Purchase this article with an account.
F. Di Matteo, M. Guareschi, I. Riva, F. Romanazzi, M. Vavassori, G. Barteselli, S. Miglior; Clinical Ability of HRT III and Stratus OCT Optic Nerve Head Algorithms in Detecting Glaucomatous Visual Field Defects. Invest. Ophthalmol. Vis. Sci. 2007;48(13):526. doi: https://doi.org/.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To evaluate the clinical ability of Stratus OCT Optic Nerve Head (ONH) Algorithms and HRT III Moorfields Regression Analysis (MRA) and Glaucoma Probability Score (GPS) in detecting glaucomatous visual field defects.
One hundred fifty nine subjects (55 controls, 58 with OHT, 46 with POAG) were examined. Healthy individuals had a normal 24/II HFA visual field (VF) and IOP <21 mmHg. OHT patients had IOP >21 mmHg on repeated measurements and normal VF. POAG patients had abnormal VF (GHT "outside normal limits" and PSD - p<0.05) with IOP >21 mmHg. HRT III and StratusOCT examinations were performed 3 times on the same day and the best quality image was considered for the analysis. Optic disc margins were determined manually on both instruments. One randomly chosen eye per patient was used for study purposes. HRT III clinical ability was based on MRA and GPS reports, whereas clinical ability of Stratus OCT was based on the evaluation of Vert. Int. Rim Area (VIRA), Horiz. Int. Rim Width (HIRW), Disk, Cup and Rim Area, C/D Area Ratio, Horiz. and Vert. C/D Ratio and Cup Volume. Three analyses were performed: 1.Normal group was compared with POAG group and OHT patients were excluded; 2.OHT patients were included in the Normal group; 3.OHT patients were included in the POAG group. Sensitivity, specificity and ROC curves were calculated by means of standard procedures.
Mean MD and PSD (±SD) were 0.02 (±1.13) and 1.44 (±0.32) in the Normal group, 0.19 (±1.15) and 1.59 (±0.23) in the OHT group and -10.24 (±7.82) and 7.91 (±3.91) in the POAG group. In analysis 1, MRA reported a sensitivity of 0.80 with a specificity of 0.93; GPS reported a sensitivity of 0.42 at the fixed specificity of 0.95 with an area under ROC curve value of 0.80. In analyses 2 and 3, MRA specificities were 0.83 and 0.93 with sensitivities of 0.80 and 0.50, whereas GPS sensitivity (at 0.95 specificity) decreased to 0.29 and 0.27. Among all the OCT ONH parameters, VIRA and HIRW showed the best discrimination ability, reporting sensitivities of 0.71 and 0.70 at the fixed specificity (0.95) in analysis 1, 0.51 and 0.68 in analysis 2 and 0.36 and 0.33 in analysis 3; the AROC in analysis 1 were 0.92 and 0.96, 0.89 and 0.95 in analysis 2 and 0.72, 0.76 in analisys 3.
HRT III and Stratus OCT ONH algorithms showed similar specificity and sensitivity in detecting glaucomatous visual field changes. The inclusion of OHT patients in the study groups strongly affected the overall clinical ability of both techniques.
This PDF is available to Subscribers Only