图像预处理方法对基于影像组学对胶质瘤MGMT基因状态预测影响的研究  

A Study on the Impact of Image Processing Methods on Predicting the MGMT Gene Status of Glioblastomas through Radiomics Analysis

作　　者：卫宏洋 宗会迁[1] 张娅 柳青[2] 史朝霞[1] 王静[1] WEI Hongyang;ZONG Huiqian;ZHANG Ya(Department of Medical Imaging,The Second Hospital of Hebei Medical University,Shijiazhuang,Hebei Province 050000,P.R.China)

机构地区：[1]河北医科大学第二医院医学影像科,石家庄050000 [2]河北医科大学第二医院医学装备部,石家庄050000

出　　处：《临床放射学杂志》2024年第9期1456-1465,共10页Journal of Clinical Radiology

基　　金：河北省卫生健康委科研基金项目(编号:20230518)。

摘　　要：目的基于随机森林分类器找到各常规MRI序列及各序列联合后最适合的图像处理方法,并联立各经过最适合的图像处理方式的序列建立一个术前预测胶质瘤MGMT基因甲基化状态的影像组学模型。方法回顾性搜集了104例(河北医科大学第二医院数据75例,TCIA公共数据集29例),经病理证实为MGMT基因甲基化(56例)和非甲基化(48例)的胶质瘤患者的T_(1)WI、T_(2)WI、FLAIR、CE-T_(1)WI图像。然后,由两名医师进行感兴趣区勾画,勾画范围包括瘤周水肿区、坏死区、实质区。之后对各序列的感兴趣区分别进行标准分数(Z-score)、Nyúl图像处理,再各自进行灰度离散化处理,并对各序列处理后的图像使用Pyradiomics包进行特征提取。对各序列进行特征筛选并进行随机森林建模,得到各序列最佳图像处理方式。再联合多序列进行特征筛选并建模,得到多序列最佳图像处理方式。最后将经过最佳处理方式的各序列特征联合进行筛选并通过随机森林建模。结果在单序列中T_(1)WI配合Nyúl序列表现最佳,训练集及验证集受试者工作特征曲线曲线下面积(AUC)值为0.98,0.85;在多序列建模中,Nyúl图像归一化方法最佳,训练集与测试集AUC值为0.94,0.89;在所有建立的模型中联合经过各序列最佳图像处理方式处理的各序列建立的模型性能最佳,训练集与验证集AUC值为0.98,0.92。结论基于联合经过各序列最佳图像处理方式处理的各序列建立的模型性能,在多序列模型中,性能最佳;在单序列中,T_(1)WI适合Nyúl,T_(2)WI适合Z-score&FBN32,FLAIR适合Z-score&FBN128,CE-T_(1)WI适合Nyúl&FBS1/128。Objective Radiomics analysis is often affected by variations introduced by different imaging devices and sites.This study aims to identify the most suitable image processing methods for individual conventional MRI sequences and their combinations using a random forest classifier.Additionally,the study aims to develop a radiomics model for preoperative prediction of MGMT gene methylation status in gliomas by combining sequences processed with their optimal image processing methods.Methods Firstly,a retrospective collection was conducted,involving 104 cases,with 75 cases from Hebei Medical University Second Hospital and 29 cases from The Cancer Imaging Archive(TCIA)public dataset.These cases were pathologically confirmed to be glioma patients with MGMT gene methylation(56 cases)and non-methylation(48 cases).The MRI images included T_(1)WI,T_(2)WI,FLAIR,and CE-T_(1)WI sequences.Subsequently,two physicians delineated regions of interest,encompassing peritumoral edema,necrotic zone,and solid region.The regions of interest for each sequence were subjected to Z-score and Nyúl image processing.Subsequently,individual sequences underwent grayscale discretization.Pyradiomics package was employed to extract features from the processed images of each sequence.Feature selection was performed for each sequence,followed by building random forest models.The optimal image processing methods for each sequence were determined through these models.Then,a combination of multiple sequences underwent feature selection and modeling to identify the optimal image processing methods for multi-sequence scenarios.Finally,the features from each sequence,processed with their optimal methods,were combined and used to build a predictive model through random forest modeling.Results In the single-sequence analysis,T_(1)WI combined with Nyúl normalization demonstrated the optimal performance,with AUC values of 0.98 for the training set and 0.85 for the validation set.In the context of multi-sequence modeling,the Nyúl image normalization method p

关 键 词：图像预处理 胶质瘤MGMT基因 标准分数 Nyúl 

分 类 号：R739.4[医药卫生—肿瘤]