Neuroanatomical Substrates and Symptoms Associated With Magnetic Resonance Imaging of Patients With Mild Traumatic Brain Injury

Abstract

Importance: Persistent symptoms after mild traumatic brain injury (mTBI) represent a major public health problem.

Objective: To identify neuroanatomical substrates of mTBI and the optimal timing for magnetic resonance imaging (MRI).

Design, setting, and participants: This prospective multicenter cohort study encompassed all eligible patients from the Collaborative European NeuroTrauma Effectiveness Research in Traumatic Brain Injury (CENTER-TBI) study (December 19, 2014, to December 17, 2017) and a local cohort (November 20, 2012, to December 19, 2013). Patients presented to the hospital within 24 hours of an mTBI (Glasgow Coma Score, 13-15), satisfied local criteria for computed tomographic scanning, and underwent MRI scanning less than 72 hours (MR1) and 2 to 3 weeks (MR2) after injury. In addition, 104 control participants were enrolled across all sites. Data were analyzed from January 1, 2019, to December 31, 2020.

Exposure: Mild TBI.

Main outcomes and measures: Volumes and diffusion parameters were extracted via automated bespoke pipelines. Symptoms were measured using the Rivermead Post Concussion Symptoms Questionnaire in the short term and the extended Glasgow Outcome Scale at 3 months.

Results: Among the 81 patients included in the analysis (73 CENTER-TBI and 8 local), the median age was 45 (interquartile range [IQR], 24-59; range, 14-85) years, and 57 (70.4%) were male. Structural sequences were available for all scans; diffusion data, for 73 MR1 and 79 MR2 scans. After adjustment for multiple comparisons between scans, visible lesions did not differ significantly, but cerebral white matter volume decreased (MR2:MR1 ratio, 0.98; 95% CI, 0.96-0.99) and ventricular volume increased (MR2:MR1 ratio, 1.06; 95% CI, 1.02-1.10). White matter volume was within reference limits on MR1 scans (patient to control ratio, 0.99; 95% CI, 0.97-1.01) and reduced on MR2 scans (patient to control ratio, 0.97; 95% CI, 0.95-0.99). Diffusion parameters changed significantly between scans in 13 tracts, following 1 of 3 trajectories. Symptoms measured by Rivermead Post Concussion Symptoms Questionnaire scores worsened in the progressive injury phenotype (median, +5.00; IQR, +2.00 to +5.00]), improved in the minimal change phenotype (median, -4.50; IQR, -9.25 to +1.75), and were variable in the pseudonormalization phenotype (median, 0.00; IQR, -6.25 to +9.00) (P = .02). Recovery was favorable for 33 of 65 patients (51%) and was more closely associated with MR1 than MR2 (area under the curve, 0.87 [95% CI, 0.78-0.96] vs 0.75 [95% CI, 0.62-0.87]; P = .009).

Conclusions and relevance: These findings suggest that advanced MRI reveals potential neuroanatomical substrates of mTBI in white matter and is most strongly associated with odds of recovery if performed within 72 hours, although future validation is required.

Figure 1.. Association of Imaging Results With Symptom Evolution Between Magnetic Resonance Imaging (MRI) Times

Patients with mild traumatic brain injury received an MRI scan within 72 hours of injury (MR1) and 2 to 3 weeks after injury (MR2). The evolution of diffusion tensor imaging (DTI) findings between scans was categorized into 3 phenotypes: progressive injury, minimal change, and pseudonormalization. A, DTI parameters at MR1 between patients with phenotypes and healthy controls, adjusted for age, sex, and scanner. The x-axis refers to the number of tracts with abnormal findings on MR1 of the short list of 13 tracts that were shown to change between MR1 and MR2. B, Evolution of DTI parameters between MR1 and MR2 (ie, the change in mean diffusivity vs the change in fractional anisotropy). Positive log ratios indicate an increase and negative log ratios a decrease in values between scans. C, Evolution of symptoms between scans measured as change in the score of the Rivermead Post Concussion Symptoms Questionnaire, with positive values indicating worsening and negative values resolving symptoms. The boxes represent the interquartile range, the middle horizontal line is the median, and the whiskers extend to the largest value no farther than 1.5 IQR from the hinge.

Figure 2.. Estimation of Recovery at 3 Months Using Magnetic Resonance Imaging (MRI)

Logistic regression was used to associate imaging with the odds of a favorable recovery at 3 months after injury, defined as a score on the extended Glasgow Outcome Scale of 8. The no-MRI model includes only age and sex. All other models contain age and sex plus imaging information. A, Models using imaging information obtained at 72 hours or 2 to 3 weeks after injury are compared. Imaging information includes both quantitative sequences (structural sequence [T1-weighted] and diffusion tensor imaging [DTI]). B, Imaging sequences obtained within 72 hours were compared. For T1, the variable used was the deviation of the patients’ cerebral white matter volume from that of healthy controls undergoing scanning on the same machine, whereby the volumes were normalized to each participant’s total intracranial volume. For DTI, variables included the number of tracts with abnormal findings with regard to fractional anisotropy, median diffusivity, or both compared with healthy controls undergoing scanning on the same machine. C, The added value of quantitative (T1-weighted and DTI) beyond qualitative information, that is, the presence or absence of any visible lesion reported by an expert who reviewed all available sequences (T1-weighted, T2-weighted, fluid-attenuated inversion recovery, and susceptibility-weighted imaging and DTI) is compared. AUC indicates area under the curve with 95% CI. P values were calculated using a paired DeLong test for comparing AUCs.