3584

Introduction: Laser photoacoustic spectroscopy can provide information on both spectroscopic as well as acoustic transmission properties of materials in a single measurement. The technique can thus be used for the dual purpose of optical pathology and optical imaging of physiological systems. Imaging techniques like x-rays, positron emission tomography, nuclear magnetic resonance and scintigraphy have disadvantages like exposure to radioactive material, insufficient resolution, inability to provide information on biochemical composition etc. Whereas, optical methods have advantages like very high sensitivity, non-invasive probing, possibility of repeat testing and also providing biochemical information. Experimental: In the present work, pulsed laser-induced photoacoustic spectroscopic studies on pathologically certified normal, benign and malignant ovarian tissues were carried out at 325nm excitation. Spectra were recorded in time domain using transient digitizer, and were Fourier transformed to frequency domain by MATLAB®. MATLAB® is a high-performance language for technical computing. It integrates computation, visualization, and programming in an easy-to-use environment. Spectral analysis and classification for discrimination among normal, malignant, and benign classes was performed using Principal Component Analysis (PCA)/cluster analysis. Results and Conclusion: Typical pulsed laser-induced photoacoustic signal of normal, benign and malignant ovarian tissues at 325 nm excitation and the corresponding Fourier Transform spectra are shown in the illustration. The photoacoustic signal spectra were recorded keeping the tissue in a 2mm wide cuvette, moist in saline, exciting it from the top and detecting the signal using PZT detector. The signal, in the time domain, was recorded using transient digitizer. For classification, principal component analysis was performed on the feature space to reduce its dimensionality. We have shown that feature vector of length 6 could be reduced to 2 components using PCA technique. It can be concluded that the PAS technique is a suitable method for discrimination of normal and malignant ovarian tissues.

[Proc Amer Assoc Cancer Res, Volume 47, 2006]