Feature analysis method and system based on incoherent motion image in voxel

Abstract

The invention discloses a feature analysis method and system based on incoherent motion images in voxels, and the method comprises the steps: firstly obtaining an IVIM image sequence of a PCa patient, delineating the region of interest of each IVIM image, and obtaining a region of interest set; performing feature extraction on the region-of-interest set through a pyradiomics toolkit to obtain a plurality of types of radiomics feature sets corresponding to the region-of-interest; for each type of image omics feature set, performing integration through an L1 regularization Logistic regression algorithm to obtain representative features of the type of image omics feature set; integrating the plurality of types of representative features, obtaining the weight of each type of representative features through an L2 regularization Logistic regression model, and obtaining the IVIM image analysis result of the PCa patient according to the weight of each representative feature. It can be seen that the invention provides a non-invasive, accurate and quantitative visualization method based on radiomics characteristics for accurate diagnosis of prostate cancer.

Description

technical field [0001] The invention relates to the field of medical image processing, in particular to a feature analysis method and system based on incoherent motion images within voxels. Background technique [0002] Prostate cancer (PCa) is the second most common tumor in men and the leading cause of cancer death in men. Clinically significant PCa (clinically significant PCa, CsPCa) refers to Gleason score ≥ 7 points, with or without volume ≥ 0.5cm 3 , PCa with or without extracapsular invasion of the prostate, these tumors are highly malignant and aggressive, and require active treatment. Clinically insignificant PCa (Clinically insignificant PCa, CiPCa) refers to tumors with a Gleason score of <7, with low invasiveness and slow progression, and can be treated with follow-up observation and active monitoring. Overdiagnosis and treatment of it will increase the burden on patients , Reduce the quality of life. Therefore, accurate preoperative identification of CsPCa...

AI Technical Summary

Problems solved by technology

Clinically insignificant PCa (Clinically insignificant PCa, CiPCa) refers to tumors with a Gleason score of <7, with low invasiveness and slow progression, and can be treated with follow-up observation and active monitoring. Overdiagnosis and treatment of it will increase the burden on patients , reduce the quality of life

Images