Rapid Prediction and Accurate Location Selection of Mild Traumatic Brain Injury (mTBI) by Using Multiple Parameter Analysis of Diffusion Tensor Imaging (DTI): Integrating Correlational and Clinical Approaches

Abstract

Background. Mild traumatic brain injury (mTBI) is a widespread and serious public health problem which also causes physical and psychological suffering to patients and their families and imposes a significant economic burden on society. But it is usually very difficult to detect and provide warning of mTBI in early stage. Therefore, a novel method is urgent for the increasing demands on the accurate and rapid prediction and feature selection of mTBI. Objectives. To establish a better idea of the performance of neuroimage biomarker in the acute phase of mTBI, our study adopts diffusion tensor imaging (DTI) which could present the pathophysiological changes of white matter through several parameters noninvasively and combined with behavioral experiments such as intelligence quotient test, memory, executive function, and motion function to find the relationship between DTI abnormal brain regions and behavioral abnormalities. Then, provide new method for rapid prediction and feature selection of mTBI. Methods. 77 mTBI patients were admitted to the Emergency and Neurosurgery Departments of the Third Xiangya Hospital of Central South University from August 2019 to July 2021; the patients (41 males and 36 females) suffered mTBI because of car accident (36), assault (11), and fall (30). All the mTBI patients were examined through MRI scan and behavioral psychology test within 3 days after injury. MRI images and behavioral psychology tests were also collected; the correlation between the DTI biomarker and the cognitive psychological outcome was analyzed. A series of integration and computational methods were also used for fusion arithmetic and result analysis. Results. Compared with the healthy control group, the patients in the acute stage of mTBI presented lower scores in the digit symbol substitution test (DSST), suggesting that mTBI patients in the acute stage had decline in information processing speed and associative learning. The difference of DTI parameters in acute stage mTBI patients was mainly manifested as increased AD and MD values in multiple brain regions, while RD and FA values have no significant difference. The most significant brain regions were bilateral corticospinal tracts (CST), bilateral posterior internal capsule lentiform nucleus, bilateral superior longitudinal fasciculus, left terminal striae, and left sagittal plane with right posterior thalamic radiation. The Pearson correlation coefficient was significantly positive correlation between AD and MD elevation in the left sagittal layer and the results of DSST and digit span in acute stage mTBI patients. Conclusions. The acute phase mTBI patients performed lower score on the DSST than those in the normal control group. This neuropsychological change was associated with increased AD value and MD value in the left sagittal layer, which indicated reduction of information processing speed in mTBI patients in the acute phase. It might be related to abnormal AD value and MD value in the upper longitudinal tract, lower longitudinal tract, lower frontal occipital tract, and sagittal layer. In this study, combined with neuropsychological test and increase of the AD value and MD value in certain brain region, neurosurgeon should pay more attention to the abnormal of the upper longitudinal tract and the patients’ information processing speed in the diagnosis and treatment of the acute phase mTBI patients. The study offers a much more secure and integrated method for rapid prediction and feature selection of mTBI, which could have broader clinical approaches and application prospects.