Share Email Print

Proceedings Paper

Statistical analysis of FTIR spectra of cervical tissues and diagnostic algorithms for cervical cancer
Author(s): Benjamin Appiah; Vengadesan Nammalvar; Rebekah Drezek
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

We have analyzed infrared spectra of microtome sections of frozen cervical tissues from the malignant (cancer) and normal histopathological groups. Intensity ratios of the different bands throughout the fingerprint region in the mid-IR are compared for the above groups. The bands are evaluated to show that the peak absorbances for the different groups exhibit statistically significant differences. Among the significant changes observed is the increase in the peak intensity of the 1400 cm-1 band of the normal tissue group. Also, the contribution of the 1240 cm-1 band which is *due to the phosphatediester group of nucleic acids was greater for the malignant tissue group compared to normal tissues. Partial-Least Squares (PLS) factors are used to further analyze the tissues to access the more subtle differences between the cancer and normal groups and to help determine the spectral bands that are most useful for cancer diagnosis. Also, we utilized Linear Discriminant Analysis (LDA) as well as the supervised learning method of Support Vector Machines (SVM) to study our tissue samples and develop diagnostic algorithms. PLS was used to decompose the spectra to reduce the spectral variables. The PLS components were then used as variables in SVM procedures to construct algorithms that produce specificity and sensitivity values greater than 90%.

Paper Details

Date Published: 22 February 2008
PDF: 8 pages
Proc. SPIE 6863, Optical Diagnostics and Sensing VIII, 68630V (22 February 2008); doi: 10.1117/12.764271
Show Author Affiliations
Benjamin Appiah, Rice Univ. (United States)
Vengadesan Nammalvar, Rice Univ. (United States)
Rebekah Drezek, Rice Univ. (United States)

Published in SPIE Proceedings Vol. 6863:
Optical Diagnostics and Sensing VIII
Gerard L. Coté; Alexander V. Priezzhev, Editor(s)

© SPIE. Terms of Use
Back to Top