HRT III glaucoma probability score and Moorfields regression across the glaucoma spectrum.

Objective

To compare the agreement, sensitivity, and specificity of the Heidelberg Retina Tomograph III Glaucoma Probability Score (GPS) and Moorfields Regression Analysis (MRA) across the spectrum of glaucomatous visual field (VF) loss.

Design

Retrospective observational study.

Methods

Data from 247 glaucoma patients and 142 controls who underwent standard achromatic perimetry (SITA-SAP) and Heidelberg Retina Tomograph III imaging within 6 months were analyzed. Sensitivity, specificity, agreement, and discrimination capability of MRA and GPS were assessed.

Results

Age-adjusted specificity was 92% and 93% and sensitivity was 76.88 and 80.85 for GPS and MRA, respectively. Sensitivity for early VF loss [mean deviation (MD) < -5 dB] (N=81) was 66.64% and 69.82%, for moderate VF loss (-5 dB <MD < -15 dB) (N=104) was 78.91% and 85.13%, and for advanced VF loss (MD> -15 dB) (N=62) was 87.70% and 86.48% (GPS and MRA, respectively). Age-specific receiver operating characteristics ranged from 0.89 to 0.92 and from 0.87 to 0.90 (GPS and MRA, respectively). Kappa ranged from 0.64 to 0.77.

Conclusions

Specificity for MRA and GPS was similar and agreement was good. GPS offered slightly higher age-specific receiver operating characteristic. GPS, which does not require contour line placement, may have a potential role in screening for severe glaucomatous damage.