Structure-Function Relationships in Glaucoma

Purpose: Standard automated perimetry (SAP) is one of the most widely used functional tests in glaucoma management, and as such, it serves as the benchmark for comparison to all other forms of glaucomatous testing. Optical coherence tomography (OCT) is also a commonly used technology in glaucoma management and allows for a quantifiable structural measurement. One of the more relevant OCT-based measurement for glaucoma is the peripapillary retinal nerve fiber layer (RNFL) thickness; however, investigations comparing OCT RNFL thickness measurements to the actual histology measurements of the RNFL have been rather limited, especially pertaining to measurements in the presence of glaucomatous disease. Flicker-defined form perimetry is a new addition to perimetric testing, and its use in glaucoma testing remains largely unexplored. The purpose of this dissertation is to investigate each of these structural and functional measurements of glaucoma as they relate to each other (SAP and OCT, OCT and histology, SAP and FDF) in order to provide a framework for understanding the optimal stage of glaucoma at which each test is valuable. Methods: In experiment 1, SAP and OCT measurements were obtained from healthy and glaucoma participants. Retinal ganglion cell estimates were calculated from the two different regions and matched to corresponding regions to determine agreement. In experiment 2, health and experimental glaucoma eyes from macaque monkeys were measured with OCT RNFL thickness protocols. The retinal tissue was then processed and imaged with confocal microscopy so that histological RNFL thickness measurements could be obtained. Comparison between similar regions were made and examined for agreement in normal and glaucomatous retina. In experiment 3, SAP and FDF measurements were obtained from healthy and individuals with early and moderate glaucomatous visual field loss. The distributions of thresholds were determined and test-retest characteristics of each form of perimetry were compared. Results: The first experiment showed with age-modification alone, RGC estimates derived from SAP and RNFL thickness values were in agreement in subjects with no visual field loss and in subjects with early to moderate glaucomatous visual field loss. Subjects with end-stage glaucomatous field loss demonstrated a systematic discordance between the perimetric and thickness-derived RGC estimates that showed a high correlation to stage of disease. Application of a stage-of-disease factor in the model produced agreement between OCT and SAP-based RGC estimates across the entire spectrum of glaucoma subjects. RNFL thickness reductions were observed in all eyes with experimental glaucoma. The limits of agreement between OCT- and histology-derived thickness measurements were quite large (95% limits of agreement: untreated = -51.17 µm to 35.15 µm and treated eyes -54.50 µm to 43.43 µm). Percent differences between measurements were much larger for the treated group than the untreated group (untreated = 11.18% ± 33.92% and treated = 40.95% ± 83.97%). Percent differences showed the highest differences when RGC counts per confocal segment (1 confocal segment = 225 µm) fell below 20 RGC/segment. The third experiment revealed different zones of visual sensitivity in glaucomatous disease between SAP and FDF perimetry. The distribution of threshold scores was less for FDF (18.53±4.32 dB) when compared to SAP (30.39±2.55), and showed considerably lower thresholds with FDF scores in glaucoma subjects (14.09±6.04 dB) when compared to the SAP scores (28.70±4.78 dB) for this group. Comparison of test-retest threshold scores reveals better repeatability in FDF perimetry than SAP for scores that fall within the middle (95% limits of agreement for SAP = -10.7 to 9.7; FDF = -6.7 to 6.5) and lower (SAP = -10.3 to 16.8 ; FDF = -8.7 to 7.9) threshold ranges for each instrument. Conclusions: Significant correlations between individuals with no glaucomatous disease and early to moderate glaucomatous disease suggest that the model is applicable to estimating ganglion cell populations from both SAP visual sensitivity values and OCT derived ganglion cell populations. Discordance between the estimates in the later stage of glaucomatous disease suggests that an additional change in the composition of the RNFL or a limitation in the measurement of the RNFL thickness may falsely inflate RGC estimates in this model. Agreement between OCT measurements of the peripapillary RNFL and histological measurements showed substantial variation. The greatest differences in measurements occurred in non-temporal segments when RGC counts fell below a critical value in eyes with glaucoma. Flicker-defined form and standard automated perimetry differ in the range of visual sensitivities that each instrument is able to measure, with FDF perimetry showing overall less sensitivity to both normal and glaucoma participants. Clinically, test-retest repeatability is improved for FDF when compared to SAP with threshold sensitivity observed in the lower- to mid- loss ranges, however this may be a product of the differences in the stimulus properties, testing strategy, dynamic range and of the each instrument, differences in the measurement scale, or a combination of one or more of these factors.Optometry, College o