Study On Pathological Evolution Of Breast Cancer By Confocal Raman Microscopy Imaging

Breast cancer is the leading cause of cancer deaths for woman in the world,and its incidence rate and death toll increase year by year.However,traditional histopathological diagnosis takes a long time,and relies heavily on the experience of pathologists,resulting in a high false positive rate of diagnosis,causing unnecessary burden to the patients.Hence,there is urgent need to develop a new technology,based on the complete and rapid characterization of molecular changes in the process of breast cancer progression,to distinguish healthy tissue from tumor tissue,and to realize cancer diagnosis accurately.Confocal Raman microscopy imaging technology can characterize and explain the composition of biological samples at the molecular level,has the advantages of high sensitivity and specificity,and does not damage the samples.Through the combination of Raman spectroscopy and microscopic images,the composition and distribution information of biochemical components in tissue samples can be clearly characterized.The detailed research work is as follows:Firstly,we describe the source and sample preparation method briefly.In order to understand the biochemical composition of different types of breast tissues,single spectral measurements were performed on healthy(H),lobular hyperplasia(LH),ductal carcinoma in situ(DCIS),invasive ductal carcinoma tissues(IDC),mucinous carcinoma tissues(MC)and solid papillary carcinoma tissues(SPC)tissue.The results of spectral analysis showed that the biochemical components of breast tissue mainly include: lipids,proteins,nucleic acids and carotenoids,which laid a foundation for future work.Secondly,after understanding the biochemical composition of different types of breast tissue,we compared and analyzed the spectral differences between DCIS and LH tissue in detail.Multivariable analysis methods,including K-means clustering analysis and principal component analysis were used to visualize the ductal structure and lobular structure.In addition,single spectral imaging was used to describe the distribution of lipids,proteins and carotenoids in tissues.This study not only enabled us to have a detailed understanding of the spectral and morphological changes caused by the process of malignant transformation and breast hyperplasia,but also provides a solid data support for the development of clinical diagnosis technology of breast cancer based on Raman spectroscopy.Finally,we determined the different biochemical information of healthy,DCIS and IDC by by ex vivo micro-Raman spectroscopy.Then,compared the performance of principal component analysis-linear discriminant analysis(PCA-LDA)and principal component analysis-support vector machine(PCA-SVM)based on three kernel functions in distinguishing spectral characteristics of different tissues.And the results show that the classification accuracy of SVM for breast tissue spectrum is the highest,reaching above 95%.This study not only enables us to have a detailed understanding of the evolution process of breast cancer,but also confirms the feasibility of Raman spectroscopy combined with multivariate analysis for breast cancer diagnosis.In conclusion,this study describes the biochemical and morphological changes of breast tissue after the occurrence of different lesions,and we have established a model that can accurately classify different types of breast tissues.This study not only demonstrated the morphological and biochemical changes in the process of breast cancer evolution,but also provides a new method for clinical diagnosis of breast cancer.