Hyperspectral imaging techniques measure spatial and spectral information, which enables discrimination between healthy tissue and lesions. Hyperspectral imaging applications in endoscopy for clinical diagnostic applications are still limited due to image distortion challenges that arise when working with a flexible endoscope, for example in the gastrointestinal tract. Here, we developed a hyperspectral endoscopy (HySE) system by exploiting a line-scanning spectrograph to measure spectral information at high spectral resolution and combining it with a CMOS camera that records wide-field images for hyperspectral image reconstruction. Moreover, we developed an image normalisation method using near-infrared light to obtain accurate hyperspectral signals by correcting uneven illumination conditions during clinical endoscopy. Our next step is to apply HySE to patients to identify abnormal features in the gastrointestinal tract in vivo.
Barrett’s oesophagus is an acquired condition that predisposes patients to the development of oesophageal adenocarcinoma through intermediate stages of dysplasia. Early detection of dysplasia allows curative endoscopic therapy, but current standard of care surveillance achieves only around 40% sensitivity for dysplasia.
Multispectral imaging (MSI) allows simultaneous collection of morphological (spatial) and biochemical (spectral) information from tissue, which can help to more effectively delineate disease. This motivated the design and construction of a compact, clinically translatable multispectral endoscope (MuSE) that can be introduced through the accessory channel of a standard gastroscope to collect multispectral images in vivo.
MuSE is based around a spectrally resolved detector array (SRDA) with 9 spectral filters (8 narrow bands; average FWHM 30nm, center wavelengths 553, 587, 629, 665, 714, 749, 791, 829nm; 1 broadband; 500–850nm). The SRDA was coupled to a clinically approved 10,000-fibre endoscope (PolyScope) for imaging. Illumination was provided by sequentially by a broadband (400–750nm) and narrowband (400–480nm) source for reflectance and autofluorescence imaging respectively.
Subjects due to undergo clinically indicated endoscopy with a previous diagnosis of dysplasia or early adenocarcinoma were enrolled for experimental imaging using MuSE in a pilot clinical study. Patients with clearly visible lesions were selected to allow co-registration of the image cubes with pathology of biopsies. Here, we present the results from these first-in-human tests of MuSE, including evaluation of the image quality and classification potential of the multispectral image cubes.
Emerging clinical interest in combining standard white light endoscopy with targeted near-infrared (NIR) fluorescent contrast agents for improved early cancer detection has created demand for multimodal imaging endoscopes. We used two spectrally resolving detector arrays (SRDAs) to realize a bimodal endoscope capable of simultaneous reflectance-based imaging in the visible spectral region and multiplexed fluorescence-based imaging in the NIR. The visible SRDA was composed of 16 spectral bands, with peak wavelengths in the range of 463 to 648 nm and full-width at half-maximum (FWHM) between 9 and 26 nm. The NIR SRDA was composed of 25 spectral bands, with peak wavelengths in the range 659 to 891 nm and FWHM 7 to 15 nm. The spectral endoscope design was based on a “babyscope” model using a commercially available imaging fiber bundle. We developed a spectral transmission model to select optical components and provide reference endmembers for linear spectral unmixing of the recorded image data. The technical characterization of the spectral endoscope is presented, including evaluation of the angular field-of-view, barrel distortion, spatial resolution and spectral fidelity, which showed encouraging performance. An agarose phantom containing oxygenated and deoxygenated blood with three fluorescent dyes was then imaged. After spectral unmixing, the different chemical components of the phantom could be successfully identified via majority decision with high signal-to-background ratio (>3). Imaging performance was further assessed in an ex vivo porcine esophagus model. Our preliminary imaging results demonstrate the capability to simultaneously resolve multiple biological components using a compact spectral endoscopy system.
A multispectral endoscope capable of simultaneous reflectance and fluorescence imaging was developed based on spectrally resolved detector arrays. The endoscope can simultaneous image and unmix oxy/deoxygenated blood and two fluorescent dyes.
Hyperspectral imaging (HSI) systems collect both morphological and chemical characteristics from a sample by simultaneously acquiring spatial and spectral information. HSI has potential to advance cancer diagnostics by characterizing reflectance and fluorescence properties of a tissue, as well as extracting microstructural in- formation, all of which are altered through the development of a tumor. Illumination uniformity is a critical pre-condition for extracting quantitative data from an HSI system. Spatial, angular, or spectral non-uniformity can cause glare, specular reflection and unwanted shading, which negatively impact statistical analysis techniques used to extract abundance of different chemical species. This is further exacerbated when imaging three-dimensional structures, such as tumors, whose appearance can cast shadows and form other occlusions. Furthermore, as HSI can be used simultaneously for white light and fluorescence imaging, a flexible system, which multiplexes narrowband and broadband illumination is necessary to fully utilize the capabilities of a biomedical HSI system. To address these challenges, we modeled illumination systems frequently used in wide-field biological imaging with the software LightTools and FRED. Each system is characterized for spectral, spatial, and angular uniformity, as well as total efficiency. While all three systems provide high spatial and spectral uniformity, the highest angular uniformity is achieved using a diffuse scattering dome, yielding a contrast of 0.503 and average deviation of 0.303 with a 3.91% model error. Nonetheless, results suggest that conventional systems may not be suitable for low-light-level applications, where tailoring illumination to match spatial and spectral requirements may be the best approach to maximize the performance.
Multispectral imaging has the potential to improve sensitivity and specificity in biomedical imaging through simultaneous acquisition of both morphological (spatial) and chemical (spectral) information. Performing multispectral imaging in real time with spectrally resolved detector arrays (SRDAs), for example in endoscopy or intraoperative imaging, requires a direct trade off between spatial and spectral resolution. We sought to quantitatively assess the impact of spectral band selection on contrast agent detection in fluorescence endoscopic imaging. As a proof of concept, we measured the ‘ground truth’ spectra from a dilution series of a single near-infrared fluorescent contrast agent using a spectrometer incorporated into the detection path of our endoscope. We then modeled the influence of an SRDA on these spectra and calculated the theoretical endmembers associated with reflectance and fluorescence signals from the pure contrast agent. To test the accuracy of our model, we incorporated into the same endoscope an off-the-shelf SRDA with a 3x3 filter deposition pattern of 9 spectral bands. After spectral unmixing using the modeled endmembers, the amplitude of the fluorescence recorded with the SRDA compared favorably with the amplitude of fluorescence derived from the ‘ground truth’ spectra recorded with the spectrometer. In the future, this approach could be used to minimize the number of spectral bands required in a given imaging system and hence maximize the spatial resolution of the multispectral camera.
Hyperspectral imaging (HSI) can combine morphological and molecular information, yielding potential for real-time and high throughput multiplexed fluorescent contrast agent imaging. Multiplexed readout from targets, such as cell surface receptors overexpressed in cancer cells, could improve both sensitivity and specificity of tumor identification. There remains, however, a need for compact and cost effective implementations of the technology. We have implemented a low-cost wide-field multiplexed fluorescence imaging system, which combines LED excitation at 590, 655 and 740 nm with a compact commercial solid state HSI system operating in the range 600 - 1000 nm. A key challenge for using reflectance-based HSI is the separation of contrast agent fluorescence from the reflectance of the excitation light. Here, we illustrate how it is possible to address this challenge in software, using two offline reflectance removal methods, prior to least-squares spectral unmixing. We made a quantitative comparison of the methods using data acquired from dilutions of contrast agents prepared in well-plates. We then established the capability of our HSI system for non-invasive in vivo fluorescence imaging in small animals using the optimal reflectance removal method. The HSI presented here enables quantitative unmixing of at least four fluorescent contrast agents (Alexa Fluor 610, 647, 700 and 750) simultaneously in living mice. A successful unmixing of the four fluorescent contrast agents was possible both using the pure contrast agents and with mixtures. The system could in principle also be applied to imaging of ex vivo tissue or intraoperative imaging in a clinical setting. These data suggest a promising approach for developing clinical applications of HSI based on multiplexed fluorescence contrast agent imaging.
Hyperspectral imaging (HSI) systems have the potential to combine morphological and spectral information to provide detailed and high sensitivity readouts in biological and medical applications. As HSI enables simultaneous detection in several spectral bands, the technology has significant potential for use in real-time multiplexed contrast agent studies. Examples include tumor detection in intraoperative and endoscopic imaging as well as histopathology. A multiplexed readout from multiple disease targets, such as cell surface receptors overexpressed in cancer cells, could improve both sensitivity and specificity of tumor identification. Here, we evaluate a commercial, compact, near-infrared HSI sensor that has the potential to enable low cost, video rate HSI for multiplexed fluorescent contrast agent studies in biomedical applications. The hyperspectral imager, based on a monolithically integrated Fabry-Perot etalon, has 70 spectral bands between 600-900 nm, making it ideal for this application. Initial calibration of the imager was performed to determine wavelength band response, quantum efficiency and the effect of F-number on the spectral response. A platform for wide-field fluorescence imaging in reflectance using fluorophore specific LED excitation was then developed. The applicability of the imaging platform for simultaneous readout of multiple fluorophore signals was demonstrated using a dilution series of Alexa Fluor 594 and Alexa Fluor 647, showing that nanomolar fluorophore concentrations can be detected. Our results show that the HSI system can clearly resolve the emission spectra of the two fluorophores in mixtures of concentrations across several orders of magnitude, indicating a high dynamic range performance. We therefore conclude that the HSI sensor tested here is suitable for detecting fluorescence in biomedical imaging applications.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.