In vivo spinal cord imaging may help to diagnose patients with neurodegenerative disorders. In addition, spinal cord quantitative MRI (qMRI) may be used to predict prognosis and monitor patients' response to treatment in clinical settings and in treatment trials. For example, European guidelines about the use of neuroimaging in the management of amyotrophic lateral sclerosis (ALS) have recommended incorporating qMRI in new clinical trials as exploratory outcomes. In addition, their use in the research field may allow the investigators to obtain new insights into disease pathophysiology. In this chapter, we will discuss the results of studies that have applied spinal cord qMRI to clinical applications, including multiple sclerosis, neuromyelitis optica, spinal cord injury, cervical spondylitic myelopathy, syringomyelia, ALS, spinal tumours, spinal vascular anomalies, and more rare spinal cord disorders.
Keywords: Diffusion weighted imaging, Diffusion tensor imaging, tractography, proton spectroscopy, magnetisation transfer, neurological disease, spinal cord, NMO, ALS, Brachial plexus injury, Cervical spondylitic myelopathy, Spinal tumours, spinal AVM.
Section 1: Introduction
What is quantitative MRI (qMRI)?
Get your grade
or your money back
using our Essay Writing Service!
As opposed to qualitative MRI (e.g., T1-, T2-weighted), quantitative MRI aims at providing values that are intrinsic to the tissue properties. Advantage is that it provides absolute and normative values that could be used for diagnosis, prognosis, multiple-site studies and ultimately in clinical trials.
The spinal cord is affected by a large number of disease processes, which lead to diverse clinical presentations and pathogenesis, affecting young and older people alike. Injury of the spinal cord often causes disabling symptoms, including limb weakness, sensory disturbance and sphincter dysfunction. The spinal cord is a more challenging region of interest than the brain to image due to the smaller cross sectional area of the spinal cord, motion artefacts from cerebrospinal fluid (CSF) flow with each cardiac and respiratory cycle and susceptibility artefact from surrounding tissues1-3. Advances in neuroimaging techniques and post-processing have allowed for recent progress to be made and a rise in the number of studies investigating spinal cord diseases using MRI.
Application of quantitative MRI (qMRI) techniques in clinical imaging studies of spinal cord disease are providing insights into the microstructural tissue damage and changes in metabolic profiles of spinal tissue in a growing number of neurological diseases, and are showing promise as potential biomarkers of disease progression.
Thoughtfully designed mechanistic MRI studies can complement histopathological studies in understanding pathophysiological processes occurring in vivo, helping to identify important mediators of disease and therefore inform rational drug design. The insights gained from recent studies into cellular and pathophysiological abnormalities in multiple sclerosis (MS) have shown promise and may be valuable in future therapeutic trials of neuroprotective agents4. It is also predicted that qMRI will play an important role in drug development, since qMRI-derived measures can be used as biomarkers of disease progression and to monitor treatment response. A recent study showing that longitudinal changes in whole-brain and tract-specific diffusion tensor imaging (DTI) indices and magnetization transfer ratio (MTR) can be reliably quantified, suggests that clinical trials using these outcome measures are feasible5. Whilst similar, longitudinal studies in patients with spinal cord diseases are currently lacking, this will no doubt be the focus of future work. In fact, it is essential that reliable imaging biomarkers of the spinal cord are validated to prepare us for the emergence of neuroprotective drugs. Such techniques should enable researchers to measure disease activity accurately and reproducibly in clinical trials, as well as risk stratify patients on entry into a trial.
This chapter will briefly review the qMRI techniques most commonly applied to the spinal cord (Diffusion Tensor Imaging, MS Spectroscopy and Magnetization Transfer Imaging, which are also discussed in following chapters), and then focus on reviewing data from qMRI studies in patients with neurodegenerative and traumatic pathologies, and in animal models (all the studies discussed are summarised in Table 1). The clinical and pathophysiological significance of the results of these studies will be discussed, and future directions of research will be proposed.
Most Commonly used qMRI techniques
Diffusion imaging is sensitive to microstructural disuse damage, axonal orientation and demyelination
Magnetization transfer can give quantitative information regarding myelination
Functional MRI measures neuronal activity by detecting associated changes in blood flow.
MR Spectroscopy able to provide metabolic profiles of tissues.
Atrophy Measurements, especially when repeated over time can give information about axonal loss
Section 2: MR Imaging Techniques
Always on Time
Marked to Standard
2.1: Diffusion Tensor Imaging
Diffusion tensor imaging (DTI) is an MRI technique capable of characterising water molecule movements in tissue. Fractional anisotropy (FA), which is derived by the DTI, can be estimated within spinal cord white matter tracts, and reflects the underlying tissue microstructure6. Furthermore animal studies have shown that demyelination often leads to an increase in other DTI-derived indices, radial diffusivity (RD), whilst axonal loss leads to a rise in axial diffusivity (AD)7 (see chapter 3.1). Obviously it is difficult to extrapolate the findings from these animal studies to patient studies, so findings on axial and radial diffusivities in patient studies must be interpreted with caution. However, the spinal cord DTI-derived parameters have the potential to be used as biomarkers of both myelin and axonal integrity, and have been demonstrated to significantly change over time after an acute injury, such as an acute demyelinating lesion8. In addition to FA, RD and AD, the mean diffusivity (MD), can be obtained. This parameter is very sensitive to a general change in tissue microstructure that allows water molecules to move less or more freely; for example, a reduction of the MD has been detected following acute neurological insults, but in chronic neurological disease, or where cellular necrosis leads to increased membrane permeability, a relative increase in the MD value can be seen9.
FA, RD, AD, and MD can be measured within the spinal cord using regions-of-interest, which can be drawn on the basis of the user's anatomical knowledge. Additionally, they can be measured within tracts, which are reconstructed using fibre tracking (FT) algorithms. Probabilistic tractography algorithms allow the user to obtain an estimate of white matter connectivity within the spinal cord tracts, which may change with pathology. Even a straightforward visualisation of major white matter tracts within the spinal cord has potential clinical applications in the pre-operative surgical planning. Detailed imaging of the fibre tracts pre-operatively can provide information regarding the integrity of the tracts and allow the clinician to predict the potential benefit from surgery more accurately. Slow growing spinal tumours may efface tracts, leaving them relatively intact, whilst more aggressive, faster growing tumours are more likely to cause significant structural disruption to the tracts. The risks and the benefits of surgery can therefore be more accurately estimated. Pre-surgical assessments of this type will empower clinicians and patients to proceed with surgery only when there is a favorable risk-benefit ratio.
Spinal cord DTI studies in patients have been performed with the assumption that DTI measures do not change across spinal cord levels. Wheeler-Kingshott et al.1 confirmed a uniform MD along the cervical spinal cord, but found a higher FA in the mid and lower section of the cervical cord compared with the upper cervical cord. Regional differences in DTI metrics of the cervical cord have also been reported by Mamata et al.,10 who reported higher mean values for ADC and FA in the upper cervical cord (C2- C3) than in the lower level (C4- C7) in healthy volunteers. These findings suggest that regional DTI measures should be compared between patients and controls whenever possible, since averaging values of DTI measures may dilute the effect of focal pathology.
2.2: Proton Magnetic Resonance Spectroscopy
Over the past decade, developments in imaging acquisition and post-processing, together with the availability of high and ultra-high field scanners, have made it possible to use MRS in clinical studies of spinal cord pathology, giving insights into pathogenesis. Reliable quantification of metabolites from the spinal cord using 1.5 tesla (T) scanners has been limited to a few metabolites. Commonly quantified metabolites in spinal cord disease include N-acetylaspartate (NAA), Choline, and Creatine. High and ultra-high field strength scanners allow increased separation of metabolite peaks, with the potential to study metabolites with more specific relevance to the pathogenesis of neurological diseases, such as glutamate, GABA, and myo-inositol. Whilst many of the changes seen in commonly measured metabolites may not be disease specific, changes in commonly measured metabolites can be seen in early neurological disease13 making them potentially useful in early diagnosis, despite lacking specificity.
N-acetylaspartate (NAA) is synthesised by neuronal mitochondria14, and commonly used as a marker of axonal integrity in neuroimaging studies. More recently, studies modelling NAA concentrations in the spinal cord with other markers of axonal integrity, such as cross sectional cord area, demonstrate that mathematically derived estimates of mitochondrial function can be made from spectroscopic data3. Development of such imaging biomarkers of mitochondrial function has wide implications for studying a large number of neurological diseases in which mitochondrial dysfunction is thought to be important.
This Essay is
a Student's Work
This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.Examples of our work
NAA concentrations can be expressed as an absolute value, but more commonly are expressed as a ratio with creatine or choline. The creatine signal which is a composite peak of creatine and phosphocreatine is often used as the reference signal as levels are thought to be quite constant in the human nervous system. In neurological disease however, changes in the resonance intensity of choline has been suggested to reflect increases in the steady state levels of membrane phospholipids released during myelin breakdown as is seen in active demyelinating disease15.
Lactate is produced as a by-product when cells within the nervous system respire under anaerobic conditions. Elevations in brain lactate have been detected using MRS in cerebral ischemia16, brain tumours17 and mitochondrial disease18. Although it is thought lactate may be a relevant marker in spinal cord pathology, it is likely to be produced in much smaller concentrations, making its detection in healthy subject difficult or not possible19. As with DTI, it is possible that the anatomical level of the region of interest within the spinal cord is an important consideration with MRS. Edden et al.20 have demonstrated that metabolite concentrations vary between the upper cervical cord and the medulla..
2.3: Magnetization Transfer Imaging
Magnetization transfer imaging (See chapter 3.4) is based on the interaction between hydrogen protons bound to macromolecules, such as those associated with lipid/lipoproteins, and the free protons normally imaged by MRI. Magnetisation Transfer Ratio (MTR), obtained by magnetisation transfer imaging, can be used as an indirect marker of demyelination, and possibly axonal loss In patients with MS, high resolution magnetization transfer measurements from the spinal cord have demonstrated that it is possible to assess demyelination of specific spinal pathways with good accuracy23. However, in neurodegenerative diseases other than MS, the use of MTR imaging has largely been confined to studies of brain pathologies, and a very few studies have focused on the spinal cord. Brain MTR has been incorporated as an exploratory end point to assess treatment efficacy in large multicentre trials, and similar studies of spinal cord disease are still awaited. Where relevant studies of spinal cord disease have used MTR, they will be discussed in this chapter.
Section 3: Application of Quantitative MRI Techniques to spinal cord disease.
This section will give an overview of some applications of qMRI in the spinal cord, including a wide spectrum of clinical applications. Given the high number of studies of qMRI in multiple sclerosis and the potential relevance of qMRI in spinal cord injury, these two applications will be expanded in subsequent chapters: 1.2 (demyelinating diseases) and 1.3 (trauma).
3.1 Multiple Sclerosis
Multiple Sclerosis (MS) is a disorder of the central nervous system with complex histopathology. It is the commonest cause of progressive neurological disability affecting young people in the western world and can present with either a relapsing, remitting clinical course, or with progressive accumulation of neurological symptoms and disability from onset26. Pathologically, MS is characterised by inflammation, demyelination, remyelination, axonal loss and gliosis.
Diagnosis is based on the presence of neurological signs and symptoms disseminated in time and space, and although the diagnosis of MS may be made on clinical grounds alone, magnetic resonance neuro-imaging has become increasingly important27.
3.1.1 Relapsing Remitting Onset Multiple Sclerosis
clinically by relapsing, remitting neurological symptoms affecting and
3.1.2 Primary Progressive Multiple Sclerosis
Primary progressive multiple sclerosis (PPMS) accounts for 10 to 15 percent of multiple sclerosis and is characterised by a clinical course which is progressive from onset without an initial period of clinical relapses with remissions28. Although PPMS patients are phenotypically different from RRMS, are more likely to be older than those with RRMS, and incidence in males and females is similar.
The relative rarity of PPMS has complicated the study
3.6 Neuromyelitis Optica
Neuromyelitis optica (NMO) is an inflammatory and demyelinating disease of the central nervous system, characterised by the preferential involvement of the spinal cord and optic nerve29. NMO is clinically and immunologically distinct from MS. Brain MTR and DTI studies on 1.5T scanners have reported abnormalities within normal appearing brain tissue of patients with NMO suggestive of microstructural tissue damage, although there is no agreement between studies, as other papers have reported normal imaging parameters in the normal-appearing brain tissue33. An investigation of the normal-appearing white matter of the spinal cord using 3T DTI demonstrated significant abnormalities in DTI metrics within the spinal cord in a small cohort of 10 NMO patients34. Compared to healthy controls, patients had significantly reduced FA, and increased MD and RD. There were strong correlations between DTI metrics and clinical scores within the NMO group, suggesting that DTI is a sensitive marker of injury within the major long spinal tracts. The increase in RD, which is commonly seen in patients with demyelinating disease7, was responsible for the decreased FA, implying that demyelination, rather than axonal injury, is the major underlying pathological process in this patient group34. However, the interpretation of the pathological correlates of RD and AD in the spinal cord of patients is not so straightforward as that in animal studies, and therefore caution is needed when discussion the underlying tissue abnormalities of the directional diffusivities35. In addition, the groups were not closely age matched in this study which may partly explain the group differences in FA36.
3.1: Acute Trauma
3.1.1: Acute spinal cord injury
Traumatic spinal cord injury (SCI) is a devastating condition which primarily affects young males with an annual incidence of 15-40 cases per million37-39. The commonest mechanisms of primary injury in human spinal cord injury (SCI) are compression, contusion, laceration, transection, and traction of the spinal cord40. Following the primary insult, several pathophysiological mechanisms produce secondary injury, including spinal cord oedema, ischaemia, free radical damage, electrolyte imbalance, excitotoxicity, inflammation and apoptosis. Qualitative spinal cord MRI findings, determined by conventional T2-weighted imaging, such as haemorrhage, oedema, and swelling, are commonly seen in complete motor and sensory SCI and are predictive of a poor neurological outcome42.
In a retrospective study of 20 victims of blunt force trauma with cervical spine injuries, diffusion imaging showed a significant reduction in whole cord ADC in patients compared with healthy controls. FA values were significantly reduced at the site of injury, but no significant difference in whole cord FA was found between patients and controls43. The greatest differences in whole cord ADC and FA were seen in those patients with haemorrhagic cord contusions. In a similar series, Cheran S et al.44 found significant reductions in MD and AD throughout all regions of the cervical cord following blunt trauma, and maximal reductions in FA at the site of injury. RD was only increased in patients with non-haemorrhagic contusions. American Spinal Injury Association (ASIA) clinical motor scores correlated positively with MD, AD and RD, and there was a negative correlation with FA (Figure 1). The data coming from both studies indicate that DTI-derived parameters are sensitive to spinal cord contusion and that MD and AD are the most sensitive markers of extent of cord injury in this group of patients. In patients with non-haemorrhagic contusions, all 4 DTI parameters were predictive of disability, with MD again being most significant (Figure 1). In both studies, the greatest DTI parameter changes were seen in patients with haemorrhagic contusion. The lack of correlation in this group with ASIA scores may be largely explained by the presence of large volumes of haemorrhage at the injury site, which may mean that the blood content of the extracellular space is contributing to diffusion anisotropy, which may not be truly reflective of axonal injury.
Both studies are limited by small cohort sizes but some findings did reach statistical significance. Larger, prospective, follow-up studies are now needed to corroborate the findings to date and assess how long term outcomes relate to changes in DTI parameters at time of injury.
Single voxel MRS of the cervical cord in patients with chronic whiplash (>6 months post-injury) was able to demonstrate reduced NAA/Cr ratios when compared with healthy controls45 suggesting underlying axonal injury. Reductions in NAA/Cr ratio of a similar magnitude in patients when compared with controls were seen in some studies of cervical myelopathy19. These findings support the hypothesis that, although neurological symptoms and signs are not a usual feature of whiplash injuries, a subclinical (spinal) insult, leading to neuronal loss in the cord, may occur. This neuronal damage observed within the spinal cord may reflect a persistently abnormal afferent input, following injury to peripheral structures45.
3.1.3: Brachial Plexus Injury
Brachial plexuses injuries can occur as a result of trauma to the shoulder, inflammation within the nerves making up the brachial plexus, or due to tumour invasion. In young adults, avulsion injuries, in which the nerve roots are torn from the spinal cord, are most commonly caused by motorcycle accidents. Injuries of this type are classed as "longitudinal spinal cord injuries", due to the damage that occurs to the cord46. The nerve roots forming the brachial plexuses are avulsed from their origin causing variable degrees of disability in the affected limb, with a plegic, anaesthetised upper limb in the most severe of cases, involving all 5 dorsal and ventral roots. In experimental models of lumbar root avulsion, motor neurone cell death and Wallerian degeneration are described47. Re-implantation of avulsed spinal ventral roots has been show to enable significant and useful regrowth of motor axons in both experimental animals and in human clinical cases, rescuing lesioned motor neurones from death48.
Kachramanoglou C et al49 carried out a 1H-MRS study of the upper (C1- C3) spinal cord, and demonstrated that patients undergoing lower cervical root re-implantation surgery following brachial plexus avulsion showed a significantly higher myo-Ins/Cr ratio when compared to healthy controls (Figure 2); this result suggests that a reactive, gliotic process may occur above the site of the lesion, probably in response to the Wallerian degeneration of avulsed neuronal fibres, which has been demonstrated in animal model of root avulsion47. Interestingly, the finding of higher myo-Ins/Cr ratio was associated with greater disability of the arm, higher pain scores and shorter time from injury. The authors proposed that gliotic changes might therefore normalise over time, reflecting re-organization within the cord.
In an animal experiment, Risling M et al48 demonstrated that gene expression varied between Sprague-Dawley rats, who underwent root implantation following avulsion of the L5 ventral root, and those that didn't (Figure 3). Whilst cell death gene expression is similar in the two groups, genes related to neurite development and neurogenesis show a much more prominent response after re-implantation than after avulsion only. Significantly, the inflammatory response is more obvious after avulsion than after replantation, which supports the explanation of the spectroscopy findings of Kachramanoglou C et al49. Although the MRS study did not demonstrate lower NAA concentrations in the patient population, reflecting axonal loss within the spinal cord, it may be that these changes would only be detectable in a control group of patients with chronic brachial plexus avulsion who did not undergo root re-implantation. Future studies will address this important question, especially in view of proposing an imaging marker for the forthcoming clinical trials with stem cells transplantation in patients with brachial plexus avulsion50.
3.2: Chronic Spondylitic Myelopathy
Cervical spondylitic myelopathy (CSM) is a common degenerative disease, which occurs most frequently in elderly patients and is characterised by intervertebral disc degeneration. As disc degeneration occurs, mechanical stresses result in osteophytic bars, which form along the ventral aspect of the spinal canal, causing narrowing of the spinal canal; this, in turn, produces myelopathic symptoms. Conventional MRI is widely used in the clinical setting, and may show narrowing in disc spaces, ligament thickening, and narrowing of the canal with T2 high signal change in the cord at the level of narrowing53. Conventional MRI, however, has the drawback of being insensitive to pathological changes and T2 hyperintensity is only seen in between 15 - 65% of patients53-56, leading in some cases to delay in diagnosis and management.
Several studies have demonstrated that diffusion imaging is able to detect abnormalities in CSM. Most studies to date have shown an increase in ADC and decrease in FA at the level of spinal canal narrowing. Sixteen patients with symptoms and signs of purely cervical myelopathy in the early stages of disease, prior to the development of signal changes on T2-weighted sequences, underwent DTI imaging at 3T60. Similar to previous studies, an increase in ADC and decrease in FA at the site of stenosis was observed in all 16 patients. However, a drawback of this study was the use of non-stenotic segments of the patients' spinal cord as controls instead of healthy controls. As discussed earlier, it has previously been reported that mean values for ADC and FA are higher in the upper cervical cord (C2- C3) than in the lower level (C4- C7) in healthy volunteers10. However, considering the large body of evidence obtained from brain DTI studies that highlights the ability of DTI to detect abnormalities which are not visible on conventional MRI, it is expected that DTI metrics of the spinal cord reflect pathological abnormalities prior to (and in the absence of) conventional MRI T2 changes in this patient group. This must now be confirmed with further follow-up work, looking for future T2 hyperintensity on follow-up scans. If it is established that the higher sensitivity seen with DTI when compared with conventional imaging is reproducible across different patient's groups and centres, it is likely to lead to earlier diagnosis and management of this disabling condition and could potentially be used to predict the outcome of conservativeÂ treatment60.
MR spectroscopy data from patients with CSM is limited. A single study performed on a 1.5T scanner reports significantly lower NAA/Cr ratios in patients with CSM than in healthy controls, suggesting significant neuroaxonal loss and dysfunction in this group of patients19. The authors also report the presence of a Lactate peak on the spectrum obtained from nearly half of patients with evidence of T2 hyperintensity on structural imaging, but only 1 (13%) of the 8 patients without a T2-weighted signal abnormality. The presence of this Lactate peak is in keeping with pathological cellular ischemia which is thought to play a role in the pathogenesis of CSM. A trend for lower NAA/Cr ratios in CSM patients with a positive Lactate signal than in those without it was detected, but it did not reach statistical significance19.
Syringomyelia is characterised by the cavitation of the central spinal canal, which slowly expands causing compression of the spinal cord parenchyma, subsequently producing neurological symptoms. Syringomyelia can affect any of the long tracts within the spinal cord, but most commonly affects the spinothalamic tracts, and typically produces a loss of pain and temperature sensation in a "cape distribution" over the back and shoulders. Expansion of the syrinx can produce symptoms of limb weakness, neuropathic pain, sensory loss, bowel, bladder and autonomic disturbance.
A small number of studies have assessed the spinal somatosensory pathways with diffusion imaging in patients with syringomyelia61-63. Hatem et al. used diffusion tensor imaging with three-dimensional fibre tracking (DTI-FT) and demonstrated that patients with cervical syrinx and established thermatosensory impairment in the hands have lower FA in the somatosensoy fibres than healthy controls61; in patients, the mean FA at two levels within the cervical cord (C3-C4 and C6-C7) correlated with both clinical and electrophysiological measures of sensory deficit61. ADC measures in the same patient group were less clinically meaningful. These findings suggest that an objective and quantitative assessment of spinal somatosensory system dysfunction is possible with advanced imaging. Similar patterns of FA reductions were found by Roser et al, who also showed that FA values normalise beyond the borders of the lesion63, suggesting the white matter tracts are preserved distally from the site of syringomyelia.
A recent study assessing the neuropathic pain and functional alterations of the sensory tracts within the spinal cord showed that patients with syringomyelia, with or without neuropathic pain were indistinguishable on the basis of quantitative sensory testing alone, and DTI-FT analyses; however, in patients with neuropathic pain, a greater pain intensity correlated with more abnormal MRI measures of diffusion indices62. In particular, higher average daily pain intensity was correlated with a lower FA and lower number of reconstructed nerve fibres from tractography. In contrast, higher paraesthesia/dysaesthesia correlated with a greater number of reconstructed nerve fibres. The authors have chosen to use a single spinal level for DTI measures (C3-C4) due to physiological variations in DTI metrics at different spinal segments, which have been mentioned above. Although methodologically sound, three of the patients studied had syrinx's which did not involve this anatomical level which may have caused some bias in the results. Allowing for this, these results do seem to suggest that the degree of damage within the spinal white matter, and not grey matter alone, is relevant to both the type and intensity of pain experienced by patients with syringomyelia.
3.4 Amyotrophic Lateral Sclerosis
Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease which leads to degeneration of the upper and lower motor neurones in the ventral horn of the spinal cord. Patients ultimately develop a combination of both upper and lower motor neurone signs with a progressive loss of bulbar and limb function. The vast majority (~ 95%) of cases, occur sporadically with an incidence in Europe of 2-3 cases per 100,000 individuals in the general population, and an overall lifetime risk of developing the disease of 1:40064. The inherited form, familial ALS, accounts for approximately 5% of all cases and is associated with hundreds of gene mutations65, with over 100 mutations of the human superoxide dismutase (SOD1) gene linked with ALS. Mitochondrial dysfunction, the generation of free radicals and impaired calcium handling are thought to be important mechanisms in ALS and more recently, two specific mutations in mitochondrial genes, COX1 and IARS2 have also been reported to cause ALS like syndromes65. The extent of mitochondrial dysfunction within spinal neurons in ALS was reported by Keeney and Bennett66 who demonstrated that within ALS spinal neurons there is reduced mtDNA gene copy numbers and increased mtDNA gene deletions.
MR spectra can reliably be obtained from the cervicalÂ spinal cordÂ inÂ ALS and have been proposed as a potential biomarker of disease progression. Clinical MRS studies on 3T clinical scanners have reported a number of neurometabolic abnormalities, giving insights into pathogenesis. Reductions in NAA/myo-Ins and NAA/Cho ratios have been reported in the cervical spines of patients with ALS, as well as pre-symptomatic individuals with a mutation in the SOD1 susceptibility gene. Similar MRS abnormalities have also been detected within the spinal cord of SOD1 mice prior to disease development69. Clinically, direct correlations between metabolites ratios and patient functions have been observed, with reduction in NAA/Myo and NAA/Cho being associated with smaller forced vital capacity (FVC) measurements68. Reduction in NAA ratios seen in the spectroscopy studies reflect the axonal loss seen histpathologically in ALS, but may also, in part at least, be related to the mitochondrial dysfunction known to occur in ALS. In the SOD1 mouse model of ALS there is a decline in estimated motor neurone numbers prior to the development of clinical signs71 and neurophysiological data from healthy SOD1 positive humans shows a reduction in estimated motor unit numbers several months in advance of symptom onset72. The presence of metabolite abnormalities in pre-symptomatic patients also supports the theory that the disease process in ALS begins before the development of clinical symptoms and signs.
Patient numbers in these MRS studies so far have been small and further longitudinal studies are required to assess the usefulness of such techniques as biomarkers of disease progression in ALS. However, the relative rarity of the disease makes studies with large numbers of patients from single centres difficult, whilst the absence of acquisition standardisation and the technical expertise needed to do high-quality magnetic resonance spectroscopy have presented barriers to multicentre collaborations73.
DTI studies in ALS have mainly focused on the brain and consistently report a reduction in FA and an increase in MD74-77. In a small, 3T DTI study of the cervical spinal cord comparing patients with ALS to healthy controls, the authors found a similar reduction in FA across the C1 - C6 spinal segments78. The reduction in FA correlated with lower average finger and foot tapping speeds. An increase in RD was seen at C3-C5 levels and a negative correlation was seen between RD and finger and foot tapping, respiratory function, as measured by FVC (Figure 4), and the ASL functional rating scale-R78. Interestingly, unlike studies in the brain, no differences in MD and AD were observed between healthy subjects and the ALS group, which is in keeping with similar finding by Cohen-Adad et al. ALS (in press). Larger differences in FA and RD between patients and controls were observed at more distal cervical segments than in the upper cervical cord, supporting the hypothesis that neurodegeneration in ALS starts distally in the motor neuron terminals and then proceeds to involve the cell body (the "dying-back" phenomenon). From a clinical perspective, the absence of lower motor neurone signs on clinical examination and longer patient survival differentiates Primary lateral sclerosis (PLS) from ALS79. Although much fewer studies have been carried out in patients with PLS, a single DTI study of the brain showed reductions in FA and increased MD in the corpus callosum and intracranial corticospinal tracts80. Whether similar changes in DTI metrics are seen in the spinal cord in PLS is unknown and still needs to be determined. PLS is histopathologically distinct from ALS in the brain and spinal cord and further clinical MRI studies of the spinal cord in this patient group are therefore needed81.
3.5 Quantitative MRI of structural spinal cord lesions.
3.5.1 Spinal cord tumours
Intramedullary spinal cord tumours are rare with an incidence of 1.1/100,000 persons. They can occur anywhere along the length of the spine, but occur most commonly in the cervical spine. The most frequent histological types are ependymomas, astrocytomas and hemangiomas82. Although they can often be visualised using conventional MRI on T2-weighted sequences, the effect of the tumours on the long tracts cannot be completely characterised using conventional imaging alone. Hemangioblastomas and ependymomas do not typically infiltrate neighbouring tissues and are, therefore, regarded as surgically resectable; conversely, fibrillary astrocytomas are infiltrative, which poses difficulties in surgical management as they cannot be fully resected. Modern surgical procedures have made fine tumour excision possible and the surgical outcome can be aided through clear tissue discrimination. Therefore the development of imaging techniques, capable of accurately predicting the course of white matter tracts within the spinal cord, has potential to improve surgical planning and patient outcomes.
Vargas MA et al demonstrated that in 3 cases of spinal ependymoma, 1 case of myeloma and 1 case of astrocytoma, DTI with tractography reconstruction showed the displacement of fibre tracts surrounding the tumour83. Three of the 5 tumours studied in the case series showed an increased value in the ADC. Values for FA derived from the site of the tumour were reduced in all cases. In a similar study of 5 patients with spinal cord astrocytomas84, FA values were calculated (on a voxel-by-voxel basis) from the tumour site and were found to be lower than in healthy controls. The FA data from both studies suggests that there is an increase in extracellular water volume. This may be secondary to vasogenic oedema or loss of axonal fibres, with subsequent loss of fibre density, causing a relative expansion of the extracellular space. Data from preliminary surgical trials suggest that DTI-derived tractography can be used to plan spinal cord tumour resection since it is capable of predicting the resectability of lesions85. However, the numbers of patients recruited in trials have so far been low and further prospective work is required.
In clinical practice, MR spectroscopy of brain tumours is now commonly used in specialist neurological units, and can be useful diagnostically, and can help avoid the need for a biopsy. However, MRS studies characterising the metabolic profiles of spinal cord tumours, or showing the reliability of the technique, are currently lacking due to the technical challenges associated with spinal cord spectroscopy discussed earlier in the chapter (also see chapter 5.1).
3.5.2 Spinal vascular anomalies
Spinal arteriovenous malformations (AVMs) are congenital abnormalities of the vasculature within the spinal cord characterised by a compact intramedullary nidus, with feeding vessels most commonly arising from the anterior or posterior spinal arteries, or both, and draining into an arterialised coronal venous plexus. Typically, spinal AVMs produce a progressive myelopathy with periods of acute neurological worsening secondary to haemorrhage. Spinal AVMs may be associated with an increase of the extracellular fluid compartment because of venous congestion and loss of white matter in the normal-appearing cord distant to the nidus. A short series by Ozanne et al demonstrated that DTI of the spinal cord is possible in the presence of AVMs88. Although their study includes a heterogeneous group of patients with AVMs of differing radiological characteristics, which makes it difficult to assign DTI changes to specific subgroups of patients, it suggests the possibility of mechanical effects of AVMs on surrounding tracts. At the level of the nidus, AVMs can interrupt, displace, or separate the fibre tracts, whereas, distant to the nidus, the effects of congestion or gliosis on the tracts can be inferred. Interruption of fibres close to the nidus, loss of fibres distant to the lesion and reduction in FA values were associated with clinical disability scores.
3.7 Application of quantitative MRI techniques as a screening tool for occult spinal cord disease
Spinal cord injury causing myelopathic symptoms can occur in a large number of systemic disorders including HIV, Syphilis, Systemic Lupus Erythematosis (SLE), Sjogrens syndrome and deficiency in Vitamins E and B12 amongst others. A small number of exploratory studies including only with few patients have attempted to apply qMRI techniques to detect pathology that is radiologically or clinically occult.
In a study of 20 patients with neuroborreliosis, where 12 patients had visible lesions in the brain on PD/T2 scans but none had lesions in the cord, MTR maps and MTR histograms found no differences between patients and controls in average cord MTR89.
In patients with neuropsychiatric systemic lupus erythematosis (NSLE), Benedetti et al. found that peak height of cervical cord MD histograms were significantly lower in 11 patients studied, compared to controls, but that average cord MD values were similar between patients and controls90. These results imply that some imaging voxels within the region studied had low MD values, but that this wasn't sufficient to affect the overall mean values significantly. There was no difference in FA histograms metrics between patients and controls. The occurrence of MD changes, in the absence of FA changes, have been suggested to reflect the consequences of Wallerian degeneration of long fibre tracts passing through diseased brain areas, because prior studies have shown that Wallerian degeneration first occurs in the axonal membranes and myelin sheaths, thus causing MD changes, but only minor or no changes to the measured FA91.
More recently, a comparison of DTI metrics of the cervical spinal cord between a cohort of asymptomatic patients with human immunodeficiency virus (HIV) infection and normal appearing spinal cords was performed92. The aim of this study was to screen for HIV-myelopathy (HIVM)-related changes associated with the disease, which only usually becomes clinically manifest in advanced stages. The authors found a trend for a lower mean FA and higher mean MD in each of the measured ROIs of the spinal cord in HIV patients compared to healthy controls, which, however, didn't reach statistical significance.
Fatemi et al. assessed men with adrenomyeloneuropathy (AMN) using MT imaging of the spinal cord, which usually shows atrophy only in the advanced stages of the disease93. The authors showed that they were able to visualise, localise and quantify damage within the dorsal and lateral columns of patients and that MT hyperintensity in the dorsal column correlated with clinical scores, as assessed by the EDSS, and vibratory sense and postural sway.
Section 5: Conclusions and future directions
Spinal cord pathology in neurological diseases often causes severe clinical disability, because important neurological functions are conveyed within a narrow canal space. Quantitative MRI techniques have huge potential to provide information about this pathology, from both microstructural and functional perspectives at the site of spinal injury and distant from it. Changes in qMRI measures have been shown to differentiate patients from controls, and in controls, they correlated well with clinical disability in a wide range of degenerative spinal cord diseases. Interestingly, in some diseases, such as spondylotic myelopathy, spinal cord diffusion may be able to predict patients who show a poor clinical recovery after surgery46, suggesting that qMRI may in future have value in predicting prognosis. However, the imaging studies in patients with spinal cord diseases that we discussed in this chapter are overall few, when compared with brain studies carried out in the same diseases, and have often included a small numbers of patients. Additionally, longitudinal spinal cord qMRI studies are lacking, and therefore it is currently unknown whether qMRI measures change over time. However, advanced spinal cord imaging potentially provides detection of subclinical disease for earlier diagnosis and management. Spinal cord clinical trials generally employ conventional spinal cord MRI to confirm the safety of the new procedure or drug. Conversely, spinal cord qMRI may provide imaging biomarkers that can be used in clinical trials as outcome measures and to monitor disease progression. After standardization of spinal cord protocols and analysis methods across centres, spinal cord qMRI may even be used in multi-centre clinical trials, which are crucial for those diseases which are rarer.
Table 1. Main features and results of the main spinal cord imaging studies discussed in this chapter.
Region of interest
Cheran S et al 2011
Reduction in FA, MD and LD throughout the cervical cord with maximal FA reduction at site of T2-weighted signal abnormality.
Shanmuganathan K et al. 2008
Reductions in whole cord ADC and ADC and FA at injury site compared with controls.
Elliot JM 2011
lower NAA/Cr ratios but not Cho/Cr ratios than healthy controls
Kachramanoglou C et al (in press)
Increased m-Ins/cr ration in patients compared to controls
Kara B et al 2011
Increase in ADC and decrease in FA at site of the stenosis.
Holly LT et al 2009
Lower NAA/Cr ratio in patients, with lactate peaks detected in nearly half of patients with T2 weighted signal abnormality
Hatem SM et al 2009
Decreased FA. Mean FA correlates with clinical scores
Nair G et al 2010
FA 12% lower than controls
RD 15% higher than controls
No difference in MD and AD between patients and controls
Carew JD et al 2011
ALS and presymptomatic SOD1 positive patients
NAA/Cr and NAA/Myo ratios reduced in SOD1+ subjects and ALS compared to controls.
Myo/Cr reduced in SOD1 + but not ALS.
NAA/cho reduced in ALS but not SOD1 + subjects.
Vargas et al. 2008
Cohort of mixed spinal tumours
Reduction in FA. ADC increased in 3 out of 5 patients.
Ducreux D et al 2006
Cervical and thoracic cord
Ozanne et al 2007
Reduced FA. FA reduction associated with clinical scores.
Qian W et al 2011
Reduced FA, Increased TD and MD. Good correlation between DTI metrics and clinical scores.
Section 6: Figure Legends
Figure 1. On the left, sagittal T2-weighted images demonstrating the three anatomic regions of the cervical cord (upper, mild and lower). Hemorrhagic (a) and non-hemorrhagic (b) cord contusions are seen. On the right, scatter graphs illustrating the relationship between DTI parameters and the severity of injury measured at the injury site for the non-haemorrhagic group: (a) mean diffusivity, (b) fractional anisotropy, (c) longitudinal diffusivity, and (d) radial diffusivity. Adapted from Cheran S et al. (2011) - currently seeking permission to reproduce from J Neurotrauma.
Figure 2. (A) T2-weighted images showing MR spectroscopic voxel placement between C1-C3 in a patient (left and centre) and a control (right). (B) Spectrum showing increase myo-ins/Cr ratio in a patient following re-implantation surgery for brachial plexus avulsion injury compared to a spectrum obtained in a healthy control (C). Courtesy of Dr C Kachramanoglou, UCL Institute of Neurology.
Figure 3. A diagram illustrating the number of significantly regulated genes in rats after root avulsion and avulsion plus re-implantation. The x-axis indicates the number of genes that have undergone significant changes in expression. Note that this number includes both up- and down-regulated genes. The number of cell death genes is similar in the two groups, while genes related to neurite development and neurogenesis show a much more prominent response after re-implantation than after avulsion only. The inflammatory response is more obvious after avulsion than after re-implantation.
Adapted from Risling M et al. (2011) - permission to reprint obtained from Frontiers in Neurology.
FigureÂ 4. (i).Â DTI of cervical spinal cord. (A) A single coronal B0 slice (2 averages) from a representative healthy control subject depicting susceptibility distortion, which causes artificial curvature of the cord in the LR plane and stretching near intervertebral discs. The cord moves out of theÂ imagingÂ slice at approximately the C6 vertebral body. (B) Distortion corrected B0 slice from the same subject showing a straight cord. (C) Representative FA (range 0-0.9), (D) MD (range 0-2.3Â Ã-Â 10âˆ’3Â mm2/s), and (E) RD (range 0-2.1Â Ã-Â 10âˆ’3Â mm2/s) maps of the brainstem and spinal cord extracted shows sufficient signal-to-noise ratio within the cord for accurate DTI quantification. (F) White matter skeleton (in green), derived using a threshold of 0.2 and FSL tools, overlaid on corresponding mean fractional anisotropy (FA) slice from all subjects after nonlinear registration. The cord and skeleton moves out of the slice plane between approximately the C1 though C3 levels and is not seen in this image. Volume rendered image (slightly magnified) showing regions with (G) a significant reduction in FA (in red,Â pÂ <Â 0.1), and (H) a significant increase in radial diffusivity (in green,Â pÂ <Â 0.1) in ALS patients compared to age-matched healthy controls. Differences in FA are more extensive along the cervical cord than differences in RD. (I) Volume rendered image showing ROIs drawn along the corticospinal tract starting superiorly at the region of the pons (yellow), pyramids (cyan), C1 and C2 region (red), C3 (blue), C4 (green), C5 (violet), and C6 (yellow). The ROIs were drawn with B0 image as reference. (nÂ =Â 14 ALS andÂ nÂ =Â 15 age-matched healthy controls).
(ii).Â The decreased diffusion anisotropy seen in the cervical cord of ALS correlates with disease severity. ROI drawn over the WM in cervical cord (C1 through C6) reveals (A) decreased fractional anisotropy (FA,Â pÂ =Â 0.003), and (B) increased radial diffusivity (RD,Â pÂ =Â 0.027) in the cervical cord of ALS patients compared to age-matched healthy controls. (C) Mean diffusivity (MD) approaches statistical significance (pÂ =Â 0.089), but (D) axial diffusivity (AD) was no different between the two groups. There is significant correlation between DTI measures in ALS patients and clinical measures of disease progression, such as (E) RD and (F) FA with average tapping speed of limbs, among others (*pÂ <Â 0.05; **Â pÂ <Â 0.01;Â nÂ =Â 14 for ALS andÂ nÂ =Â 15 for age-matched healthy controls).
Reproduced from Nair G et al. 2010 - permission to reprint obtained by Neuroimage.