Abstract
Purpose: :
Conventional Frequency Doubling Technology (FDT) perimeters have a large stimulus size, and localized visual field abnormalities in early-stage glaucoma may be missed with the use of such perimeters. It has been suggested that the new perimeter, the Humphrey Matrix (HM), may be useful for the detection of glaucoma in the earlier stage, on account of the reduced stimulus size and increased number of measurement points. We examined the efficacy of HM in outpatients as a screening tool for glaucoma.
Methods: :
The subjects were 80 eyes of 47 glaucoma or preperimetric glaucoma patients examined at the Tokyo Rosai Hospital between December 2002 and April 2004. All patients provided consent for this study, and assessment was conducted using the reliable Humphrey perimeter 30-2 (HFA), the HM24-2-1 screening program (HM24), and the N-30-1 screening program (N30) with the conventional stimulus size. The mean age of the patients was 61.4 years (range, 34-84 years). Diagnosis of glaucoma was based on Anderson criteria in HFA. The cutoff values for evaluation of abnormal findings in HM24 were analyzed from the receiver operating characteristic (ROC) curve, and comparatively assessed with those for N30.
Results: :
The area under the ROC curve for the detection of glaucoma, which corresponded to the number of abnormal points in HM24, was 0.876 [95% confidence interval (CI): 0.793-0.959]. When cases with 2 cutoff values calculated from the ROC curve, i.e., cases with 2 abnormal points, were labeled as cases of glaucoma, the sensitivity and specificity for the diagnosis were 91.2% and 84.8%, respectively. When glaucoma was detected during the measurement time in HM24, the area under ROC was 0.880 (95% CI: 0.805-0.956). When the measurement time of 107 seconds or more was evaluated as being positive, the sensitivity and specificity for the diagnosis were 88.2% and 80.4%, respectively. The corresponding sensitivity and specificity for N30 were 76.5% and 91.3%, respectively. When the evaluation was conducted according to the measurement time in N30, the sensitivity and specificity for the diagnosis were 85.3% and 87.0%, respectively.
Conclusions: :
HM24 was efficient for the screening of glaucoma. It was considered that the detection of 2 or more abnormal points and a measurement time of 107 seconds or more were valid diagnostic criteria for glaucoma.
Keywords: perimetry • visual fields • clinical (human) or epidemiologic studies: systems/equipment/techniques