Quantitative imaging in medicine and surgery · 2025
Investigation of quantitative synthetic MRI in the evaluation of attention-deficit/hyperactivity disorder.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Li Yufen, Lin Liping, Qiu Huaqiong, Hou Weifeng, Gao Xiang, Dai Yan, Xu Boyan, Chen Yingqian, Zhang Hongyu, Yang Zhiyun, Liu Meina, Su Shu
Аннотация
BACKGROUND: Currently, synthetic magnetic resonance imaging (SyMRI) has been considered as a promising way to characterize intrinsic microstructural tissue alterations by quantitatively measuring tissue relaxation times. This study aimed to explore the utility of combining machine learning algorithms with SyMRI for identifying attention-deficit/hyperactivity disorder (ADHD) and further revealing its underlying brain pathology.
METHODS: Multiple quantitative parameters (T1 and T2 relaxometry values and myelin volume fraction) of whole-brain gray and white matter derived from SyMRI data on 50 individuals with ADHD and 50 age-, sex-matched healthy controls (HCs) were used as original features. Feature selection was performed by a two-sample t-test and the least absolute shrinkage and selection operator (LASSO). Machine learning models were built based on each single feature and combined features. The accuracy, sensitivity, specificity and area under the curve were used to evaluate the performance of classifiers.
RESULTS: Machine learning model based on combined features demonstrated stable performance across different classifiers [support vector machine, random forest and linear regression yielding accuracies of 0.780, 0.770, and 0.780 (all P<0.001), respectively] and showed superior performance compared to those based on single feature [accuracy of 0.730 (gray matter) and 0.770 (white matter) with support vector machine (both P<0.001)]. The most discriminative features mainly involved the macrostructure of the frontal, temporal lobes and the cerebellum, and the microstructure of the corpus callosum, frontostriatal tracts, corticospinal tract and cingulum. Moreover, the T2 relaxation time of the left cingulum was associated with symptom severity in ADHD (r=0.285, P=0.050).
CONCLUSIONS: Our study demonstrated that multi-parametric quantitative metrics derived from SyMRI combined with machine learning not only achieved good classification accuracy for distinguishing ADHD from HCs but also revealed widespread microscopic changes of gray and white matter underlying ADHD. These findings highlight the potential of using SyMRI to provide a novel perspective for exploring the disease-related microstructural processes and detecting the promising imaging marker for early diagnosis of ADHD.
METHODS: Multiple quantitative parameters (T1 and T2 relaxometry values and myelin volume fraction) of whole-brain gray and white matter derived from SyMRI data on 50 individuals with ADHD and 50 age-, sex-matched healthy controls (HCs) were used as original features. Feature selection was performed by a two-sample t-test and the least absolute shrinkage and selection operator (LASSO). Machine learning models were built based on each single feature and combined features. The accuracy, sensitivity, specificity and area under the curve were used to evaluate the performance of classifiers.
RESULTS: Machine learning model based on combined features demonstrated stable performance across different classifiers [support vector machine, random forest and linear regression yielding accuracies of 0.780, 0.770, and 0.780 (all P<0.001), respectively] and showed superior performance compared to those based on single feature [accuracy of 0.730 (gray matter) and 0.770 (white matter) with support vector machine (both P<0.001)]. The most discriminative features mainly involved the macrostructure of the frontal, temporal lobes and the cerebellum, and the microstructure of the corpus callosum, frontostriatal tracts, corticospinal tract and cingulum. Moreover, the T2 relaxation time of the left cingulum was associated with symptom severity in ADHD (r=0.285, P=0.050).
CONCLUSIONS: Our study demonstrated that multi-parametric quantitative metrics derived from SyMRI combined with machine learning not only achieved good classification accuracy for distinguishing ADHD from HCs but also revealed widespread microscopic changes of gray and white matter underlying ADHD. These findings highlight the potential of using SyMRI to provide a novel perspective for exploring the disease-related microstructural processes and detecting the promising imaging marker for early diagnosis of ADHD.