Cancer clinical index evaluation system constructed based on radiomics qualitative algorithm

Abstract

The invention discloses a cancer clinical index evaluation system constructed based on a radiomics qualitative algorithm. The system comprises a data input module which is used for inputting multi-dimensional patient clinical data comprising a preoperative multi-modal medical image spectrum and clinical data of a patient; a feature extraction and optimization module connected with the data input module and used for extracting radiomics quantitative feature values and clinical data features and performing dimensionality reduction optimization on the quantitative feature values; a qualitative model construction module connected with the feature extraction and optimization module and used for constructing, through the radiomics qualitative algorithm, an evaluation model for predicting various clinical indexes of the cancer of the patient; and an evaluation module connected with the qualitative model construction module and used for providing noninvasive, accurate and non-subjective evaluation for early diagnosis, prognosis and drug effect subtypes of cancers.

Description

technical field [0001] The present invention relates to the technical field of radiomics detection, in particular to a radiomics qualitative algorithm based on the construction of a cancer clinical index evaluation system, and more specifically to the application of radiomics quantitative eigenvalue relative size relationship, that is, radiomics qualitative method Build a cancer clinical index evaluation system to provide auxiliary evaluation for patient treatment. Background technique [0002] Medical imaging can non-invasively and cost-effectively display the characteristics of the entire tumor, and provide dynamic information for monitoring the occurrence and development of tumors. Studies in recent years have found that multimodal medical imaging atlases, such as computerized tomography (CT), magnetic resonance imaging (Magnetic Resonance Imaging, MRI) and positron emission tomography (Positron Emission Tomography-Computed Tomography) , PET-CT), etc., can extract charac...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the above problems, the present invention provides a method for constructing various clinical index evaluation systems for cancer based on an algorithm based on radiomics qualitative features. Data and cancer-related clinical data are collected and feature extracted, multi-dimensional input feature vectors are established, and various clinical index evaluation models of cancer are constructed through radiomics qualitative algorithms, and then the evaluation model is used as the basis for early diagnosis, prognosis and drug treatment of cancer. It provides a non-invasive, accurate and non-subjective assessment of effective subtypes, so as to solve the defect that the currently developed subtype prediction and other markers cannot be clinically applied due to the batch effect of imaging feature data in different medical centers

Images