Ophthalmic Optical Coherence Tomography Imaging Research Articles

Artificial intelligence applications in ophthalmic optical coherence tomography: a 12-year bibliometric analysis

AIM: To explore the current application and research frontiers of global ophthalmic optical coherence tomography (OCT) imaging artificial intelligence (AI) research. METHODS: The citation data were downloaded from the Web of Science Core Collection database (WoSCC) to evaluate the articles in application of AI in ophthalmic OCT published from January 1, 2012 to December 31, 2023. This information was analyzed using CiteSpace 6.2.R2 Advanced software, and high-impact articles were analyzed. RESULTS: In general, 877 articles from 65 countries were studied and analyzed, of which 261 were published by the United States and 252 by China. The centrality of the United States is 0.33, the H index is 38, and the H index of two institutions in England reaches 20. Ophthalmology, computer science, and AI are the main disciplines involved. Hot keywords after 2018 include deep learning (DL), AI, macular degeneration, and automatic segmentation. CONCLUSION: The annual number of articles on AI applications in ophthalmic OCT has grown rapidly. The United States holds a prominent position. Institutions like the University of California System and the University of London are spearheading advancements. Initial researches centered on the automatic recognition and diagnosis of ocular diseases leveraging traditional machine learning (ML) technology and OCT images. Nowadays, the imaging process algorithm selection has shifted its focus towards DL. Concurrently, optical coherence tomography angiography (OCTA) and computer-aided diagnosis (CAD) have emerged as key areas of contemporary research.

Investigating Ganglion Cell Complex Thickness in Children with Chronic Heart Failure due to Dilated Cardiomyopathy.

Purpose: To assess ganglion cell complex (GCC) thickness in children with chronic heart failure (CHF) due to dilated cardiomyopathy (DCM) using optical coherence tomography (OCT). Methods: Sixty eyes of 30 patients with chronic heart failure (CHF) due to dilated cardiomyopathy (DCM) and 60 eyes of 30 age- and sex-matched healthy volunteers (control group) were enrolled. The mean age of the patients and controls was 9.9 ± 3.57 (range 5–17) years and 10.08 ± 3.41 (range 4–16) years, respectively. All patients underwent a complete ophthalmic assessment and OCT imaging using RTVue XR Avanti (Optovue). The following OCT-based parameters were analysed: average ganglion cell complex thickness (avgGCC), superior ganglion cell complex thickness (supGCC), inferior ganglion cell complex thickness (infGCC), global loss of volume (GLV) and focal loss of volume (FLV). Results: There were no significant differences in avgGCC (98.13 μm vs. 99.96 μm, p = 0.21), supGCC (97.17 μm vs. 99.29 μm, p = 0.13), infGCC (99.03 μm vs. 100.71 μm, p = 0.25), FVL (0.49% vs. 0.4%, p = 0.25) and GVL (2.1% vs. 1.3%, p = 0.09) between patients with chronic heart failure due to dilated cardiomyopathy and healthy children. There was no correlation between avgGCC, supGCC, infGCC, FLV, GLV and ocular biometry, refractive errors or age. There was no correlation between avgGCC, supGCC, infGCC, FLV, GLV and NT-proBNP or LVEF. There were no significant differences in the studied parameters between the sexes. There were no significant differences in the studied parameters between the left and right eye. Conclusion: Our study seems to be the first to analyse ganglion cell complex in paediatric patients with dilated cardiomyopathy. We have demonstrated no changes in the ganglion cell complex thickness parameters in children with chronic heart failure due dilated cardiomyopathy, as compared to their healthy peers.

Deep OCT image compression with convolutional neural networks.

We report an end-to-end image compression framework for retina optical coherence tomography (OCT) images based on convolutional neural networks (CNNs), which achieved an image size compression ratio as high as 80. Our compression scheme consists of three parts: data preprocessing, compression CNNs, and reconstruction CNNs. The preprocessing module was designed to reduce OCT speckle noise and segment out the region of interest. Skip connections with quantization were developed and added between the compression CNNs and the reconstruction CNNs to reserve the fine-structure information. Two networks were trained together by taking the semantic segmented images from the preprocessing module as input. To train the two networks sensitive to both low and high frequency information, we leveraged an objective function with two components: an adversarial discriminator to judge the high frequency information and a differentiable multi-scale structural similarity (MS-SSIM) penalty to evaluate the low frequency information. The proposed framework was trained and evaluated on ophthalmic OCT images with pathological information. The evaluation showed reconstructed images can still achieve above 99% similarity in terms of MS-SSIM when the compression ratio reached 40. Furthermore, the reconstructed images after 80-fold compression with the proposed framework even presented comparable quality with those of a compression ratio 20 from state-of-the-art methods. The test results showed that the proposed framework outperformed other methods in terms of both MS-SSIM and visualization, which was more obvious at higher compression ratios. Compression and reconstruction were fast and took only about 0.015 seconds per image. The results suggested a promising potential of deep neural networks on customized medical image compression, particularly valuable for effective image storage and tele-transfer.

Wide-field Ophthalmic Space-Division Multiplexing Optical Coherence Tomography.

High-speed ophthalmic optical coherence tomography systems are of interest because they allow rapid, motion-free, and wide-field retinal imaging. Space-division multiplexing optical coherence tomography (SDM-OCT) is a high-speed imaging technology which takes advantage of the long coherence length of microelectromechanical vertical cavity surface emitting laser (MEMs VCSEL) sources to multiplex multiple images along a single imaging depth. We demonstrate wide-field retinal OCT imaging, acquired at an effective A-scan rate of 800,000 A-scans/sec with volumetric images covering up to 12.5 mm × 7.4 mm on the retina acquired in less than 1 second. A clinical feasibility study was conducted to compare the ophthalmic SDM-OCT with commercial OCT systems, illustrating the high-speed capability of SDM-OCT in a clinical setting.

Deep learning based noise reduction method for automatic 3D segmentation of the anterior of lamina cribrosa in optical coherence tomography volumetric scans

A deep-learning (DL) based noise reduction algorithm, in combination with a vessel shadow compensation method and a three-dimensional (3D) segmentation technique, has been developed to achieve, to the authors best knowledge, the first automatic segmentation of the anterior surface of the lamina cribrosa (LC) in volumetric ophthalmic optical coherence tomography (OCT) scans. The present DL-based OCT noise reduction algorithm was trained without the need of noise-free ground truth images by utilizing the latest development in deep learning of de-noising from single noisy images, and was demonstrated to be able to cover more locations in the retina and disease cases of different types to achieve high robustness. Compared with the original single OCT images, a 6.6 dB improvement in peak signal-to-noise ratio and a 0.65 improvement in the structural similarity index were achieved. The vessel shadow compensation method analyzes the energy profile in each A-line and automatically compensates the pixel intensity of locations underneath the detected blood vessel. Combining the noise reduction algorithm and the shadow compensation and contrast enhancement technique, medical experts were able to identify the anterior surface of the LC in 98.3% of the OCT images. The 3D segmentation algorithm employs a two-round procedure based on gradients information and information from neighboring images. An accuracy of 90.6% was achieved in a validation study involving 180 individual B-scans from 36 subjects, compared to 64.4% in raw images. This imaging and analysis strategy enables the first automatic complete view of the anterior LC surface, to the authors best knowledge, which may have the potentials in new LC parameters development for glaucoma diagnosis and management.

Retinal changes in diabetic patients without diabetic retinopathy

The purpose of this study was to measure retinal vessel caliber and to examine early changes in macular thickness using optical coherence tomography (OCT). We evaluated to what extend vascular caliber and macular thickness differed between patients with type 2 diabetes mellitus without diabetic retinopathy compared with healthy individuals. 26 diabetic patients without diabetic retinopathy and 26 normal participants without any retinal and optic nerve diseases underwent ophthalmic examination, fundus photography, and OCT imaging. Temporal inferior retinal vessel diameters were measured using OCT. Also, we measured macular thickness in nine ETDRS subfields using Cirrus OCT. The mean age in the diabetic group was 61.5 years and in the control group, 55.5 years. Wider retinal arterioles and venules were found in patients with diabetes compared with healthy subjects (120 µm versus 96 µm, p<0.005 and 137 µm versus 120.5 µm, p value <0.001, respectively). In patients with type 2 diabetes mellitus, central macular thickness was significantly thinner than that of control eyes (243.5 µm versus 269.9 µm, p value <0.001). Our results support the hypothesis that the association between vascular damage and structural changes of the neuroretina is an early indicator of retinal impairment in patients with diabetes without diabetic retinopathy.

Speckle reduction and cartoon-texture decomposition of ophthalmic optical coherence tomography images by variational image decomposition

Optical coherence tomography (OCT) provides high resolution, cross-sectional tomographic images of internal structures of specimens, with wide application in ophthalmology. OCT images, however, are usually degraded by significant speckle noise. Development of successful speckle noise reduction algorithms while preserving fine details of image for ophthalmic OCT is particularly challenging, because of the difficulty in separating the noise and the information parts of a speckle pattern. In this paper we presented a new speckle reduction and cartoon-texture decomposition method of ophthalmic OCT images by using variational image decomposition. Experimental results show that the proposed method could decompose the ophthalmic OCT images into cartoon part, texture part and speckle noise part. Further it is also shown the advantage of variational image decomposition in the speckle reduction of ophthalmic OCT images.

Wavelet domain compounding for speckle reduction in optical coherence tomography

Visibility of optical coherence tomography (OCT) images can be severely degraded by speckle noise. A computationally efficient despeckling approach that strongly reduces the speckle noise is reported. It is based on discrete wavelet transform (DWT), but eliminates the conventional process of threshold estimation. By decomposing an image into different levels, a set of sub-band images are generated, where speckle noise is additive. These sub-band images can be compounded to suppress the additive speckle noise, as DWT coefficients resulting from speckle noise tend to be approximately decorrelated. The final despeckled image is reconstructed by taking the inverse wavelet transform of the new compounded sub-band images. The performance of speckle reduction and edge preservation is controlled by a single parameter: the level of wavelet decomposition. The proposed technique is applied to intravascular OCT imaging of porcine carotid arterial wall and ophthalmic OCT images. Results demonstrate the effectiveness of this technique for speckle noise reduction and simultaneous edge preservation. The presented method is fast and easy to implement and to improve the quality of OCT images.

Optic disc hamartomas in a family with heterozygous mutation in the VMD2 gene: Clinical, diagnostic and molecular genetic findings

AbstractPurpose To report a new form of juvenile‐onset Best disease associated with optic disc deposits and autosomal dominant inheritance in a three‐generation family.Methods Five affected individuals aged from 5 to 76 years were assessed. Detailed ophthalmic examination including refraction, color fundus photographs, fundus autofluorescence imaging (FAF), optical coherence tomography (OCT), ISCEV‐standard ffERG, mfERG, EOG, as well as ultrasound imaging were performed. Direct sequencing of all exons and intron‐exon junction of the BEST1 gene was conducted.Results The diagnosis of vitelliform maculopathy was confirmed by clinical and electrophysiological findings supported by genetic analysis. Ophthalmic imaging, OCT and ultrasound imaging showed in four out of five affected members deep subretinal deposits in the macula. Unexpected when the macula was affected, deposits located on the surface of the optic disc, showing the same echography and FAF characteristics, were observed. All five subjects had an abnormal EOG and a normal ffERG. The individual mfERG responses measured in the central, but also in the optic disc areas were reduced, the latencies were delayed. Sequencing revealed a heterozygous c.[670 C&gt;A], coding for a L224M mutation in BEST1 in all affected subjects.Conclusion Through comparison of the morphological tests: OCT, the FAF ultrasound images and the functional measure: mfERG, we deduct that the observed optic disc deposits are part of generalized disease and have the same origin as the subretinal vitelliform macular and multifocal deposits. This suggests that the described optic disc lesions are vitelliform‐like optic disc hamartomas.

Intracoronary imaging using attenuation-compensated optical coherence tomography allows better visualisation of coronary artery diseases

To allow an accurate diagnosis of coronary artery diseases by enhancing optical coherence tomography (OCT) images of atheromatous plaques using a novel automated attenuation compensation technique. One of the major drawbacks of coronary OCT imaging is the rapid attenuation of the OCT signal, limiting penetration in tissue to only few millimetres. Visualisation of deeper anatomy is however critical for accurate assessment of plaque burden in-vivo. A compensation algorithm, previously developed to correct for light attenuation in soft tissues and to enhance contrast in ophthalmic OCT images, was applied to intracoronary plaque imaging using spectral-domain OCT. Application of the compensation algorithm significantly increased tissue contrast in the vessel wall and atherosclerotic plaque boundaries. Contrast enhancement allows a better differentiation of plaque morphology, which is particularly important for the identification of lipid rich fibro atheromatous plaques and to guide decision on treatment strategy. The analysis of arterial vessel structure clinically captured with OCT is improved when used in conjunction with automated attenuation compensation. This approach may improve the OCT-based interpretation of coronary plaque morphology in clinical practice.

Ultrahigh speed Spectral / Fourier domain OCT ophthalmic imaging at 70,000 to 312,500 axial scans per second

We demonstrate ultrahigh speed spectral / Fourier domain optical coherence tomography (OCT) using an ultrahigh speed CMOS line scan camera at rates of 70,000 - 312,500 axial scans per second. Several design configurations are characterized to illustrate trade-offs between acquisition speed, resolution, imaging range, sensitivity and sensitivity roll-off performance. Ultrahigh resolution OCT with 2.5 - 3.0 micron axial image resolution is demonstrated at approximately 100,000 axial scans per second. A high resolution spectrometer design improves sensitivity roll-off and imaging range performance, trading off imaging speed to 70,000 axial scans per second. Ultrahigh speed imaging at >300,000 axial scans per second with standard image resolution is also demonstrated. Ophthalmic OCT imaging of the normal human retina is investigated. The high acquisition speeds enable dense raster scanning to acquire densely sampled volumetric three dimensional OCT (3D-OCT) data sets of the macula and optic disc with minimal motion artifacts. Imaging with approximately 8 - 9 micron axial resolution at 250,000 axial scans per second, a 512 x 512 x 400 voxel volumetric 3D-OCT data set can be acquired in only approximately 1.3 seconds. Orthogonal registration scans are used to register OCT raster scans and remove residual axial eye motion, resulting in 3D-OCT data sets which preserve retinal topography. Rapid repetitive imaging over small volumes can visualize small retinal features without motion induced distortions and enables volume registration to remove eye motion. Cone photoreceptors in some regions of the retina can be visualized without adaptive optics or active eye tracking. Rapid repetitive imaging of 3D volumes also provides dynamic volumetric information (4D-OCT) which is shown to enhance visualization of retinal capillaries and should enable functional imaging. Improvements in the speed and performance of 3D-OCT volumetric imaging promise to enable earlier diagnosis and improved monitoring of disease progression and response to therapy in ophthalmology, as well as have a wide range of research and clinical applications in other areas.

State-of-the-art retinal optical coherence tomography

OCT functions as a type of optical biopsy, providing information on retinal pathology in situ and in real time, with resolutions approaching that of excisional biopsy and histopathology. The development of ultrabroad-bandwidth and tunable light sources, as well as high-speed Fourier detection techniques, has enabled a significant improvement in ophthalmic optical coherence tomography (OCT) imaging performance. Three-dimensional, ultrahigh-resolution OCT (UHR OCT) can provide information on intraretinal morphology that is not available from any other non-invasive diagnostic. High-speed imaging facilitates the acquisition of three-dimensional data sets (3D-OCT), thus enabling volumetric rendering and the generation of OCT fundus images that precisely and reproducibly register OCT images to fundus features. The development of broadband light sources emitting at new wavelengths, e.g., approximately 1050 nm, has enabled not only 3D-OCT imaging with enhanced choroidal visualization, but also reduced scattering losses and improved OCT performance in cataract patients. Adaptive optics using high-stroke, deformable mirror technology to correct higher order aberrations in the human eye, in combination with specially designed optics to compensate chromatic aberration along with three-dimensional UHR OCT, has recently enabled in vivo cellular-resolution retinal imaging. In addition, extensions of OCT have been developed to enhance image contrast and to enable non-invasive depth-resolved functional imaging of the retina, thus providing blood flow, spectroscopic, polarization-sensitive and physiological information. Functional OCT promises to enable the differentiation of retinal pathologies via localized, functional retinal response or metabolic properties. These advances promise to have a powerful impact on fundamental as well as clinical studies.

Speckle reduction in optical coherence tomography images by use of a spatially adaptive wavelet filter

A spatially adaptive two-dimensional wavelet filter is used to reduce speckle noise in time-domain and Fourier-domain optical coherence tomography (OCT) images. Edges can be separated from discontinuities that are due to noise, and noise power can be attenuated in the wavelet domain without significantly compromising image sharpness. A single parameter controls the degree of noise reduction. When this filter is applied to ophthalmic OCT images, signal-to-noise ratio improvements of >7 dB are attained, with a sharpness reduction of <3%.

Enhanced visualization of choroidal vessels using ultrahigh resolution ophthalmic OCT at 1050 nm

In this article the ability of ultrahigh resolution ophthalmic optical coherence tomography (OCT) to image small choroidal blood vessels below the highly reflective and absorbing retinal pigment epithelium is demonstrated for the first time. A new light source (lambdac= 1050 nm, Deltalambda = 165 nm, Pout= 10 mW), based on a photonic crystal fiber pumped by a compact, self-starting Ti:Al2O3 laser has therefore been developed. Ex-vivo ultrahigh resolution OCT images of freshly excised pig retinas acquired with this light source demonstrate enhanced penetration into the choroid and better visualization of choroidal vessels as compared to tomograms acquired with a state-of-the art Ti:Al2O3 laser (Femtolasers Compact Pro, lc= 780 nm, Deltalambda= 160 nm, Pout= 400 mW), normally used in clinical studies for in vivo ultrahigh resolution ophthalmic OCT imaging. These results were also compared with retinal tomograms acquired with a novel, spectrally broadened fiber laser (MenloSystems, lambdac= 1350 nm, Deltalambda= 470 nm, Pout = 4 mW) permitting even greater penetration in the choroid. Due to high water absorption at longer wavelengths retinal OCT imaging at ~1300 nm may find applications in animal ophthalmic studies. Detection and follow-up of choroidal neovascularization improves early diagnosis of many retinal pathologies, e.g. age-related macular degeneration or diabetic retinopathy and can aid development of novel therapy approaches.

Histologic correlation of pig retina radial stratification with ultrahigh-resolution optical coherence tomography.

To compare ultrahigh-resolution optical coherence tomography (OCT) cross-sectional images of the pig retina with histology, to evaluate the potential of ultrahigh-resolution OCT for enhanced visualization of intra- and subretinal structures. Ultrahigh-resolution OCT images were acquired with 1.4- micro m axial x 3- micro m transverse resolution from in vitro posterior eyecup preparations of the domestic pig. Frozen sections were obtained in precise alignment with OCT tomograms, by using major blood vessels as orientation markers and were counterstained with cresyl violet or unstained and examined by differential interference contrast microscopy. Micrographs from histologic sections were linearly scaled to correct for tissue shrinkage and compared with OCT tomograms. In the proximal retina, ultrahigh-resolution OCT signal bands directly corresponded to the main retinal layers. For the wavelength region used ( approximately 800 nm), axodendritic layers (nerve fiber layer, inner and outer plexiform layers) were more reflective than cell body layers (ganglion cell layer, inner nuclear layer, outer nuclear layer). In the distal retina, substructures of the photoreceptor layer such as the interface between inner and outer segments were visualized, and the retinal pigment epithelium, the choriocapillaris, and superficial choroid layers were resolved. In addition, the time sequence of a retinal detachment event was monitored by ultrahigh-resolution OCT. In vitro ophthalmic ultrahigh-resolution OCT imaging reveals retinal morphology with unprecedented detail. The specific assignment of OCT signal patterns to retinal substructures provides a basis for improved interpretation of in vivo ophthalmic OCT tomograms of high clinical relevance.