We report the development of a probability-based multi-class diagnostic algorithm to simultaneously distinguish highgrade
dysplasia from low-grade dysplasia, squamous metaplasia as well as normal human cervical tissues using nearinfrared
Raman spectra acquired in-vivo from the cervix of patients at the Vanderbilt University Medical Center.
Extraction of diagnostic features from the Raman spectra uses the recently formulated theory of nonlinear Maximum
Representation and Discrimination Feature (MRDF), and classification into respective tissue categories is based on the
theory of Sparse Multinomial Logistic Regression (SMLR), a recent Bayesian machine-learning framework of statistical
pattern recognition. The algorithm based on MRDF and SMLR was found to provide very good diagnostic performance
with a predictive accuracy of ~90% based on leave-one-out cross validation in classifying the tissue Raman spectra into
the four different classes, using histology as the "gold standard". The inherently multi-class nature of the algorithm
facilitates a rapid and simultaneous classification of tissue spectra into various tissue categories without the need to train
and heuristically combine multiple binary classifiers. Further, the probabilistic framework of the algorithm makes it
possible to predict the posterior probability of class membership in discriminating the different tissue types.
Ovarian cancer is the fifth leading cause of cancer death in women, in part because of the limited knowledge about early stage disease. We develop a novel rat model of ovarian cancer and perform a pilot study to examine the harvested ovaries with complementary optical imaging modalities. Rats are exposed to repeated daily dosing (20 days) with 4-vinylcyclohexene diepoxide (VCD) to cause early ovarian failure (model for postmenopause), and ovaries are directly exposed to 7,12-dimethylbenz(a)anthracene (DMBA) to cause abnormal ovarian proliferation and neoplasia. Harvested ovaries are examined with optical coherence tomography (OCT) and light-induced fluorescence (LIF) at one, three, and five months post-DMBA treatment. VCD causes complete ovarian follicle depletion within 8 months after onset of dosing. DMBA induces abnormal size, cysts, and neoplastic changes. OCT successfully visualizes normal and abnormal structures (e.g., cysts, bursa, follicular remnant degeneration) and the LIF spectra show statistically significant changes in the ratio of average emission intensity at 390:450 nm between VCD-treated ovaries and both normal cycling and neoplastic DMBA-treated ovaries. Overall, this pilot study demonstrates the feasibility of both the novel animal model for ovarian cancer and the ability of optical imaging techniques to visualize ovarian function and health.
Ovarian cancer is relatively rare but is the fifth leading cause of death from cancer in women. Little is known about the precursors and early stages of ovarian cancer partially due to the lack of a realistic animal model. A cohesive model that incorporates ovarian cancer induction into a menopausal rodent would be well suited for comprehensive studies of ovarian cancer, and non-destructive imaging would allow carcinogenesis to be followed. Optical Coherence Tomography (OCT) and Light-Induced Fluorescence (LIF) are minimally invasive optical modalities that allow both structural and biochemical changes to be noted. Rat ovaries were exposed to 4-vinylcyclohexene diepoxide (VCD) for 20 days in order to destroy the primordial follicles. Sutures coated with 7,12-dimethylbenz(a)anthracene (DMBA) were implanted in the right ovary, in order to produce epithelial based ovarian cancers. Rats were sacrificed at 1, 3, and 5 months and ovaries were harvested and imaged with a combined OCT/LIF system. Histology was preformed on the harvested ovaries and any pathology determined. OCT was able to visualize follicle loss and DMBA-induced abnormalities. LIF spectra were also different between cycling, follicle deplete, and DMBA-exposed ovaries. Overall this pilot study demonstrated the feasibility of both the animal model and optical imaging.
Ovarian cancer is not a common cancer-approximately 25,000 new cases in 2004-but it is the fifth leading cause of death from cancer in women (over 16,000 in 2004). Little is known about the precursors and early stages of ovarian cancer partially due to the lack of human samples at the early stages. A cohesive model that incorporates ovarian cancer induction into a menopausal rodent would be well suited for comprehensive studies of ovarian cancer. Non-destructive imaging would allow carcinogenesis to be followed. Optical Coherence Tomography (OCT), Optical Coherence Microscopy (OCM) and Light-Induced Fluorescence (LIF) are minimally invasive optical modalities that allow both structural and biochemical changes to be noted. Rat ovaries were exposed to 4-vinylcyclohexene diepoxide (VCD) for 20 days in order to destroy the primordial follicles. Plain sutures and sutures coated with 7,12-dimethylbenz(a)anthracene (DMBA) were implanted in the right ovary, in order to produce epithelial based ovarian cancers (a plain suture was inserted in the control). Rats were sacrificed at 4 weeks and ovaries were harvested and imaged with a combined OCT/LIF system and with the OCM. Histology was preformed on the harvested ovaries and any pathology determined. Two of the ovaries were visually abnormal; the OCT/LIF imaging confirmed these abnormalities. The normal ovary OCM and OCT images show the organized structure of the ovary, the follicles, bursa and corpus lutea are visible. The OCM images show the disorganized structure of one of the abnormal ovaries. Overall this pilot study demonstrated the feasibility of both the animal model and optical imaging.
Optical coherence tomography (OCT) is an imaging modality capable of acquiring cross-sectional images of tissue using back-reflected light. Conventional OCT images have a resolution of 10-15μm, and are thus best suited for visualizing tissue layers and structures. OCT images of collagen (with and without endothelial cells) have no resolvable features and may appear to simply show an exponential decrease in intensity with depth. However, examination of these images reveals that they display a characteristic repetitive structure due to speckle.
The purpose of this study is to evaluate the application of statistical and spectral texture analysis techniques for differentiating living and non-living tissue phantoms containing various sizes and distributions of scatterers based on speckle content in OCT images. Statistically significant differences between texture parameters and excellent classification rates were obtained when comparing various endothelial cell concentrations ranging from 0 cells/ml to 25 million/ml. Statistically significant results and excellent classification rates were also obtained using various sizes of microspheres with concentrations ranging from 0 microspheres/ml to 500 million microspheres/ml.
This study has shown that texture analysis of OCT images may be capable of differentiating tissue phantoms containing various sizes and distributions of scatterers.