In-vivo Optical Coherence Tomography Imaging Research Articles

SSN2V: unsupervised OCT denoising using speckle split

Denoising in optical coherence tomography (OCT) is important to compensate the low signal-to-noise ratio originating from laser speckle. In recent years learning algorithms have been established as the most powerful denoising approach. Especially unsupervised denoising is an interesting topic since it is not possible to acquire noise free scans with OCT. However, speckle in in-vivo OCT images contains not only noise but also information about blood flow. Existing OCT denoising algorithms treat all speckle equally and do not distinguish between the noise component and the flow information component of speckle. Consequently they either tend to either remove all speckle or denoise insufficiently. Unsupervised denoising methods tend to remove all speckle but create results that have a blurry impression which is not desired in a clinical application. To this end we propose the concept, that an OCT denoising method should, besides reducing uninformative noise, additionally preserve the flow-related speckle information. In this work, we present a fully unsupervised algorithm for single-frame OCT denoising (SSN2V) that fulfills these goals by incorporating known operators into our network. This additional constraint greatly improves the denoising capability compared to a network without. Quantitative and qualitative results show that the proposed method can effectively reduce the speckle noise in OCT B-scans of the human retina while maintaining a sharp impression outperforming the compared methods.

Cervical optical coherence tomography image classification based on contrastive self-supervised texture learning.

Cervical cancer (CC) seriously affects the health of the female reproductive system. Optical coherence tomography (OCT) emerged as a noninvasive, high-resolution imaging technology for cervical disease detection. However, OCT image annotation is knowledge-intensive and time-consuming, which impedes the training process of deep-learning-based classification models. This study aims to develop a computer-aided diagnosis (CADx) approach to classifying in-vivo cervical OCT images based on self-supervised learning. In addition to high-level semantic features extracted by a convolutional neural network (CNN), the proposed CADx approach designs a contrastive texture learning strategy to leverage unlabeled cervical OCT images' texture features. We conducted 10-fold cross-validation on the OCT image dataset from a multicenter clinical study on 733 patients from China. In a binary classification task for detecting high-risk diseases, including high-grade squamous intraepithelial lesion and CC, our method achieved an area-under-the-curve value of 0.9798±0.0157 with a sensitivity of 91.17%±4.99% and a specificity of 93.96%±4.72% for OCT image patches; also, it outperformed two out of four medical experts on the test set. Furthermore, our method achieved 91.53% sensitivity and 97.37% specificity on an external validation dataset containing 287 three-dimensional OCT volumes from 118 Chinese patients in a new hospital using a cross-shaped threshold voting strategy. The proposed contrastive-learning-based CADx method outperformed the end-to-end CNN models and provided better interpretability based on texture features, which holds great potential to be used in the clinical protocol of "see-and-treat."

Optical coherence tomography for identification and quantification of human airway wall layers.

BackgroundHigh-resolution computed tomography has limitations in the assessment of airway wall layers and related remodeling in obstructive lung diseases. Near infrared-based optical coherence tomography (OCT) is a novel imaging technique that combined with bronchoscopy generates highly detailed images of the airway wall. The aim of this study is to identify and quantify human airway wall layers both ex-vivo and in-vivo by OCT and correlate these to histology.MethodsPatients with lung cancer, prior to lobectomy, underwent bronchoscopy including in-vivo OCT imaging. Ex-vivo OCT imaging was performed in the resected lung lobe after needle insertion for matching with histology. Airway wall layer perimeters and their corresponding areas were assessed by two independent observers. Airway wall layer areas (total wall area, mucosal layer area and submucosal muscular layer area) were calculated.Results13 airways of 5 patients were imaged by OCT. Histology was matched with 51 ex-vivo OCT images and 39 in-vivo OCT images. A significant correlation was found between ex-vivo OCT imaging and histology, in-vivo OCT imaging and histology and ex-vivo OCT imaging and in-vivo OCT imaging for all measurements (p < 0.0001 all comparisons). A minimal bias was seen in Bland-Altman analysis. High inter-observer reproducibility with intra-class correlation coefficients all above 0.90 were detected.ConclusionsOCT is an accurate and reproducible imaging technique for identification and quantification of airway wall layers and can be considered as a promising minimal-invasive imaging technique to identify and quantify airway remodeling in obstructive lung diseases.

In vivo deep tissue imaging using wavefront shaping optical coherence tomography.

Multiple light scattering in tissue limits the penetration of optical coherence tomography (OCT) imaging. Here, we present in vivo OCT imaging of a live mouse using wavefront shaping (WS) to enhance the penetration depth. A digital micromirror device was used in a spectral-domain OCT system for complex WS of an incident beam which resulted in the optimal delivery of light energy into deep tissue. Ex vivo imaging of chicken breasts and mouse ear tissues showed enhancements in the strength of the image signals and the penetration depth, and in vivo imaging of the tail of a live mouse provided a multilayered structure inside the tissue.

Peri-strut low-intensity areas in optical coherence tomography correlate with peri-strut inflammation and neointimal proliferation

Peri-strut low-intensity area (PLI) is a common imaging finding during the evaluation of in-stent neointima using optical coherence tomography (OCT). We aimed to determine the biological significance of PLI by comparing in-vivo OCT images with the corresponding histological sections obtained from the familial hypercholesterolemic swine model of coronary stenosis. A total of 26 coronary vessels of nine familial hypercholesterolemic swine were injured with 30% balloon overstretch and then immediately followed by everolimus eluting or bare metal stent placement at 20% overstretch. At 30 days, all stented vessels were subjected to in-vivo OCT analysis and were harvested for histological evaluation. For OCT analysis, stent cross-sections (three per stent) were categorized into presence (PLI+) or absence (PLI-) of PLI. In histology, inflammation and fibrin deposition were scored semiquantitatively from 0 (none) to 3 (severe). PLI was found in 64.9% of stent sections. Peri-strut inflammation was more frequently observed in OCT sections PLI (+) compared with PLI (-) (56.0 vs. 7.4%, P=0.01). In contrast, peri-strut fibrin deposits was similar in both groups (PLI+=58.0% vs. PLI-=59.3%, P=0.94). Histological neointimal thickness was significantly higher in PLI (+) sections (mean±SE: 0.68±0.06 vs. 0.34±0.02 mm; P<0.01), yielding a higher percent area stenosis compared with PLI (-) (mean±SE: 59.0±4.4 vs. 34.1±2.2%, P<0.01). The PLI diagnostic sensitivity and specificity for inflammation were 80 and 76.1%, respectively (>56% PLI, area under the curve=0.86, P<0.01), whereas for fibrin deposition, the sensitivity and specificity were 42.2 and 76.1%, respectively (area under the curve=0.56, P=NS). Area under the receiver operating characteristic curve was significantly higher for identifying inflammation than fibrin (0.86 vs. 0.56, P<0.01). The severity of PLI correlated with the neointimal thickness when assessed by OCT (R=0.79, P<0.001). The presence of PLI in OCT correlates with neointimal thickness and appears to have a diagnostic value in the recognition of peri-strut inflammation, therefore possibly serving as a surrogate for in-vivo assessment of stent efficacy.

Linearly Wavenumber-Swept Active Mode Locking Short-Cavity Fiber Laser for In-Vivo OCT Imaging

We demonstrate a highly linearly wavenumber-swept active mode locking (AML) fiber laser in the 1.3 μm region for in-vivo imaging in optical coherence tomography (OCT) without wavenumber space resampling. In this all-electric AML wavenumber-swept mechanism, the conventional wavelength selection filter is eliminated, and instead a suitable programmed electric modulation signal is applied directly to the gain medium. For a high sweep rate (up to 1 MHz) along the wavenumber, the fiber cavity structure is made as short as possible (0.88 m in air). A 15 ps/nm chirped fiber Bragg grating and a circulator are used for a shorter ring cavity configuration. A linewidth of 0.1 nm and tuning range of 42 nm are obtained under the mode-locking condition. Various types of wavenumber (or wavelength) tunings can be implemented because of the filterless cavity configuration. Therefore, we successfully demonstrate a linearly wavenumber-swept AML fiber laser with 26.5 mW of output power for obtaining in-vivo OCT images at a sweep rate of 100 kHz.

Gradient-index lens rod based probe for office-based optical coherence tomography of the human larynx

Optical coherence tomography (OCT) is an evolving noninvasive imaging modality that has been used to image the human larynx during surgical endoscopy. The design of a long gradient-index lens-based probe capable of capturing images of the human larynx by use of spectral domain OCT during a typical office-based laryngoscopy examination is presented. An optical-ballast-based 4f optical relay system is proposed to realize variable working distance with a constant optical delay. In-vivo OCT imaging of the human larynx is demonstrated. Office-based OCT is a promising imaging modality for early laryngeal cancer diagnosis.