complete agreement between global classifications of GPS and MRA was found in 70% of POAG and 65% of normal eyes whereas in 30% of POAG and in 27.5% of normal eyes only a partial agreement was found.
Conclusions: :
GPS and MRA showed similar diagnostic performances and a good agreement in most cases, discriminating glaucomatous from normal eyes with good sensitivity and specificity.CR: none
Keywords: imaging/image analysis: clinical • optic disc • image processing