Volumetric Spectral-domain Optical Coherence Tomography Research Articles

A Deep-Learning Algorithm to Predict Short-Term Progression to Geographic Atrophy on Spectral-Domain Optical Coherence Tomography

The identification of patients at risk of progressing from intermediate age-related macular degeneration (iAMD) to geographic atrophy (GA) is essential for clinical trials aimed at preventing disease progression. DeepGAze is a fully automated and accurate convolutional neural network-based deep learning algorithm for predicting progression from iAMD to GA within 1 year from spectral-domain optical coherence tomography (SD-OCT) scans. To develop a deep-learning algorithm based on volumetric SD-OCT scans to predict the progression from iAMD to GA during the year following the scan. This retrospective cohort study included participants with iAMD at baseline and who either progressed or did not progress to GA within the subsequent 13 months. Participants were included from centers in 4 US states. Data set 1 included patients from the Age-Related Eye Disease Study 2 AREDS2 (Ancillary Spectral-Domain Optical Coherence Tomography) A2A study (July 2008 to August 2015). Data sets 2 and 3 included patients with imaging taken in routine clinical care at a tertiary referral center and associated satellites between January 2013 and January 2023. The stored imaging data were retrieved for the purpose of this study from July 1, 2022, to February 1, 2023. Data were analyzed from May 2021 to July 2023. A position-aware convolutional neural network with proactive pseudointervention was trained and cross-validated on Bioptigen SD-OCT volumes (data set 1) and validated on 2 external data sets comprising Heidelberg Spectralis SD-OCT scans (data sets 2 and 3). Prediction of progression to GA within 13 months was evaluated with area under the receiver-operator characteristic curves (AUROC) as well as area under the precision-recall curve (AUPRC), sensitivity, specificity, positive predictive value, negative predictive value, and accuracy. The study included a total of 417 patients: 316 in data set 1 (mean [SD] age, 74 [8]; 185 [59%] female), 53 in data set 2, (mean [SD] age, 83 [8]; 32 [60%] female), and 48 in data set 3 (mean [SD] age, 81 [8]; 32 [67%] female). The AUROC for prediction of progression from iAMD to GA within 1 year was 0.94 (95% CI, 0.92-0.95; AUPRC, 0.90 [95% CI, 0.85-0.95]; sensitivity, 0.88 [95% CI, 0.84-0.92]; specificity, 0.90 [95% CI, 0.87-0.92]) for data set 1. The addition of expert-annotated SD-OCT features to the model resulted in no improvement compared to the fully autonomous model (AUROC, 0.95; 95% CI, 0.92-0.95; P = .19). On an independent validation data set (data set 2), the model predicted progression to GA with an AUROC of 0.94 (95% CI, 0.91-0.96; AUPRC, 0.92 [0.89-0.94]; sensitivity, 0.91 [95% CI, 0.74-0.98]; specificity, 0.80 [95% CI, 0.63-0.91]). At a high-specificity operating point, simulated clinical trial recruitment was enriched for patients progressing to GA within 1 year by 8.3- to 20.7-fold (data sets 2 and 3). The fully automated, position-aware deep-learning algorithm assessed in this study successfully predicted progression from iAMD to GA over a clinically meaningful time frame. The ability to predict imminent GA progression could facilitate clinical trials aimed at preventing the condition and could guide clinical decision-making regarding screening frequency or treatment initiation.

Probabilistic Forecasting of Anti-VEGF Treatment Frequency in Neovascular Age-Related Macular Degeneration.

PurposeTo probabilistically forecast needed anti-vascular endothelial growth factor (anti-VEGF) treatment frequency using volumetric spectral domain–optical coherence tomography (SD-OCT) biomarkers in neovascular age-related macular degeneration from real-world settings.MethodsSD-OCT volume scans were segmented with a custom deep-learning-based analysis pipeline. Retinal thickness and reflectivity values were extracted for the central and the four inner Early Treatment Diabetic Retinopathy Study (ETDRS) subfields for six retinal layers (inner retina, outer nuclear layer, inner segments [IS], outer segments [OS], retinal pigment epithelium-drusen complex [RPEDC] and the choroid). Machine-learning models were probed to predict the anti-VEGF treatment frequency within the next 12 months. Probabilistic forecasting was performed using natural gradient boosting (NGBoost), which outputs a full probability distribution. The mean absolute error (MAE) between the predicted versus actual anti-VEGF treatment frequency was the primary outcome measure.ResultsIn a total of 138 visits of 99 eyes with neovascular AMD (96 patients) from two clinical centers, the prediction of future anti-VEGF treatment frequency was observed with an accuracy (MAE [95% confidence interval]) of 2.60 injections/year [2.25–2.96] (R2 = 0.390) using random forest regression and 2.66 injections/year [2.31–3.01] (R2 = 0.094) using NGBoost, respectively. Prediction intervals were well calibrated and reflected the true uncertainty of NGBoost-based predictions. Standard deviation of RPEDC-thickness in the central ETDRS-subfield constituted an important predictor across models.ConclusionsThe proposed, fully automated pipeline enables probabilistic forecasting of future anti-VEGF treatment frequency in real-world settings.Translational RelevancePrediction of a probability distribution allows the physician to inspect the underlying uncertainty. Predictive uncertainty estimates are essential to highlight cases where human-inspection and/or reversion to a fallback alternative is warranted.

Quantitative Analysis of En Face Spectral-Domain Optical Coherence Tomography Imaging in Polypoidal Choroidal Vasculopathy.

Spectral-domain optical coherence tomography (SD-OCT) imaging can be used to visualize polypoidal choroidal vasculopathy (PCV) lesions in the en face plane. Here, the authors describe a novel lesion quantification technique and compare PCV lesion area measurements and morphology before and after anti-vascular endothelial growth factor (VEGF) treatment. Volumetric SD-OCT scans in eyes with PCV before and after induction anti-VEGF therapy were retrospectively analyzed. En face SD-OCT images were generated and a pixel intensity thresholding process was used to quantify total lesion area. Thirteen eyes with PCV were analyzed. En face SD-OCT PCV lesion area quantification showed good intergrader reliability (intraclass correlation coefficient = 0.944). Total PCV lesion area was significantly reduced after anti-VEGF therapy (2.22 mm2 vs. 2.73 mm2; P = .02). The overall geographic pattern of the branching vascular network was typically preserved. PCV lesion area analysis using en face SD-OCT is a reproducible tool that can quantify treatment related changes. [Ophthalmic Surg Lasers Imaging Retina. 2017;48:126-133.].

Application of Novel Software Algorithms to Spectral-Domain Optical Coherence Tomography for Automated Detection of Diabetic Retinopathy.

To present novel software algorithms applied to spectral-domain optical coherence tomography (SD-OCT) for automated detection of diabetic retinopathy (DR). Thirty-one diabetic patients (44 eyes) and 18 healthy, nondiabetic controls (20 eyes) who underwent volumetric SD-OCT imaging and fundus photography were retrospectively identified. A retina specialist independently graded DR stage. Trained automated software generated a retinal thickness score signifying macular edema and a cluster score signifying microaneurysms and/or hard exudates for each volumetric SD-OCT. Of 44 diabetic eyes, 38 had DR and six eyes did not have DR. Leave-one-out cross-validation using a linear discriminant at missed detection/false alarm ratio of 3.00 computed software sensitivity and specificity of 92% and 69%, respectively, for DR detection when compared to clinical assessment. Novel software algorithms applied to commercially available SD-OCT can successfully detect DR and may have potential as a viable screening tool for DR in future. [Ophthalmic Surg Lasers Imaging Retina. 2016;47:410-417.].

Sequential Chorioretinal Changes in Presumed Ocular Histoplasmosis Syndrome Analyzed Using Spectral Domain Optical Coherence Tomography

ABSTRACTPurpose: To analyze sequential chorioretinal changes in presumed ocular histoplasmosis syndrome (POHS) using volumetric spectral-domain optical coherence tomography (SD-OCT).Methods: A total of 28 treatment-naive eyes of 20 POHS patients were characterized as: no choroidal neovascular membrane (CNVM) at initial SD-OCT but CNVM development during follow-up (primary group); CNVM at initial SD-OCT (progressive group); and contralateral asymptomatic eyes (control group).Results: The primary group (4/28 eyes) initially demonstrated external limiting membrane (ELM), ellipsoid zone and retinal pigment epithelium (RPE)/Bruch’s disruption in 100%, and pigment epithelial detachment (PED) in 75% of eyes. CNVM developed in 100% of eyes during follow-up. Progressive group (8/28 eyes) showed CNVM in 100% of eyes. Control group (16/28 eyes) showed ELM disruption in two and PED in three additional eyes on follow-up from initial analysis.Conclusions: SD-OCT is a useful tool to sequentially follow eyes with POHS for subtle chorioretinal changes that may suggest impending CNVM development, and thus benefit from closer clinical monitoring.

The Spatial Profile of Macular Pigments Is Related to the Topological Characteristics of the Foveal Avascular Zone.

Macular pigments are preferentially concentrated in the central fovea, an area devoid of vasculature. We hypothesized that there may be a link between the macular pigment profile and the size and structural characteristics of the foveal avascular zone (FAZ). Two-wavelength autofluorescence method was used to quantify macular pigment optical density (MPOD) and the radius at half peak of MPOD, which was defined as the retinal eccentricity where the MPOD value was 50% of the peak value. Volumetric spectral-domain optical coherence tomography (OCT) images of the macula were obtained from 32 subjects. The equivalent radius of the FAZ was determined using data generated from OCT angiography. Generalized estimating equations were used to test the hypothesis that there are interrelationships among the central foveal thickness, peak MPOD, the radius at half peak of MPOD and the equivalent radius of the FAZ. The equivalent radius of the FAZ was highly correlated with the radius at half peak of MPOD (P < .001). The equivalent radius of the FAZ was a significant predictor for central foveal thickness (P < .001). The significant predictor for peak MPOD was central foveal thickness (P = .004). Eyes with larger FAZs were more likely to have a secondary peak in their MPOD spatial profile in a zone ranging from 0.5 to 1.0 degrees from the foveal center. The spatial distribution of macular pigment is related to the size of the FAZ, in addition to the central foveal thickness. It is possible that xanthophyll pigment accumulation in the macula serves functions, such as attenuation of shorter wavelengths of light, that would have been provided by the light-filtering characteristics of blood vessels.

In vivo three-dimensional characterization of the adult zebrafish brain using a 1325 nm spectral-domain optical coherence tomography system with the 27 frame/s video rate.

In this study, a spectral-domain optical coherence tomography (SD-OCT) system was used for noninvasive imaging of the adult zebrafish brain. Based on a 1325 nm light source and two high-speed galvo mirrors, our SD-OCT system can offer a large field of view of brain morphology with high resolution (12 μm axial and 13 μm lateral) at video rate (27 frame/s). In vivo imaging of both the control and injured brain was performed using adult zebrafish model. The recovered results revealed that olfactory bulb, optic commissure, telencephalon, tectum opticum, cerebellum, medulla, preglomerular complex and posterior tuberculum could be clearly identified in the cross-sectional SD-OCT images of the adult zebrafish brain. The reconstructed results also suggested that SD-OCT can be used for diagnosis and monitoring of traumatic brain injury. In particular, we found the reconstructed volumetric SD-OCT images enable a comprehensive three-dimensional characterization of the control or injured brain in the intact zebrafish.

Characterizing the Impact of Off-Axis Scan Acquisition on the Reproducibility of Total Retinal Thickness Measurements in SDOCT Volumes.

Off-axis acquisition of spectral domain optical coherence tomography (SDOCT) images has been shown to increase total retinal thickness (TRT) measurements. We analyzed the reproducibility of TRT measurements obtained using either the retinal pigment epithelium (RPE) or Bruch's membrane as reference surfaces in off-axis scans intentionally acquired through multiple pupil positions. Five volumetric SDOCT scans of the macula were obtained from one eye of 25 normal subjects. One scan was acquired through a central pupil position, while subsequent scans were acquired through four peripheral pupil positions. The internal limiting membrane, the RPE, and Bruch's membrane were segmented using automated approaches. These volumes were registered to each other and the TRT was evaluated in 9 Early Treatment of Diabetic Retinopathy Study (ETDRS) regions. The reproducibility of the TRT obtained using the RPE was computed using the mean difference, coefficient of variation (CV), and the intraclass correlation coefficient (ICC), and compared to those obtained using Bruch's membrane as the reference surface. A secondary set of 1545 SDOCT scans was also analyzed in order to gauge the incidence of off-axis scans in a typical acquisition environment. The photoreceptor tips were dimmer in off-axis images, which affected the RPE segmentation. The overall mean TRT difference and CV obtained using the RPE were 7.04 ± 4.31 μm and 1.46%, respectively, whereas Bruch's membrane was 1.16 ± 1.00 μm and 0.32%, respectively. The ICCs at the subfoveal TRT were 0.982 and 0.999, respectively. Forty-one percent of the scans in the secondary set showed large tilts (> 6%). RPE segmentation is confounded by its proximity to the interdigitation zone, a structure strongly affected by the optical Stiles-Crawford effect. Bruch's membrane, however, is unaffected leading to a more robust segmentation that is less dependent upon pupil position. The way in which OCT images are acquired can independently affect the accuracy of automated retinal thickness measurements. Assessment of scan angle in a clinical dataset demonstrates that off-axis scans are common, which emphasizes the need for caution when relying on automated thickness parameters when this component of scan acquisition is not controlled for.

A Pilot Quantitative Study of Topographic Correlation between Reticular Pseudodrusen and the Choroidal Vasculature Using En Face Optical Coherence Tomography

PurposeTo analyze the topographic correlation between reticular pseudodrusen (RPD) visualized on infrared reflectance (IR) and choroidal vasculature using en-face volumetric spectral-domain optical coherence tomography (SD-OCT).MethodsA masked observer marked individual RPD on IR images using ImageJ (NIH, Bethesda, MD). Using the macular volume scan (Cirrus, Carl Zeiss Meditec Inc, Dublin, CA), the RPE slab function was used to generate a C-scan of the most superficial choroidal vasculature. An independent masked grader created a topographic binary map of the choroidal vasculature by thresholding the en-face image, which was overlaid onto the IR map of RPD. For each IR image, ImageJ was used to generate a random set of dots as “control lesions”.Results17 eyes of 11 patients (78±13.7 years) with RPD were analyzed. The average number of RPD lesions identified on IR images was 414±71.5, of which 49.6±4.3% were located overlying the choroidal vasculature, compared to 45.4±4.0% in controls (p = 0.014). 50.4±4.3% of lesions overlay the choroidal stroma, of which 76.5±3.1% were ≤3 pixels from the choroidal vessels. The percentage of RPD lesions located within ≤3 pixels from the choroidal vasculature was significantly greater than the percentage located ≥7 pixels away. (p<0.0001). Compared to controls (71.6±3.8%), RPD were more likely to be located ≤3 pixels away from choroidal vessels (p = 0.014). In contrast, control lesions were more likely to be ≥7 pixels away from choroidal vessels than RPD (9.1±1.9% vs. 4.8±1.2%, respectively, p = 0.002).ConclusionsOur analysis shows that RPD lesions follow the underlying choroidal vasculature. Approximately half the RPD directly overlay the choroidal vessels and the majority of the remaining lesions were ≤3 pixels (≤30microns) from the vessel edge, supporting the hypothesis that RPD maybe related to pathologic changes at the choroidal level.

Segmentation of the Geographic Atrophy in Spectral-Domain Optical Coherence Tomography and Fundus Autofluorescence Images

Geographic atrophy (GA) is the atrophic late-stage manifestation of age-related macular degeneration (AMD), which may result in severe vision loss and blindness. The purpose of this study was to develop a reliable, effective approach for GA segmentation in both spectral-domain optical coherence tomography (SD-OCT) and fundus autofluorescence (FAF) images using a level set-based approach and to compare the segmentation performance in the two modalities. To identify GA regions in SD-OCT images, three retinal surfaces were first segmented in volumetric SD-OCT images using a double-surface graph search scheme. A two-dimensional (2-D) partial OCT projection image was created from the segmented choroid layer. A level set approach was applied to segment the GA in the partial OCT projection image. In addition, the algorithm was applied to FAF images for the GA segmentation. Twenty randomly chosen macular SD-OCT (Zeiss Cirrus) volumes and 20 corresponding FAF (Heidelberg Spectralis) images were obtained from 20 subjects with GA. The algorithm-defined GA region was compared with consensus manual delineation performed by certified graders. The mean Dice similarity coefficients (DSC) between the algorithm- and manually defined GA regions were 0.87 ± 0.09 in partial OCT projection images and 0.89 ± 0.07 in registered FAF images. The area correlations between them were 0.93 (P < 0.001) in partial OCT projection images and 0.99 (P < 0.001) in FAF images. The mean DSC between the algorithm-defined GA regions in the partial OCT projection and registered FAF images was 0.79 ± 0.12, and the area correlation was 0.96 (P < 0.001). A level set approach was developed to segment GA regions in both SD-OCT and FAF images. This approach demonstrated good agreement between the algorithm- and manually defined GA regions within each single modality. The GA segmentation in FAF images performed better than in partial OCT projection images. Across the two modalities, the GA segmentation presented reasonable agreement.

Distributed scanning volumetric SDOCT for motion corrected corneal biometry

We present a method, termed distributed scanning OCT (DSOCT), which reduces the effects of patient motion on corneal biometry utilizing current-generation clinically available spectral domain optical coherence tomography (SDOCT) systems. We first performed a pilot study of the power spectrum of normal patient axial eye motion based on repeated (M-mode) SDOCT. Using DSOCT to reduce the effects of patient motion, we conducted a preliminary patient study comparing the measured anterior and posterior corneal curvatures and the calculated corneal power to both corneal topography and Scheimpflug photography in normal subjects. The repeatability for the measured radius of curvature of both anterior and posterior surfaces as well as calculated corneal refractive power using DSOCT was comparable to those of both topography and Scheimpflug photography.

Corneal biometry from volumetric SDOCT and comparison with existing clinical modalities

We present a comparison of corneal biometric values from dense volumetric spectral domain optical coherence tomography (SDOCT) scans to reference values in both phantoms and clinical subjects. We also present a new optically based “keratometric equivalent power” formula for SDOCT that eliminates previously described discrepancies between corneal power form SDOCT and existing clinical modalities. Phantom objects of varying radii of curvature and corneas of normal subjects were imaged with a clinical SDOCT system. The optically corrected three-dimensional surfaces were used to recover radii of curvature and power as appropriate. These were then compared to the manufacturer’s reference values in phantoms and to measurements from topography and Scheimpflug photography in subjects. In phantom objects, paired differences between SDOCT and reference values for radii of curvature were not statistically significant. In subjects, there were no significant paired differences between SDOCT and reference values from the other modalities for anterior radius and corneal keratometric power. In contrast to other studies, we found that dense volumetric scans with available SDOCT can be used to recover corneal biometric values—including power—that correspond well with existing clinical measurements.