RobOCTNet: Robotics and Deep Learning for Referable Posterior Segment Pathology Detection in an Emergency Department Population.

Purpose: To evaluate the diagnostic performance of a robotically aligned optical coherence tomography (RAOCT) system coupled with a deep learning model in detecting referable posterior segment pathology in OCT images of emergency department patients. Methods: A deep learning model, RobOCTNet, was trained and internally tested to classify OCT images as referable versus non-referable for ophthalmology consultation. For external testing, emergency department patients with signs or symptoms warranting evaluation of the posterior segment were imaged with RAOCT. RobOCTNet was used to classify the images. Model performance was evaluated against a reference standard based on clinical diagnosis and retina specialist OCT review. Results: We included 90,250 OCT images for training and 1489 images for internal testing. RobOCTNet achieved an area under the curve (AUC) of 1.00 (95% confidence interval [CI], 0.99-1.00) for detection of referable posterior segment pathology in the internal test set. For external testing, RAOCT was used to image 72 eyes of 38 emergency department patients. In this set, RobOCTNet had an AUC of 0.91 (95% CI, 0.82-0.97), a sensitivity of 95% (95% CI, 87%-100%), and a specificity of 76% (95% CI, 62%-91%). The model's performance was comparable to two human experts' performance. Conclusions: A robotically aligned OCT coupled with a deep learning model demonstrated high diagnostic performance in detecting referable posterior segment pathology in a cohort of emergency department patients. Translational Relevance: Robotically aligned OCT coupled with a deep learning model may have the potential to improve emergency department patient triage for ophthalmology referral.

NOVEL METHOD FOR VISUALIZING PERIPHERAL RETINAL STRUCTURES WITH MICROSCOPE-INTEGRATED OPTICAL COHERENCE TOMOGRAPHY.

BACKGROUND/PURPOSE: Visualization of peripheral retinal structures with optical coherence tomography (OCT) can be challenging but can offer valuable clinical information. We describe a method for intraoperative OCT of the peripheral retina. METHODS: An investigational microscope-integrated OCT system with real-time 4D volumetric imaging was used in conjunction with a Goldmann style mirrored contact lens intraoperatively to capture peripheral images in three patients. RESULTS: We identified retinoschisis, a retinal break, and areas of focal retinal detachment using our peripheral OCT method. CONCLUSION: Use of a Goldmann lens in conjunction with intraoperative OCT offers surgeons the ability to resolve peripheral pathology that cannot be easily evaluated with OCT otherwise.

Robotic Optical Coherence Tomography Retinal Imaging for Emergency Department Patients: A Pilot Study for Emergency Physicians' Diagnostic Performance.

STUDY OBJECTIVE: To evaluate the diagnostic performance of emergency physicians' interpretation of robotically acquired retinal optical coherence tomography images for detecting posterior eye abnormalities in patients seen in the emergency department (ED). METHODS: Adult patients presenting to Duke University Hospital emergency department from November 2020 through October 2021 with acute visual changes, headache, or focal neurologic deficit(s) who received an ophthalmology consultation were enrolled in this pilot study. Emergency physicians provided standard clinical care, including direct ophthalmoscopy, at their discretion. Retinal optical coherence tomography images of these patients were obtained with a robotic, semi-autonomous optical coherence tomography system. We compared the detection of abnormalities in optical coherence tomography images by emergency physicians with a reference standard, a combination of ophthalmology consultation diagnosis and retina specialist optical coherence tomography review. RESULTS: Nine emergency physicians reviewed the optical coherence tomography images of 72 eyes from 38 patients. Based on the reference standard, 33 (46%) eyes were normal, 16 (22%) had at least 1 urgent/emergency abnormality, and the remaining 23 (32%) had at least 1 nonurgent abnormality. Emergency physicians' optical coherence tomography interpretation had 69% (95% confidence interval [CI], 49% to 89%) sensitivity for any abnormality, 100% (95% CI, 79% to 100%) sensitivity for urgent/emergency abnormalities, 48% (95% CI, 28% to 68%) sensitivity for nonurgent abnormalities, and 64% (95% CI, 44% to 84%) overall specificity. In contrast, emergency physicians providing standard clinical care did not detect any abnormality with direct ophthalmoscopy. CONCLUSION: Robotic, semi-autonomous optical coherence tomography enabled ocular imaging of emergency department patients with a broad range of posterior eye abnormalities. In addition, emergency provider optical coherence tomography interpretation was more sensitive than direct ophthalmoscopy for any abnormalities, urgent/emergency abnormalities, and nonurgent abnormalities in this pilot study with a small sample of patients and emergency physicians.

Optical coherence tomography refraction and optical path length correction for image-guided corneal surgery.

Optical coherence tomography (OCT) may be useful for guidance of ocular microsurgeries such as deep anterior lamellar keratoplasty (DALK), a form of corneal transplantation that requires delicate insertion of a needle into the stroma to approximately 90% of the corneal thickness. However, visualization of the true shape of the cornea and the surgical tool during surgery is impaired in raw OCT volumes due to both light refraction at the corneal boundaries, as well as geometrical optical path length distortion due to the group velocity of broadband OCT light in tissue. Therefore, uncorrected B-scans or volumes may not provide an accurate visualization suitable for reliable surgical guidance. In this article, we introduce a method to correct for both refraction and optical path length distortion in 3D in order to reconstruct corrected OCT B-scans in both natural corneas and corneas deformed by needle insertion. We delineate the separate roles of phase and group index in OCT image distortion correction, and introduce a method to estimate the phase index from the group index which is readily measured in samples. Using the measured group index and estimated phase index of human corneas at 1060 nm, we demonstrate quantitatively accurate geometric reconstructions of the true cornea and inserted needle shape during simulated DALK surgeries.

Practice Patterns of Fundoscopic Examination for Diabetic Retinopathy Screening in Primary Care.

Importance: Primary care professionals (PCPs) have a central role in screening for diabetic retinopathy (DR), especially in settings where access to specialty eye care is limited. Data on current DR screening practice patterns in primary care are needed to inform screening strategies. Objectives: To assess the practice patterns of fundoscopic examination for DR screening in a large primary care network and to evaluate the sensitivity and accuracy of PCP fundoscopy for detecting DR. Design, Setting, and Participants: A retrospective cohort study was performed using random sampling and manual review of electronic health records of PCP fundoscopic examination documentation compared with documentation of an examination performed by an eye care professional (ophthalmologist or optometrist) within 2 years before or after primary care encounters. From a single-institution primary care network of 28 clinics, 7449 adult patients with diabetes seen at least once in the primary care network in 2019 were eligible for this study. Data from 2001 encounters were abstracted from the electronic health record for a random sample of 767 patients. Data analysis was performed from January 2021 to May 2022. Exposure: Fundoscopic examination by PCPs. Main Outcomes and Measures: The frequency of PCPs performing fundoscopy at least once in the calendar year for patients with diabetes. Univariate and multivariable logistic regression analyses were performed to identify patient, clinician, and clinic factors associated with PCPs performing fundoscopy at least once in the calendar year. The PCP examination results were compared with diagnoses made by eye care professionals to assess the sensitivity and accuracy of the findings from PCP examinations. Results: Among the 767 adult patients with diabetes included in the analysis, 387 (50.5%) were female, and the median age was 64 years (IQR, 54-71 years). Primary care professionals documented a fundoscopic examination for 93 patients (12.1%); all results were documented as normal. When eye care professional examination results were used as the reference standard, the accuracy of PCP fundoscopic examination was 62.7% (95% CI, 50.0%-73.9%) and sensitivity for detecting disease was 0.0% (95% CI, 0.0%-14.9%). No patient demographic or clinical characteristics were associated with PCPs performing fundoscopy. In multivariable logistic regression, the number of PCP years in practice was associated with greater odds of patients receiving fundoscopy at least once in the year (adjusted odds ratio per 10 years in practice, 1.26; 95% CI, 1.01-1.59; P = .04); having nurse practitioner credentials was associated with lower odds of receiving fundoscopy (adjusted odds ratio, 0.23; 95% CI, 0.04-0.79; P = .049; compared with having physician credentials); after adjusting for rural clinic location, clinic location in a primary care shortage area, and documentation of an up-to-date eye care professional examination by a PCP in the study year. Conclusions and Relevance: In this cohort study, fundoscopic examination was rarely performed and was not sensitive for detecting DR in primary care practice. Because the rate of DR screening by eye care professionals remains low, research to explore and break down barriers to the implementation of effective primary care-based DR screening strategies, such as teleretinal imaging, is needed to prevent vision loss from undiagnosed DR.

Nonlinear distortion correction for posterior eye segment optical coherence tomography with application to tree shrews.

We propose an empirical distortion correction approach for optical coherence tomography (OCT) devices that use a fan-scanning pattern to image the posterior eye segment. Two types of reference markers were used to empirically estimate the distortion correction approach in tree shrew eyes: retinal curvature from MRI images and implanted glass beads of known diameter. Performance was tested by correcting distorted images of the optic nerve head. In small animal eyes, our purposed method effectively reduced nonlinear distortions compared to a linear scaling method. No commercial posterior segment OCT provides anatomically correct images, which may bias the 3D interpretation of these scans. Our method can effectively reduce such bias.

Computational 3D microscopy with optical coherence refraction tomography.

Optical coherence tomography (OCT) has seen widespread success as an in vivo clinical diagnostic 3D imaging modality, impacting areas including ophthalmology, cardiology, and gastroenterology. Despite its many advantages, such as high sensitivity, speed, and depth penetration, OCT suffers from several shortcomings that ultimately limit its utility as a 3D microscopy tool, such as its pervasive coherent speckle noise and poor lateral resolution required to maintain millimeter-scale imaging depths. Here, we present 3D optical coherence refraction tomography (OCRT), a computational extension of OCT which synthesizes an incoherent contrast mechanism by combining multiple OCT volumes, acquired across two rotation axes, to form a resolution-enhanced, speckle-reduced, refraction-corrected 3D reconstruction. Our label-free computational 3D microscope features a novel optical design incorporating a parabolic mirror to enable the capture of 5D plenoptic datasets, consisting of millimetric 3D fields of view over up to ±75° without moving the sample. We demonstrate that 3D OCRT reveals 3D features unobserved by conventional OCT in fruit fly, zebrafish, and mouse samples.

Ultrawide field, distortion-corrected ocular shape estimation with MHz optical coherence tomography (OCT).

Ocular deformation may be associated with biomechanical alterations in the structures of the eye, especially the cornea and sclera in conditions such as keratoconus, congenital glaucoma, and pathological myopia. Here, we propose a method to estimate ocular shape using an ultra-wide field MHz swept-source optical coherence tomography (SS-OCT) with a Fourier Domain Mode-Locked (FDML) laser and distortion correction of the images. The ocular biometrics for distortion correction was collected by an IOLMaster 700, and localized Gaussian curvature was proposed to quantify the ocular curvature covering a field-of-view up to 65°×62°. We achieved repeatable curvature shape measurements (intraclass coefficient = 0.88 ± 0.06) and demonstrated its applicability in a pilot study with individuals (N = 11) with various degrees of myopia.

In vivo quantitative analysis of anterior chamber white blood cell mixture composition using spectroscopic optical coherence tomography.

Anterior uveitis is the most common form of intraocular inflammation, and one of its main signs is the presence of white blood cells (WBCs) in the anterior chamber (AC). Clinically, the true composition of cells can currently only be obtained using AC paracentesis, an invasive procedure to obtain AC fluid requiring needle insertion into the AC. We previously developed a spectroscopic optical coherence tomography (SOCT) analysis method to differentiate between populations of RBCs and subtypes of WBCs, including granulocytes, lymphocytes and monocytes, both in vitro and in ACs of excised porcine eyes. We have shown that different types of WBCs have distinct characteristic size distributions, extracted from the backscattered reflectance spectrum of individual cells using Mie theory. Here, we further develop our method to estimate the composition of blood cell mixtures, both in vitro and in vivo. To do so, we estimate the size distribution of unknown cell mixtures by fitting the distribution observed using SOCT with a weighted combination of reference size distributions of each WBC type calculated using kernel density estimation. We validate the accuracy of our estimation in an in vitro study, by comparing our results for a given WBC sample mixture with the cellular concentrations measured by a hemocytometer and SOCT images before mixing. We also conducted a small in vivo quantitative cell mixture validation pilot study which demonstrates congruence between our method and AC paracentesis in two patients with uveitis. The SOCT based method appears promising to provide quantitative diagnostic information of cellular responses in the ACs of patients with uveitis.

Robotically aligned optical coherence tomography with 5 degree of freedom eye tracking for subject motion and gaze compensation.

Optical coherence tomography (OCT) has revolutionized diagnostics in ophthalmology. However, OCT requires a trained operator and patient cooperation to carefully align a scanner with the subject's eye and orient it in such a way that it images a desired region of interest at the retina. With the goal of automating this process of orienting and aligning the scanner, we developed a robot-mounted OCT scanner that automatically aligned with the pupil while matching its optical axis with the target region of interest at the retina. The system used two 3D cameras for face tracking and three high-resolution 2D cameras for pupil and gaze tracking. The tracking software identified 5 degrees of freedom for robot alignment and ray aiming through the ocular pupil: 3 degrees of translation (x, y, z) and 2 degrees of orientation (yaw, pitch). We evaluated the accuracy, precision, and range of our tracking system and demonstrated imaging performance on free-standing human subjects. Our results demonstrate that the system stabilized images and that the addition of gaze tracking and aiming allowed for region-of-interest specific alignment at any gaze orientation within a 28° range.

QUANTITATIVE TOPOGRAPHIC CURVATURE MAPS OF THE POSTERIOR EYE UTILIZING OPTICAL COHERENCE TOMOGRAPHY.

PURPOSE: Deformations of the retina such as staphylomas in myopia or scleral flattening in high intracranial pressure can be challenging to quantify with en face imaging. We describe an optical coherence tomography-based method for the generation of quantitative posterior eye topography maps in normal and pathologic eyes. METHODS: Using "whole eye" optical coherence tomography, we corrected for subjects' optical distortions to generate spatially accurate posterior eye optical coherence tomography volumes and created local curvature (KM, mm-1) topography maps for each consented subject. We imaged nine subjects, three normal, two with myopic degeneration, and four with papilledema including one that was imaged longitudinally. RESULTS: Normal subjects mean temporal KM was 0.0923 mm-1, nasal KM was 0.0927 mm-1, and KM local variability was 0.0162 mm-1. In myopic degeneration, subjects KM local variability was higher at 0.0836 mm-1. In papilledema subjects nasal KM was flatter compared with temporal KM (0.0709 vs. 0.0885 mm-1). Mean intrasession KM repeatability for all subjects was 0.0036 mm-1. CONCLUSION: We have developed an optical coherence tomography based method for quantitative posterior eye topography that offers the ability to analyze local curvature with micron scale resolution and offers the potential to help clinicians and researchers characterize subtle, local retinal deformations earlier in patients and follow their development over time.

Multimodal handheld adaptive optics scanning laser ophthalmoscope.

Non-confocal adaptive optics scanning laser ophthalmoscopy (AOSLO) has enhanced the study of human retinal photoreceptors by providing complementary information to standard confocal AOSLO images. Previously we developed the first confocal handheld AOSLO (HAOSLO) capable of in vivo cone photoreceptor imaging in supine and non-cooperative patients. Here, we introduce the first multimodal (M-)HAOSLO for confocal and non-confocal split-detection (SD) imaging to allow for more comprehensive patient data collection. Aside from its unprecedented miniature size and weight, M-HAOSLO is also the first system to perform sensorless wavefront-corrected SD imaging of cone photoreceptors.

Topography and pachymetry maps for mouse corneas using optical coherence tomography.

The majority of the eye's refractive power lies in the cornea, and pathological changes in its shape can affect vision. Small animal models offer an unparalleled degree of control over genetic and environmental factors that can help elucidate mechanisms of diseases affecting corneal shape. However, there is not currently a method to characterize the corneal shape of small animal eyes with topography or pachymetry maps, as is done clinically for humans. We bridge this gap by demonstrating methods using optical coherence tomography (OCT) to generate the first topography and pachymetry (thickness) maps of mouse corneas. Radii of curvature acquired using OCT were validated using calibration spheres as well as in vivo mouse corneas with a mouse keratometer. The resulting topography and pachymetry maps are analogous to those used diagnostically in clinic and potentially allow for characterization of genetically modified mice that replicate key features of human corneal disease.

PPIP5K2 and PCSK1 are Candidate Genetic Contributors to Familial Keratoconus.

Keratoconus (KC) is the most common corneal ectatic disorder affecting >300,000 people in the US. KC normally has its onset in adolescence, progressively worsening through the third to fourth decades of life. KC patients report significant impaired vision-related quality of life. Genetic factors play an important role in KC pathogenesis. To identify novel genes in familial KC patients, we performed whole exome and genome sequencing in a four-generation family. We identified potential variants in the PPIP5K2 and PCSK1 genes. Using in vitro cellular model and in vivo gene-trap mouse model, we found critical evidence to support the role of PPIP5K2 in normal corneal function and KC pathogenesis. The gene-trap mouse showed irregular corneal surfaces and pathological corneal thinning resembling KC. For the first time, we have integrated corneal tomography and pachymetry mapping into characterization of mouse corneal phenotypes which could be widely implemented in basic and translational research for KC diagnosis and therapy in the future.

Ocular anterior chamber blood cell population differentiation using spectroscopic optical coherence tomography.

There is potential clinical significance in identifying cellular responses in the anterior chamber (AC) of the eye, which can indicate hyphema (an accumulation of red blood cells [RBCs]) or aberrant intraocular inflammation (an accumulation of white blood cells [WBCs]). In this work, we developed a spectroscopic OCT analysis method to differentiate between populations of RBCs and subtypes of WBCs, including granulocytes, lymphocytes and monocytes, both in vitro and in ACs of porcine eyes. We developed an algorithm to track single cells within OCT data sets, and extracted the backscatter reflectance spectrum of each single cell from the detected interferograms using the short-time Fourier transform (STFT). A look-up table of Mie back-scattering spectra was generated and used to correlate the backscatter spectral features of single cells to their characteristic sizes. The extracted size distributions based on the best Mie spectra fit were significantly different between each cell type. We also studied theoretical backscattering models of single RBCs to further validate our experimental results. The described work is a promising step towards clinically differentiating and quantifying AC blood cell types.

Advances in Whole-Eye Optical Coherence Tomography Imaging.

Contemporary anterior segment and retinal optical coherence tomography (OCT) systems only image their particular designated region of the eye and cannot image both areas of the eye at once. This separation is due to the differences in optical system design needed to properly image the front or back of the eye and also due to limitations in the imaging depth of current commercial OCT systems. More recently, research and commercial OCT systems capable of "whole-eye" imaging have been described. These whole-eye OCT systems enable applications such as ocular biometry for cataract surgery, ocular shape analysis for myopia, and others. Further, these whole-eye OCT systems allow us to image the eye as an integrated whole rather than as separate, independent divisions.

Wide-field whole eye OCT system with demonstration of quantitative retinal curvature estimation.

Current conventional clinical OCT systems image either only the anterior or the posterior eye during a single acquisition. This localized imaging limits conventional OCT's use for characterizing global ocular morphometry and biometry, which requires knowledge of spatial relationships across the entire eye. We developed a "whole eye" optical coherence tomography system that simultaneously acquires volumes with a wide field-of-view for both the anterior chamber (14 x 14 mm) and retina (55°) using a single source and detector. This system was used to measure retinal curvature in a pilot population and compared against curvature of the same eyes measured with magnetic resonance imaging.

Constant linear velocity spiral scanning for near video rate 4D OCT ophthalmic and surgical imaging with isotropic transverse sampling.

Ultrahigh speed optical coherence tomography (OCT) systems with >100 kHz A-scan rates can generate volumes rapidly with minimal motion artifacts and are well suited for 4D imaging (volumes through time) applications such as intra-operative imaging. In such systems, high OCT data acquisition efficiency (defined as the fraction of usable A-scans generated during the total acquisition time) is desired to maximize the volumetric frame rate and sampling pitch. However, current methods for beam scanning using non-resonant and resonant mirror scanners can result in severe scan distortion and transverse oversampling as well as require acquisition dead times, which limit the acquisition efficiency and performance of ultrahigh speed 4D OCT. We introduce constant linear velocity spiral scanning (CLV-SC) as a novel beam scanning method to maximize the data acquisition efficiency of ultrahigh speed 4D OCT systems. We demonstrate that CLV-SC does not require acquisition dead times and achieves more uniform transverse sampling compared to raster scanning. To assess its clinical utility, we implement CLV-SC with a 400 kHz OCT system and image the anterior eye and retina of healthy adults at up to 10 volumes per second with isotropic transverse sampling, allowing B-scans with equal sampling pitch to be extracted from arbitrary locations within a single volume. The feasibility of CLV-SC for intra-operative imaging is also demonstrated using a 800 kHz OCT system to image simulated retinal surgery at 15 volumes per second with isotropic transverse sampling, resulting in high quality volume renders that enable clear visualization of surgical instruments and manipulation of tissue.

Volumetric Measurement of Subretinal Blebs Using Microscope-Integrated Optical Coherence Tomography.

PURPOSE: We advance studies of subretinal treatments by developing a microscope-integrated optical coherence tomography (MIOCT) image-based method for measuring the volume of therapeutics delivered into the subretinal space. METHODS: A MIOCT image-based volume measurement method was developed and assessed for accuracy and reproducibility by imaging an object of known size in model eyes. This method then was applied to subretinal blebs created by injection of diluted triamcinolone. Bleb volumes obtained from MIOCT were compared to the intended injection volume and the surgeon's estimation of leakage. RESULTS: Validation of the image-based volume measurement method showed accuracy to ±1.0 µL (6.0% of measured volume) with no statistically significant variation under different imaging settings. When this method was applied to subretinal blebs, four of 11 blebs without surgeon-observed leakage yielded a mean volume of 32 ± 12.5 µL, in contrast to the intended 50 µL volume injected from the delivery device. This constituted a mean difference of -18 µL (mean percent error, 36 ± 25%). For all 11 blebs, the surgeon's estimations of leakage were significantly different from and showed no correlation with the volume loss based on image-based volume measurements (P < 0.001, paired t-test; intraclass correlation = 0). CONCLUSIONS: We validated an accurate and reproducible method for measuring subretinal volumes using MIOCT. Use of this method revealed that the intended volume might not be delivered into the subretinal space. MIOCT can allow for accurate assessment of subretinal dose delivered, which may have therapeutic implications in evaluating the efficacy and toxicity of subretinal therapies. TRANSLATIONAL RELEVANCE: Use of MIOCT can provide feedback on the accuracy of subretinal injection volumes delivered.

Retinal imaging in human autopsy eyes using a custom optical coherence tomography periscope.

Age-related macular degeneration (AMD) is a major cause of vision loss in the elderly. To better study the pathobiology of AMD, postmortem eyes offer an excellent opportunity to correlate optical coherence tomography (OCT) imaging characteristics with histopathology. However, postmortem eyes from autopsy present challenges to standard OCT imaging including opaque anterior segment structures and standard of care autopsy processing resulting in oblique views to the macula. To overcome these challenges, we report a custom periscope attached by a standard mount to an OCT sample arm and demonstrate high quality macular OCT acquisitions in autopsy-processed eyes.