Faculty Bibliography

PURPOSE/OBJECTIVE:To determine which spectral domain optical coherence tomography biomarkers of idiopathic macular hole (MH) correlate with the postoperative best-corrected visual acuity (BCVA) in anatomically closed MH. METHODS:Retrospective analysis of spectral domain optical coherence tomography scans of 44 patients presenting with MH followed for a mean of 17 months. Widths of MH aperture, base, and ellipsoid zone disruption were calculated from presenting foveal spectral domain optical coherence tomography B-scans. Macular hole base area and ellipsoid zone disruption area were calculated through the custom in-house software. RESULTS:Poorer postoperative BCVA correlated with increased preoperative choroidal hypertransmission (r = 0.503, P = 0.0005), minimum diameter (r = 0.491, P = 0.0007), and base diameter (r = 0.319, P = 0.0348), but not with preoperative ellipsoid zone width (r = 0.199, P = 0.2001). Applying en-face analysis, the BCVA correlated weakly with preoperative ellipsoid zone loss area (r = 0.380, P = 0.013), but not with preoperative MH base area (r = 0.253, P = 0.1058). CONCLUSION/CONCLUSIONS:Increased MH minimum diameter, base diameter, base area, and choroidal hypertransmission are correlated with a poorer postoperative BCVA. Ellipsoid zone loss measurements were not consistently correlated with a BCVA. Choroidal hypertransmission width may be an easy-to-visualize predictive imaging biomarker in MH surgery.

AIMS/OBJECTIVE:To evaluate the associations between retinal blood flow (RBF) and optical coherence tomography (OCT) structural measurements in normal-tension glaucoma (NTG) eyes with single-hemifield visual field (VF) damage by the Doppler OCT. METHODS:The Doppler OCT was used to measure temporal artery (TA) RBF and temporal vein (TV) RBF. Retinal nerve fibre layer thickness (RNFLT) was measured by spectral-domain OCT. RESULTS:Forty-three consecutive eyes of 43 patients with NTG with VF defect confined to a single hemifield and 24 eyes of 24 age-matched healthy subjects were studied. TA and TV RBF and RNFLT were reduced in the damaged hemisphere compared with the normal hemisphere (mean (SD), 3.61 (1.68) vs 5.86 (2.59) µL/min, p<0.001; 5.61 (2.51) vs 6.94 (2.83) µL/min, p=0.010; 69.0 (19.7) vs 99.7 (22.8) µm, p<0.001). Those values in the normal hemisphere of NTG eyes also decreased compared with the healthy hemisphere of the healthy eyes (8.40 (3.36) µL/min, p<0.001; 9.28 (4.47) µL/min, p<0.002; 122.8 (20.2) µm, p<0.001). Multivariate model showed that normal and damaged hemispheres and RNFLT were associated with RBF reduction. In addition, the RBF in the normal hemisphere was lower than that in the healthy hemisphere even after adjusting for RNFLT. CONCLUSION/CONCLUSIONS:In NTG eyes with single-hemifield damage, the RBF was significantly reduced in the damaged hemisphere compared with the normal one. The RBF decreased in the normal and damaged hemispheres of NTG eyes compared with the healthy hemisphere independent from RNFLT.

PURPOSE/OBJECTIVE:To evaluate the accuracy at which visual field global indices could be estimated from OCT scans of the retina using deep neural networks and to quantify the contributions to the estimates by the macula (MAC) and the optic nerve head (ONH). DESIGN/METHODS:Observational cohort study. PARTICIPANTS/METHODS:A total of 10 370 eyes from 109 healthy patients, 697 glaucoma suspects, and 872 patients with glaucoma over multiple visits (median = 3). METHODS:Three-dimensional convolutional neural networks were trained to estimate global visual field indices derived from automated Humphrey perimetry (SITA 24-2) tests (Zeiss, Dublin, CA), using OCT scans centered on MAC, ONH, or both (MAC + ONH) as inputs. MAIN OUTCOME MEASURES/METHODS:Spearman's rank correlation coefficients, Pearson's correlation coefficient, and absolute errors calculated for 2 indices: visual field index (VFI) and mean deviation (MD). RESULTS:The MAC + ONH achieved 0.76 Spearman's correlation coefficient and 0.87 Pearson's correlation for VFI and MD. Median absolute error was 2.7 for VFI and 1.57 decibels (dB) for MD. Separate MAC or ONH estimates were significantly less correlated and less accurate. Accuracy was dependent on the OCT signal strength and the stage of glaucoma severity. CONCLUSIONS:The accuracy of global visual field indices estimate is improved by integrating information from MAC and ONH in advanced glaucoma, suggesting that structural changes of the 2 regions have different time courses in the disease severity spectrum.

The direct analysis of 3D Optical Coherence Tomography (OCT) volumes enables deep learning models (DL) to learn spatial structural information and discover new bio-markers that are relevant to glaucoma. Downsampling 3D input volumes is the state-of-art solution to accommodate for the limited number of training volumes as well as the available computing resources. However, this limits the network's ability to learn from small retinal structures in OCT volumes. In this paper, our goal is to improve the performance by providing guidance to DL model during training in order to learn from finer ocular structures in 3D OCT volumes. Therefore, we propose an end-to-end attention guided 3D DL model for glaucoma detection and estimating visual function from retinal structures. The model consists of three pathways with the same network architecture but different inputs. One input is the original 3D-OCT cube and the other two are computed during training guided by the 3D gradient class activation heatmaps. Each pathway outputs the class-label and the whole model is trained concurrently to minimize the sum of losses from three pathways. The final output is obtained by fusing the predictions of the three pathways. Also, to explore the robustness and generalizability of the proposed model, we apply the model on a classification task for glaucoma detection as well as a regression task to estimate visual field index (VFI) (a value between 0 and 100). A 5-fold cross-validation with a total of 3782 and 10,370 OCT scans is used to train and evaluate the classification and regression models, respectively. The glaucoma detection model achieved an area under the curve (AUC) of 93.8% compared with 86.8% for a baseline model without the attention-guided component. The model also outperformed six different feature based machine learning approaches that use scanner computed measurements for training. Further, we also assessed the contribution of different retinal layers that are relevant to glaucoma. The VFI estimation model achieved a Pearson correlation and median absolute error of 0.75 and 3.6%, respectively, for a test set of size 3100 cubes.

Purpose: Rigorous clinical testing has established that Schlemm"™s canal cross-sec-tional area (SC-CSA) is reduced in glaucomatous eyes. However, to date, it is unclear whether trabecular bypass procedures impact the morphology of the proximal aqueous outflow tract, or if the introduction of a local region of low outflow resistance adversely affects SC-CSA elsewhere, specifically presenting as SC diminution. This study quantifies changes in the morphology of the distal outflow pathway after iStent Trabecular Micro-Bypass stent (Glaukos Corp, Laguna Hills, CA, USA) implantation in living eyes by anterior segment optical coherence tomography (OCT). Design: This was a prospective observational study. Subjects: This study included six patients (eight eyes) with primary-open angle glaucoma. Methods: Patients underwent iStent placement in the nasal anterior chamber angle quadrant. OCT imaging was obtained of both nasal and temporal eye quadrants before and after surgery. For each SC parameter, an average of ten consecutive, evenly spaced measurements were manually obtained over a 1 mm segment of SC on FIJI ImageJ. Linear mixed effects modeling quantified the effect of the iStent on these parameters. Main outcome measures: Main outcome measures were changes in SC-CSA, inner-to-outer wall distance (IOD), and trabecular meshwork (TM) thickness following iStent placement. Results: Following iStent placement, total SC-CSA increased an average of 1,039.12 µm² (P = 0.05). Individually, there were no significant changes in SC-CSA in the nasal or temporal quadrants. Total SC-IOD and nasal SC-IOD increased an average of 2.35 µm (P = 0.01) and 2.96 µm (P = 0.04), respectively. There were no significant changes in temporal quadrant SC-IOD. There were no significant changes in TM thickness in either quadrant. Conclusions: Implantation of the iStent Trabecular Micro-Bypass stent significantly increases SC-IOD in the nasal quadrant at the location of implant, with no evidence of SC diminution in the temporal quadrant. It remains unclear how these observa-tions relate to the surgical efficacy of trabecular bypass procedures.

Purpose/UNASSIGNED:The purpose of this study was to develop a machine learning model to forecast future circumpapillary retinal nerve fiber layer (cpRNFL) thickness in eyes of healthy, glaucoma suspect, and glaucoma participants from multimodal temporal data. Design/UNASSIGNED:Retrospective analysis of a longitudinal clinical cohort. Participants/UNASSIGNED:Longitudinal clinical cohort of healthy, glaucoma suspect, and glaucoma participants. Methods/UNASSIGNED:(LTBE). Main Outcome Measures/UNASSIGNED:The mean absolute difference and Pearson's correlation coefficient between the true and forecasted values of the cpRNFL in the healthy, glaucoma suspect, and glaucoma patients. Results/UNASSIGNED:< 0.01) for the 3 groups, respectively. Conclusions/UNASSIGNED:The performance of the proposed forecasting model for cpRNFL is consistent across glaucoma suspect and glaucoma patients, which implies the robustness of the developed model against the disease state. These forecasted values may be useful to personalize patient care by determining the most appropriate intervisit schedule for timely interventions.

Purpose : To compare the discrimination accuracy for glaucoma diagnosis using the OCT RNFL clock hours compared with average RNFL. Methods : In a large, ongoing, longitudinal cohort of healthy subjects and subjects with glaucoma, all subjects underwent visual field (VF) and OCT testing. Principal component (PC) analysis was used to reduce the dimensionality of clock hour measurements while maintaining maximum information variability for diagnostic performance. The first four PCs with linear regression were used as predictors of VF mean deviation (MD) and to classify glaucoma diagnosis. The prediction accuracy and discrimination power using cross validation were compared to the models using only average RNFL as a predictor. All models were adjusted for age, signal strength, and intra-subject correlation. Results : 1317 healthy and glaucomatous eyes (717 subjects) were included in the study. A PC analysis was built on the 9 clock hours while excluding non-informative sectors (clock hours 3, 4, and 9). The first PC explained 51% of the total variance, and the first four PCs explained 82% of the total variance and thus were used for subsequent regression models. A PC regression for glaucoma discrimination showed that clock hours 1, 5, 6, 7, 10, 11, 12 were significantly association with diagnosis. The PC showed better glaucoma diagnosis performance compared to average RNFL, with 10-fold cross-validation AUCs of 0.898 and 0.877, respectively (p<0.001). The PC regression for MD improved the model fit measured by R² by 9% compared to a regression using average RNFL. PC showed that clock hours 2, 5, 6, 7, 10, 11, 12 were significantly associated with MD. Conclusions : Using PCs with RNFL clock hours improved classification performance for glaucoma diagnosis and model fit for MD, compared to using average RNFL. This method improves discrimination performance by both considering all sectoral RNFL information and removing locations with low diagnostic yield