Investigation of structural, functional, and metabolic magnetic resonance biomarkers in the spinal cord and brain after spinal cord injury

Abstract

Spinal cord injury (SCI) is a devastating and life-changing incident that leads to immediate and usually persisting motor, sensory, and autonomic dysfunctions, such as impairments in bladder and bowel function. These are often accompanied by the development of chronic pain and psychological concomitants, for instance depression or cognitive decline, which are negatively associated with rehabilitation and quality of life after SCI. Trauma-induced mechanical disruption and dislocation of the spinal vertebral column lead to an immediate damage of the spinal cord which is characterized by neuronal and glial cell death and disruption of axons and the spinal vasculature. This primary injury initiates a multifaceted cascade of secondary injury processes which is hallmarked by the recruitment of reactive astrocytes forming a glial scar to encapsulate the toxic environment of the lesion core that represents an inhibitory barrier to axonal regrowth. Injury-induced neurodegenerative and neuroinflammatory processes can last for years, spreading to supralesional levels and exacerbating the initial damage. Early post-injury treatment includes neuroprotective and neuroregenerative medication, aiming at preventing secondary damage and increasing the intramedullary regenerative potential, beyond endogenous plasticity and attempts of repair. Evaluation of the efficacy of distinct treatment approaches in interventional trials calls for assessments that can sensitively detect changes of clinically relevant markers. Clinical examinations usually lack the necessary sensitivity as they use ordinal scales for the classification of patients, and are also unspecific to structural correlates underlying the clinical deficits. Magnetic resonance imaging (MRI) is a non-invasive and objective method which enables the qualitative and quantitative assessment of macroscopic and microscopic changes within the neural tissue after SCI. It can thus improve our understanding of pathological primary and secondary injury processes along the whole neuraxis post-SCI and how these change during the course of recovery and relate to clinical outcomes. Previous studies found MRI-derived cervical spinal cord atrophy above the lesion level of up to 30% at 5 years post-injury. Volumetric changes in clinically relevant motor and sensory regions in the brain were also detected, which were paralleled by decreased tract integrity, pathological reductions in myelin-sensitive markers, and increases in neuroinflammatory metabolites in chronic SCI. The overall goal of this thesis was to improve our understanding of conventional and advanced MRI-derived structural, molecular, and functional changes at the lesion site and above the lesion following SCI and to explore their value as biomarkers for characterization of lesion severity and neurological disability, prediction of clinical outcome, and potential implementation as surrogate markers in interventional trials. In the first study, we investigated the longitudinal spatiotemporal dynamics of the lesion extent and the width of injury-spared tissue bridges in paraplegic SCI patients by means of conventional MRI and how baseline lesion characteristics relate to clinical outcome. At 1 month following SCI, after widespread edema and hemorrhage have resolved, midsagittal lesion segmentation was reliably and accurately performable. Over the course of 2 years, the lesion extent slightly decreased while tissue bridges' width increased, possibly due to the cystic cavity getting fully demarcated at around three to four weeks post-SCI. Crucially, patients of traumatic or ischemic etiology showed similar lesion evolutions and courses of neurological recovery, its magnitude being associated with the size of the lesion and spared tissue bridges in both patient groups. The second study aimed at exploring the value of conventional MRI-derived spared tissue bridges for prediction of neurological recovery at 1 year post-SCI, focusing not only on the width, as in most previous studies, but also on the localization of tissue bridges. Larger 1-month ventral and dorsal tissue bridges demonstrated a predictive relationship with higher American Spinal Injury Association Impairment Scale (AIS) grade conversion rates at 1 year follow-up. Using unbiased recursive partitioning conditional inference tree (URP-CTREE) analysis, ventral tissue bridges were identified as predictors of pin prick and motor scores and dorsal tissue bridges as predictors of functional independence scores, adding value to baseline clinical scores for improved stratification of patients into more homogenous and recovery-specific sub-groups. In the third study, we explored the role of spared tissue bridges for the development of neuropathic pain (NP), which represents one of the main burdens following SCI and is often refractory to treatment. We found that 1-month ventral tissue bridges were larger in patients with NP when compared to pain-free SCI patients. Representing a proxy for spinothalamic tract function, ventral tissue bridges were furthermore predictive of 1-year pin prick score and pain intensity. Crucially, the URP-CTREE algorithm identified a critical cut-off in ventral tissue bridges’ width (≤2.1mm or >2.1mm), dividing patients into sub-groups of higher and lower scores of pin prick sensation and pain intensity. These two studies highlight the potential of spared tissue bridges to improve prediction of outcome and optimize stratification of SCI patients in interventional trials. The fourth study aimed at identifying neurochemical biomarkers of chronic NP after SCI above the lesion level by means of advanced and quantitative proton magnetic resonance spectroscopy (MRS). At the remote cervical spinal cord at level C2/3, we revealed a pain-specific molecular fingerprint of increased neuroinflammatory but similar neurodegenerative metabolite concentrations in patients with NP when compared to pain-free SCI patients. Increased neuroinflammatory metabolite levels were furthermore associated with less cord atrophy and higher pin prick scores, which might represent molecular surrogate markers of NP and potentially provide new NP treatment targets. In the fifth study, we investigated whether the right hippocampus is pathologically affected in chronic SCI patients, using a multimodal approach including structural MRI, quantitative MRS, and memory performance. Previous studies reported that SCI patients suffer from cognitive impairments including deficits in learning and memory, possibly caused by brain inflammation. In this study, no secondary supralesional changes were detected in terms of an altered metabolic profile indicative of potential neurodegeneration or neuroinflammation, hippocampal atrophy, or memory impairments. SCI thus does not seem to directly affect the hippocampal integrity and post-SCI cognitive deficits might be attributable to secondary concomitants of SCI or methodological study differences. In conclusion, the studies in this thesis made use of conventional and advanced MRI for characterization of the primary injury at the focal lesion site and investigation of far-reaching structural, molecular, and functional biomarkers above the lesion level which are indicative of remote secondary injury processes following SCI, such as neurodegeneration, neuroinflammation, or reorganization. We were able to identify preserved tissue bridges adjacent to the lesion as predictive neuroimaging biomarkers of tract-specific neurological recovery and pathological somatosensory function in terms of NP. This thesis expands previous knowledge of secondary neurodegenerative and demyelinating processes at the remote cervical spinal cord and reveals a NP-related neuroinflammatory metabolic profile in the upper cervical cord in chronic SCI patients by means of advanced quantitative MRS. At the same time, the hippocampus was not found to be pathologically affected on a structural or molecular level after SCI. Conventional and advanced MRI-derived quantitative biomarkers have the potential to detect subclinical changes post-SCI and to be used as surrogate markers in interventional trials for monitoring of recovery processes and treatment effects with the ultimate goal of improving personalized patient care and achieving clinically meaningful recovery following SCI. Adding complementary value to neurological and electrophysiological assessments, clinically applicable conventional and advanced MRI might be especially powerful when used synergistically and acquired longitudinally across different centers. This could help us to improve our understanding of neuropathological and reorganizational processes along the neuraxis post-injury and to identify the most reliable and sensitive neuroimaging biomarkers to evaluate the efficacy of interventions aiming at reducing neurodegeneration and enhancing recovery following SCI.