THIRD PRIZE, COS AWARDS FOR EXCELLENCE IN OPHTHALMIC RESEARCH

Authors: Jennifer L. Gao, Jack Quach, Marcelo T. Nicolela, Lesya M. Shuba, Balwantray C. Chauhan, Jayme R. Vianna
Author Disclosure Block: J.L. Gao: None. J. Quach: None. M.T. Nicolela: Consultant and Lecturer: Allergan, Alcon, Thea, Bausch & Lomb. L.M. Shuba: Consultant and Lecturer: Allergan, Alcon. B.C. Chauhan: Consultant and Lecturer, Equipment support: Allergan, Heidelberg Engineering, Topcon. J.R. Vianna: None.

Abstract Body:

Purpose: Technological advances in spectral-domain optical coherence tomography (OCT) have provided clinicians with multiple diagnostic parameters for detecting glaucoma, including the macular ganglion cell layer thickness (GCLT), peripapillary retinal nerve fibre layer thickness (RNFLT), and Bruch membrane opening-minimum rim width (MRW). It remains equivocal, however, whether a particular test is diagnostically more sensitive and whether some glaucomatous eyes are uniquely identified by any one test. We conducted a retrospective study to determine and compare the diagnostic sensitivity of the GCLT, RNFLT, and MRW in detecting glaucomatous damage. This study also aims to determine the proportion of glaucomatous eyes that are uniquely identified by each type of OCT parameter.
Study Design: Retrospective observational study
Methods: This study included 639 eyes of 639 patients with glaucoma. The reference standard for glaucomatous damage was a 24-2 standard automated perimetry test on the Humphrey Field Analyzer that was reliable (less than 15% false positives) and abnormal, defined as having a glaucoma hemifield test outside normal limits and either an abnormal mean deviation (MD) or pattern standard deviation at P < 5%. The visual field must have been performed within 24 months prior to or 6 months after the OCT exam. For patients with bilateral glaucoma, one eye was randomly selected for the study. Patients with macular disease, non-glaucomatous optic neuropathies, and cerebrovascular accidents were excluded. All patients underwent OCT imaging of the GCLT, RNFLT, and MRW by the Spectralis OCT (Heidelberg Engineering, Germany) at our centre from April 2017 to September 2018. Each parameter was computed globally and in 6 sectors. The GCLT, RNFLT, and MRW tests were considered abnormal if the global or sectoral values were abnormal (P < 0.01) compared to normative database values, adjusted for age and BMO area. The primary outcome measure was the sensitivity of each OCT parameter in diagnosing glaucoma and comparisons between the tests were done using the McNemar test. A Venn diagram was created to show the ability of each OCT parameter to uniquely identify eyes with glaucoma.
Results: The median (interquartile range) age of study patients was 70.6 (62.9-78.2) years and visual field mean deviation was -4.81 (-8.79 to -2.15) DB. The OCT parameter with the highest diagnostic sensitivity for glaucoma was RNFLT at 77% (95% confidence interval, 74-80%) sensitivity, followed by MRW at 63% (59-66%) and GCLT at 61% (57-65%). The RNFLT was statistically significant in having a greater sensitivity compared to the other parameters (P<0.001), while the sensitivity of the MRW and GCLT were similar (P = 0.41). Of all study eyes, 297 (46%) were identified to have glaucoma by all three OCT parameters. A number of glaucoma cases were uniquely detected by each test (and undetected by the other tests): 26 (4%) by GCLT, 56 (8%) by RNFLT, and 15 (2%) by MRW.
Conclusions: From the OCT parameters, RNFLT had higher diagnostic sensitivity than MRW and GCLT in identifying glaucomatous eyes. However, some glaucoma patients were uniquely identified by the MRW or GCLT, suggesting that all parameters have some level of diagnostic value.