Abstract
Purpose: :
Evaluation of the diagnostic benefit of a fast comprehensive multi parameter test on disc morphology and its relationship to RNFL thickness measurements in preperimetric and early perimetric glaucoma.
Methods: :
64 controls, 77 preperimetric primary open angle glaucoma (POAG) and 63 early perimetric POAG patients were recruited from the Erlangen Glaucoma Registry. Heidelberg Retina Tomograph (HRT 3) and Spectralis HRA+OCT (both Heidelberg Engineering, Germany) scans were obtained. Retinal nerve fibre layer (RNFL) measurements (circular B-scans: 3.4 mm diameter, 768 A-scans, 25 sweeps) were used as observer independent reference. HRT images were anonymized, masked and evaluated by three experienced investigators. HRT qualitative variables were reflectivity of retinal surface, configuration of double-hump pattern, shape of the cup and fulfilled ISNT-rule. The standard quantitative variables were rim-volume, cup shape measure, FSM, RB, Moorfields analysis (MRA) and glaucoma probability score (GPS). A new additive score for glaucoma was created ranging from 0 to 10. Area under ROC (AUROC) and specificities at a fixed sensitivity have been calculated (SPSS, Version 18.0; Chicago, IL, USA) to evaluate predictive value for glaucoma.
Results: :
The overall HRT score and RNFL-thickness were significantly different from normals in both diagnostic groups. AUROC curves for preperimetric POAG were: averaged score 0.789, MRA 0.744, GPS 0.686 and for RNFL-thickness 0.916. For perimetric POAG: averaged score 0.920, MRA 0.853, GPS 0.826 and for RNFL-thickness 0.965. Sensitivities at fixed specificity of at least 90% were for preperimetric POAG: averaged score 41.7 %, MRA 32.9%, GPS 28.9 % and for RNFL-thickness 78.9%. For perimetric PAOG: averaged score 73.1 %, MRA 58.1%, GPS 48.4 % and for RNFL-thickness 93.5%. The cut off point for the averaged score was 3.
Conclusions: :
To differentiate normal from glaucoma eyes RNFL-thickness with Spectralis was the variable with the highest sensitivity followed by the score. Multi parameter analysis improves glaucoma classification based on HRT.
Keywords: imaging/image analysis: clinical • optic disc • nerve fiber layer