Identification Of Lymphoma And Cluster Analysis Of Breast Cancer Tissues Using Laser-induced Breakdown Spectroscopy

Early detection and diagnosis of cancer is of great significance to improve prognosis and reduce mortality.LIBS is a new cancer diagnosis technology because of its advantages of real-time online detection and multi-element simultaneous detection.The study of stage and type based on serum samples can greatly expand the application of LIBS in cancer diagnosis,but it has not been reported publicly.The combination of LIBS and cluster analysis can realize the automatic pathological diagnosis of cancer tissue,but its accuracy still needs to be improved.On the basis of previous work,this paper carried out the study of lymphoma differential diagnosis and breast cancer tissues cluster analysis based on LIBS,in order to expand the application of LIBS in serum cancer diagnosis and improve the accuracy of tumor tissues LIBS pathological diagnosis.This paper first introduces the research progress of LIBS technology in the field of cancer diagnosis and tissue element imaging,analyzes the current situation of cluster analysis in the field of cancer,and puts forward the research content of this paper.Secondly,the physical mechanism of LIBS is introduced,and the commonly used data mining methods and evaluation indicators are given,focusing on such classification algorithms as k neighbors,support vector machines,neural networks and such as k mean clustering,k-medoids clustering,hierarchical clustering,fuzzy C mean clustering and so on.Thirdly,the differential diagnosis of lymphoma based on LIBS spectrum is studied.The LIBS experimental parameters for serum sample analysis are optimized,including sample substrate,serum smear count,ambient gas flow,and detection delay.The LIBS spectra of 153 serum samples from normal people and patients with different types and stages of lymphoma are obtained.The differences of serum spectral characteristics between normal and lymphoma patients,diffuse large B-cell lymphoma(DLBCL)subtype and follicular cell type,DLBCL endogenous and extracellular subtype,DLBCL endogenous stage and DLBCL extracellular stage are analyzed.Based on k-nearest neighbor,support vector machine and neural network,the diagnosis,classification and staging models of lymphoma are established,and the application of LIBS in cancer classification and staging is expanded.The highest accuracy rates of lymphoma diagnosis,classification and staging are 95.1%,96.4% and 100%,respectively,which provides a fast,accurate and economic technical means for differential diagnosis of lymphoma.Finally,the cluster analysis of breast cancer is carried out.In order to improve the accuracy of pathological diagnosis,based on LIBS spectral data,k-means clustering,k-medoids clustering,hierarchical clustering and fuzzy c-means clustering pathological diagnosis models are established.The results of cluster analysis are evaluated by matching the chromatogram and element distribution of samples,dividing the normal area and cancer area.The accuracy of pathological diagnosis of breast cancer is improved by optimizing the model parameters such as spectral data,distance function and principal component selection.Through the analysis of seven samples,it is found that the overall performance of hierarchical clustering and k-center clustering model is high,with the highest clustering accuracy of 85.7%,sensitivity of 91.7%,specificity of 79.8% and accuracy of 94.1%,and the highest accuracy of 15 percentage points higher than the previous work.On the basis of cluster analysis,by calculating the correlation between K,Ca,Na,Mg element distribution map and cluster result map,the correlation between element distribution and tumor distribution is analyzed.It is found that Ca element is most correlated with tumor distribution,and the correlation coefficient is up to 0.894,followed by Na,Mg element,and K element.The research results in this paper will play an important role in early screening and intelligent pathological diagnosis of cancer,and it has important clinical and social significance for reducing cancer incidence rate and mortality rate and improving prognosis of cancer patients.