Deep learning model based on integrated 18F-FDG PET/MRI for evaluating cerebral metabolism around cerebral infarction
10.13929/j.issn.1672-8475.2024.11.004
- VernacularTitle:基于一体化18F-FDG PET/MRI深度学习模型评估脑梗死周围脑代谢
- Author:
Yuxin LIANG
1
,
2
;
Bixiao CUI
;
Yi SHAN
;
Jie MA
;
Miao ZHANG
;
Jie LU
Author Information
1. 首都医科大学宣武医院放射与核医学科,北京 100053
2. 磁共振成像脑信息学北京市重点实验室,北京 100053
- Keywords:
brain infarction;
radionuclide imaging;
magnetic resonance imaging
- From:
Chinese Journal of Interventional Imaging and Therapy
2024;21(11):665-669
- CountryChina
- Language:Chinese
-
Abstract:
Objective To investigate the value of deep learning(DL)model based on integrated 18F-FDG PET/MRI for evaluating cerebral metabolic status around cerebral infarction.Methods A total of 46 patients with cerebral infarction caused by unilateral internal carotid artery(ICA)or middle cerebral artery(MCA)steno-occlusion were retrospectively collected.Based on integrated 18F-FDG PET/MRI,DL model was used to automatically segment cerebral infarction area.Asymmetry index(AI)was used to evaluate the volume of reduced metabolic areas in the segmented affected frontal lobe,temporal lobe,parietal lobe,occipital lobe and cerebral hemisphere of cerebral infarction area as well as their proportions,while their correlations with National Institutes of Health stroke scale(NIHSS)score of neurological function were analyzed.Results Among 46 patients,the volume of decreased metabolism in the affected temporal lobe,parietal lobe and cerebral hemisphere was(41.35±10.52)ml,(65.58±14.82)ml and(178.89±34.23)ml,respectively,all positively correlated with NIHSS scores(rs=0.359,0.343,0.362,all P＜0.05).The proportion of the reduced metabolic volume in the affected frontal lobe,temporal lobe,parietal lobe and cerebral hemisphere was(45.68±10.35)%,(42.32±10.19)%,(45.05±9.44)%and(44.11±8.63)%,respectively,all positively correlated with NIHSS scores(rs=0.344,0.340,0.439,0.393,all P＜0.05).Conclusion DL model based on integrated 18F-FDG PET/MRI was of important clinical value for evaluating cerebral metabolic state around cerebral infarction.
