Effect of normal variations on disease classification of Raman spectra from cervical tissue.

In this paper, we examine how variations in normal tissue can influence disease classification of Raman spectra. Raman spectra from normal areas may be affected by previous disease or proximity to areas of dysplasia. Spectra were acquired in vivo from 172 patients and classified into five tissue categories: true normal (no history of disease), previous disease normal (history of disease, current normal diagnosis), adjacent normal (disease on cervix, spectra acquired from visually normal area), low grade, and high grade. Taking into account the various "normal" states of the tissue before statistical analysis led to a disease classification accuracy of 97%. These results indicate that abnormal changes significantly affect Raman spectra, even when areas are histopathologically normal. The sensitivity of Raman spectroscopy to subtle biochemical differences must be considered in order to successfully implement it in a clinical setting for diagnosing cervical dysplasia and cancer.

Application of Raman spectroscopy for cervical dysplasia diagnosis.

Cervical cancer is the second most common malignancy among women worldwide, with over 490 000 cases diagnosed and 274 000 deaths each year. Although current screening methods have dramatically reduced cervical cancer incidence and mortality in developed countries, a "See and Treat" method would be preferred, especially in developing countries. Results from our previous work have suggested that Raman spectroscopy can be used to detect cervical precancers; however, with a classification accuracy of 88%, it was not clinically applicable. In this paper, we describe how incorporating a woman's hormonal status, particularly the point in menstrual cycle and menopausal state, into our previously developed classification algorithm improves the accuracy of our method to 94%. The results of this paper bring Raman spectroscopy one step closer to being utilized in a clinical setting to diagnose cervical dysplasia.

Effect of hormonal variation on Raman spectra for cervical disease detection.

OBJECTIVE: The objective of the study was to characterize the variations in normal cervical spectra because of menopausal status and location within the menstrual cycle. Using the information obtained, the accuracy of Raman spectroscopy to diagnose low-grade squamous intraepithelial lesion (LGSIL) will be improved. STUDY DESIGN: A total of 133 patients undergoing either colposcopy or Papanicolaou smear were recruited from either Vanderbilt University or Tri-State Women's Health. Raman spectra were collected from both normal and diseased areas. The data were processed and analyzed using a multiclass discrimination and classification algorithm to determine whether the spectra were correctly classified. RESULTS: Stratifying the data by menopausal state resulted in correctly classifying LGSIL 97% of the time (from 74%). CONCLUSION: This study brings Raman spectroscopy one step closer to clinical use by improving the sensitivity to differentiate LGSIL from normal.

Multiclass discrimination of cervical precancers using Raman spectroscopy.

Raman spectroscopy has the potential to differentiate among the various stages leading to high-grade cervical cancer such as normal, squamous metaplasia, and low-grade cancer. For Raman spectroscopy to successfully differentiate among the stages, an applicable statistical method must be developed. Algorithms like linear discriminant analysis (LDA) are incapable of differentiating among three or more types of tissues. We developed a novel statistical method combining the method of maximum representation and discrimination feature (MRDF) to extract diagnostic information with sparse multinomial logistic regression (SMLR) to classify spectra based on nonlinear features for multiclass analysis of Raman spectra. We found that high-grade spectra classified correctly 95% of the time; low-grade data classified correctly 74% of the time, improving sensitivity from 92 to 98% and specificity from 81 to 96% suggesting that MRDF with SMLR is a more appropriate technique for categorizing Raman spectra. SMLR also outputs a posterior probability to evaluate the algorithm's accuracy. This combined method holds promise to diagnose subtle changes leading to cervical cancer.

Comparison of Raman spectrograph throughput using two commercial systems: transmissive versus reflective.

A common goal of most Raman spectroscopists is the acquisition of clear Raman spectra in short integration times. It is paramount, then, that the hardware used in a Raman system be optimized for throughput, signal-to-noise ratio, and resolution. To this end, we performed a systematic comparison of two Raman spectrographs: the widely used Kaiser HoloSpec and a relatively new commercial offering, the Raman Explorer by Headwall Photonics. Both strong and weak Raman scattering samples were measured using various launching conditions. When resolution-matched, the throughputs of both spectrographs were found to be roughly similar over the central range of the fingerprint region (approximately 800 to 1140 cm(-1), using 785 nm illumination), with the Raman Explorer demonstrating a slight throughput advantage outside this range. Other factors are also considered such that end users may better select the optimum spectrograph for their particular application.

Detecting temporal and spatial effects of epithelial cancers with Raman spectroscopy.

Epithelial cancers, including those of the skin and cervix, are the most common type of cancers in humans. Many recent studies have attempted to use Raman spectroscopy to diagnose these cancers. In this paper, Raman spectral markers related to the temporal and spatial effects of cervical and skin cancers are examined through four separate but related studies. Results from a clinical cervix study show that previous disease has a significant effect on the Raman signatures of the cervix, which allow for near 100% classification for discriminating previous disease versus a true normal. A Raman microspectroscopy study showed that Raman can detect changes due to adjacent regions of dysplasia or HPV that cannot be detected histologically, while a clinical skin study showed that Raman spectra may be detecting malignancy associated changes in tissues surrounding nonmelanoma skin cancers. Finally, results of an organotypic raft culture study provided support for both the skin and the in vitro cervix results. These studies add to the growing body of evidence that optical spectroscopy, in this case Raman spectral markers, can be used to detect subtle temporal and spatial effects in tissue near cancerous sites that go otherwise undetected by conventional histology.

Dual modality imaging of a novel rat model of ovarian carcinogenesis.

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.

Texture analysis of speckle in optical coherence tomography images of tissue phantoms.

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 10-15 microm, 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 cells/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.