You Position: Home > Paper

Application of MR radiomics in predicting the prognosis of amnestic mild cognitive impairment patients

( views:0, downloads:0 )
Author:
No author available
Journal Title:
Chinese Journal of Radiology
Issue:
10
DOI:
10.3760/cma.j.cn112149-20211008-00646
Key Word:
认知障碍;影像组学;磁共振成像;列线图;决策曲线;Cognition disorders;Radiomics;Magnetic resonance imaging;Nomogram;Decision curve

Abstract: Objective:To investigate the efficacy of MR radiomics in predicting the transformation of amnesic mild cognitive impairment (aMCI) into Alzheimer′s disease (AD).Methods:The clinical and imaging data of 30 patients with aMCI diagnosed in the First Hospital of Shanxi Medical University from December 2018 to December 2020 and 190 aMCI cases from public databases were retrospectively analyzed. The study population included 120 males and 110 females, aging from 50 to 80 years old. Thirty-nine cases progressed to AD, and 181 cases did not progress to AD. They were randomly divided into a training set (154 cases) and a validation set (66 cases) with a ratio of 7∶3. The edges of the bilateral hippocampus were manually delineated layer by layer on the transverse, coronal, sagittal images of T 1WI three-dimensional magnetization preparatory gradient echo sequence to obtain a three-dimensional labeled image, and then the radiomics features of the left hippocampus, right hippocampus, and combined bilateral hippocampus were extracted and screened, respectively. Finally, a radiomics label was constructed. Receiver operating characteristic (ROC) curves were used to evaluate the performance of each radiomics signature. The optimal radiomics label was selected to calculate the radiomics score (Rad-score), and constructed a radiomics (R) nomogram. According to the Rad-score threshold of the optimal radiomics signature, high-risk and low-risk groups for aMCI converting to AD were distinguished, and Kaplan-Meier survival curves were drawn to predict conversion to AD in aMCI patients. The clinical features affecting the progression of aMCI patients were determined by univariate logistic regression analysis, the nomogram of clinical features (C) was established, and finally the nomogram of radiomics & clinical features (R+C) was established. The efficacy of R nomogram, C nomogram, and R+C nomogram in predicting the progression of AD in aMCI patients at 3 years and 5 years was evaluated by C index. The decision curve was drawn to determine the application value of R nomogram, C nomogram and R+C nomogram in the clinical diagnosis. Results:The ROC curve showed that the right hippocampal radiomics signature had the best performance in predicting the conversion of aMCI to AD (the area under the curve was 0.989 for the training set and 0.897 for the validation set). Taking the Rad-score=21.374 of the right hippocampus radiomics label as the threshold, there were 70 cases in the high-risk group and 84 cases in the low-risk group in the training set, and 34 cases in the high-risk group and 32 cases in the low-risk group in the validation set. The mean time for conversion to AD in aMCI patients in the low-risk and high-risk groups predicted by the right hippocampal radiomics label in the training set was 42.5 and 25.0 months, and 41.3 and 22.4 months in the validation set, respectively. Both the nomogram and the decision curve indicated that the R+C nomogram showed greater benefit in predicting the conversion of aMCI to AD than the R nomogram and the C nomogram.Conclusions:The right hippocampal radiomics signature has good performance in predicting whether aMCI patients will progress to AD. The right hippocampal radiomics signature combined with the nomogram of clinical features can be used to predict the progression to AD in aMCI patients, which may help clinicians make decisions.

  • This article has no references!
WanfangData CO.,Ltd All Rights Reserved
About WanfangData | Contact US
Healthcare Department, Fuxing Road NO.15, Haidian District Beijing, 100038 P.R.China
Tel:+86-010-58882616 Fax:+86-010-58882615 Email:yiyao@wanfangdata.com.cn