Identifying early glaucoma with optical coherence tomography.

PURPOSE To evaluate performance of optical coherence tomography (OCT) for detection of early glaucoma. DESIGN Observational case-control study. METHODS SETTING University-based tertiary care center. STUDY POPULATION One eye from 50 normals, 42 glaucoma suspects, and 59 early glaucoma patients meeting the following criteria: visual field (VF) mean deviation > or =-6.00 dB, age > or =40 years, spherical refractive error < or =5 diopters, astigmatism < or =3 diopters, and visual acuity > or =20/30. OBSERVATION PROCEDURE Early glaucoma by VF (EGVF) was described as repeatable abnormal achromatic VFs based on predefined criteria. Glaucoma suspects (GS) were defined as presence of glaucomatous disk appearance with normal achromatic VFs. OUTCOME MEASURES Average nerve fiber layer thickness (NFLT) and NFLT in each of four quadrants and 12 clock-hour sectors. Receiver operating characteristic curves and sensitivity and specificity were used to assess the performance of OCT. RESULTS Average NFLT was 128.4 +/- 15.4, 102.0 +/- 25.4, and 86.5 +/- 31.5 microm in normal, GS, and EGVF eyes, respectively. Normal eyes were different from both glaucoma groups (P <.001); NFLT in the superior quadrant and at the 11 o'clock position had the highest area under the receiver operating characteristic curve (.840 and.933) in the GS and EGVF groups (P =.03). The sensitivity of the OCT for detection of glaucoma was 71% and 85% for the GS and EGVF groups with specificity fixed at 90%. CONCLUSION The OCT discriminates well between eyes with early perimetric glaucoma and normal eyes. However, its performance is less adequate in eyes with suspicious disk and normal VFs.