Optical coherence tomography (OCT) is a commonly used imaging modality in the evaluation of glaucomatous damage. The commercially available spectral domain (SD)-OCT offers benefits in glaucoma assessment over the earlier generation of time domain-OCT due to increased axial resolution, faster scanning speeds and has been reported to have improved reproducibility but similar diagnostic accuracy. The capabilities of SD-OCT are rapidly advancing with 3D imaging, reproducible registration, and advanced segmentation algorithms of macular and optic nerve head regions. A review of the evidence to date suggests that retinal nerve fibre layer remains the dominant parameter for glaucoma diagnosis and detection of progression while initial studies of macular and optic nerve head parameters have shown promising results. SD-OCT still currently lacks the diagnostic performance for glaucoma screening.

Purpose. Machine-learning classifiers are trained computerized systems with the ability to detect the relationship between multiple input parameters and a diagnosis. The present study investigated whether the use of machine-learning classifiers improves optical coherence tomography (OCT) glaucoma detection. methods. Forty-seven patients with glaucoma (47 eyes) and 42 healthy subjects (42 eyes) were included in this cross-sectional study. Of the glaucoma patients, 27 had early disease (visual field mean deviation [MD] ≥ −6 dB) and 20 had advanced glaucoma (MD < −6 dB). Machine-learning classifiers were trained to discriminate between glaucomatous and healthy eyes using parameters derived from OCT output. The classifiers were trained with all 38 parameters as well as with only 8 parameters that correlated best with the visual field MD. Five classifiers were tested: linear discriminant analysis, support vector machine, recursive partitioning and regression tree, generalized linear model, and generalized additive model. For the last two classifiers, a backward feature selection was used to find the minimal number of parameters that resulted in the best and most simple prediction. The cross-validated receiver operating characteristic (ROC) curve and accuracies were calculated. results. The largest area under the ROC curve (AROC) for glaucoma detection was achieved with the support vector machine using eight parameters (0.981). The sensitivity at 80% and 95% specificity was 97.9% and 92.5%, respectively. This classifier also performed best when judged by cross-validated accuracy (0.966). The best classification between early glaucoma and advanced glaucoma was obtained with the generalized additive model using only three parameters (AROC = 0.854). conclusions. Automated machine classifiers of OCT data might be useful for enhancing the utility of this technology for detecting glaucomatous abnormality.

Background/aims: To investigate retinal nerve fibre layer (RNFL) thickness measurement reproducibility using conventional time-domain optical coherence tomography (TD-OCT) and spectral-domain OCT (SD-OCT), and to evaluate two methods defining the optic nerve head (ONH) centring: Centred Each Time (CET) vs Centred Once (CO), in terms of RNFL thickness measurement variability on SD-OCT. Methods: Twenty-seven eyes (14 healthy subjects) had three circumpapillary scans with TD-OCT and three raster scans (three-dimensional or 3D image data) around ONH with SD-OCT. SD-OCT images were analysed in two ways: (1) CET: ONH centre was defined on each image separately and (2) CO: ONH centre was defined on one image and exported to other images after scan registration. After defining the ONH centre, a 3.4 mm diameter virtual circular OCT was resampled on SD-OCT images to mimic the conventional circumpapillary RNFL thickness measurements taken with TD-OCT. Results: CET and CO showed statistically significantly better reproducibility than TD-OCT except for 11:00 with CET. CET and CO methods showed similar reproducibility. Conclusions: SD-OCT 3D cube data generally showed better RNFL measurement reproducibility than TD-OCT. The choice of ONH centring methods did not affect RNFL measurement reproducibility.

Aims To evaluate the glaucoma discriminating ability of macular retinal layers as measured by spectral domain optical coherence tomography (SD-OCT). Methods Healthy, glaucoma suspect and glaucomatous subjects had a comprehensive ocular examination, visual field testing and SD-OCT imaging (Cirrus HD-OCT; Carl Zeiss Meditec, Dublin, California, USA) in the macular and optic nerve head regions. OCT macular scans were segmented into macular nerve fibre layer (mNFL), ganglion cell layer with inner plexiform layer (GCIP), ganglion cell complex (GCC) (composed of mNFL and GCIP), outer retinal complex and total retina. Glaucoma discriminating ability was assessed using the area under the receiver operator characteristic curve (AUC) for all macular parameters and mean circumpapillary retinal nerve fibre layer (cpRNFL). Results Analysis was performed on 51 healthy, 49 glaucoma suspect and 63 glaucomatous eyes. The median visual field MD was −2.21 dB (IQR: −6.92 to −0.35) for the glaucoma group, −0.32 dB (IQR: −1.22 to 0.73) for the suspect group and −0.18 dB (IQR: −0.92 to 0.71) for the healthy group. Highest age adjusted AUCs were found for average GCC and GCIP (AUC=0.901 and 0.900, respectively) and their sectoral measurements: infero-temporal (0.922 and 0.913), inferior (0.904 and 0.912) and supero-temporal (0.910 and 0.897). These values were similar to the discriminating ability of the mean cpRNFL (AUC=0.913). Comparison of these AUCs did not yield any statistically significant difference (all p>0.05). Conclusions SD-OCT GCIP and GCC measurements showed similar glaucoma diagnostic ability and were comparable with that of cpRNFL.

Purpose. To investigate the effect on optical coherence tomography (OCT) retinal nerve fiber layer (RNFL) thickness measurements of varying the standard 3.4-mm-diameter circle location. methods. The optic nerve head (ONH) region of 17 eyes of 17 healthy subjects was imaged with high-speed, ultrahigh-resolution OCT (hsUHR-OCT; 501 × 180 axial scans covering a 6 × 6-mm area; scan time, 3.84 seconds) for a comprehensive sampling. This method allows for systematic simulation of the variable circle placement effect. RNFL thickness was measured on this three-dimensional dataset by using a custom-designed software program. RNFL thickness was resampled along a 3.4-mm-diameter circle centered on the ONH, then along 3.4-mm circles shifted horizontally (x-shift), vertically (y-shift) and diagonally up to ±500 μm (at 100-μm intervals). Linear mixed-effects models were used to determine RNFL thickness as a function of the scan circle shift. A model for the distance between the two thickest measurements along the RNFL thickness circular profile (peak distance) was also calculated. results. RNFL thickness tended to decrease with both positive and negative x- and y-shifts. The range of shifts that caused a decrease greater than the variability inherent to the commercial device was greater in both nasal and temporal quadrants than in the superior and inferior ones. The model for peak distance demonstrated that as the scan moves nasally, the RNFL peak distance increases, and as the circle moves temporally, the distance decreases. Vertical shifts had a minimal effect on peak distance. conclusions. The location of the OCT scan circle affects RNFL thickness measurements. Accurate registration of OCT scans is essential for measurement reproducibility and longitudinal examination (ClinicalTrials.gov number, NCT00286637).

Purpose.: The lamina cribrosa (LC) is a prime location of glaucomatous damage. The purpose of this study was to compare LC 3-dimensional micro-architecture between healthy and glaucomatous eyes in vivo by using optical coherence tomography (OCT). Methods.: Sixty-eight eyes (19 healthy and 49 glaucomatous) from 47 subjects were scanned in a 3.5 × 3.5 × 3.64-mm volume (400 × 400 × 896 pixels) at the optic nerve head by using swept-source OCT. The LC micro-architecture parameters were measured on the visible LC by an automated segmentation algorithm. The LC parameters were compared to diagnosis and visual field mean deviation (VF MD) by using a linear mixed effects model accounting for age. Results.: The average VF MD for the healthy and glaucomatous eyes was −0.50 ± 0.80 dB and −7.84 ± 8.75 dB, respectively. Beam thickness to pore diameter ratio (P = 0.04) and pore diameter standard deviation (P < 0.01) were increased in glaucomatous eyes. With worse MD, beam thickness to pore diameter ratio (P < 0.01), pore diameter standard deviation (P = 0.05), and beam thickness (P < 0.01) showed a statistically significant increase while pore diameter (P = 0.02) showed a significant decrease. There were no significant interactions between any of the parameters and age (all P > 0.05). Conclusions.: Glaucomatous micro-architecture changes in the LC, detected by OCT analysis, reflect beams remodeling and axonal loss leading to reduction in pore size and increased pore size variability.