The initial insult after traumatic brain injury (TBI) is followed by a range of secondary tissue-level changes associated with a broad spectrum of symptoms and disabilities¹. Quantitative susceptibility mapping (QSM) is sensitive to the tissue magnetic susceptibility and can reveal changes in tissue composition of diamagnetic and paramagnetic substances. After TBI, iron accumulates with bleeding², calcifications appear in areas of extensive neurodegeneration³, and white matter undergoes demyelination². Therefore, the purpose of this study was to investigate changes in the susceptibility distribution in the rat brain after TBI.Lateral fluid percussion injury was performed in adult male Sprague-Dawley rats (n(TBI)=5; n(sham)=4). Six months after TBI, rats were perfused with 4% paraformaldehyde. The rat brains were imaged ex vivo at 9.4 T using an Agilent console and a 3D multi-echo GRE sequence with TR=130 ms, flip angle=25˚, BW=50 kHz, 6 echoes with TE=3:4:23 ms, matrix size=150×192×256, 100 μm isotropic resolution, and a scan time of 106 min. For QSM, field maps were estimated by fitting the complex data over TEs⁴, followed by Laplacian unwrapping and SHARP filtering (threshold=0.05, kernel diameter=9 voxels) to remove residual wraps and background fields. Susceptibility maps were then calculated using TKD⁵ (threshold=2/3) and correction for susceptibility underestimation⁶. Regions of interest (ROIs) were manually defined on magnitude images of 3 coronal slices in selected brain areas. In a separate group of rats 6 months post-TBI (n=8) typical alterations in myelin content, cellularity, iron content and calcifications were evaluated using gold chloride, Nissl, Perls´ and Alizarin Red stainings, respectively.

Fig. 1 shows ex vivo QSM maps in a sham and a TBI rat.

White Matter

The ipsilateral corpus callosum (cc) in TBI rats showed more diamagnetic susceptibility values than sham rats (Fig.2A). Changes in susceptibility were also found in the contralateral cc due to the continuity of this structure between hemispheres (Fig. 3A). Histology revealed a decreased cc thickness due to loss of myelinated axons. However, in the remaining cc myelin content was not markedly changed. We also detected gliosis and sparse microbleeds along the ipsilateral and contralateral cc in most of the animals after TBI (Fig.3G). One animal showed increased cc susceptibility due to a higher number of bleeds (Fig.2A).

The ipsilateral internal capsule in TBI rats showed more paramagnetic susceptibility values than in sham rats (Fig. 2B). Histology revealed thinning of this structure together with decreased myelin density (Fig.3B) as well as iron deposits in the ipsilateral internal capsule and neighboring areas (Fig.3D).

Cortex

Surprisingly, the cortex close to the lesion site in TBI rats appeared to be more diamagnetic than both the contralateral hemisphere in the same animals and in shams (Fig.2C). As we frequently observed paramagnetic microbleeds below the perilesional cortex, we cannot completely rule out the contribution of residual non-local field effect that was not fully resolved by the QSM processing. There was loss of myelinated axons throughout the cortex despite a low overall cortical myelin content (Fig.3B). There was diffuse iron accumulation, mainly in deep cortical layers (Fig.3D). Nissl staining also showed increased cellularity in perilesional cortex, which can be attributed to gliosis.

Thalamic Nuclei

The two thalamic nuclei included in our analysis, ventrobasal complex and laterodorsal nucleus, are composed of a mixture of white matter fibre bundles and grey matter. Most of the TBI rats showed more paramagnetic susceptibility values in the ipsilateral ventrobasal complex compared to the contralateral ROI (Fig.2D). There was one animal which showed more diamagnetic values. In this thalamic area the contribution of myelin loss (Fig. 3B), iron accumulation (Fig.3D), and calcifications (Fig.3F) are expected to influence the susceptibility. Susceptibility values were variable in the laterodorsal thalamic nucleus (Fig.2E). This area showed extensive neurodegeneration, gliosis and calcifications (Fig.3F).

We tested the sensitivity of QSM to pathological processes including demyelination, iron accumulation and calcifications in a rat model of TBI. As expected, several pathological processes with potentially counteracting influences on magnetic susceptibility overlap in the same tissue. Furthermore, caution should be exercised in reconstruction and interpretation of QSM maps as TBI pathology involves focal spots of drastically altered magnetic susceptibility, which may lead to residual non-local effects. Validation of the results requires detailed histological characterization at the level of individual animals. This kind of comparative study will lead to a better understanding of the contributions to altered magnetic susceptibility during progressive brain pathology and will pave the way for utilization of QSM as a part of the non-invasive imaging toolbox available to study the complex pathology of TBI.