Early Detection of Colorectal Cancer Based on Competitive Adaptive Reweighted Sampling with Human Plasma Fluorescence Spectrum and PLS£­LDA

KeyWord:fluorescence spectrum  colorectal cancer  competitive adaptive reweighted sampling £¨CARS£©  partial least squares £­linear discrimination analysis £¨PLS£­LDA£©

Abstract:

      Metabolism of cancer cells is quite different from that of normal cells£¬which results in great differences appearing in the fluorescence spectra of human plasma£®By using this difference£¬the possibility for using fluorescence spectrum as a tool for the detection of colorectal cancer could be explored£®In order to improve the detection accuracy£¬the competitive adaptive reweighted sampling£¨CARS£© method was used to screen important human plasma fluorescence spectrum variables in this paper£¬then a partial least squares£­linear discrimination analysis£¨PLS£­LDA£© was applied to establish a classification model for colorectal cancer patients and non-cancer patients£¬which was compared with the full-wavelength model and parallel factor analysis model£¨PARAFAC£©£®From the evaluation indexes of these three models£¬the performance of CARS£­PLS£­LDA model was significantly better than those of the latter two models£®The area under curve£¨AUC£© values for the fluorescence spectra of high-wave undiluted group and low-wave diluted group combined with the CARS£­PLS£­LDA classification model were higher than 0.9£®In addition£¬the sensitivity£¬specificity and AUC values of the classification model for cancer and adenoma patients in the high wave undiluted group were all 1.000 0£®Results showed that the CARS£­PLS£­LDA model could be used to effectively identify the patients with colorectal cancer£®Furthermore£¬from the data of high wave dilution group and low wave undiluted group£¬the CARS method preferably obtained the important variables£¬which could be traced back to the original fluorescence data£¬determining the corresponding excitation/emission wavelength information£¬and analyzing the chemical mechanism behind the identification£®The results also showed that these excitation/emission wavelength regions were consistent with porphyrin£¬tryptophan£¬tyrosine and NADH£¬indicating that the colorectal cancer samples could be effectively distinguished from the other control samples in these material dimensions£®Based on the parallel factor analysis model£¬it was considered that porphyrin has nothing to do with cancer£®The influence variables obtained from the high wave undiluted group data of CARS model mainly concentrated in the excitation wavelength range of 400£­420 nm and the emission wavelength range of 610£­625 nm£¬which just accorded with the wavelength range of porphyrin fluorescence£®This result suggested that there was a significant difference between the plasma porphyrin concentrations of cancer patients and healthy people£®Compared with the former£¬the results are consistent with relevant experimental studies£®The influence variables obtained from the low wave dilution group data of CARS model concentrated in the excitation wavelength of 250£­260 nm and the emission wavelength of 310£­360 nm£¬which corresponded to free tryptophan£¬bound tryptophan and tyrosine in blood£®Meanwhile£¬there were also another excitation wavelength of 260£­272 nm and another emission wavelength of 450£­500 nm£¬which was consistent with the fluorescence characteristics of NADH£®The results of this study showed that CARS variable screening could significantly improve the performance of colorectal cancer classification model and contribute to the development and research of follow-up cancer clinical diagnosis tools.