Answers are grounded in this paper and cite their sources.
ples.Our findings, and those of others [5], raised the possibil-Fourier-transform infrared (FT-IR) spectra of malignant and ity that FT-IR spectroscopy may have a potential as a diagdysplastic cervical scrapings were abnormal, as first described in nostic method.our study of a limited number of samples, where the spectra were Our initial study [4], aimed at the evaluation of structural evaluated by visual inspection and peak intensity ratios.We have changes associated with cervical carcinogenesis, was based expanded our study to evaluate more cervical conditions, and to on a limited number of samples.Moreover, discrimination analyze the spectra by a chemometric approach (principal compobetween spectra was achieved by visual inspection and comnent analysis [PCA]).Cervical samples from 436 females were parison of peak intensity ratios.However, visual inspection evaluated by FT-IR and Papanicolaou testing; 40/436 spectra is not the ideal method to compare spectra.This approach were nonanalyzable.The remaining were as follows: normal, 174; lends itself to subjective bias and provides no quantitative malignant, 19; dysplasia, 8; atypia, 113; atrophy, 19; inflammatory, information.Using peak intensity ratios to discriminate be-47; bloody smear, 12; hypocellular, 4. PCA analysis followed byx 2 test revealed that statistically significant frequencies of being tween spectra also has several disadvantages.This method is predicted malignant by FT-IR were associated with samples diaginsensitive to subtle differences between spectra and cannot nosed as malignant (P õ 0.0001), and also those diagnosed as model for the possible interference from extraneous factors.''atrophy'' (P õ 0.001), ''atypical with bloody smear'' (P õ 0.05), For these reasons, we expanded our study (a) to evaluate by ''atypical with atrophic pattern'' (P õ 0.05), and ''dysplasia'' (P FT-IR spectroscopy a wider range of cervical conditions õ 0.05).Based on these findings, for the diagnosis of cervical and (b) to analyze the FT-IR spectra by a more comprehencancer by FT-IR, as defined here, the sensitivity is 79%, the specisive approach. ficity is 77%, the positive predictive value is 15%, and the negativeA thorough discrimination between spectra of cervical predictive value is 98.6%.Our findings (a) demonstrate the applispecimens that have subtle variations requires the use of cation of a chemometric approach to the study of cervical FT-IR robust and sensitive methods of analysis.These methods spectra; (b) assess its potential diagnostic role; (c) suggest that must model for the nonlinearities that may arise due to variatrophic and neoplastic samples share structural features; and (d) suggest that blood may interfere with such spectroscopic evalua-ous causes as well as account for the day to day drifts in tion.These findings warrant further evaluation of FT-IR spectrosinstrument settings.Sample processing errors, baseline copy in cervical and other malignancies.᭧ 1997 Academic Pressshifts, batch to batch variations, the presence of nondiagnostic debris, and all other factors that adversely affect discrimination must be also accounted adequately and modeled for.
No takes yet. Share an insight, caveat, or question.
Cohenford et al. (1997) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: