In vivo assessment of inflammatory bowel disease in rats with ultrahigh-resolution colonoscopic OCT

Abstract

A technology capable of high-resolution, label-free imaging of subtle pathology in vivo during colonoscopy is imperative for the early detection of disease and the performance of accurate biopsies. While colonoscopic OCT has been developed to visualize colonic microstructures beyond the mucosal surface, its clinical potential remains limited by sub-optimal resolution (∼6.5 µm in tissue), inadequate imaging contrast, and a lack of high-resolution OCT criteria for lesion detection. In this study, we developed an ultrahigh-resolution (UHR) colonoscopic OCT and evaluated its ability to volumetrically visualize and identify the pathological features of inflammatory bowel disease (IBD) in a rat model. Owing to its improved resolution (∼1.7 µm in tissue) and enhanced contrast, UHR colonoscopic OCT can accurately delineate fine colonic microstructures and identify the pathophysiological characteristics of IBD in vivo. By using a quantitative optical attenuation map, UHR colonoscopic OCT is able to differentiate diseased tissue (such as crypt distortion and microabscess) from normal colonic mucosa over a large field of view in vivo. Our results suggest the clinical potential of UHR colonoscopic OCT for in vivo assessment of IBD pathology.

LiDeOCTNet: A Lightweight OCT-aware Framework for Segmentation of Irregularly Layered Tissue Structures

“…Three different OCT datasets were used in our experiment: AROI retinal dataset [8], airway endoscopy dataset [5], and colon endoscopy dataset [6]. The range of data sources provided a comprehensive domain coverage, enabling us to verify the efficacy of LiDeOCTNet.…”

“…Accurate segmentation and quantification of the layered structures in OCT images is essential for precise diagnosis and treatment of eye diseases [1], [3]. Moreover, OCT has been increasingly used for endoscopic imaging in luminal organs [2], [4]- [6]. Delineation of tissue layers in endoscopic OCT images also helps identify crucial histopathological biomarkers for diagnosis, including smooth muscle layer in airway wall [5] and mucosa layer of colon [6].…”

“…Moreover, OCT has been increasingly used for endoscopic imaging in luminal organs [2], [4]- [6]. Delineation of tissue layers in endoscopic OCT images also helps identify crucial histopathological biomarkers for diagnosis, including smooth muscle layer in airway wall [5] and mucosa layer of colon [6]. However, special training is required for manual annotation and analysis of OCT images, which is timeconsuming and subjective [7], [8].…”

“…Semantic segmentation aims to assign class labels to pixels from input images, resulting in a densely structured prediction task [6], [9]- [11]. Benefiting from advances in convolutional neural networks (CNNs) [12], accurate semantic segmentation has become a fundamental perception component in many computer vision applications, such as image-guided intervention [13], medical robotic navigation [14], and medical imaging diagnosis [15], [16].…”

Visualizing enteric nervous system activity through dye-free dynamic full-field optical coherence tomography

“…This imaging strategy is most widely used to diagnose disease in the human eye 9 , but when combined with a fiber optic probe, the OCT could acquire images in the cardiovascular, respiratory, or digestive system 10 . In the gut, endoscopic OCT has been typically used for the diagnosis of Barrett's esophagus 11 and preliminary data was obtained for other indications like inflammatory bowel diseases and colorectal cancer 12,13 . In vivo endomicroscopic OCT with a typical resolution of tens of microns 14 has a too low resolution to visualize the ENS, but higher resolution can be obtained with two implementations of this technique: FFOCT 15,16 and high-resolution SD-OCT [17][18][19] .…”