Infrared micro-spectroscopy is a useful tool for basic research and biomedical applications. Conventional microspectroscopic imaging apparatuses use thermal sources for sample illumination, which have low brightness, low optical spectral intensity, and high noise. This work evaluates the system engineering advantages of using mid-infrared semiconductor lasers that offer orders-of magnitude higher brightness, spectral intensity, and lower noise. A laser-based microscopic spectral imaging system with focal plane array detectors demonstrated a high signal-to-noise ratio (>20 dB) at video frame rate for a large illuminated area. Microscopic spectral imaging with fixed-wavelength and tunable lasers of 4.6, 6, and 9.3-μm wavelength was applied to a number of representative samples that consist of biological tissues (plant and animal) and solid material (a stack of laminated polymers). Transmission spectral images with ~30-dB dynamic range were obtained with clear evidence of spectral features for different samples. The potential of more advanced systems with a wide coverage of spectral bands is discussed.
We have constructed an adaptive digitally tuned light source in the form of a de-dispersive imaging spectrograph in both the visible and near infrared spectral regions capable of illuminating a sample with appropriate energy weighted spectral bands or spatio-spectral bands that relate only to the constituents of interest to the investigator. The energy from each of the spectral resolution elements can be digitally modulated to provide a tuned weighted spectral output. A tuned light source based on this technology was adapted for use in a conventional imaging microscope system to enable direct measure of spatio-spectral features of interest. Some imagery resulting from preliminary tests on colon tissue biopsies are presented.
Placing a spatial light modulator, such as the Texas Instruments Digital Micromirror Device (DMD), in the light path of a microscope enables a variety of novel applications. One application enables reflectance in vivo confocal imaging of cells and tissue structure through a fiber-optic image guide. While multi-wavelength reflectance confocal microendoscopy with optical sectioning is a requirement for a clinically useful device, some form of axial scanning is also necessary. This is readily achieved using a multi-element lens system with some form of mechanical translation, however, this generally results in large probes and high cost. These limitations can be overcome using a two-element GRIN lens system in which the traditionally undesirable chromatic aberration of such a system can be exploited to allow for color-encoded optical sectioning. In our system a wavelength encoding range of 200 nm permits a sectioning range of 40 μm from the tip of the probe into the tissue.
Molecular medicine now requires molecular pathology. While fluorescence has traditionally been used for high-resolution multiplexed molecular imaging, clinical practitioners prefer non-fluorescent multicolor methods. However, in brightfield, typically, only one color is used at a time, which precludes assessment of co-expression on a cell-by-cell basis. Similar constraints apply to brightfield in-situ hybridization techniques. Double- and triple-staining procedures are rarely performed in non-research settings not only because the wet chemistry can be difficult, but also because it can be challenging or impossible to determine visually where and to what extent different chromogens may physically overlap. Spectral imaging can be useful in this context. Two methods of acquiring spectral images are described, along with their application to multicolor immunohistochemistry and transmission in-situ hybridization (TRISH): 1) liquid crystal tunable filters; and 2) a novel, spectrally agile light source. This source emits white light of any desired color temperature, or single 10-nm wavelength bands in the range 420 to 700 nm, or any combination of wavelengths with individual intensity control. Both methods are allied with a grayscale camera and appropriate algorithms to analyze multicolor samples of clinical significance. Spectrally unmixed images clearly separate signals linked to different chromogens, even with spectral and spatial overlap. Intriguing challenges in matching mathematical algorithms to these specific problems remain: how many bands are enough? What are the optimal unmixing procedures? What automated tools can be applied to speed and simplify the procedures?
Complete infiltrating brain tumor margin resection continually eludes neurosurgeons due to inherent limitations of current margin localization techniques. A need exists for an objective, on-site, real-time imaging system which can accurately localize brain tumor margins and therefore be used as a basis for image-guided surgery. Optical biopsy methods are a proven means for successful brain tissue discrimination, indicating promise for spectral imaging to fill such a need. Before testing spectral imaging for surgical guidance, various spectral imaging modalities must be systematically compared to determine the modality most conducive to the clinical setting. A liquid crystal tunable filter spectral imaging system was characterized for field of view, spatial and spectral resolution, and ability to retain spectral features acquired from a clinical single-pixel spectroscopy system. For a 35-mm diameter field of view, the system possessed a spatial resolution of 50 μm in both image dimensions and a spectral resolution which monotonically increased from 10 to 30 nm over the tuning range of the filter. Differences between imaging and single-pixel spectra for location and FWHM of fluorescence peaks from two fluorescent dye targets were summarily less than 3 nm. However, two remediable artifacts were introduced to imaging system spectra during spectral sensitivity correction.
Multivariate curve resolution (MCR) using constrained alternating least squares algorithms represents a powerful analysis capability for a quantitative analysis of hyperspectral image data. We will demonstrate the application of MCR using data from a new hyperspectral fluorescence imaging microarray scanner for monitoring gene expression in cells from thousands of genes on the array. The new scanner collects the entire fluorescence spectrum from each pixel of the scanned microarray. Application of MCR with nonnegativity and equality constraints reveals several sources of undesired fluorescence that emit in the same wavelength range as the reporter fluorphores. MCR analysis of the hyperspectral images confirms that one of the sources of fluorescence is due to contaminant fluorescence under the printed DNA spots that is spot localized. Thus, traditional background subtraction methods used with data collected from the current commercial microarray scanners will lead to errors in determining the relative expression of low-expressed genes. With the new scanner and MCR analysis, we generate relative concentration maps of the background, impurity, and fluroescent labels over the entire image. Since the concentration maps of the fluorescent labels are relativly uaffected by the presence of background and impurity emissions, the accuracy and useful dynamic range of the gene expression data are both greatly improved over those obtained by commercial microarray scanners.
The use of fluorescence and reflectance spectroscopy in the analysis of cervical histopathology is a growing field of research. The majority of this research is performed with point-like probes. Typically, clinicians select probe sites visually, collecting a handful of spectral samples. An exception to this methodology is the Hyperspectral Diagnostic Imaging (HSDI) instrument developed by Science and Technology International. This non-invasive device collects contiguous hyperspectral images across the entire cervical portio. The high spatial and spectral resolution of the HSDI instruments make them uniquely well suited for addressing the issues of coupled spatial and spectral variability of tissues in vivo. Analysis of HSDI data indicates that tissue spectra vary from point to point, even within histopathologically homogeneous regions. This spectral variability exhibits both random and patterned components, implying that point monitoring may be susceptible to significant sources of noise and clutter inherent in the tissue. We have analyzed HSDI images from clinical CIN (cervical intraepithelial neoplasia) patients to quantify the spatial variability of fluorescence and reflectance spectra. This analysis shows the spatial structure of images to be fractal in nature, in both intensity and spectrum. These fractal tissue textures will limit the performance of any point-monitoring technology.
Hyperspectral imaging of skin combines the spectral information of diffuse reflectance spectroscopy with the spatial information of 2D imaging. Skin chromophore maps can be reconstructed in which features such as pigmented lesions, diffuse and localized erythema, areas of increased blood stasis, etc. could be identified and the relative
parameters quantified. Hyperspectral imaging is the only reliable method to produce a quantitative distribution map of chromophores contributing to the color appearance of the skin.
Multispectral imaging is receiving attention in medical color imaging, as high-fidelity color information can be acquired by the multispectral image capturing. On the other hand, as color enhancement in medical color image is effective for distinguishing lesion from normal part, we apply a new technique for color enhancement using multispectral image to enhance the features contained in a certain spectral band, without changing the average color distribution of original image. In this method, to keep the average color distribution, KL transform is applied to spectral
data, and only high-order KL coefficients are amplified in the enhancement.
Multispectral images of human skin of bruised arm are captured by 16-band multispectral camera, and the proposed color enhancement is applied. The resultant images are compared with the color images reproduced assuming CIE D65 illuminant (obtained by natural color reproduction technique). As a result, the proposed technique successfully visualizes unclear bruised lesions, which are almost invisible in natural color images. The proposed technique will
provide support tool for the diagnosis in dermatology, visual examination in internal medicine, nursing care for preventing bedsore, and so on.
The present and accepted standard for determining the status of tissue relies on visual inspection of the tissue. Based on the surface appearance of the tissue, medical personnel will make an assessment of the tissue and proceed to a course of action or treatment. Visual inspection of tissue is central to many areas of clinical medicine, and remains a cornerstone of dermatology, reconstructive plastic surgery, and in the management of chronic wounds, and burn injuries. Near infrared spectroscopic imaging holds the promise of being able to monitor the dynamics of tissue physiology in real-time and detect pathology in living tissue. The continuous measurement of metabolic, physiological, or structural changes in tissue is of primary concern in many clinical and biomedical domains. A near infrared hyperspectral imaging system was constructed for the assessment of burn injuries and skin flaps or skin grafts. This device merged basic science with engineering and integrated manufacturing to develop a device suitable to detect ischemic tissue. This device has the potential of providing measures of tissue physiology, oxygen delivery and tissue hydration during patient screening, in the operating room or during therapy and post-operative/treatment monitoring. Results from a pre-clinical burn injury study will be presented.
An imaging spectrograph, designed and built by Science and Technology International (STI), and a point monitoring system, developed at the Lund Institute of Technology, have been used to measure the fluorescence and reflectance of cervical tissue in vivo. The instruments have been employed in a clinical trial in Vilnius, Lithuania, where 111 patients were examined. Patients were initially screened by Pap smear, examined by colposcopy and a tissue sampling procedure was performed. Detailed histopathological assessments were performed on the biopsies, and these assessments were correlated with spectra and images. The results of the spectroscopic investigations are illustrated by a thorough discussion of a case study for one of the patients, suggesting that the techniques are useful in the management of cervical malignancies.
Craniosynostosis, the premature fusion of the skull bones at the sutures, is the second most common human birth defect that affects the face and skull. The top most flat bones that comprise the skull, or calvaria, are most often affected. We previously showed that treatment of mouse calvaria with FGF2-soaked beads leads to craniosynostosis. In this study we treated mouse calvaria with FGF2-soaked beads and then used Raman imaging to demonstrate the spatial distribution of apatitic mineral and matrix in the sutures. There was no difference between FGF2 treated and control calvaria in the type of mineral produced (a lightly carbonated apatite), however we did observe increased mineral deposition in FGF2 treated calvaria. Raman imaging has great promise to detect the earliest mineral and matrix changes that occur in craniosynostosis.
We describe demonstrations of two new techniques for snapshot pectral imaging in two dimensions. The first, based on a generalisation of the Lyot filter, we believe to be the first technique able to spectrally image in snapshot mode with modest resolution and without the need for data inversion. The second demonstration is of a biologically inspired foveal hyperspectral imager, which mitigates the data acquisition and processing bottleneck encountered in traditional hyperspectral imaging approaches.