医学
小儿癫痫
儿科
癫痫
梅德林
急诊医学
小儿中风
厄尔尼诺现象
重症监护医学
放射科
作者
Rushuai Li,Feng Wang,Rui Yang,Rui Luo,Xiaoyuan Li,Runze Wu,Jiaxin Hao,Liangshun Xiong,Heping Ju,艾书跃,Lei Xu,Qingle Meng
标识
DOI:10.2967/jnmt.126.272511
摘要
To minimize radiation risks in pediatric PET imaging, injected dose should be kept as low as reasonably achievable. The aim of this study was to explore whether simulated low-count [18F]FDG PET/MR could preserve diagnostic quality in pediatric epilepsy, thereby supporting the design of future lower-activity validation studies. Methods: Twenty-seven pediatric patients with epilepsy were enrolled. List-mode PET data acquired over 20 min were truncated to 10, 7, and 4 min to simulate count reductions of 50%, 65%, and 80%, respectively. Image quality of the PET images was assessed visually using a 5-point Likert scale and quantitatively using contrast-to-noise ratio and coefficient of variation. The intraclass correlation coefficient was used to evaluate SUV ratio (SUVR) consistency across the simulated count reduction levels in different brain regions. Two computer-assisted methods, region of interest (ROI)–based methods and voxel-based methods, were used for epileptogenic zone (EZ) lateralization. For the ROI-based method, a machine learning model using regional SUVRs as features was applied, and performance was evaluated via receiver-operating-characteristic curve analysis and the area under the curve, with statistical significance assessed by the DeLong test. For the voxel-based method, EZ lateralization was determined on the basis of the hemisphere of hypometabolic clusters identified through asymmetry analysis. Lateralization agreement with the clinical consensus was compared across simulated levels of count reduction. Results: Quantitative analysis showed that the contrast-to-noise ratio decreased and the coefficient of variation increased significantly with simulated count reduction. Visual assessment indicated that, although overall image quality declined, count reduction up to 65% remained clinically acceptable under these simulated conditions. High consistency in mean SUVRs was observed between full-count and simulated low-count PET images across brain regions. Both ROI-based and voxel-based methods demonstrated that EZ lateralization agreement with the clinical consensus did not decline with simulated count reduction. Conclusion: Our study in pediatric epilepsy suggests that a 65% simulated count reduction preserved clinically acceptable [18F]FDG PET image quality, stable SUVR quantification, and concordance of EZ lateralization with the clinical consensus. These findings support the need for prospective validation of a low-activity protocol in pediatric epilepsy.
科研通智能强力驱动
Strongly Powered by AbleSci AI