Method for analyzing correlation between CT imaging genomics characteristics and gene expression in lung cancer

Abstract

The invention relates to a method for analyzing the correlation between CT imaging genomics characteristics and gene expression in lung cancer. The method includes: using a semi-automatic segmentationmethod to extract CT imaging genomics characteristics of a segmented tumor were; performing clustering analysis on the basis of preprocessed gene data, and taking a first principal component as a representative of a gene clustering result with a similar expression profile; and finally, using a gene chip saliency analysis algorithm to find the correlation between CT imaging genomics characteristics and gene expression in lung cancer, and verifying and analyzing the result. The invention provides a novel scheme for exploring the relationship between the image characteristics and the gene data,and attempts to find an imaging substitute of the gene, interprets the image characteristics from the gene level, and assists the personalized treatment of the tumor well.

Description

technical field [0001] The invention relates to a computer-aided correlation analysis method, in particular to a correlation analysis method based on CT radiomics features and non-small cell lung cancer gene expression. Background technique [0002] Lung cancer has a high incidence rate, high mortality rate, and low 5-year survival rate. It is the number one killer of cancer deaths in the world, of which non-small cell lung cancer accounts for about 80%. In recent years, imaging genomics, an emerging direction, has developed rapidly in the field of tumor research. It is a technology that combines medical imaging features with gene expression. [0003] Genomic information is characterized by high specificity. At present, genetic testing samples are mainly obtained through biopsy or surgery. However, due to the high spatial heterogeneity of tumors, the gene information detected by these local tissues cannot fully represent the global information of the entire tumor lesion. I...

AI Technical Summary

Problems solved by technology

If most studies adopt manual segmentation, the segmentation results will inevitably be affected by differences in radiologist experience; the extracted qualitative features of the image are poor in repeatability and strong subjectivity, while the extracted quantitative features of the image are not perfect, and important features may be missed. Characteristic information; the processing results of genetic data lack the analysis of biological significance

Images