Modern technologies for retinal scanning and imaging: an introduction for the biomedical engineer

CSANet: a lightweight channel and spatial attention neural network for grading diabetic retinopathy with optical coherence tomography angiography

Background

Diabetic retinopathy (DR) is one of the most common eye diseases. Convolutional neural networks (CNNs) have proven to be a powerful tool for learning DR features; however, accurate DR grading remains challenging due to the small lesions in optical coherence tomography angiography (OCTA) images and the small number of samples.

Methods

In this article, we developed a novel deep-learning framework to achieve the fine-grained classification of DR; that is, the lightweight channel and spatial attention network (CSANet). Our CSANet comprises two modules: the baseline model, and the hybrid attention module (HAM) based on spatial attention and channel attention. The spatial attention module is used to mine small lesions and obtain a set of spatial position weights to address the problem of small lesions being ignored during the convolution process. The channel attention module uses a set of channel weights to focus on useful features and suppress irrelevant features.

Results

The extensive experimental results for the OCTA-DR and diabetic retinopathy analysis challenge (DRAC) 2022 data sets showed that the CSANet achieved state-of-the-art DR grading results, showing the effectiveness of the proposed model. The CSANet had an accuracy rate of 97.41% for the OCTA-DR data set and 85.71% for the DRAC 2022 data set.

Conclusions

Extensive experiments using the OCTA-DR and DRAC 2022 data sets showed that the proposed model effectively mitigated the problems of mutual confusion between DRs of different severity and small lesions being neglected in the convolution process, and thus improved the accuracy of DR classification.

Measuring the Impact of Data Sharing: From Author-Level Metrics to Quantification of Economic and Non-tangible Benefits

In early 2023, the National Institutes of Health (NIH) implemented its Data Management and Sharing (DMS) Policy, requiring researchers to share scientific data produced with NIH funding. The policy's objective is to amplify the benefits of public investment in research by promoting the dissemination and reusability of primary data. Given this backdrop, identifying a robust methodology to assess the impact of data sharing across diverse research domains is essential. In this review, we adopted two methodological paradigms, the bottom-up and top-down strategies, and employed content analysis to pinpoint established methodologies and innovative practices within this intricate field. Although numerous author-level metrics are available to gauge the impact of data sharing, their application is still limited. Non-traditional metrics, encompassing economic (e.g., cost savings) and intangible benefits, presently appear to hold more potential for evaluating the impact of primary data sharing. Finally, we address the primary obstacles encountered by open data policies and introduce an innovative "Shared model for shared data" framework to bolster data sharing practices and refine evaluation metrics.

An innovative multi-modal retinal imaging system for in vivo retinal detection in small animals

This paper presents an innovative retinal imaging system tailored for in vivo fundus detection in small animals. This system integrates Scanning Laser Ophthalmoscopy (SLO) and optical Coherence Tomography (OCT) techniques, enabling the simultaneous generation of images from various modalities, including SLO reflectance, SLO fluorescein angiogram, OCT, and OCT angiogram. The existing multi-modal retinal imaging systems generally encounter limitations such as the inability to detect peripheral lesion areas attributed to small Field of View (FOV) design and susceptibility to sample motion due to slow data acquisition speed. To address these challenges, it's essential to underscore that this proposed system offers a range of notable advantages, including its compact design, the capacity for widefield imaging with a FOV of up to 100°, and a rapid OCT A-scan rate of 250 kHz, notably exceeding the capabilities of pre-existing multi-modal retinal imaging systems. Validation of the system involved imaging the eyes of normal wild-type mice and diseased mice afflicted with retinal detachment and choroidal neovascularization (CNV). The favorable imaging results demonstrate the system's reliability in identifying retinal lesions in small animals.

Optimizing the OCTA layer fusion option for deep learning classification of diabetic retinopathy

The purpose of this study is to evaluate layer fusion options for deep learning classification of optical coherence tomography (OCT) angiography (OCTA) images. A convolutional neural network (CNN) end-to-end classifier was utilized to classify OCTA images from healthy control subjects and diabetic patients with no retinopathy (NoDR) and non-proliferative diabetic retinopathy (NPDR). For each eye, three en-face OCTA images were acquired from the superficial capillary plexus (SCP), deep capillary plexus (DCP), and choriocapillaris (CC) layers. The performances of the CNN classifier with individual layer inputs and multi-layer fusion architectures, including early-fusion, intermediate-fusion, and late-fusion, were quantitatively compared. For individual layer inputs, the superficial OCTA was observed to have the best performance, with 87.25% accuracy, 78.26% sensitivity, and 90.10% specificity, to differentiate control, NoDR, and NPDR. For multi-layer fusion options, the best option is the intermediate-fusion architecture, which achieved 92.65% accuracy, 87.01% sensitivity, and 94.37% specificity. To interpret the deep learning performance, the Gradient-weighted Class Activation Mapping (Grad-CAM) was utilized to identify spatial characteristics for OCTA classification. Comparative analysis indicates that the layer data fusion options can affect the performance of deep learning classification, and the intermediate-fusion approach is optimal for OCTA classification of DR.

Accuracy and Time Comparison Between Side-by-Side and Artificial Intelligence Overlayed Images

BACKGROUND AND OBJECTIVE

The purpose of this study was to evaluate the accuracy and the time to find a lesion, taken in different platforms, color fundus photographs and infrared scanning laser ophthalmoscope images, using the traditional side-by-side (SBS) colocalization technique to an artificial intelligence (AI)-assisted technique.

PATIENTS AND METHODS

Fifty-three pathological lesions were studied in 11 eyes. Images were aligned using SBS and AI overlaid methods. The location of each color fundus lesion on the corresponding infrared scanning laser ophthalmoscope image was analyzed twice, one time for each method, on different days, for two specialists, in random order. The outcomes for each method were measured and recorded by an independent observer.

RESULTS

The colocalization AI method was superior to the conventional in accuracy and time (P < .001), with a mean time to colocalize 37% faster. The error rate using AI was 0% compared with 18% in SBS measurements.

CONCLUSIONS

AI permitted a more accurate and faster colocalization of pathologic lesions than the conventional method. [Ophthalmic Surg Lasers Imaging Retina 2023;54:108-113.].

Lateral Resolution of a Commercial Optical Coherence Tomography Instrument

Purpose

The lateral resolution of an optical coherence tomography (OCT) instrument was considered to be equal to the illumination spot size on the retina. To evaluate the potential lateral resolution of the Spectralis OCT, an instrument calculated to have a 14 µm resolution.

Methods

The lateral point spread function (PSF) was evaluated using diamond abrasive powder 0 to 1 µm in diameter in silicone elastomer and a validated target with 800 nm FeO particles in urethane. The amplitude transfer function was calculated from human OCT images. Finally, resolution was measured using the 1951 USAF target.

Results

Measurement of the lateral PSF from 1215 diamond particle images yielded a full-width half maximum (FWHM) to be 5.11 µm and for 732 FeO particles, 4.9 µm. From the amplitude transfer function, the FWHM of the diffraction limited PSF was calculated to be 5.0 µm. The USAF target imaging showed a lateral resolution of 4.6 µm.

Conclusions

Although a calculation of the spot size of the illumination beam was reported in the past as the lateral resolution of the OCT instrument, the actual lateral resolution is better by a factor of at least 2.5 times. The clinically used A-scan spacing was derived from the calculated, and not the true resolution, and results in under sampling. This set of findings likely apply to all commercial clinical instruments.

Translational Relevance

The scan density parameters of past and present commercial OCT instruments were based on earlier translational concepts, which now appear to have been incorrect.

Optical coherence tomography in patients with Wilson's disease

OBJECTIVES

Wilson disease (WD) is an autosomal recessive disorder that leads to copper accumulation and deposition in different organs, frequently affecting visual pathways. Recent studies have detected morphological changes of the retina in patients with WD using optical coherence tomography (OCT). Measuring the thickness of the retinal nerve fibre layer (RNFL) with OCT provides an objective assessment of integrity and morphological abnormalities of the retina. The aim of this study was to evaluate the relationship between OCT parameters and form of the disease, therapy and symptoms duration, as well as severity of neurological impairment.

METHODS

The study comprised of 52 patients with WD and 52 healthy controls (HC). All the patients were on a regular and stable chelation therapy and/or zinc salts. Patients were divided into two groups, with neurological (NWD) or hepatic form of the disease (HWD). OCT was performed to assess the RNFL thickness.

RESULTS

The WD patients had significantly lower intraocular pressure in both eyes and lower RNFL thickness than the HC. There were no differences between NWD and HWD in any of the ophthalmologically tested parameters. No significant correlations were found between clinical features and retinal thickness parameters. Stratification of the cohort according to the disease duration showed that disease duration did not influence the RNFL thickness.

CONCLUSION

We found that involvement of the retina represented a subclinical finding in neurologically intact patients in the HWD group. Nevertheless, the value of OCT as a biomarker for the assessment of the clinical course and progression of WD still remains uncertain.

CLINICAL APPLICATION OF MULTICOLOR IMAGING IN CENTRAL SEROUS CHORIORETINOPATHY

PURPOSE

To characterize features of central serous chorioretinopathy (CSC) using multicolor (MC) imaging and to compare the efficacy of using MC imaging and traditional color fundus photography (CFP) for detecting features of CSC.

METHODS

A retrospective review of 75 eyes of 69 participants with CSC from the Eye Center of the Renmin Hospital of Wuhan University. The patients underwent same-day CFP, fundus fluorescein angiography (FFA), spectral domain optical coherence tomography (SD-OCT), and MC imaging (including infrared reflectance (IR), green reflectance (GR), blue reflectance (BR), combined standard MC image, and green-blue-enhanced image). Positive CSC lesions were evaluated using FFA and SD-OCT. Features in images of CFP, standard MC, green-blue-enhanced, IR, GR, and BR were analyzed and calculated.

RESULTS

Among the 75 eyes of 69 patients, 75 eyes with subretinal fluid (SRF) and retinal pigment epithelial (RPE) leakage point, and 43 eyes with RPE damage were observed by FFA and SD-OCT. The detection rate of SRF was significantly higher with the standard MC image (92.0%), green-blue-enhanced image (92.0%), IR (88.0%), and GR (88.0%) than that with CFP (66.7%) (P < 0.05). Blue reflectance (45.3%) was associated with lower rate of SRF detection compared to CFP (P < 0.05). The standard MC image (84.0%), green-blue-enhanced image (84.0%), IR (78.7%), and GR (80%) delineated the boundaries of SRF more effectively than CFP (44%). The abnormal areas corresponding to RPE leakage points on FFA were observed with the standard MC image, green-blue-enhanced image, and IR at detection rates of 90.7, 82.7, and 90.7%, respectively, which were significantly higher than with CFP (29.3%) (P < 0.05). However, the detection rates of the abnormalities corresponding to leakage on FFA were lower with GR (5.3%) and BR (1.3%) than those with CFP (P < 0.05). Areas of RPE damage on FFA were observed at the same locations with the standard MC image, green-blue-enhanced image, and IR at detection rates of 97.7, 93.0, and 95.3%, respectively, which were significantly higher than with CFP (41.9%) (P < 0.05). Compared with CFP, the detectable rates of RPE damage on GR (14.0%) and BR (9.3%) were lower (P < 0.05).

CONCLUSION

As an adjunct to SD-OCT, the MC image can delineate the extent or boundaries of SRF more effectively than CFP. As an adjunct to FFA, the MC image can identify foci of leakage more effectively than CFP. The MC image, particularly the IR channel, can identify areas of RPE damage more effectively than CFP. Therefore, the MC image may be a useful adjunct to FFA and OCT for detecting or monitoring CSC.

ULTRAWIDEFIELD PSEUDOCOLOR RETINAL IMAGING VERSUS REAL-COLOR FUNDUS PHOTOGRAPHY FOR DETECTION OF INTRARETINAL PIGMENT MIGRATION IN AGE-RELATED MACULAR DEGENERATION

PURPOSE

To compare pseudocolor Optos ultrawidefield (UWF) retinal images with conventional real-color fundus photography (CFP) for detecting macular hyperpigmentary changes in intermediate age-related macular degeneration.

METHODS

This retrospective study included 50 patients diagnosed with intermediate age-related macular degeneration. All patients underwent Optos imaging and CFP. The overall accuracy to visualize hyperpigmentation and its morphologic features was graded by two independent readers using a standardized grid. Structural and en face optical coherence tomography images were correlated with UWF and CFP images to determine spatial correspondence of pigment clumping on fundus images and hyperreflective foci on optical coherence tomography.

RESULTS

One hundred eyes of 50 patients had hyperpigmentary changes on funduscopic examination and were included. The intragraders and intergraders agreements were high for all measurements (P < 0.001). At least one hyperpigmentary changes within the standardized grid was detected in 93% using CFP and 100% using UWF camera (P = 0.02). The total area of hyperpigmentation measured on UWF images was significantly higher than on CFP images (P < 0.001). There was a significant correlation between the presence of hyperpigmentary changes on both CFP and UWF images and hyperreflective foci on structural optical coherence tomography (P < 0.001).

CONCLUSION

Ultrawidefield fundus images allow high detection and accurate quantification of macular hyperpigmentary changes in intermediate age-related macular degeneration compared with conventional CFP.

Association of retinal venular tortuosity with impaired renal function in the Northern Ireland Cohort for the Longitudinal Study of Ageing

BACKGROUND

Previous studies have identified retinal microvascular features associated with renal dysfunction. Biopsies are necessary to confirm kidney microvascular damage and retinal imaging may enable evaluation of microangiopathic characteristics reflecting renal changes associated with chronic kidney disease (CKD). We evaluated retinal microvascular parameters (RMPs) for associations with renal function in a cross-sectional analysis of the Northern Ireland Cohort for the Longitudinal Study of Ageing.

METHODS

RMPs (central retinal arteriolar/ venular equivalents [CRAE/CRVE], arteriolar to venular ratio [AVR], fractal dimension and tortuosity) were measured from optic disc centred fundus images using semi-automated software. Associations were assessed with multivariable regression analyses between RMPs and estimated glomerular filtration rate (eGFR) defined by serum creatinine (eGFRscr) and cystatin C (eGFRcys) and also CKD status characterised by eGFR < 60 mL/min/1.73m2. Regression models were adjusted for potential confounders including age, sex, diabetes, smoking status, educational attainment, cardiovascular disease, body mass index, antihypertensive medication, systolic blood pressure, triglycerides, high- and low-density lipoprotein levels.

RESULTS

Data were included for 1860 participants that had measures of renal function and retinal fundus images of sufficient quality for analysis. Participants had a mean age of 62.0 ± 8.5 yrs. and 53% were female. The mean eGFR for scr and cys were 82.2 ± 14.9 mL/min/1.73m2 and 70.7 ± 18.6 mL/min/1.73m2 respectively. eGFRcys provided lower estimates than eGFRscr resulting in a greater proportion of participants categorised as having CKD stages 3-5 (eGFRcys 26.8%; eGFRscr 7.9%). Multivariable regression analyses showed that increased venular tortuosity (OR = 1.30; 95%CI: 1.10, 1.54; P < 0.01) was associated with CKD stages 3-5 characterised by eGFRscr < 60 mL/min/1.73 m2. No additional associations between CKD status characterised by eGFRscr or with eGFRcys, were detected (P > 0.05). Multivariable regression failed to detect associations between CRAE, CRVE, AVR, fractal dimension or tortuosity and eGFRscr or eGFRcys (P > 0.05).

CONCLUSION

Increased retinal venular tortuosity was associated with CKD stages 3-5 defined by eGFRscr < 60 mL/min/1.73 m2, in an older population independent of potential confounding factors. These retinal measures may provide non-invasive microvascular assessment of associations with CKD.

Retinal oximetry in glaucoma: investigations and findings reviewed

Abnormalities of the retinal blood supply have been widely implicated in primary open-angle glaucoma (POAG). Impaired blood supply to the retina and optic nerve head (ONH) may be a primary pathophysiologic mechanism contributing to POAG ('vascular hypothesis'). However, the decreased metabolic activity of atrophic tissue is itself known to induce both vascular changes and decreased blood flow due to reduced oxygen demand. Therefore, primary nonvascular factors could potentially induce glaucomatous atrophy, with subsequent secondary vascular pathology ('mechanical hypothesis'). Retinal oximetry holds great promise in the investigation of glaucoma pathogenesis, as it can provide useful data on retinal metabolic oxygen demand, especially when combined with measurements of retinal blood flow. This review surveys the research on retinal metabolism in POAG using spectroscopic retinal oximetry. The use of mathematical models in combination with oximetric data to investigate the role of retinal metabolism and oxygen supply in POAG is also discussed.

AV-Net: deep learning for fully automated artery-vein classification in optical coherence tomography angiography

This study is to demonstrate deep learning for automated artery-vein (AV) classification in optical coherence tomography angiography (OCTA). The AV-Net, a fully convolutional network (FCN) based on modified U-shaped CNN architecture, incorporates enface OCT and OCTA to differentiate arteries and veins. For the multi-modal training process, the enface OCT works as a near infrared fundus image to provide vessel intensity profiles, and the OCTA contains blood flow strength and vessel geometry features. A transfer learning process is also integrated to compensate for the limitation of available dataset size of OCTA, which is a relatively new imaging modality. By providing an average accuracy of 86.75%, the AV-Net promises a fully automated platform to foster clinical deployment of differential AV analysis in OCTA.

Transfer Learning for Automated OCTA Detection of Diabetic Retinopathy

Purpose

To test the feasibility of using deep learning for optical coherence tomography angiography (OCTA) detection of diabetic retinopathy.

Methods

A deep-learning convolutional neural network (CNN) architecture, VGG16, was employed for this study. A transfer learning process was implemented to retrain the CNN for robust OCTA classification. One dataset, consisting of images of 32 healthy eyes, 75 eyes with diabetic retinopathy (DR), and 24 eyes with diabetes but no DR (NoDR), was used for training and cross-validation. A second dataset consisting of 20 NoDR and 26 DR eyes was used for external validation. To demonstrate the feasibility of using artificial intelligence (AI) screening of DR in clinical environments, the CNN was incorporated into a graphical user interface (GUI) platform.

Results

With the last nine layers retrained, the CNN architecture achieved the best performance for automated OCTA classification. The cross-validation accuracy of the retrained classifier for differentiating among healthy, NoDR, and DR eyes was 87.27%, with 83.76% sensitivity and 90.82% specificity. The AUC metrics for binary classification of healthy, NoDR, and DR eyes were 0.97, 0.98, and 0.97, respectively. The GUI platform enabled easy validation of the method for AI screening of DR in a clinical environment.

Conclusions

With a transfer learning process for retraining, a CNN can be used for robust OCTA classification of healthy, NoDR, and DR eyes. The AI-based OCTA classification platform may provide a practical solution to reducing the burden of experienced ophthalmologists with regard to mass screening of DR patients.

Translational Relevance

Deep-learning-based OCTA classification can alleviate the need for manual graders and improve DR screening efficiency.

IDENTIFICATION OF POSTERIOR SEGMENT PATHOLOGY WITH EN FACE RETINAL IMAGING USING MULTICOLOR CONFOCAL SCANNING LASER OPHTHALMOSCOPY

PURPOSE

To assess posterior segment findings on multicolor confocal scanning laser ophthalmoscopy by correlation with spectral domain optical coherence tomography (SD-OCT) and to quantify agreement between these imaging modalities.

METHODS

Retrospective review of 159 eyes of 96 consecutive patients who underwent concurrent imaging with multicolor confocal scanning laser ophthalmoscopy and SD-OCT. Positive percent agreement and negative percent agreement were calculated for each finding identified on infrared, green, blue, and multicolor reflectance images using SD-OCT as a comparator.

RESULTS

Infrared reflectance best detected outer retinal and choroidal findings such as choroidal lesions, retinal pigment epithelium atrophy, peripapillary atrophy, and drusen (positive percent agreement 100, 92, 92, and 67%, respectively). Inner retinal changes including epiretinal membrane, lamellar macular hole, and inner retinal alterations were best detected on blue reflectance (positive percent agreement 94, 50, and 100%, respectively). Composite multicolor reflectance most effectively detected conditions with retinal elevation, including pigment epithelial detachment, intraretinal fluid, and subretinal fluid (positive percent agreement 65, 49, and 54%, respectively). Multicolor confocal scanning laser ophthalmoscopy detected intraretinal and subretinal hemorrhages, which were not detected on SD-OCT (negative percent agreement 87 and 97%, respectively).

CONCLUSION

Multicolor confocal scanning laser ophthalmoscopy is capable of identifying posterior segment pathology at various anatomical depths and may be a useful adjunct to SD-OCT for detecting or monitoring certain retinal conditions.

Computer-aided fixation detection using retinal birefringence in multi-modal ophthalmic systems: Computer, electronics, algorithms

Many diagnostic and some therapeutic ophthalmic devices require a reliable complementing method to track the direction of gaze or just to validate fixation of the eye on a presented target. This would allow acquisition of artefact-free robust images of the fovea and the surrounding macula. So far, there have been only few attempts to provide fast and dependable fixation information to an optical imaging system in real time, to guide image acquisition. The author's lab has developed several instruments that detect the location of the fovea using retinal birefringence scanning (RBS), proven to be very effective. Here, an RBS-based fixation detection subsystem is proposed, designed to operate conjointly with a number of ophthalmic imaging technologies. Combining RBS with such technologies is not trivial, because RBS uses polarized light and polarization-sensitive optics, while most other modalities don't. The polarization optics was optimized by means of enhanced computer modeling. Both the electronic hardware and the software were designed for fast and reliable performance. Because many retinal imaging systems are used in pediatric settings, extensive audio-visual circuitry was employed for efficient attention/fixation attraction. The optomechanics has been optimized for robust data acquisition. This computer-aided conjoint system employs true anatomical information from the back of the eye and needs no calibration. The prototype instrument uses a decision-making logic based on four frequencies generated during scanning. The results reveal the applicability of RBS as an adjunct fixation monitoring modality, showing promise to remove the limitation imposed by eye movements upon advanced ophthalmic imaging technologies.

Modeling retinal detachment associated with hemorrhage by Monte Carlo simulation

In this work we study fundus reflection in the case of retinal detachment. Vitreous humor builds up beneath the retina through a break in the retina or when there is traction on the retina. Further, we assume that this detachment is associated with hemorrhage in regions both above and below the retina. To investigate hemorrhage, we assume erythrocytes to be spheres of different radii. Then we apply Mie scattering theory to these spheres and calculate the scattering coefficient, the absorption coefficient, and the anisotropy factor. Using these parameters, we apply Monte Carlo simulation to calculate reflection. In order to model fundus reflection under retinal detachment, we define three types of reflection: reflection from the vitreous body, retinal reflection where photons are reflected from the retinal vitreous interface once, and retinal reflection where photons are reflected many times from the retinal vitreous interface.

Multimodal Imaging Correlation of the Concentric Macular Rings Sign in Foveal Hypoplasia: A Distinctive Henle Fiber Layer Geometry

PURPOSE

To expand the multimodal imaging correlation of the concentric macular rings (CMR) sign seen on ultra-widefield fundus photography in patients with foveal hypoplasia.

DESIGN

Retrospective case series.

PARTICIPANTS

Thirty-two patients with foveal hypoplasia who demonstrated the CMR sign on ultra-widefield fundus photography.

METHODS

Inclusion criterion was the presence of a CMR sign detected on ultra-widefield fundus photography.

MAIN OUTCOME MEASURES

Noninvasive multimodal retinal imaging, including ultra-widefield fundus photography, structural OCT, near-infrared reflectance, and blue fundus autofluorescence, were investigated. Horizontal dense B-scans and en face OCT images were acquired.

RESULTS

Evaluation of all patients (n = 32 patients) demonstrated a CMR sign on ultra-widefield fundus photography. Structural OCT scans were consistent with foveal hypoplasia in all patients. En face OCT images acquired at the level of the Henle fiber layer highlighted similar concentric rings around the location of the incipient fovea. The series of concentric rings was not visible at any other level of the macula. A significant correlation was found between the horizontal diameter of the largest outer ring and foveal hypoplasia grades (P < 0.0001).

CONCLUSIONS

The CMR sign seen on ultra-widefield fundus imaging may be a distinctive feature of foveal hypoplasia and can support this diagnosis, especially in patients in whom OCT cannot be performed (patients with poor fixation or nystagmus or young children). Multimodal imaging correlation suggests that these concentric rings occur exclusively at the level of the Henle fiber layer. This distinctive Henle fiber layer geometry may reflect an arrested development stage in the timeline of foveal maturation.

Methods to label, image, and analyze the complex structural architectures of microvascular networks

Microvascular networks play key roles in oxygen transport and nutrient delivery to meet the varied and dynamic metabolic needs of different tissues throughout the body, and their spatial architectures of interconnected blood vessel segments are highly complex. Moreover, functional adaptations of the microcirculation enabled by structural adaptations in microvascular network architecture are required for development, wound healing, and often invoked in disease conditions, including the top eight causes of death in the Unites States. Effective characterization of microvascular network architectures is not only limited by the available techniques to visualize microvessels but also reliant on the available quantitative metrics that accurately delineate between spatial patterns in altered networks. In this review, we survey models used for studying the microvasculature, methods to label and image microvessels, and the metrics and software packages used to quantify microvascular networks. These programs have provided researchers with invaluable tools, yet we estimate that they have collectively attained low adoption rates, possibly due to limitations with basic validation, segmentation performance, and nonstandard sets of quantification metrics. To address these existing constraints, we discuss opportunities to improve effectiveness, rigor, and reproducibility of microvascular network quantification to better serve the current and future needs of microvascular research.

Color Fundus Image Guided Artery-Vein Differentiation in Optical Coherence Tomography Angiography

Purpose

This study aimed to develop a method for automated artery-vein classification in optical coherence tomography angiography (OCTA), and to verify that differential artery-vein analysis can improve the sensitivity of OCTA detection and staging of diabetic retinopathy (DR).

Methods

For each patient, the color fundus image was used to guide the artery-vein differentiation in the OCTA image. Traditional mean blood vessel caliber (m-BVC) and mean blood vessel tortuosity (m-BVT) in OCTA images were quantified for control and DR groups. Artery BVC (a-BVC), vein BVC (v-BVC), artery BVT (a-BVT), and vein BVT (a-BVT) were calculated, and then the artery-vein ratio (AVR) of BVC (AVR-BVC) and AVR of BVT (AVR-BVT) were quantified for comparative analysis. Sensitivity, specificity, and accuracy were used as performance metrics of artery-vein classification. One-way, multilabel ANOVA with Bonferroni's test and Student's t-test were employed for statistical analysis.

Results

Forty eyes of 20 control subjects and 80 eyes of 48 NPDR patients (18 mild, 16 moderate, and 14 severe NPDR) were evaluated in this study. The color fundus image-guided artery-vein differentiation reliably identified individual arteries and veins in OCTA. AVR-BVC and AVR-BVT provided significant (P < 0.001) and moderate (P < 0.05) improvements, respectively, in detecting and classifying NPDR stages, compared with traditional m-BVC analysis.

Conclusions

Color fundus image-guided artery-vein classification provides a feasible method to differentiate arteries and veins in OCTA. Differential artery-vein analysis can improve the sensitivity of OCTA detection and classification of DR. AVR-BVC is the most-sensitive feature, which can classify control and mild NPDR, providing a quantitative biomarker for objective detection of early DR.

Fully automated geometric feature analysis in optical coherence tomography angiography for objective classification of diabetic retinopathy

This study is to establish quantitative features of vascular geometry in optical coherence tomography angiography (OCTA) and validate them for the objective classification of diabetic retinopathy (DR). Six geometric features, including total vessel branching angle (VBA: θ), child branching angles (CBAs: α1 and α2), vessel branching coefficient (VBC), and children-to-parent vessel width ratios (VWR1 and VWR2), were automatically derived from each vessel branch in OCTA. Comparative analysis of heathy control, diabetes with no DR (NoDR), and non-proliferative DR (NPDR) was conducted. Our study reveals four quantitative OCTA features to produce robust DR detection and staging classification: (ANOVA, P<0.05), VBA, CBA1, VBC, and VWR1.

OCT feature analysis guided artery-vein differentiation in OCTA

Differential artery-vein analysis promises better sensitivity for retinal disease detection and classification. However, clinical optical coherence tomography angiography (OCTA) instruments lack the function of artery-vein differentiation. This study aims to verify the feasibility of using OCT intensity feature analysis to guide artery-vein differentiation in OCTA. Four OCT intensity profile features, including i) ratio of vessel width to central reflex, ii) average of maximum profile brightness, iii) average of median profile intensity, and iv) optical density of vessel boundary intensity compared to background intensity, are used to classify artery-vein source nodes in OCT. A blood vessel tracking algorithm is then employed to automatically generate the OCT artery-vein map. Given the fact that OCT and OCTA are intrinsically reconstructed from the same raw spectrogram, the OCT artery-vein map is able to guide artery-vein differentiation in OCTA directly.

Multimodal Imaging of Nonneovascular Age-Related Macular Degeneration

Nonneovascular (dry) AMD is a retinal disease with potential for significant central visual impairment. The hallmarks of this disease are macular drusen, RPE alterations, and geographic atrophy (GA). Classification schemes for nonneovascular AMD have evolved over the years as major advances in retinal imaging have enabled a greater understanding of disease pathophysiology. The original classifications of nonneovascular AMD were based on color fundus photography (CFP), while more modern schemes rely on a multimodal imaging approach. Effective diagnosis and management of nonneovascular AMD requires a thorough understanding of its multimodal imaging features as detailed in this review. Future imaging modalities and imaging biomarkers that may aid in diagnosis and management are also discussed.

Assessing blood vessel perfusion and vital signs through retinal imaging photoplethysmography

One solution to the global challenge of increasing ocular disease is a cost-effective technique for rapid screening and assessment. Current ophthalmic imaging techniques, e.g. scanning and ocular blood flow systems, are expensive, complex to operate and utilize invasive contrast agents during assessment. The work presented here demonstrates a simple retinal imaging photoplethysmography (iPPG) system with the potential to provide screening, diagnosis, monitoring and assessment that is non-invasive, painless and radiationless. Time series of individual retinal blood vessel images, captured with an eye fundus camera, are processed using standard filtering, amplitude demodulation and principle component analysis (PCA) methods to determine the values of the heart rate (HR) and respiration rate (RR), which are in compliance with simultaneously obtained measurements using commercial pulse oximetry. It also seems possible that some information on the dynamic changes in oxygen saturation levels (SpO₂) in a retinal blood vessel may also be obtained. As a consequence, the retinal iPPG modality system demonstrates a potential avenue for rapid ophthalmic screening, and even early diagnosis, against ocular disease without the need for fluorescent or contrast agents.

In vivo photoacoustic imaging of chorioretinal oxygen gradients

Chorioretinal imaging has a crucial role for the patients with chorioretinal vascular diseases, such as neovascular age-related macular degeneration. Imaging oxygen gradients in the eye could better diagnose and treat ocular diseases. Here, we describe the use of photoacoustic ocular imaging (PAOI) in measuring chorioretinal oxygen saturation (CR  -  sO2) gradients in New Zealand white rabbits (n  =  5) with ocular ischemia. We observed good correlation (R2  =  0.98) between pulse oximetry and PAOI as a function of different oxygen percentages in inhaled air. We then used an established ocular ischemia model in which intraocular pressure is elevated to constrict ocular blood flow, and notice a positive correlation (R2  =  0.92) between the injected volume of phosphate buffered saline (PBS) and intraocular pressure (IOP) as well as a negative correlation (R2  =  0.98) between CR  -  sO2 and injected volume of PBS. The CR  -  sO2 was measured before (baseline), during (ischemia), and after the infusion (600-μL PBS). The ischemia-reperfusion model did not affect the measurement of the sO2 using a pulse oximeter on the animal's paw, but the chorioretinal PAOI signal showed a nearly sixfold decrease in CR  -  sO2 (n  =  5, p  =  0.00001). We also observe a sixfold decrease in CR  -  sO2 after significant elevation of IOP during ischemia, with an increase close to baseline during reperfusion. These data suggest that PAOI can detect changes in chorioretinal oxygenation and may be useful for application to imaging oxygen gradients in ocular disease.

Multi-modal and multi-vendor retina image registration

Multi-modal retinal image registration is often required to utilize the complementary information from different retinal imaging modalities. However, a robust and accurate registration is still a challenge due to the modality-varied resolution, contrast, and luminosity. In this paper, a two step registration method is proposed to address this problem. Descriptor matching on mean phase images is used to globally register images in the first step. Deformable registration based on modality independent neighbourhood descriptor (MIND) method is followed to locally refine the registration result in the second step. The proposed method is extensively evaluated on color fundus images and scanning laser ophthalmoscope (SLO) images. Both qualitative and quantitative tests demonstrate improved registration using the proposed method compared to the state-of-the-art. The proposed method produces significantly and substantially larger mean Dice coefficients compared to other methods (p<0.001). It may facilitate the measurement of corresponding features from different retinal images, which can aid in assessing certain retinal diseases.

Arteriovenous ratio and papilledema based hybrid decision support system for detection and grading of hypertensive retinopathy

BACKGROUND AND OBJECTIVES

Hypertensive Retinopathy (HR) is a retinal disease which happened due to consistent high blood pressure (hypertension). In this paper, an automated system is presented that detects the HR at various stages using arteriovenous ratio and papilledema signs through fundus retinal images.

METHODS

The proposed system consists of two modules i.e. vascular analysis for calculation of arteriovenous ratio and optic nerve head (ONH) region analysis for papilledema.  First module uses a set of hybrid features in Artery or Vein (A/V) classification using support vector machine (SVM) along with its radial basis function (RBF) kernel for arteriovenous ratio. In second module, proposed system performs analysis of ONH region for possible signs of papilledema. This stage utilizes different features along with SVM and RBF for classification of papilledema.

RESULTS

The first module of proposed method shows average accuracies of 95.10%, 95.64% and 98.09%for images of INSPIRE-AVR, VICAVR, and local dataset respectively. The second module of proposed method achieves average accuracies of 95.93% and 97.50% on STARE and local dataset respectively.

CONCLUSIONS

The system finally utilizes results from both modules to grade HR with good results. The presented system is a novel step towards automated detection and grading of HR disease and can be used as clinical decision support system.

Optical coherence tomography as a marker of neurodegeneration in patients with Wilson's disease

Wilson’s disease (WD) is an inherited autosomal recessive disorder that leads to pathological copper accumulation in different organs. Optical coherence tomography (OCT) is proposed as a marker of neurodegeneration in many neurological diseases. Thinning of the total retinal nerve fiber layer (RNFL) and macular thickness (Mth) examined by OCT was detected in patients with WD, especially those with brain magnetic resonance imaging changes. The aim of this study was to evaluate the relationship between OCT parameters and the progression of neurological signs measured by the Unified Wilson’s Disease Rating Scale (UWDRS) in patients with WD. Consecutive patients with WD admitted to the Department of Neurology underwent OCT to assess the thickness of the macula and total RNFL. Patients also had neurologic assessments according to the UWDRS part III. Patients were divided into two groups based on the presence (UWDRS+) and absence (UWDRS−) of neurological symptoms. Fifty-eight patients (34 females, 24 males) were enrolled. Mean duration of treatment was 9 years (standard deviation [SD], ±10.8). The mean UWDRS score at the time of study was 8.4 (range 1–52; SD ±13.9) points. Total RNFL as well as macula thickness were significantly decreased in the UWDRS+ group versus the UWDRS− group. A significant negative correlation was found between OCT parameters (RNFL and Mth measurements) and neurological impairment according the UWDRS scale. This study confirms that OCT may be a useful tool for measuring the degree of neurodegeneration in patients with WD, and may play role in monitoring disease progression.

Understanding Retinal Changes after Stroke

Stroke is the fifth leading cause of death and disability in the United States. According to World Heart Federation, every year, 15 million people suffer from stroke worldwide out of which nearly 6 million people die and another 5 million people are disabled. Out of many organs affected after stroke, one of them is eye. Majority of the stroke victims suffer vision loss due to stroke-induced retinal damage. However, stroke-induced retinal damage and microvascular changes have not been given paramount importance in understanding stroke pathophysiology and predicting its occurrence. Retinal imaging can be a very powerful tool to understand and predict stroke. This review will highlight the importance of retinal changes in predicting occurrence of stroke, major retinal changes, the relationship between retinal diseases and stroke and moreover, molecular mechanisms delineating the stroke induced-retinal changes and therapeutics associated with it.

Structural and component mining of nails using bioengineering techniques

The human nail is one of the challenging membranes for the scientists to target and to improve the clinical efficacy of ungual formulations. The understanding of nail physiology, impact of hydration on its properties and presence of trace elements in nails as biomarkers has been explored by various researchers in clinical studies. Despite the importance of biophysical techniques for the assessment of structure and physiology of nail, minimum literature analyses biophysical, biochemical and bioanalytical approaches. However, nowadays scientists in bioengineering field are keen in developing non-invasive, reliable and reproducible techniques for the assessment of different anatomical and functional parameters of nails for testing of ungual products.

Photoacoustic imaging of the eye: A mini review

The eye relies on the synergistic cooperation of many different ocular components, including the cornea, crystalline lens, photoreceptors, and retinal neurons, to precisely sense visual information. Complications with a single ocular component can degrade vision and sometimes cause blindness. Immediate treatment and long-term monitoring are paramount to alleviate symptoms, restore vision, and cure ocular diseases. However, successful treatment requires understanding ocular pathological mechanisms, precisely detecting and monitoring the diseases. The investigation and diagnosis of ocular diseases require advanced medical tools. In this mini review, we discuss non-invasive photoacoustic (PA) imaging as a potential research tool and medical screening device. In the research setting, PA imaging can provide valuable information on the disease progression. In the clinical setting, PA imaging can potentially aid in disease detection and treatment monitoring.

Microvascular endothelial function and severity of primary open angle glaucoma

PurposeThe role of microvascular endothelial dysfunction on severity of primary open angle glaucoma (POAG) was investigated in this study.Patients and methodsA prospective cohort study was conducted. One hundred and fourteen ethnically Malay patients (114 eyes) with POAG treated at the eye clinic of Hospital University Sains Malaysia between April 2012 and December 2014 were recruited. Patients aged between 40 and 80 years with two consecutive reliable and reproducible Humphrey visual field 24-2 analyses were selected. Patients who were diagnosed with any other type of glaucoma, previous glaucoma-filtering surgery, or other surgeries except uncomplicated cataract and pterygium surgery were excluded. Humphrey visual field analysis 24-2 was used to stratify the severity of glaucoma using Advanced Glaucoma Intervention Study (AGIS) score at the time of recruitment. Microvascular endothelial function was assessed using Laser Doppler fluximetry and iontophoresis. Iontophoresis process with acetylcholine (ACh) and sodium nitroprusside (SNP) was used to measure microvascular endothelium-dependent and -independent vasodilatation, respectively.ResultsBased on the AGIS score, 55 patients showed mild glaucoma, with 29 moderate and 30 severe. There was statistically significant difference in microvascular endothelial function (ACh% and ACh_max) between mild and moderate POAG cases (P=0.023) and between mild and severe POAG cases (P<0.001). There was negative correlation between microvascular endothelial function and severity of POAG (r=-0.457, P<0.001).ConclusionMicrovascular endothelial dysfunction may have a role in influencing the severity of POAG in Malay patients.

Attention attraction in an ophthalmic diagnostic device using sound-modulated fixation targets

This study relates to eye fixation systems with combined optical and audio systems. Many devices for eye diagnostics and some devices for eye therapeutics require the patient to fixate on a small target for a certain period of time, during which the eyes do not move and data from substructures of one or both eyes are acquired and analyzed. With young pediatric patients, a monotonously blinking target is not sufficient to retain attention steadily. We developed a method for modulating the intensity of a point fixation target using sounds appropriate to the child's age and preference. The method was realized as a subsystem of a Pediatric Vision Screener which employs retinal birefringence scanning for detection of central fixation. Twenty-one children, age 2-18, were studied. Modulation of the fixation target using sounds ensured the eye fixated on the target, and with appropriate choice of sounds, performed significantly better than a monotonously blinking target accompanied by a plain beep. The method was particularly effective with children of ages up to 10, after which its benefit disappeared. Typical applications of target modulation would be as supplemental subsystems in pediatric ophthalmic diagnostic devices, such as scanning laser ophthalmoscopes, optical coherence tomography units, retinal birefringence scanners, fundus cameras, and perimeters.

Optical coherence tomography and electrophysiology of retinal and visual pathways in Wilson's disease

We evaluated correlations between positive findings of changes on brain magnetic resonance imaging (MRI) and selected morphological and electrophysiological parameters of the retinal and visual systems in Wilson's disease. Fifty-eight Wilson's disease symptomatic patients were divided according to whether they displayed brain changes on MRI (positive, n = 39; negative, n = 19). All participants and healthy control group (n = 30), underwent retinal optical coherence tomography to assess the thickness of macula and the total retinal nerve fiber layer. Visual evoked potentials were measured and electroretinography was performed. Macular and retinal nerve fibers were thinner in participants with changes on MRI than in participants without changes. Electrophysiological parameters were markedly different in the MRI positive group compared with the negative group and 30 healthy controls; however, some abnormalities were evident in cases without visible brain pathology. Morphological and electrophysiological changes of retinal and visual pathways are associated with MRI visualized brain injury in Wilson's disease and may be useful for detecting the degree of neurodegeneration.

Detection of central fixation using short-time autoregressive spectral estimation during retinal birefringence scanning

The manuscript reports on the implementation of autoregressive spectral estimation aimed at improving the accuracy of detecting short-lasting events in signals acquired by a retinal birefringence scanning device that was reported earlier. A signal consisting of two frequency components is generated, where each frequency is a multiple of the scanning frequency. One frequency is produced during central fixation, while another one prevails during off-central fixation. These components may be of a very short duration, presenting a challenge for the FFT to identify them with sufficient time- and frequency resolution. Autoregressive spectral estimation using the Burg algorithm provided a satisfactory solution, capable of reliably differentiating between the two frequency components (96 and 192 Hz) on signal segments of duration as short as 5 ms. The device and the signal analysis methods were developed originally with the purpose of checking for eye alignment and strabismus - a major risk factor for amblyopia. The method enables the technology to work with less-cooperative patients, such as young children. Other medical and non-medical applications are possible.

Optimal timing of retinal scanning during dark adaptation, in the presence of fixation on a target: the role of pupil size dynamics

While validating our newly developed vision screener based on a double-pass retinal scanning system, we noticed that in all patients the signals from the retina were significantly higher when measurements were performed within a certain time interval referenced to the initial moment when the lights were dimmed and the test subject was asked to fixate on a target. This appeared to be most likely attributable to pupil size dynamics and triggered the present study, whose aim was to assess the pupillary “lights-off” response while fixating on a target in the presence of an accommodative effort. We found that pupil size increases in the first 60 to 70 s after turning off the room lights, and then it decreases toward the baseline in an exponential decay. Our results suggest that there is an optimal time window during which pupil size is expected to be maximal, that is during the second minute after dimming the room lights. During this time, window retinal diagnostic instruments based on double-pass measurement technology should deliver an optimal signal-to-noise ratio. We also propose a mathematical model that can be used to approximate the behavior of the normalized pupil size.