论文部分内容阅读
Purpose: To evaluate the ability of scanning laser polarimetry parameters and a novel deviation map algorithm to discriminate between healthy and early glaucomatous eyes with localized visual field (VF) defects confined to one hemi-field. Design: Prospective case-control study. Participants: Seventy glaucomatous ey es with localized VF defects and 66 normal controls. Methods: A Humphrey field analyzer 24-2 full-threshold test and scanning laser polarimetry with variable corneal compensation were used. Main Outcome Measures: We assessed the sensitivity and specificity of scanning laser polarimetry parameters, sensitivity and cutoff values for scanning laser polarimetry deviation map algorithms at different specificity values (80%, 90%, and 95%) in the detection of glaucoma, and correlations between the algorithms of scanning laser polarimetry and of the pattern deviation derived from Humphrey field analyzer testing. Results: There were significant differences between the glaucoma group and normal subjects in the mean parametric values of the temporal, superior, nasal, inferior, temporal (TSNIT) average, superior average, inferior average, and TSNIT standard deviation (SD) (P < 0.05). The sensitivity and specificity of each scanning laser polarimetry variable was as follows: TSNIT, 44.3%(95%confidence interval CI-, 39.8%-49.8 %) and 100%(95.4%-100%); superior average, 30%(25.5%-34.5%) and 97%(93.5%-100%); inferior average, 45.7%(42.2%-49.2%) and 100%(95.8%-100%); and TSNIT SD, 30%(25.9%-34.1%) and 97%(93.2%-100%), respectively (when abnormal was defined as P < 0.05). Based on nerve fiber indicator cutoff values of < 30 and < 51 to indicate glaucoma, sensitivities were 54.3%(50.1%-58.5%) and 10%(6.4%-13.6%), and specificities were 97%(93.2%-100%)-and 100%( 95.8%-100%), respectively. The range of areas under the receiver operating ch aracteristic curves using the scanning laser polarimetry deviation map algorithm was 0.790 to 0.879. Overall sensitivities combining each probability scale and severity score at 80%, 90%, and 95%specificities were 90.0%(95%CI, 86.4%- 93.6%), 71.4%(67.4%-75.4%), and 60.0%(56.2%-63.8%), respectively. There was a statistically significant correlation between the scanning laser polarime try severity score and the VF severity score (R2=0.360, P < 0.001). Conclusions: Scanning laser polarimetry parameters may not be sufficiently sensitive to dete ct glaucomatous patients with localized VF damage. Our algorithm using the scann ing laser polarimetry deviation map may enhance the understanding of scanning la ser polarimetry printouts in terms of the locality, deviation size, and severity of localized retinal nerve fiber layer defects in eyes with localized VF loss.
Purpose: To evaluate the ability of scanning laser polarimetry parameters and a novel deviation map algorithm to discriminate between healthy and early glaucomatous eyes with localized visual field (VF) defects confined to one hemi-field. Design: Prospective case-control study. Methods: A Humphrey field analyzer 24-2 full-threshold test and scanning laser polarimetry with variable corneal compensation were used. Main Outcome Measures: We assessed the sensitivity and specificity of scanning laser polarimetry parameters, sensitivity and cutoff values for scanning laser polarimetry deviation map algorithms at different specificity values (80%, 90%, and 95%) in the detection of glaucoma, and correlations between the algorithms of scanning laser polarimetry and of the pattern deviation derived from Humphrey field analyzer testing. Results: There were significant differences between the glaucoma group and normal subjects in the mean parametric values of the temporal, superior, nasal, inferior, temporal (TSNIT) average, superior average, inferior average, and TSNIT standard deviation (SD) 44.5% (95% confidence interval CI-, 39.8% -49.8%) and 100% (95.4% -100%); superior average, 30% (25.5% -34.5%) and 97 (93.5% -100%); inferior average 45.7% (42.2% -49.2%) and 100% (95.8% -100%); and TSNIT SD 30% (25.9% -34.1%) and 97% Based on nerve fiber indicator cutoff values of <30 and <51 to indicate glaucoma, sensitivities were 54.3% (50.1% -58.5%) and 10% (% -100%), respectively 6.4% -13.6%), and specificities were 97% (93.2% -100%) - and 100% (95.8% -100%), respectively. The range of areas under the receiver operating chracteristic curves using the scanning laser polarimetry deviation map algorithm was 0.790 to 0.879. Overall sensitivities combining each(95% CI, 86.4% -93.6%), 71.4% (67.4% -75.4%), and 60.0% (56.2% -63.8%) at 80%, 90%, and 95% There was a significant correlation between the scanning laser polarime try severity score and the VF severity score (R2 = 0.360, P <0.001). Conclusions: Scanning laser polarimetry parameters may not be sufficient sensitive to dete ct glaucomatous patients with localized VF damage. Our algorithm using the scann ing laser polarimetry deviation map may enhance the understanding of scanning la ser polarimetry printouts in terms of the locality, deviation size, and severity of localized retinal nerve fiber layer defects in eyes with localized VF loss.