Optical diagnosis of cervical cancer by fluorescence spectroscopy technique

Abstract

In the present work, we examine normal and malignant stage IIIB cervical tissue by laser induced fluorescence, with 2 different objectives. (i) Development of the fluorescence spectroscopy technique as a standard optical method for discrimination of normal and malignant tissue samples and, (ii) Optimization of the technique by the method of matching of a sample spectrum with calibration sets of spectra of pathologically certified samples. Laserinduced fluorescence spectra were measured using samples from 62 subjects at different excitation wavelengths. Principal component analysis (PCA) of spectra and intensity ratios of curveresolved fluorescence peaks were tested for discrimination. It was found that PCA of total fluorescence at 325 nm excitation gives specificity and sensitivity over 95%. Use of calibration sets of spectra of histo-pathologically certified samples combined with PCA for matching and pass/fail classification of test samples is shown to have high sensitivity/specificity for routine diagnostic purposes as well as for possible staging of the disease. Further, the multi-component origin of the fluorescence spectra is illustrated by curve resolution and fluorescence spectra of separated proteins of tissue homogenates. ' 2006 Wiley-Liss, Inc. Key words: laser induced fluorescence; optical diagnosis; principal component analysisCervical Cancer is the second leading cancer in females all over the world, standing first in many of the developing countries.1,2 Most patients are diagnosed with cervical cancer through symptoms such as abnormal vaginal bleeding and vaginal discharge. Investigation by Pap smear/colposcopy followed by histopathology confirms the diagnosis. Pap smear is reported to have high false negative (15-40%) rates. 3,4 The extraordinary monotony of the work, large volume of smears in screening applications, inadequate sampling and comparatively smaller cellular changes being overlooked because of fatigue are all reasons ascribed to the high rate of false-negative results. At present, liquid based cytological screening technique ''Thin prep'' aided by Pap Net, an automated computer based visualization technique, seems to have increased the sensitivity of cytological screening. However, the efficiency of thin prep and Pap net are still under scrutiny. 5,6

Investigation of the influence of high-risk human papillomavirus on the biochemical composition of cervical cancer cells using vibrational spectroscopy

“…Identifying the presence of HPV is a very important factor in the prevention of cervical cancer. Technologies such as 55 fluorescence or vibrational spectroscopies are being investigated as potential high sensitivity and specificity methodologies for cancer screening and diagnosis [8][9][10][11][12][13] . Vibrational spectroscopy is a molecular specific spectroscopy in which normal vibrational modes specific to molecular 60 bonds as well as more complex molecular fingerprints are probed.…”

Laser-induced fluorescence: Progress and prospective for in vivo cancer diagnosis

“…Generally, the predictive accuracy of spectra is better than the prediction based on biopsy solely. Laser-induced fluorescence spectroscopy can minimize the need for repetitive biopsy [56][57][58][59].…”

A pilot study on colonic mucosal tissues by fluorescence spectroscopy technique: Discrimination by principal component analysis (PCA) and artificial neural network (ANN) analysis

“…Development of various multivariable statistical tools such as principal component analysis (PCA) [19,20], support vector machine (SVM), etc. enhanced the reliability and accuracy of classification of normal and malignant patients based on collected autofluorescence spectra.…”

“…enhanced the reliability and accuracy of classification of normal and malignant patients based on collected autofluorescence spectra. In this work, we focus on the classification of normal and malignant colon tissues excited at 325 nm using PCA and artificial neural network (ANN) [21][22][23][24] analysis. As the choice of appropriate algorithm and wavelength have great influence on the specificity and sensitivity of the analysis based on LIF technique, the present work has great practical significance, especially from the point of view of development of accurate optical diagnostic tool.…”

“…For each sample, spectra were recorded from several points, separated by a few hundred microns to a few millimetres. Details of the sample preparation and spectral recording are mentioned elsewhere [19][20][21][22][23][24]. The sample details are given in Table I.…”