It is important to detect cervical dysplasia, Cervical Intraepithelial Neoplasia (CIN). CIN is the potentially premalignant and abnormal squamous cells on surface of cervix. In this study, the spatial frequency spectra of pre-cancer cervical tissues are used to detect differences among different grades of human cervical tissues. Seven sets of thick tissue sections of human cervix of normal, CIN 1, CIN 2, and CIN 3 tissues are studied. The confocal microscope images of the stromal region of normal and CIN human tissues were analyzed using Fast Fourier Transform (FFT) to generate the spatial spectra. It is observed that higher frequency components exist in CIN tissues than those in normal tissue, as well as those in higher grade CIN tissue than those in lower grade CIN tissue. The width of the spatial frequency of different types of tissues is used to create a criterion for CIN grading by training a support vector machine (SVM) classifier. The results show that the randomness of tissue structures from normal to different stages of precancer in cervical tissue can be recognized by fingerprints of the spatial frequency. The efficacy of spatial frequency analysis for CIN grading is evaluated as excellent since high AUC (area under the ROC curve), sensitivity and specificity are obtained by the statistics study. This works lays the foundation of using spatial frequency spectra for a histology evaluation.

Date Published: 17 March 2014
PDF: 9 pages
Proc. SPIE 8940, Optical Biopsy XII, 894016 (17 March 2014); doi: 10.1117/12.2035498

Published in SPIE Proceedings Vol. 8940:
Optical Biopsy XII
Robert R. Alfano; Stavros G. Demos, Editor(s)