This PDF file contains the front matter associated with SPIE Proceedings Volume 10472, including the Title Page, Copyright information, Table of Contents, and Conference Committee listing.

Ovarian cancer has the lowest survival rate among all gynecologic cancers due to predominantly late diagnosis. Early detection of ovarian cancer can increase 5-year survival rates from 40% up to 92%, yet no reliable early detection techniques exist. Optical coherence tomography (OCT) is an emerging technique that provides depthresolved, high-resolution images of biological tissue in real time and demonstrates great potential for imaging of ovarian tissue. Mouse models are crucial to quantitatively assess the diagnostic potential of OCT for ovarian cancer imaging; however, due to small organ size, the ovaries must rst be separated from the image background using the process of segmentation. Manual segmentation is time-intensive, as OCT yields three-dimensional data. Furthermore, speckle noise complicates OCT images, frustrating many processing techniques. While much work has investigated noise-reduction and automated segmentation for retinal OCT imaging, little has considered the application to the ovaries, which exhibit higher variance and inhomogeneity than the retina. To address these challenges, we evaluated a set of algorithms to segment OCT images of mouse ovaries. We examined ve preprocessing techniques and six segmentation algorithms. While all pre-processing methods improve segmentation, Gaussian filtering is most effective, showing an improvement of 32% +/- 1.2%. Of the segmentation algorithms, active contours performs best, segmenting with an accuracy of 0.948 +/- 0.012 compared with manual segmentation (1.0 being identical). Nonetheless, further optimization could lead to maximizing the performance for segmenting OCT images of the ovaries.

Preterm birth (PTB) presents a serious medical heath concern throughout the world and maintains a high incidence rate in both developed and developing countries ranging between 11-15%, respectively. PTB can be caused by many different morbidities and ultimately results in the disorganization of cervical collagen and the premature alteration of the cervix mechanical properties. Changes in cervical collagen orientation and distribution may prove to be a predictor of PTB. Polarization imaging is an effective means to measure optical anisotropy in birefringent materials such as those rich in collagen. Non-invasive, in-vivo full-field Mueller Matrix polarimetry (MMP) imaging was conducting using a modified colposcope in a clinical study comparing collagen orientation and distribution between non-pregnant and pregnant patients. Six patients threatening PTB were imaged at the Jackson Memorial Hospital Triage Unit and six non-pregnant patients were image at Florida International University STAR center. In pregnant women collagen distributions changed depending on patient age and number of pregnancies in the non-pregnant population age played an important role in collagen organization.

The early detection of cervical dysplasia enables early treatment, a critical factor in cancer prevention. In the United States, cervical cancer screening is age-based and includes cervical cytology with human papilloma virus (HPV) testing with referral to colposcopy for abnormal results. Colposcopy is used to visualize changes in the appearance of the transformation zone to direct biopsies which can confirm a diagnosis of dysplasia or cancer. Directed biopsies can be limited in detection of abnormalities because they represent a small area of the transformation zone and can be limited by provider expertise. Additionally, biopsies contribute to patient discomfort and anxiety awaiting for results. We recently reported the first in vivo cervical data from angle-resolved low-coherence interferometry (a/LCI), an optical technique that measures nuclear size as a biomarker for dysplasia, which is well-suited for screening due to its high sensitivity and specificity and its non-invasive utilization. However, in order to target the single-point measurements of the a/LCI instrument, we aimed to construct a probe capable of mapping the cervical epithelium to identify the transformation zone between the ectocervical and endocervical epithelia, the location at which dysplasia is most likely to develop. We termed this complementary technology multiplexed low-coherence interferometer (m/LCI). Thirty-six parallel fiber-optic interferometers were constructed to obtain optical depth profiles using spectral-domain LCI. Light from each channel is delivered to the cervix via a 6x6 fiber-optic bundle and a custom endoscopic probe. The depth-profile from each optical channel enables the identification of the ectocervix and endocervix. A pilot study at Duke University (n=5) was followed by an ongoing clinical study at New York City Health + Hospitals/Jacobi (Bronx, New York) (current n=20, target n=50). We present the results from these first studies to demonstrate the feasibility of m/LCI as a means of identifying the transformation zone for screening of dysplasia.

Spatial Frequency Domain Imaging (SFDI) is a Diffuse Optical Imaging (DOI) technique that is well suited for preclinical functional imaging. Recently, we have shown that SFDI can successfully be used for longitudinal monitoring of a prostate subcutaneous tumor xenograft, where we have applied a look-up-table (LUT) based approach to extract tissue absorption (μa) and scattering properties (μs’). This LUT assumes a semi-infinite homogeneous medium and simulates reflectance (Rd) in spatial domain, and scales Rd for all μa and μs’ of interest from a single Monte Carlo simulation. However, converting Rd to spatial frequency domain (SFD) and scaling for μs’ may introduces unacceptable errors. Most importantly, the homogeneous model fails to mimic the actual physiology of a subcutaneous tumor, which can be described as a two-layer medium with a thin skin layer above the tumor layer. To overcome these limitations, we have developed a Monte Carlo based two-layer LUT with a wide range of tumor (bottom) layer optical properties, and fixed skin (top) properties. The two-layer LUT will be validated by two-layer silicone phantoms and tested for sensitivity to inaccurate layer assumptions. Additionally, the homogeneous and two-layer LUTs will be used on a large mouse tumor database (n=54 mice monitored over 3 months) to identify how the two-layer LUT can improve accuracy of SFDI by more accurately reflecting in vivo physiology, and reducing discretization and scaling errors. Improved SFDI findings in small animals, in the long run, will help establish clinical DOI tools for early detection of chemotherapy efficacy during treatment.

Prostate cancer comprises the second most common cancer in men. One of the most powerful and established prognostic indicators of adenocarcinoma of the prostate is the Gleason score, a subjective assessment of the pattern of tumor growth and extent of glandular differentiation in H&E stained histology slides. Despite being the most dominant prostate grading method in use, the Gleason score suffers from high variability between grading pathologists, and due to its 2D nature, fails to effectively capture potentially prognostic information contained in 3D glandular growth patterns. We have previously demonstrated that persistent homology, a subset of topological data analysis (TDA), is effective in generating a quantitative morphological descriptor capable of differentiating Gleason grade in 2D. By capturing glands as loops in 2D, and voids in 3D, persistent homology lends itself naturally to the assessment of 3D glandular growth patterns while maintaining a correspondence to their 2D analogue. Dual-view inverted selective plane illumination microscopy (diSPIM) with a fluorescent H&E analogue was leveraged for volumetric imaging of optically-cleared prostate biopsies. The two orthogonal views of the diSPIM system yielded isotropic resolution in all dimensions, facilitating reconstruction of tissue histology in 3D for quantitative morphological assessment by persistent homology. The use of a nuclei specific hematoxylin analog (DRAQ5), in addition to the isotropic resolution of the system, enabled accurate 3D nuclei segmentation, thereby facilitating application of persistent homology to the corresponding nuclei 3D point clouds. Through TDA a quantitative, reproducible descriptor for 3D prostate cancer morphology will be demonstrated.

Breast cancer is the most common cancer in women and usually originates from the epithelial cells of the mammary duct. During the earliest stage, the tumor cells remain trapped inside the duct, generating an indolent “Ductal Carcinoma In Situ” (DCIS). However, DCIS cells can break the wall of the mammary duct, invade the surrounding stromal tissue, and metastasize to other organs. Unfortunately, the mechanisms that trigger this invasion remain elusive. One hypothesis is that the harsh microenvironment (i.e., hypoxia, nutrient-starvation) of DCIS leads to epithelial cell invasion into the stroma. In this work, a microfluidic DCIS model was developed including DCIS and normal epithelial cells, fibroblasts, and blood vessel-like structures. Optical metabolic imaging (OMI) of the metabolic co-factors NAD(P)H and FAD was used to assess spatial gradients in metabolism within the microfluidic model. We observed that the epithelial cells hindered the penetration of nutrients inside the lumen, and led to severe hypoxia. This hypoxia exerted OMI-measured metabolic changes in both normal and DCIS cells. Nuclear magnetic resonance (NMR) metabolomics analysis showed DCIS-specific metabolic differences compared with normal cells, in agreement with OMI results. Metabolic changes included an increase in glycolysis products and production of cancer-associated metabolites. A hypoxia-activated prodrug (Tirapazamine) selectively destroyed the hypoxic tumor cells inside the lumen, without affecting cells at the lumen surface or the fibroblasts in the matrix. In the future, OMI of this microfluidic model will be used to test metabolic therapies that prevent the growth of DCIS into an invasive tumor.

Emerging studies increasingly suggest an active role for the primary tumor-mediated pre-metastatic niche at the secondary organ in supporting future cancer cell infiltration. However, owing to the limited set of tools that can map subtle differences in molecular mediators in organ-specific microenvironments, etiology of such pre-metastatic niches remains poorly understood. Novel non-perturbative tools that can provide quantitative insights into subtle changes in biochemical composition of secondary sites prior to morphological manifestations are urgently needed. To address this unmet need, we have developed an approach to detect pre-metastatic changes in the lung microenvironment, in response to primary breast tumors, using a combination of metastatic mouse models, Raman spectroscopy and multivariate analysis. Specifically, we employed orthotopic breast cancer xenografts comprised of high (MDA-MB-231) and low (MCF-7) metastatic tdTomato fluorescent protein expressing cells and performed label-free Raman spectroscopic mapping to record the molecular content of pre-metastatic lungs. Our study reveals discriminative Raman features, characteristic of collageneous stroma as well as proteoglycans, that uniquely identify the metastatic potential of primary tumor based on the compositional changes in their lungs. The spectroscopic findings are in agreement with Masson’s trichrome staining observations and gene expression analysis of microarray data of pre-metastatic lung samples from mice harboring breast tumor xenografts of varying metastatic potential. The presented data are both unique and complementary to that acquired using conventional analytical tools such as downstream genomic tests and fluorescence-tagged cell tracking. Overall, our findings create a new landscape for spectroscopic monitoring of pre-metastatic niche by uncovering discriminative stromal spectroscopic features in secondary sites.

Early detection of metastatic cancer can reduce patient mortality and decrease cost of cancer treatment. However, current methods of prognosis or genetic screening are expensive and might not be applicable to all tumors. Although previous studies indicated that cancer cells are glycolytic, the link between metabolism and metastatic progression is not fully understood. To better understand the tumor bioenergetics, we investigated in vivo the vascular oxygenation, glucose intake, and optical redox ratio between a metastatic breast cancer cell line (4T1), a non-metastatic isogenic cell line (168FARN), and a non-metastatic derivative of 4T1 (TWIST gene knockout). The vascular oxygenation was measured by injecting 10,000 cells into mouse dorsal window chambers and acquiring and processing trans-illumination images of the tumor from 520 nm-620 nm light wavelength in 10 nm intervals. Glucose intake was measured by continuous fluorescent imaging of the glucose analog, 2-NBDG, for 90 minutes. Optical redox ratio was measured by intrinsic fluorescence imaging of electron carrying intermediates, NADH and FAD, where an increase in the ratio (FAD/FAD+NADH) meant increased oxidative phosphorylation. Our data show that the optical redox ratio and vascular oxygenation are higher and glucose intake is lower in metastatic tumors compared to non-metastatic tumors, suggesting that metastatic tumors display decreased glycolysis and increased oxidative phosphorylation. We observed a similar trend in vitro, where the redox ratio increased as the cell metastatic potential increased, indicating that metastatic cells can efficiently produce energy. These findings indicate that optical redox ratio can be a potential prognosis tool for detecting malignant tumors.

The lack of sufficient margin assessment during breast conserving surgery results in up to 40% of the cases in incomplete tumor removal. We evaluate the feasibility of hyperspectral imaging as an intra-operative margin assessment technique. Hyperspectral imaging rapidly collects diffuse reflected light with a large field of view over a broad wavelength range (900-1700 nm). Thereby a 3D hypercube is created that contains both spectral and spatial information of the imaged scene. Measurements are performed on 20 freshly excised breast specimen with a pushbroom camera (900-1700 nm). The specimen is sliced according to standard protocol and one slice, that contains both tumour and healthy tissue, is selected for optical measurements. Histopathology of the measured surface of this slice is obtained afterwards and used for hyperspectral data labelling. We use a spectral-spatial classifier to discriminate tumorous tissue from surrounding healthy tissue. First, we apply a linear Support Vector Machine (SVM) to obtain a pixel-based spectral classification. As output, we obtain classified pixels and their probability estimates. Second, we use this output as input for the spatial regulation, which is based on Markov Random Fields. This results in a spectral-spatial classification accuracy of 91%. Spatial regulation mainly affects pixels with a tissue type classification dissimilar to its neighbourhood. Thereby, the spectral classification accuracy is not significantly increased but the ‘pepper-and-salt’ effect, observed after pixel-based classification, is reduced.

Breast cancer is the second most common cancer worldwide and by far the most frequent cancer among women. A major limiting factor for complete surgical resection is the physician’s ability to intraoperatively assess presence of tumor positive resection margins. Many surgeons still rely on visual or tactile guidance and this leads in incomplete cancer resection rate that ranges between 20% and 50%. In this study we use multi-spectral Time-Resolved Fluorescence Spectroscopy (ms-TRFS), allowing for dynamic raster tissue scanning by merging a 450 nm aiming beam with the pulsed fluorescence excitation light in a single fiber collection. We developed a device that combines multispectral time resolved fluorescence lifetime with state-of-the-art machine learning techniques to delineate tumor margins of excised breast cancer specimen in real-time. In order to train the classifier, we precisely registered ex-vivo specimen with histology slides using fiducial markers and piecewise shape matching. A probabilistic random forest classifier was trained to rapidly delineate tumor regions. Moreover, the system not only provides binary output on tumor regions, but also quantifies the classifier’s certainty of each prediction. This allows the surgeon to either rescan the ambiguous area to increase certainty or extend the resection area to decrease the probability of positive tumor margins. The outcome is visualized by a simple color scheme showing tumor in red and adipose and fibrous tissue in blue and green and the certainty is encoded in color saturation. The system has been evaluated for n=10 lumpectomy specimen showing promising agreement between the classifier’s predictions and histology.

1. Background

Tourette syndrome (TS) is a neurological disorder characterized by repetitive, stereotyped, involuntary movements and vocalizations called tics. Near-Infrared Spectroscopy (NIRS) can assess brain function non-invasively by detecting changes in blood hemoglobin concentrations associated with neural activity with tasks like Posner’s paradigm (concerning response inhibition and attention shifts).

2. Objective

To develop a possible noninvasive objective neuroimaging protocol with a wearable wireless device for assessment of brain activities in children with Tourette syndrome.

3. Method

Children aged 6-15 years, with TS or healthy control, received functional NIRS (task-based) with the Posner paradigm after informed consent and neuropsychiatric tests (including WISC-IV test, SNAP-IV rating scale, Yale Global Tic Severity Scale Score). Behavioral data (reaction time and error rates (omission, anticipation, orientation) and NIRS data for neural changes by changes in oxy-hemoglobin and deoxy-hemoglobin levels were recorded and statistically analyzed using the SPSS software.

4. Results

20 subjects were included, 13 male and 7 female (mean age: 9.79 years; all right-handed). No significant differences in reaction time and error rate between Tourette subjects and control. For the NIRS data, more dominant activation at left prefrontal area with increasing flow with task was seen in control subjects while no dominant activation or flow increase with task was noted in Tourette subjects.

5. Conclusion

NIRS with prefrontal channels with the wearable wireless device can effectively assess the frontal activation differences and thus probably act as promising neurofeedback tools for TS or other developmental disorders like autism or attention deficit hyperactivity disorder.

The study of the mechanical properties of embryonic tissues has become an area of increasing need as more relationships between structure and function are discovered. However, there are no appropriate tools currently available to study mechanical properties of soft, millimeter-scale structures. Here, we present work on a micro-scale compression optical coherence elastography (C-OCE) system which can make quantitative measurements of the mechanical properties of small biological tissues. Inspired by "sensor" C-OCE, in which a hydrogel is laid over the sample of interest to calibrate the deformation, here we fully embed a small sample in hydrogel, then apply compressive or tensile force. Meanwhile, a phase-stable optical coherence tomography (OCT) system images the sample. Nanometer-scale displacements are extracted from the phase signal and used to generate a cross-sectional strain map. The strain map is then interpreted to provide information about the absolute mechanical properties of the sample at micron-level resolution.  To date, we have demonstrated sensitivity to 20 kPa differences in Young's modulus in soft gelatin phantoms. Additionally, we have used this method to measure the mechanical properties of de-lamination of endocardial cushions as they develop into cardiac valves in quail embryos. Further work will include 3D property characterizations, use of finite element modeling to calculate absolute mechanical properties of complex structures, and deeper investigation into the role of mechanics during valvulogenesis and other developmental processes. 

Congenital coronary anomalies can result in severe consequences such as arrhythmias and sudden death. However, the etiology of abnormal embryonic coronary microvasculature development is understudied. Using a novel contrast-agent-based optical coherence tomography (OCT) technique, scatter labeled imaging of microvasculature in excised tissue (SLIME), we compared diseased and normal embryonic quail coronary microvasculature in 3D. Congenital heart defects associated with fetal alcohol syndrome (FAS) were induced in a quail model by injecting 40 uL of 50% ethanol solution into eggs during gastrulation. These and saline-injected quail eggs were incubated until stage 36. SLIME contrast agent was perfused through the aortas of embryos and fixed in the vessels with a crosslinking agent. Dissected hearts were treated with a CUBIC-1 clearing agent and the scattering contrast labeled vasculature were imaged using customized spectral domain OCT systems. SLIME data revealed that coronary microvasculature of the control group was organized as parallel bundles over the left ventricle near the apex, whereas in a majority of the ethanol-treated embryos coronary microvasculature had chaotic patterns in similar regions. These differences in alignment of microvasculature have not been previously described in this disease model. Quantitative and statistical assessment will aid in evaluating the significance of this coronary defect. Future investigations will determine whether coronary mispatterning may reflect misalignment of cardiomyocytes that could lead to other negative consequences.

Biodynamic imaging (BDI) is capable of capturing the intracellular dynamics of blastocysts within a relatively short time. Spectroscopic signatures of embryos in the 0.01 Hz - 1 Hz range display responses to external factors before morphology changes take place. Viability evaluation is consistent with results from other non-invasive methods. Biodynamic imaging is a potential tool for selecting high quality embryos in clinical IVF practices.

Optical clearing is an effective tool for investigating spatial distribution of molecules in embryonic tissue. Unfortunately, it has not been broadly adapted in the field of development biology. One reason is that most current optical clearing methods involve complicated procedures that are more difficult compared to common lab procedures. To address this problem, we developed an easy and convenient optical clearing method, termed lipid-preserving index matching for prolonged imaging depth (LIMPID), that involves only one major step in the entire procedure. Since all LIMPID ingredient are water-soluble, it can directly diffuse into the fixed tissue and match the refractive index. Because no dehydration, organic solvent exchange or lipid extraction is required, LIMPID also well preserves fluorescent signal and tissue morphology while maintaining high clearing capability. In addition, LIMPID clearing solution has low viscosity that allows fast diffusion of the chemicals into the tissue. We have tested LIMPID using embryonic quail tissue at various developmental stages. By simply immersing the fixed and stained tissue in excess LIMPID solution, it is capable of clearing whole-mount stage 20 quail embryos in 10 minutes, stage 36 quail hearts overnight, and stage 36 quail brains in 24 hours. Verified by confocal microscopy, fluorescent signals from eYFP, DAPI, LysoTracker and several Alexa Fluor tagged primary antibodies were all well preserved. Imaging depth of LIMPID is only limited by the working distance of the optical system, up to multiple millimeters. Many small embryo tissues can be imaged all the way through using common confocal setups.

Accurately quantifying embryonic cardiac electrophysiology is important not only for understanding the complex interplay between the conduction system and development of other cardiac tissues, but also for assessing how the conduction system is altered in disease and how this alteration, in turn, can cause congenital heart defects. Optical mapping (OM) using fluorescent voltage-sensitive dyes is a powerful tool for measuring electrical activities of early intact beating embryonic hearts with a large field-of-view. However, conventional OM provides a two-dimensional (2-D) representation of a three-dimensional (3-D) sample, and thus information acquired is incomplete and dependent upon the orientation of the sample, which makes it hard to compare one heart to another. To overcome this limitation, volumetric OM using light-sheet microscopy has been developed. However, the methods for calculating conduction velocity (CV) from volumetric data are lacking. Here, we extend this approach to measure 3-D CV. First, the activation time for each voxel is determined by fitting the upstroke of the action potential trace in time. From this, we find the spatial gradient of activation time. Finally, CV is calculated based on this spatial gradient of activation time. This approach was validated using a 3-D simulated digital phantom, and embryonic quail hearts at looping stages. The conduction patterns in looping hearts are consistent with previously observed features: conduction is faster in the ventricle region and slower at the atrioventricular junction and outflow tract; the inner curvature of the ventricle has slower conduction than the outer curvature.

Altered hemodynamics in developing embryonic hearts lead to congenital heart diseases, motivating close monitoring of blood flow over several stages of development. Doppler OCT can assess blood flow in tubular hearts where blood velocity increases drastically during the period of cardiac cushion (valve precursors) formation. The blood-induced shear stress undergoes dramatic changes as well, which affects gene expression by the endothelial cells. Previously, we built a high-speed OCT system using an FDML laser (Optores GmbH, Germany) at a sweep rate of 1.68 MHz (axial resolution - 12 μm, sensitivity - 105 dB, phase stability - 96 mrad). The ultra-fast A-line rate of this laser may be used to collect real-time volumetric images of the heart, or can be traded off to obtain dense B-scans for more accurate Doppler measurements with a larger dynamic range using Doppler complex regression. Since we cannot achieve volumetric imaging with dense B-scans, an image-based retrospective gating technique was developed to register the asynchronously acquired dense B-scans to the 4D volumes. The direction of flow was determined by finding the centroid of the Doppler signal from the rearranged B-scans along the heart tube to compute absolute velocity. Subsequently, the cross-section from which the shear stress is calculated was realigned orthogonal to the direction of blood flow to approximate the velocity gradient normal to the wall. In conclusion, our high-speed OCT system will enable semi-automated measurement of the absolute blood velocity, as well as mapping the shear stress exerted on the inner walls of the embryonic hearts.

Hemodynamic load, contractile forces, and tissue elasticity are regulators of cardiac development and contribute to the mechanical homeostasis of the developing vertebrate heart. Congenital heart disease (CHD) is a prevalent condition in the United States that affects 8 in 1000 live births[1], and has been linked to disrupted cardiac biomechanics[2-4]. Therefore, it is important to understand how these forces integrate and regulate vertebrate cardiac development to inform clinical strategies to treat CHD early on by reintroducing proper mechanical load or modulating downstream factors that rely on mechanical signalling. Toward investigation of biomechanical regulation of mammalian cardiovascular dynamics and development, our methodology combines live mouse embryo culture protocols, state-of-the-art structural and functional Optical Coherence Tomography (OCT), second harmonic generation (SHG) microscopy, and computational analysis. Using these approaches, we assess functional aspects of the developing heart and characterize how they coincide with a determinant of tissue stiffness and main constituent of the extracellular matrix (ECM)—type I collagen. This work is bringing us closer to understanding how cardiac biomechanics change temporally and spatially during normal development, and how it regulates ECM to maintain mechanical homeostasis for proper function.

We previously showed that optical redox imaging (ORI) of snap-frozen breast biopsies by the Chance redox scanner readily discriminates cancer from normal tissue. Moreover, indices of redox heterogeneity differentiate among tumor xenografts with different metastatic potential. These observations suggest that ORI of fluorescence of NADH and oxidized flavoproteins (Fp) may provide diagnostic/prognostic value for clinical applications. In this work, we investigate whether ORI of formalin-fixed-paraffin-embedded (FFPE) unstained clinical tissue slides of breast tumors is feasible and comparable to ORI of snap-frozen tumors. If ORI of FFPE is validated, it will enhance the versatility of ORI as a novel diagnostic/prognostic assay as FFPE samples are readily available. ORI of fixed tissue slides was performed using a fluorescence microscope equipped with a precision automated stage and appropriate optical filters. We developed a vignette correction algorithm to remove the tiling effect of stitched-images. The preliminary data from imaging fixed slides of breast tumor xenografts showed intratumor redox heterogeneity patterns similar to that of the frozen tissues imaged by the Chance redox scanner. From ORI of human breast tissue slides we identified certain redox differences among normal, ductal carcinoma in situ, and invasive carcinoma. We found paraformaldehyde fixation causes no change in NADH signals but enhances Fp signals of fresh muscle fibers. We also investigated the stability of the fluorescence microscope and reproducibility of tissue slide fluorescence signals. We plan to validate the diagnostic/prognostic value of ORI using clinically annotated breast cancer sample set from patients with long-term follow-up data.

Accurate methods for determining metastatic risk from the primary tumor are crucial for patient survival. Cell metabolism could potentially be used as a marker of metastatic risk. Optical imaging of the endogenous fluorescent molecules nicotinamide adenine dinucleotide (NADH) and flavin adenine dinucleotide (FAD) provides a non-destructive and label-free method for determining cell metabolism. The optical redox ratio (FAD/FAD+NADH) is sensitive to the balance between glycolysis and oxidative phosphorylation (OXPHOS). We have previously established that hypoxia-reoxygenation stress leads to metastatic potential-dependent changes in optical redox ratio. The objective of this study was to monitor the changes in optical redox ratio in breast cancer cells in response to different periods of hypoxic stress as well various levels of hypoxia to establish an optimal protocol. We measured the optical redox ratio of highly metastatic 4T1 murine breast cancer cells under normoxic conditions and after exposure to 30, 60, and 120 minutes of 0.5% O2. This was followed by an hour of reoxygenation. We found an increase in the optical redox ratio following reoxygenation from hypoxia for all durations. Statistically significant differences were observed at 60 and 120 minutes (p˂0.01) compared with normoxia, implying an ability to adapt to OXPHOS after reoxygenation. The switch to OXPHOS has been shown to be a key promoter of cell invasion. We will present our results from these investigations in human breast cancer cells as well as non-metastatic breast cancer cells exposed to various levels of hypoxia.