Gene heterogeneity visual quantification method in glioma based on pyradiomics and system

Abstract

The invention belongs to the technical field of medical treatment and pyradiomics, and particularly relates to a gene heterogeneity visual quantification method in glioma based on pyradiomics and a system. The method of the invention comprises the steps of segmenting a glioma magnetic resonance image by means of an image segmenting network 3D U-net; performing predictive modeling on the integral glioma IDH (isocitrate dehydrogenase), namely performing high-flux characteristic extraction and characteristic screening on the image, and screening a characteristic combination which is most sensitive and most effective to gene expression; performing heterogeneous modeling on the glioma IDH based on an image block, extracting a multi-dimensional data block of the glioma image, obtaining the IDH expression strength of each data block based on the integral predicting model; and finally forming the IDH distribution visualization and quantitative expression of the whole tumor. The method and thesystem have advantages of more accurately determining the prognosis and radiotherapy and chemotherapy sensitivity of the patient, realizing surgery resection and targeting treatment in heterogeneous atlas navigation, and realizing high clinical value in improving treatment effect of the patient and improving survival prognosis.

Description

technical field [0001] The invention belongs to the technical field of medical treatment and imaging omics, and in particular relates to a method and a system for visualizing and quantifying gene heterogeneity in glioma. Background technique [0002] In recent years, with the development and popularization of gene sequencing technology, more and more evidence shows that glioma may be a "genetic disease". In 2010, glioma, as the first cancer selected into the National Cancer Genome Program (TCGA) in the United States, attracted worldwide attention, and transcriptome-based genotyping and molecular typing of gene mutations were widely recognized , it is more representative to divide glioblastoma into four gene subtypes of "Neural", "Proneural", "Mesenchymal" and "Classical". The three genes of IDH1 / 2, TERT and 1p19q will be lower Grade gliomas are divided into five molecular subtypes, and three subtypes in the WHO 2016 classification: IDH mutant, wild type, and IDH to be class...

AI Technical Summary

Problems solved by technology

[0005]Currently, the research bottleneck for glioma heterogeneity lies in two aspects: the research method is relatively complicated, especially for the study of in vivo tumors, it must be based on MRI navigation Point puncture sampling, so that tissue samples can be accurately obtained, and more accurate gene sequencing can be performed. However, multi-point puncture takes a long time, and the most critical risk is bleeding and trauma; most of them are "point-to-point" research and lack "face-to-face" research. Research

The current gene heterogeneity is all reflected in the level that tumor tissues in different parts belong to different gene subtypes. There is no way to draw a gene heterogeneity map on the same tumor, resulting in most of the studies being single-site studies. Form regionalized or even integrated research results

[0006] Existing radiomics studies focus on the prediction and evaluation of the overall gene expression, treatment options, and prognosis of individual tumors, and there are few studies on the differences within individual tumors

Images