3001

QSM Detects Post-Concussion Changes in Subcortical Gray Matter Susceptibility
Kevin Koch1, Brad Swearingen2, Robin Karr1, Andrew Nencka1, L. Tugan Muftuler2, Timothy Meier2, and Michael McCrea2

1Radiology, Medical College of Wisconsin, Milwaukee, WI, United States, 2Neurosurgergy, Medical College of Wisconsin, Milwaukee, WI, United States

Synopsis

A longitudinal QSM study of sports concussion in 80 injured and control athletes is presented. Regional ROI analysis demonstrated group susceptibility effects that reproduced a previous smaller cohort study finding that QSM diffusely increased in the white matter after sports concussion. In addition, this larger cohort study identified a significant acute trend of decreased susceptibility in sub-cortical gray matter, which is indicative of the calcium influx that is known to occur during the neurometabolic cascade following brain injury. The subcortical gray matter QSM decrease correlated strongly with clinical injury severity metrics.

Introduction

Quantitative MRI is well-suited for potential use as a probe of the subtle neurological changes resulting from mTBI. Previous studies have explored mTBI-induced changes in MR imaging diffusion tensor imaging [1], diffusional kurtosis imaging[2], arterial spin-labeling[3], functional connectivity[4], and quantitative susceptibility mapping[5,6,7].

A recently published study presented widespread acute increases in white matter QSM after sports concussion in cohorts of 27 injured and control subjects[7]. Here, we present results of a similar longitudinal sports concussion study on a larger, independent cohort of 80 injured players and 80 matched controls. The new results reproduce the findings of broad white-matter QSM increases and also identify significant acute decreases in gray matter QSM. The latter finding provides preliminary evidence for a diagnostic imaging indicator of concussion reflecting the well-known calcium influx that occurs within the neurological cascade after traumatic brain injury[8]

Methods

Collegiate and high school football athletes provided written consent (or assent and parental consent if minors) for a study approved by the local institutional human research review board. 78 athletes enrolled in the study were imaged within 24-48 hours after injury, followed by examinations at 8 days, 15 days, and 45 days post-injury.

Imaging was performed on a clinical 3T MR imaging scanner using a 32-channel head receive array. QSM data were collected by saving the raw k-space data from a commercially available SWI application. SWI data acquisition parameters were as follows: in-plane data matrix, 320 256; FOV, 24 cm; slice thickness, 2 mm; TR, 58.6 ms, which allowed for the collection of 4 echoes. QSM were computed using RESHARP[9] background removal and VP-QSM[10].

Subject recruitment for the longitudinal exams was not perfect and quality control efforts excluded several datasets from analysis. The final analysis cohort had injured subject sub-cohorts of [78 82 78 56] at the four time-points, while the control cohort had sub-cohorts of [75 58 58 54].

All QSM maps were registered to a single QSM reference map in MNI coordinate space. A templated neurological QSM stability map was utilized from a previous, independent longitudinal control subject cohort[7] to restrict analysis to regions that had reliable and stable QSM estimates.

ROI analysis within white and subcortical gray matter (including global white and gray) compartments was performed using anatomic segmentations extracted from the Johns Hopkins University and Harvard MNI space atlases.

Mean susceptibility values were computed within each gray and white matter ROI for each subject. Group differences between the means were computed using 2-tailed independent samples t-tests with unequal variances at each visit. In addition, effect sizes between the groups were estimated by computing the Cohen D at each visit. Symptom duration (as a proxy of injury severity) measures were also correlated with the 24-hour susceptibility measurement.

Results and Discussion

Figure 1 provides representative mean QSM maps across the entire control group for two axial slices. Table 1 provides the results of the statistical ROI analyses. Substantial group differences were found in the white matter and gray matter compartments. Several of the white matter compartment changes (gray highlights) were reproductions of a previous similar study on a smaller cohort (N=27). The green highlights were regions that showed new effects in the current larger cohort study, with the gray matter findings being a crucial new discovery. White matter and gray matter directional trends were all consistent, with increases in white matter and decreases in gray matter. Figure 2 provides a box-plot summarizing the longitudinal trends.

Figure 3 provides a scatter plot of the symptom duration metric (subjects > 6 days of symptoms ) vs QSM in the global white matter, while Figure 4 provides the same data for the global gray matter compartment (restricted by the aforementioned externally-derived QSM stability map). Correlations with injury severity were strongest using the displayed window (> 6 days). Lower severity injuries showed increased variation relative to susceptibility measurements. This may be indicative of the limits of the susceptibility measurement's sensitivity to injury severity. In addition to the correlation plots in Figures 2 and 3, QSM maps of the control mean (top) and injury difference (from the control mean) of one subject (circled in scatter plot) are presented. The Pearson correlations were 0.31 (p=0.05) and -0.46 (p=0.003) for the white and gray matter compartments, respectively.

In summary, this study has 1) reproduced identified post-concussion QSM increases in white matter and 2) acute QSM decreases in deep gray matter structures that correlate strongly with injury severity metrics (duration of symptoms). The former finding can be explained by the well-known calcium influx that results from neurological trauma[8].

Acknowledgements

Funded in part by the Department of Defense Head to Head 2 Project.

References

1. Shenton ME, Hamoda HM, Schneiderman JS, et al. A review of magnetic resonance imaging and diffusion tensor imaging findings in mild traumatic brain injury. Brain Imaging Behav 2012;6:137–92 CrossRef Medline

2. Lancaster MA, Olson DV, McCrea MA, et al. Acute white matter changes following sport-related concussion: a serial diffusion tensor and diffusion kurtosis tensor imaging study. Hum Brain Mapping 2016;37:3821–34 CrossRef Medline

3. Wang Y, Nelson LD, LaRoche AA, et al. Cerebral blood flow alterations in acute sport-related concussion. J Neurotrauma 2016;33: 1227–36 CrossRef Medline

4. Mayer AR, Bellgowan PS, Hanlon FM. Functional magnetic resonance imaging of mild traumatic brain injury. Neurosci Biobehav Rev 2015;49:8 –18 CrossRef Medline

5. Liu W, Soderlund K, Senseney JS, et al. Imaging cerebral microhemorrhages in military service members with chronic traumatic brain injury. Radiology 2016;278:536 – 45 CrossRef Medline

6 Lin HH, Liu HS, Tsai PH, et al. Quantitative susceptibility mapping in mild traumatic brain injury. In: Proceedings of the Annual Meeting and Exhibition of the International Society for Magnetic Resonance in Medicine, Honolulu, Hawaii. April 22–27, 2017:2395

7. Koch, K.M., Meier, T.B., Karr, R., Nencka, A.S., Muftuler, L.T. and McCrea, M., 2018. Quantitative susceptibility mapping after sports-related concussion. American Journal of Neuroradiology, 39(7), pp.1215-1221.

8. Giza, C. C., & Hovda, D. A. (2001). The Neurometabolic Cascade of Concussion. Journal of athletic training, 36(3), 228-235.

9. Sun H, Wilman AH. Background field removal using spherical mean value filtering and Tikhonov regularization. Magn Reson Med 2014;71:1151–57 CrossRef Medline

10. Anderson C, Nencka A, Muftuler T, et al. Volume-parcellated quantitative susceptibility mapping.In: Proceedings of the Annual Meeting and Exhibition of the International Society for Magnetic Resonance in Medicine, Singapore. May 7–13, 2016:1108

Figures

Figure 1: Mean QSM maps averaged over the control cohort (all time points) for two axial slices.

Table 1: p-values of student t-tests comparing the injured and control groups and Cohen’s D effect sizes (injured vs control) for each time point across regional brain ROIs. Gray highlighted ROIs are those with reproduced effects from the previous smaller cohort study [7]. All of the ROI trends with significant group differences from the previous smaller cohort study [7] were reproduced in the present study. Green highlighted ROIs are those which have added effects in the new study. The most significant group differences are in boldface.

Figure 2: Box plots of the QSM values within global white and subcortical gray matter ROIs, along with two representative ROIs from each compartment. Injured and control groups are shown for each of the 4 time points (24-48 hours, 8 days, 15 days, 45 days). Note the progression of injured QSM means towards the control equilibrium value with increasing time from injury.

Figure 3. Correlation of symptom duration metric with acute (1st time point) white matter susceptibility for individual subjects. Regression analysis showed a moderate correlation between these two metrics within the group. QSM are shown to reflect the control average within the white matter ROI and the difference of a subject with a long days-injured metric. The QSM increases within the white matter compartment are clearly evident.

Figure 4. Correlation of symptom severity metric with acute (1st time point) deep gray matter susceptibility for individual subjects. Regression analysis showed a strong negative correlation between these two metrics within the group, indicating a reduction of susceptibility with increasing injury severity. QSM are shown to reflect the control average within the gray matter ROI and the difference of a subject with a long days-injured metric (same subject as Figure 3). The strong QSM decreases within the gray matter compartment are clearly evident.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)
3001