Real time FTIR spectroscopy of tissue smears from gynecologic cancer tissues: A preliminary study.

Abstract

e17089 Background: Women with suspected Gynecologic cancer undergo surgical procedures during which tissue from suspected areas is excised. Fast histopathology analysis is performed intra-operatively using frozen section (FS) analysis, results of which are available within less than an hour. However, the accuracy of the FS test ranges between 75% and 100% when compared to final histopathology diagnoses. Fourier transform Infrared (FTIR) spectroscopy, utilized for classification of tissue samples into malignant and benign tumors, has shown comparable results to those of FS histopathology analysis. However, the sample preparation time and the effects of tissue preparation on the measured spectra have been a concern for the utilization of this technique in clinical practice. In this study we used attenuated total reflection (ATR) FTIR spectroscopy to examine fresh tissue impression smears as an alternative to the FS technique for rapid classification of tissue samples obtained during surgery. Methods: The study was approved by relevant ethics committees and was conducted in accordance with the Declaration of Helsinki. All patients provided written, informed consent. In total, 23 biopsies (ovarian and uterine) were extracted from suspected tumor sites during surgical procedures and sent to the histopathology laboratory for both pathological and FTIR analyses. Results of the histopathology analysis classified 15 samples as benign and 8 samples as malignant. Prior to the histopathologic analysis, tissue samples from these tumors were lightly pressed against the surface of an ATR crystal, leaving on its' surface impression smears. These smears were air dried for ~5 minutes. Mid-IR absorbance spectra were collected using an ATR-FTIR spectrometer. Machine learning techniques (PCA-LDA and SVM) were utilized to build discrimination models from the absorbance data of the measured smears. Sensitivity and specificity were calculated. Results: IR absorbance spectra of malignant smears were consistently higher from spectra of benign smears in the 850cm-1 to 1450 cm-1 range and they were consistently lower in the 3200cm-1 to 3600cm-1 range. The PCA-LDA discrimination model correctly classified the samples with a sensitivity and specificity of 100%, and the SVM showed a training accuracy of 100% and a cross validation accuracy of 91.3%. Conclusions: These preliminary results suggest that ATR-FTIR spectroscopy of tissue smears may have an important role in the development of next-generation techniques for intra-operative tumor classification.