A radiomics prediction method for the expression level of ki-67 in glioma

Abstract

The invention discloses a radiomics prediction method for the expression level of glioma Ki-67. The key molecular marker Ki-67 corresponding to patients with low-grade glioma samples is collected in the glioma database. Expression status; extract the radiomics characterization of each lower-grade glioma sample; screen out the differential radiomics characteristics of patients with different Ki-67 expression status, and establish a linear regression model; according to the established linear regression model, statistical Calculate and screen out the radiomics features related to the expression level of Ki-67 in gliomas to form a prediction model; screen out the radiomics features of patients, and predict the molecular marker Ki-67 of glioma in patients according to the prediction model The expression level. Based on the existing glioma imaging and genomics database, the present invention adopts the method of linear regression model, and obtains the imaging omics characterization that can reflect the genetic characteristics of the tumor through rapid and accurate analysis of the established prediction model, and more accurately predicts Ki‑67 expression levels in glioma patients.

Description

technical field [0001] The invention belongs to the field of gliomas, and particularly relates to a radiomics prediction method for the expression level of Ki-67 in gliomas. Background technique [0002] Gliomas are the most common brain malignancies, and about 30% of them are low-grade gliomas. Although low-grade gliomas have a relatively good prognosis, almost all low-grade gliomas develop with High-grade glioma with high mortality. Among them, the expression level of Ki-67 in glioma can be used to assist in judging the grade of glioma, and can more accurately reflect the reproduction level of glioma tumor cells, which has an important reference value for judging the prognosis of glioma. [0003] The genetic background of lower-grade gliomas determines biological behaviors such as tumor growth and invasion, which can be visualized on magnetic resonance imaging. Due to the marked heterogeneity of the genetic background of lower-grade gliomas, the MR appearances of the tum...

AI Technical Summary

Problems solved by technology

Chinese patent document CN 106683081 discloses a nondestructive prediction method and prediction system for glioma molecular markers based on radiomics. The automatic segmentation method of 3D MRI images, registering the segmented Saturdays to the standard brain atlas, obtaining multiple features of tumor distribution, extracting 110 related features, and then using the support vector machine and AdaBoost classifier to obtain the IDH1 The classification with a prediction accuracy of 80%; at the same time, since IDH1 belongs to the two classifications, mutation or non-mutation, there is a large deviation in the prediction accuracy

Images