Recently microendoscopes are being developed as a tool to detection cancer or pre-cancerous lesions in the milk ducts of the human breast. The microendoscope can be inserted into the duct through the nipple. Integration of fluorescence spectroscopy into microendoscopy can provide an improved platform for real-time cancer detection followed by immediate intervention. Typically, the optical fibers employed by existing microendoscope systems transmit in the 450 to 900 nm range. A prototype system combining fluorescence spectroscopy with visible imaging by microendoscopy is described and preliminary measurements on ex vivo human breast tissues are presented. Image resolution and distortion are discussed.

Currently no clinical tool exists that measures the degree of ischemic injury incurred in tissue and assesses tissue function following transplantation. In response to this clinical problem, we explore optical spectroscopy to quantitatively assess ischemic injury. In our method we monitor the autofluorescence intensities under excitation suitable to excite specific tissue fluorophores. Specifically, a first excitation probes NADH, a biomolecule known to change its emission properties depending on the tissue's metabolic state. A second excitation is used to mainly probe tryptophan, a biomolecule expected to be minimally affected by metabolism. We postulate that the ratio of the two autofluorescence signals can be used to monitor tissue behavior during ischemia and reperfusion. To evaluate this approach, we acquire autofluorescence images of the injured and contralateral control kidney in vivo in a rat model under excitation at both wavelengths during injury and reperfusion. Our results indicate that this approach has the potential to provide real-time monitoring of organ function during transplantation.

Extraction of quantitative biochemical information from measured fluorescence spectra is hindered by the presence of potentially significant distortions introduced by tissue scattering and absorption. Such distortions can be removed by extracting the intrinsic fluorescence spectra from the measured fluorescence spectra. This paper explores the potential applicability of spatially resolved fluorescence technique for simultaneous extraction of intrinsic fluorescence and evaluation of optical transport parameters, namely, reduced scattering coefficient (μ_s|), absorption coefficient (μ_a) from tissue mimicking model systems and human breast tissues. A hybrid diffusion theory-Monte Carlo simulation based theoretical model was used to estimate the values for μ_s| and μ_a and to recover intrinsic fluorescence from the measured spatially resolved fluorescence from the samples. The agreement between the values for μ_s' and μ_a estimated for tissue mimicking phantoms using the spatially resolved fluorescence measurement technique and the corresponding calculated values are seen to be satisfactory with a maximum percentage error of ≤ 10 % and also line shape and intensity of intrinsic fluorescence recovered using this approach was observed to be free from the disentangling effects of absorption and scattering properties of the medium. Intrinsic fluorescence spectra of breast tissues show a distinct difference between malignant and its normal counterpart. A narrowing of the line shape is also observed as compared to the bulk fluorescence spectrum.

Native fluorescence of tissues in the UV and visible spectral regions has been investigated for over two decades. Native fluorescence has been demonstrated to be an accurate tools for distinguish normal tissue from malignant and pre-malignant. Prior investigations have demonstrated that there are several ratio-based algorithms, which can distinguish malignant tissue from normal with high sensitivity and specificity.¹ The wavelength combinations used in these ratios isolate the contributions from pairs of tissue fluorophors, one of which is frequently tryptophan (trp), the predominant tissue fluorophore with excitation in the UV (250-300 nm). In this work, algorithms using a combination of native fluorescence and trp phosphorescence were developed which show promise for providing enhanced detection accuracy. Using optical fibers to collect the emission from the specimen allowed interrogation of small regions of tissue, providing precise spatial information. Using a specially designed setup, specimens were excited in the UV and spectra were collected in the range of 300 to 700 nm. Three main emission bands were selected for analysis: 340 nm (trp fluorescence); 420 - 460 nm band (fluorescence from the extra cellular matrix); and 500 - 520 nm (trp phosphorescence). Normal specimens consistently exhibited a low ratio (<10) of 345 to 500 nm emission intensity while this same ratio was consistently high (>15) for cancer specimens. Creating intensities ratio maps from the tissue allows one to localize the malignant regions with high spatial precision. The study was performed on ex vivo human breast tissues. The ratio analysis correlated well with histopathology.

Early detection and precise excision of neoplasms are imperative requirements for successful cancer treatment. In this study we evaluated the use of dye-enhanced confocal microscopy as an optical pathology tool in the ex vivo trial with fresh thick non-melanoma skin cancer excisions and in vivo trial with B16F10 melanoma cancer in mice. For the experiments the tumors were rapidly stained using aqueous solutions of either toluidine blue or methylene blue and imaged using multimodal confocal microscope. Reflectance images were acquired at the wavelengths of 630nm and 650 nm. Fluorescence was excited at 630 nm and 650 nm. Fluorescence emission was registered in the range between 680 nm and 710 nm. The images were compared to the corresponding en face frozen H&E sections. The results of the study indicate confocal images of stained cancerous tissue closely resemble corresponding H&E sections both in vivo and in vitro. This remarkable similarity enables interpretation of confocal images in a manner similar to that of histopathology. The developed technique may provide an efficient real-time optical tool for detecting skin pathology.

Enlargement of mammalian cells nuclei due to the cancerous inflammation can be detected early through noninvasive optical techniques. We report on the results of cellular experiments, aimed towards the development of a fiber optic endoscopic probe used for precancerous detection of Barrett's esophagus. We previously presented white light scattering results from tissue phantoms (polystyrene polybead microspheres). In this paper, we discuss light scattering properties of epithelial MDCK (Madine-Darby Canine Kidney) cells and cell nuclei suspensions. A bifurcated optical fiber is used for experimental illumination and signal detection. The resulting scattering spectra from the cells do not exhibit the predicted Mie theory oscillatory behavior inherent to ideally spherical scatterers, such as polystyrene microspheres. However, we are able to demonstrate that the Fourier transform spectra of the cell suspensions are well correlated with the Fourier transform spectra of cell nuclei, concluding that the dominate scatterer in the backscattering region is the nucleus. This correlation experimentally illustrates that in the backscattering region, the cell nuclei are the main scatterer in the cells of the incident light.

The backscattering of circularly polarized light at normal incidence to a half-space composed of two index matched layers with different absorption coefficients is studied using the Electric Field Monte Carlo method. The top layer, of thickness L₁ = 2.5[l_s], where l_s is the scattering length, is non-absorbing and is composed of particles suspended in water with anistropy factor g = 0.8. The bottom layer, of thickness L₂ = 25[l_s], is composed of absorbing particles with g = 0.8. The backscattered light with the same helicity (co-polarized) as the incident beam emerging from the top surface is analyzed in the time domain as absorption in the second layer increases from 1% to 10% of the scattering coefficient, μs. For the case of a homogenous half-space, composed of non absorbing particles with anisotropy factor g = 0.8, a ring-peak is known to be observed in the time-resolved co-polarized backscattered light intensity. For the two layer geometry tested here, a similar ring structure is found and used to determine the path length of photons traveling in the second layer. In recent studies, the ring-peak was postulated to be comprised of photons undergoing semi-circular trajectories as a result of near forward scattering events in the forward scattering media. This ideal picture of photon trajectories is tested and found to be an accurate characterization of photon trajectories in forward scattering media. Specifically, it is shown that time-sliced measurements of the backscattered co-polarized intensity at the ring-peak and path lengths of photons determined from the segment of arc of their idealized semi-circluar trajectories in the second layer can be used in conjunction with Beer's law to reproduce the known absorption coefficient of the second layer. This is a first indication that photons contributing to the ring-peak in co-polarized backscatter follow semi-circular trajectories. Moreover, it demonstrates that ring-structure can be used to determine subsurface features such as absorption coefficients in layered structures.

In this paper we present a polarization based technique for optical sectioning and imaging of multi-layer cell patterns separated by a weakly diffused media. Multi-layer cell pattern is important to study because this type of structure is often used for heterogeneous three dimensional cell culture and bio-chips applications, where information at different depths would be crucial. Functioning of this type of bi-layer or multilayer cell patterns can easily be monitored using polarization based imaging techniques. For polarization based imaging, samples are excited by white light source with different set of band-pass filter and linear polarizer, and images are collected through corresponding long-pass filters and analyzer by CCD camera. Preliminary experiments are carried out using absorption inhomogeneity separated by a weakly diffused thin polymer layers. Polarized images at various angles are collected at a set of excitation wavelength. Such measurements can identify 3x3 sub-matrix elements out of the full 4x4 sixteen elements of Mueller matrix. In order to enhance the image contrast, the 3x3 Mueller components are further decomposed into diattenuation and depolarization power images. Superficial layer image information is found to be more prominent in the depolarization power images, and diattenuation images provide sub layer information. By comparing the decomposition images at various wavelengths, we can observe sub-layer structures at different depths.

Multiphoton microscopy and second harmonic generation microscopy were used to image epithelial changes in a hamster model for oral malignant transformation. In vivo imaging was performed to characterize morphometric alterations in normal and precancerous regions. Morphometric measurements such as cell nucleus area and epithelial thicknesses obtained from MPM-SHGM were in excellent agreement with histology obtained after in vivo imaging. MPM-SHGM was highly sensitive to spectroscopic and architectural alterations throughout carcinogenesis, showing statistically significant changes in morphology. MPM revealed hyperkeratosis, nuclear enlargement/crowding in dysplasia, and immune cell infiltration. SHGM revealed alterations in submucosal architecture, with a decrease in SHG density evident during early stages of precancer. By combining MPM with SHGM, the basement membrane could be identified in normal, hyperplasia, and dysplasia samples and in some cases of early invasion. The combined technique of MPM-SHGM has the potential to serve as an adjunct to biopsy for assessing precancerous changes and will be investigated further for that purpose. Additionally, the method can provide spatiotemporal assessment of early neoplastic changes in order to elucidate the stages of transformation in vivo and could be used to assess therapeutic efficacy of agents being tested for the treatment of epithelial precancers/cancer.

The capability of Brillouin spectroscopy for diagnosis of elasticity has been investigated. Although detection of weak signal under noise is a barrier to overcome, we have shown that the stimulated scattering can be enhanced by two orders of magnitude using stimulated thermal Brillouin scattering excited at a wavelength of 1550 nm where water has a moderate absorption of 12 cm^-1. While thermal lens effect might cause excess noise in a certain experimental setup, we have shown it can be circumvented by polarization modulation technique.

The interaction of light with tissue and cells is the underlying mechanism for optical biomedical imaging and spectroscopy to detect pathology changes. We use fractal continuous random media to model visible and near infrared light scattering by biological tissue and cell suspensions, which provides a simple relation between the morphological features of the sample and its optical properties (absorption, scattering and anisotropic factor of light scattering). Good agreement with experimental results are found for this fractal continuous random medium model. A novel optical biopsy scheme of oblique incidence Fourier reflectometry is proposed for spectroscopic tissue diagnosis based on the model.

We explore imaging of tissue microstructures using autofluorescence and light scattering methods implemented through a hyperspectral microscope design. This system utilizes long working distance objectives that enable off-axis illumination of tissue thereby allowing for excitation at any optical wavelength without requiring change of optical elements within the microscope. Spectral and polarization elements are easily and rapidly incorporated to take advantage of spectral variations of spectroscopic optical signatures for enhanced contrast. The collection efficiency has been maximized such that image acquisition may be acquired within very short exposure times, a key feature for transferring this technology to a clinical setting. Preliminary studies using human and animal tissues demonstrate the feasibility of this approach for real-time imaging of intact tissues as they would appear in the operating room.

Development of the optical biopsy system for experimental small animals is in progress. A prototype of the system which consists of a miniaturized gastro endoscope unit and Raman probes has been completed by now. The system is developed to study a gastric cancer rat model. The endoscope is 2.5 mm in diameter and is equipped an imaging bundle fiber, illumination fibers, a channel and a mechanism to angle the probe head. The head of the Raman probe comes out through the channel and it is possible to aim the probe to the target watching on the monitor. The endoscope was inserted into the anaesthetized healthy rat under the breathing support. It was successfully observed inside of the stomach of the living rat and measured Raman spectra. The spectrum of blood vessels contains the strong contribution from lipids. The present results demonstrate high potential of the system in the in vivo Raman study using the rat model.

We apply a unique micro-optoelectromechanical tuned light source and new algorithms to the hyper-spectral microscopic analysis of human colon biopsies. The tuned light prototype (Plain Sight Systems Inc.) transmits any combination of light frequencies, range 440nm 700nm, trans-illuminating H and E stained tissue sections of normal (N), benign adenoma (B) and malignant carcinoma (M) colon biopsies, through a Nikon Biophot microscope. Hyper-spectral photomicrographs, randomly collected 400X magnication, are obtained with a CCD camera (Sensovation) from 59 different patient biopsies (20 N, 19 B, 20 M) mounted as a microarray on a single glass slide. The spectra of each pixel are normalized and analyzed to discriminate among tissue features: gland nuclei, gland cytoplasm and lamina propria/lumens. Spectral features permit the automatic extraction of 3298 nuclei with classification as N, B or M. When nuclei are extracted from each of the 59 biopsies the average classification among N, B and M nuclei is 97.1%; classification of the biopsies, based on the average nuclei classification, is 100%. However, when the nuclei are extracted from a subset of biopsies, and the prediction is made on nuclei in the remaining biopsies, there is a marked decrement in performance to 60% across the 3 classes. Similarly the biopsy classification drops to 54%. In spite of these classification differences, which we believe are due to instrument and biopsy normalization issues, hyper-spectral analysis has the potential to achieve diagnostic efficiency needed for objective microscopic diagnosis.