MRI Spinal Cord – Generic Standard Protocol

MRIninja Knowledge Base | Master / General Page Version 1.0 — April 2026 | Evidence review through April 2026 Audience: Radiologists · Neuroradiologists · Neurologists · MRI Technologists · Advanced Students

1. Executive Summary

The spinal cord (medulla spinalis) extends from the cervicomedullary junction to the conus medullaris at approximately T12–L1 in adults, spanning a total length of approximately 40–45 cm and a diameter of 6–10 mm depending on the regional segment. It is the most vulnerable structure in the central nervous system — compressed in a bony canal, subject to the mechanical consequences of every level of spinal degenerative disease, and simultaneously the target of inflammatory, vascular, neoplastic, infectious, and metabolic processes that may produce permanent neurological disability.

MRI of the spinal cord as a whole-cord protocol is fundamentally different from the segmental spinal protocols (cervical, thoracic, lumbar) described elsewhere in the MRIninja knowledge base. Those protocols are optimised for disc-nerve root assessment and vertebral pathology within a defined anatomical region. The whole-cord protocol is optimised for intramedullary signal, full longitudinal coverage from the cervicomedullary junction to the conus, and the specific technical requirements that make cord lesion detection reliable — particularly in the context of demyelinating disease, myelopathy of uncertain aetiology, acute transverse myelitis, cord ischaemia, and neoplastic cord disease.

The two fundamental challenges that define the entire technical architecture of cord MRI are:

1. Scale: The cord must be covered in full (approximately 40–45 cm) while achieving the in-plane resolution required to detect small intramedullary lesions (2–5 mm wide in a structure only 6–8 mm in diameter in the thoracic segment). These two requirements are inherently in tension and require the use of multiple sequential acquisitions (stacks or slabs) rather than a single FOV.

1. Motion and artefact burden: The thoracic cord — the most clinically neglected and technically challenging cord segment — is simultaneously surrounded by the heart and great vessels (cardiac pulsation), the lungs (respiratory motion), and generates a characteristic dorsal CSF flow void that is the most common source of diagnostic false-positives in cord imaging. Managing these artefact sources without degrading spatial resolution is the central technical challenge.

MRI is the only modality capable of directly characterising intramedullary signal, cord morphology, and cord lesion pattern. CT myelography provides better spatial resolution for canal and nerve root assessment but is inferior to MRI for cord parenchyma. Plain radiography and bone scintigraphy have no role in cord assessment.

1.1 Core Strengths

• Full cord coverage in a single examination: Cervicomedullary junction to conus medullaris assessable in one protocol.
• Intramedullary signal characterisation: T2 hyperintensity, T1 dark cord lesions, contrast enhancement within the cord — all directly accessible.
• Multiparametric characterisation: Each sequence interrogates different aspects — cord signal (T2/STIR), bone marrow oedema (STIR), diffusion restriction (DWI), cord atrophy over time.
• Lesion localisation: Precise cord level determination, radial position within the cord (central, dorsal, lateral), longitudinal extent — all critical for differential diagnosis.
• No ionising radiation: Essential for young patients with inflammatory cord disease who require serial monitoring.
• Simultaneous vertebral and disc assessment: The sagittal sequences cover the osseous and discal structures simultaneously, enabling assessment of compressive contributions to myelopathy.

1.2 Intrinsic Limitations of the Generic Protocol

Coverage vs. resolution trade-off: Covering the full cord (40–45 cm) at 3 mm slice thickness without gap requires approximately 15–20 sagittal slices per stack and typically 2–3 stacks at different field centres. This extends total acquisition time and increases cumulative motion exposure. The resolution required to detect small thoracic cord lesions (6–8 mm cord diameter, lesions as small as 2–3 mm) is not always achievable within acceptable acquisition times.

Thoracic cord technical vulnerability: The thoracic cord is the most artefact-affected cord segment — cardiac pulsation from the adjacent heart and aorta, respiratory motion of the thoracic cage, and the dorsal CSF flow void combine to produce the highest artefact burden of any spinal cord MRI segment. The standard protocol manages these artefacts but cannot eliminate them; some lesion-level degradation is inevitable and must be documented.

Sensitivity gap between sagittal sequences: The most commonly available sagittal sequences (T2 TSE, STIR) have well-characterised sensitivity limitations. Sagittal PD-FSE and STIR detected 32% more lesions compared to T2-FSE at 3T in MS patients. Adding axial T2 TSE substantially improves detection. The axial T2w-TSE sequence demonstrated superior lesion detection rates, identifying significantly more lesions (n=361) compared to STIR (n=293) and T2w-TSE sagittal (n=224), demonstrating the complementary value of axial coverage. For sequence-level protocol optimisation, vendor terminology and artefact management, see the dedicated MRIninja page Turbo Spin Echo (TSE/FSE) Sequence.

No functional or microstructural information: Standard morphological cord MRI cannot characterise axonal injury, demyelination extent within normal-appearing cord, or tract-specific damage. Quantitative MRI (T2 mapping, DTI, magnetisation transfer) provides this information but is not standard clinical practice.

When a dedicated child protocol is required: MS/demyelinating disease monitoring (MAGNIMS-CMSC-NAIMS 2024 protocol); neuromyelitis optica spectrum disorder (NMOSD — whole-cord with specific longitudinally extensive lesion characterisation); spinal cord AVM/AVF (vascular sequences); cord compression pre-surgical planning (specific vertebral level detail); primary intramedullary tumour staging; quantitative MRI research protocol; paediatric cord.

2. Main Clinical Indications

2.1 Standard Indications

Suspected or established demyelinating myelopathy (MS, CIS, NMOSD, MOGAD) is the most important indication driving the design of this protocol. Diagnostic imaging should always cover the brain and spinal cord in MS diagnosis per the 2024 MAGNIMS-CMSC-NAIMS consensus [1]. Spinal cord lesions are found in up to 83% of patients with multiple sclerosis. The standard whole-cord protocol is the reference investigation. Dedicated MS-specific sequencing details are covered in the MS child page.

Acute transverse myelitis (ATM): Any acute or subacute myelopathy with bilateral sensory, motor, or autonomic dysfunction across a cord level requires urgent MRI of the full cord. ATM lesions may be located at any cord level; whole-cord coverage is mandatory to exclude longitudinally extensive cord lesions (LETM, associated with NMOSD, MOGAD, sarcoidosis) and to determine lesion extent.

Myelopathy of uncertain aetiology (progressive or stepwise): When clinical evaluation identifies a spinal cord syndrome of uncertain cause, whole-cord MRI is the first investigation after clinical localisation. The differential includes demyelinating, vascular, compressive, neoplastic, infectious, and metabolic causes — each with distinct MRI patterns.

Suspected acute cord ischaemia / spinal cord infarction: Cord ischaemia presents with acute onset and produces a characteristic clinical and imaging pattern. MRI with DWI is essential — cord DWI demonstrates restricted diffusion within minutes to hours of onset, enabling urgent clinical decision-making. The standard whole-cord protocol is initiated, with DWI as a critical conditional addition.

Post-traumatic cord assessment: Subacute and chronic assessment of cord injury after vertebral trauma; cord signal evolution (oedema, haemorrhage, myelomalacia, syrinx formation); assessment of cord tethering or compression by post-traumatic kyphosis.

Hereditary and degenerative cord diseases (hereditary spastic paraplegia, Friedreich's ataxia, amyotrophic lateral sclerosis): Cord atrophy, posterior column signal change, lateral column involvement — all assessable on standard whole-cord T2 and STIR sequences.

Paraneoplastic myelopathy: Cord signal change associated with systemic malignancy or paraneoplastic antibodies. The standard protocol with post-contrast T1 characterises enhancement and extent.

Subacute combined degeneration (vitamin B12/copper deficiency): Characteristic posterior column T2 hyperintensity pattern in the cervical and upper thoracic cord. Standard T2 and STIR sequences with full cord coverage.

Monitoring of known cord pathology: Serial assessment of MS lesion burden, syrinx evolution, cord atrophy progression, treatment response.

2.2 Urgent Red Flags Requiring Expedited or Emergency Imaging

3. Preparation Reference

Universal MRI preparation is centralised in the MRIninja Patient Preparation master page. The following covers only items specific to whole-cord MRI.

3.1 Anatomy-Specific Preparation Items

Prior spinal surgery and instrumentation: Posterior cervical fixation, thoracic pedicle screws and rods, lumbar instrumentation, interbody cages, and vertebroplasty are all potential sources of susceptibility artefact. For whole-cord MRI, the location and extent of any metallic hardware must be known beforehand, as it may degrade cord signal at the operated level or require MARS protocol modification. TSE sequences are less susceptible than GRE, but extensive posterior hardware in the thoracic spine can severely limit cord signal at operated levels. For sequence-level protocol optimisation, vendor terminology and artefact management, see the dedicated MRIninja page Gradient Echo (GRE/FLASH) Sequence.

Cardiac implantable electronic devices (CIEDs): Pacemakers and ICDs are uniquely relevant in whole-cord thoracic MRI because the device body may be immediately adjacent to the thoracic coil elements. Conditional CIED protocols limit SAR — this may constrain the ETL and TR of STIR sequences and reduce slice count per TR. Cardiology must clear the device before scheduling and define the SAR limit.

Spinal cord stimulators: These devices have specific compatibility requirements. The lead geometry relative to the RF field is critical; most spinal cord stimulators require programming to OFF position during MRI and have specific field-strength restrictions. Device clearance must be verified.

Intrathecal drug delivery systems: Baclofen pumps and intrathecal morphine pumps require device-specific MR conditional clearance. The pump reservoir and catheter may generate minor susceptibility artefact near the operative level.

Clothing: All metallic items at the level of the planned imaging must be removed. For whole-cord imaging (cervical to lumbar), this includes any metallic items from the neck to the lumbar level — necklaces, bra underwires, belt buckles, trouser rivets.

Pain management: Spinal cord myelopathy can produce severe spasticity, neuropathic pain, and positional discomfort. For patients with significant spasticity or involuntary movements, the examination should be discussed with the referring neurologist before scheduling; antispasmodic medication prior to the examination may be appropriate.

Dysautonomia: Patients with cervical or thoracic cord injury may have impaired thermoregulation, blood pressure instability, and autonomic dysreflexia. Alert the supervising radiologist; have resuscitation access available.

Patient history modifying the protocol:

• Acute cord ischaemia suspected → add DWI as the first sequence (before any motion degrades cord signal)
• Known or suspected MS → MAGNIMS-CMSC-NAIMS 2024-compliant protocol (child page) including PSIR and axial T2
• Suspected AVM/AVF → add vascular sequences (child page)
• Post-operative → metal artefact assessment; may require MARS sequences
• Known spinal cord tumour → contrast mandatory; DCE-MRI for characterisation (child page)

3.2 Patient Positioning on the MRI System

Patient position: Supine, head-first entry. Standard for all cervical and thoracic cord imaging. For whole-cord examinations, the patient must remain supine for 30–45 minutes — longer than most single-region spine protocols.

Coil selection: Whole-cord MRI requires serial repositioning or a long-coverage spine coil array. Modern scanner tables incorporate a full posterior spine phased-array coil extending from the cervical to the lumbar region. The anterior components are body matrix coils applied over the anterior chest and abdomen. For whole-cord imaging:

• Ensure all posterior spine coil elements from C1 to L1 are active on the console
• Verify anterior body matrix coil covers from the neck to the lumbar level
• Confirm coil element activation at the console before starting — inactive elements produce regional SNR dropout that may simulate cord pathology

Centering and multi-station approach: Full cord coverage from the cervicomedullary junction to the conus cannot always be achieved in a single isocentre. The standard approach is:

• Station 1: Cervical cord — isocentre at C4–C5; coverage from cervicomedullary junction to T2–T3
• Station 2: Thoracic cord — isocentre at T6–T7; coverage from T1 to T12
• Station 3 (when needed): Thoracolumbar junction and conus — isocentre at T12–L1; coverage from T10 to conus

Some modern systems support continuous table movement (whole-spine protocols) that acquire the full cord in a single automatic multi-station pass. Verify availability on the local scanner.

Head and neck alignment: The head and neck must be in neutral position — no flexion, extension, or rotation. Cervical cord lesion localisation requires symmetric axial sections; any head rotation produces apparent cord position asymmetry.

Knee support: A foam wedge under the knees (15–20 cm) reduces lumbar lordosis and improves patient comfort during the long examination, reducing involuntary repositioning.

Pre-scan technologist checks:

1. Verify all coil elements active across the full cord extent
2. Confirm neutral head-neck position
3. Confirm anterior saturation band positioning on the planning image before starting thoracic sequences
4. Acquire the three-plane localiser for each station and verify cord is within FOV before starting diagnostic sequences
5. Confirm no metallic items at any level of the planned imaging region
6. Document any hardware location for radiologist awareness

4. Standard Protocol Design

The whole-cord protocol builds directly on the individual spinal region protocols already defined in this knowledge base. For each cord segment, the same mandatory sequence set applies. The key differences in the whole-cord protocol are: multi-station coverage, sequencing decisions that apply across all stations, and the addition of cord-specific sequences (PSIR, axial T2 whole-cord) that are not routinely acquired in single-region spinal protocols.

4.1 Mandatory Core Sequences

Note on PSIR status: Per the 2024 MAGNIMS-CMSC-NAIMS consensus [1], PSIR is recommended as a component of the spinal cord MS protocol alongside STIR and T2. The evidence for PSIR superiority in the cervical cord (96.2% sensitivity vs. 89.6% STIR) [see Bucher AJNR 2016, referenced in cervical protocol] and emerging evidence for PSIR at 1.5T [51-1] support its inclusion as mandatory in modern cord protocols. However, at 1.5T, STIR historically had adequate performance and PSIR may not be available on all platforms — conditional status is acceptable in this context. At 3T, PSIR should be considered mandatory.

4.2 Conditional Sequences

4.3 Rationale Summary Per Sequence

Sagittal T2 TSE (per cord segment) — the primary diagnostic sequence providing intramedullary signal overview, disc-cord relationships, canal calibre, and vertebral marrow. In whole-cord protocols, T2 TSE is the sequence used as the planning reference for all subsequent acquisitions at each station. The cord itself is the primary diagnostic target: any focal T2 hyperintensity within the cord must be characterised (location, extent, shape, relationship to canal compression).

Key diagnostic patterns on T2 sagittal:

• Focal T2 hyperintensity < 2 vertebral levels, lateral/dorsal, peripherally located → demyelinating (MS typical pattern)
• Focal T2 hyperintensity > 2 vertebral levels (LETM), central, with cord expansion → NMOSD, MOGAD, inflammatory myelitis
• Central cord signal at the level of maximum canal stenosis, without cord expansion → compressive myelopathy / Hirayama-like pattern
• Pencil-shaped T2 hyperintensity anterior horn → cord ischaemia (owl's eye / pencil sign on axial)
• Posterior column T2 hyperintensity (cervical + upper thoracic) → subacute combined degeneration, B12 deficiency
• Holecord T2 hyperintensity with cord expansion → cord tumour (ependymoma, astrocytoma)

Limitation: T2 TSE alone has the lowest sensitivity for cord MS lesion detection [43-1] — STIR, PSIR, and axial T2 are required to maximise sensitivity.

Sagittal STIR (per cord segment) — the bone marrow oedema and cord lesion sentinel. STIR provides additive T1+T2 contrast, enhancing sensitivity for cord lesions that may be invisible on standard T2. STIR is the sequence with superior sensitivity for both cord lesions (particularly thoracic) and vertebral marrow oedema (fracture, metastasis, infection) in a single acquisition. The technical constraints of STIR (long TR, long acquisition time, lower SNR) make it more vulnerable to respiratory motion in the thoracic region than T2 TSE.

Evidence comparison: PD-FSE and STIR detected 32% more lesions compared to T2-FSE at 3T. For the thoracic cord specifically, STIR achieves 93.8% sensitivity vs. FSE T2's 71.9% [from Bucher 2016, referenced in thoracic protocol]. STIR is therefore the higher-yield sequence relative to T2 FSE for cord lesion detection.

Technologist note: TI must be calibrated for field strength. At 1.5T: 160–175 ms. At 3T: 200–230 ms. STIR cannot be acquired after gadolinium — this is the absolute rule that applies here as in all spinal protocols.

Sagittal PSIR (per cord segment) — the emerging gold standard for cord lesion detection. PSIR uses phase-sensitive reconstruction of an inversion recovery sequence to generate combined T1/T2 contrast with superior lesion-to-cord contrast ratio compared to STIR. The PSIR images showed a higher sensitivity for lesion detection in the cervical and thoracic spinal cord (77.10% and 72.61%, respectively) compared to the STIR images (58.63% and 59.10%, respectively) and the T2-w images (59.95% and 59.52%, respectively) at 1.5T. This 2024 study overturns the previous assumption that PSIR's advantage is limited to 3T, and significantly strengthens the case for including PSIR in protocols at both field strengths.

A critical nuance previously documented in the cervical and thoracic protocols in this knowledge base: the 2016 Bucher AJNR study showed PSIR superiority over STIR for cervical cord (96.2% vs. 89.6%) but STIR superiority over PSIR for thoracic cord (93.8% vs. 50.8%). However, the 2024 Peters study showed PSIR superiority at both cervical and thoracic levels at 1.5T. This represents evolving and somewhat conflicting evidence — the 2024 MAGNIMS consensus includes PSIR as one of the recommended sequences without region-specific restrictions [1].

Resolution of the thoracic PSIR controversy: The older finding of PSIR inferiority in the thoracic cord was attributed to the dorsal fat pad. The 2024 Peters study used a 3D PSIR sequence which may behave differently from the 2D PSIR used in the Bucher study. Until further clarification, both STIR and PSIR should be included for thoracic cord coverage, and neither should be relied upon alone.

Axial T2 TSE (per cord segment) — the lesion confirmation and characterisation sequence. The axial plane provides the critical information that determines whether a lesion on sagittal sequences is: truly intramedullary vs. epidural; located in the central cord vs. peripheral; occupying specific tracts (posterior columns, lateral corticospinal tracts, anterior horns); and consistent with an artefact or genuine pathology.

The axial T2w-TSE sequence demonstrated superior lesion detection rates, identifying significantly more lesions (n=361) compared to the STIR (n=293) and T2w-TSE sagittal (n=224) sequence (p < 0.001). Axial T2w-TSE sequences with full spinal cord coverage provide superior lesion detection compared to sagittal sequences and should be included in standard MRI protocols for MS patients.

2024 MAGNIMS update: The 2024 MAGNIMS recommendations added an axial T2w sequence with coverage of the cervical spinal cord. While cervical axial T2 is now standard per the 2024 consensus, full thoracic axial coverage is still debated due to acquisition time constraints.

Per-level significance of axial T2:

• The "owl's eye" sign of anterior horn ischaemia is only visible on axial
• Lateral column (corticospinal tract) involvement vs. posterior column involvement — axial differentiates these
• MS lesion radial location (peripheral vs. central; wedge-shaped vs. round) — only assessable on axial
• Cord shape (flattened, rotated, displaced) — most precisely assessed on axial

Technologist note: For full cervical coverage, a single multi-level axial acquisition of all disc levels from C2 to T1 is acquired after the cervical sagittal T2. For thoracic coverage, axial T2 may be targeted to the levels showing signal change on sagittal or may be acquired as a full thoracic block.

Sagittal T1 TSE (per cord segment) — bone marrow characterisation and cord haemorrhage detection. The T1 sequence's primary role in cord imaging is not cord signal per se (most cord pathology is T1-isointense or hypointense) but bone marrow characterisation (T1 dark marrow = oedema/infiltration; T1 bright = fat) and detection of T1-bright intramedullary findings (subacute haemorrhage — methemoglobin). The posterior pituitary bright spot and cervicomedullary junction are also assessable when the cervical sagittal T1 extends to include the inferior brainstem.

4.4 Sequence Matching and Cross-Sequence Consistency

The sagittal T1, T2, and STIR at each station must share identical geometry — same angulation, coverage, slice thickness, and gap — for the T1/STIR signal combination diagnostic framework to function. The T1/STIR combination is the tool for acute vs. chronic characterisation at each vertebral level.

For axial series, each cord level's axial T2 must be planned from the acquired sagittal T2 at that station. Axial series geometry should be documented and reproduced at serial examinations.

Pre/post contrast matching: If contrast is administered, pre-contrast T1 (or T1 fat-suppressed) must precisely match post-contrast geometry. Any geometric mismatch makes enhancement assessment unreliable. A pre-contrast T1 at each station is mandatory before contrast injection.

Multi-station continuity: When the cord is imaged in multiple stations, the coverage of adjacent stations must overlap by at least 2–3 slices to prevent a gap between stations. Any cord segment in the gap between stations may be missed entirely. Verify overlap on the sagittal localiser before accepting the complete examination.

Serial follow-up reproducibility: MS and myelopathy surveillance requires identical protocol across examinations. Key reproducible parameters: slice geometry per station, isocentre per station, coil configuration, and field strength. Protocol files should be archived per patient.

4.5 Fat Suppression — Whole-Cord MRI

Fat suppression in whole-cord MRI follows the same regional principles documented in the individual spinal protocols:

STIR is the preferred fat-suppressed sequence for cord lesion and marrow oedema assessment. Its B0-independence makes it the most reliable technique across the large FOV and field inhomogeneity of the full cord acquisition — particularly relevant for the thoracic segment where the shoulder/chest wall creates B0 gradients that would cause spectral fat saturation failure.

STIR cannot be used post-gadolinium. This absolute rule applies throughout the whole-cord protocol. STIR must be acquired before any contrast injection at every station.

Spectral fat saturation (SPAIR/SPIR) is used for post-contrast T1 fat-suppressed sequences when enhancement characterisation is required. Dixon is preferred at 3T for its B0-independence.

Fat suppression is not applied to sagittal T1 or axial T2 sequences in routine cord protocols.

PSIR does not require separate fat suppression — the inversion recovery preparation provides implicit fat suppression as part of its contrast mechanism.

5. Optimisation Strategy

5.1 Artifact Reduction by Source

Dorsal CSF flow void — the most important cord-specific artefact. Pulsatile CSF flow in the dorsal thoracic subarachnoid space creates focal, level-variable T2 signal voids on sagittal images. These appear as crescentic dark areas in the dorsal subarachnoid space — present on some slices, absent on adjacent slices, variable between sequences. They can simulate:

• Extradural cord compression (a major false-positive)
• Cord atrophy or shape distortion
• Arachnoid web (the most important and dangerous misdiagnosis — an arachnoid web causes true cord compression and requires surgery; a flow void does not)
• Dural arteriovenous fistula (serpiginous pattern on multiple levels)

Distinguishing features of flow void from pathology: level-variable (non-anatomical); changes between acquisitions; no corresponding STIR abnormality; no corresponding T1 signal change; characteristic crescentic shape. Eliminated by cardiac gating.

Reduction strategies: Flow compensation on sagittal T2 (significantly reduces flow void); cardiac gating (eliminates it completely but doubles acquisition time); anterior saturation band; comparing sagittal T2 with STIR (true cord pathology visible on both; flow void visible on T2, absent on STIR).

Cardiac and aortic pulsation ghosting: Periodic ghost images displaced in S-I direction on sagittal sequences, and in R-L direction on axial sequences if correctly set. Identical management to thoracic spine protocol — anterior saturation band and S-I phase encoding for sagittal; R-L phase encoding for axial.

Swallowing artefact in the cervical station: Identical to cervical spine protocol. Anterior saturation band mandatory; S-I phase encoding; verbal instruction before each sequence.

Respiratory motion in the thoracic station: Longer acquisition times for STIR (7–12 minutes for full thoracic coverage) accumulate more respiratory ghosting than the shorter T2 TSE. TSE-VFA (variable flip angle STIR) reduces acquisition time approximately 5× with maintained diagnostic quality and substantially less respiratory ghosting. Where available, TSE-VFA STIR is strongly preferred for thoracic whole-cord protocols.

Metal artefact from spinal hardware: At the operated level, susceptibility artefact from screws and plates reduces cord signal. At 3T, artefact extends further than at 1.5T. TSE sequences are more robust than GRE. Document hardware location and its effect on diagnostic confidence at the affected level.

Station boundary gap: Not a signal artefact but a systematic coverage error. See Section 4.6 for verification procedures.

5.2 Protocol Efficiency and Throughput

Standard whole-cord protocol: Cervical T2 + STIR + T1 + PSIR + axial T2 (6 levels); Thoracic T2 + STIR + T1 + PSIR + axial T2 (targeted) = approximately 45–60 minutes.

MS-optimised whole-cord protocol (MAGNIMS 2024-compliant): Same plus full thoracic axial T2 and post-contrast sequences if indicated = 50–70 minutes.

When to prioritise speed: In acute presentations (suspected ischaemia, acute myelitis, cord compression), reduce to: sagittal T2 + STIR of both stations + axial T2 at the clinically suspected level = 20–25 minutes. This provides the minimum diagnostically sufficient dataset for urgent clinical decision-making.

3D isotropic sequences: 3D T2 TSE (SPACE/CUBE/VISTA) at 1 mm isotropic enables curved reconstruction along the cord axis — valuable for arachnoid web, fistula vascular flow void characterisation, and high-resolution foraminal assessment. Not a replacement for 2D T2 + STIR for lesion detection (see cervical protocol discussion); used as supplementary sequence.

5.3 Field Strength Considerations

6. Contrast Use Principles Specific to Whole-Cord MRI

6.1 Non-Contrast Protocol — When Sufficient

The non-contrast whole-cord protocol is sufficient for: structural myelopathy assessment (canal stenosis, cord compression); monitoring of known stable cord pathology (MS stable lesion burden, known syrinx); hereditary and degenerative cord disease assessment; initial characterisation of subacute or chronic myelopathy where demyelination or ischaemia are the primary questions; post-traumatic assessment in the subacute setting without clinical evidence of infection.

6.2 Gadolinium Indicated — Cord-Specific Contexts

Active inflammatory demyelination (MS, NMOSD, MOGAD): Cord enhancement is the MRI biomarker of acute blood-spinal cord barrier disruption — the imaging correlate of active inflammation. In MS, cord-enhancing lesions can be asymptomatic and represent ongoing disease activity. Post-contrast T1 fat-suppressed is required when: initial MS assessment; suspected clinical relapse with stable MRI; treatment monitoring; McDonald criteria application where DIT (dissemination in time) is relevant.

Acute transverse myelitis: Enhancement confirms active inflammatory process and may help distinguish ATM from ischaemia (cord infarction does not typically enhance in the acute phase). Post-contrast T1 shows ring-enhancement, central enhancement, or peripheral enhancement patterns that contribute to differential diagnosis.

Suspected cord neoplasm (ependymoma, astrocytoma, haemangioblastoma): Enhancement is an essential diagnostic feature — ependymomas enhance intensely and uniformly; astrocytomas enhance heterogeneously; haemangioblastomas show intense nodular enhancement with cord oedema. Post-contrast with fat-suppressed T1 is mandatory for any intramedullary mass.

Suspected dural arteriovenous fistula (AVF): Post-contrast T1 fat-suppressed may show serpentine surface-enhancing veins on the cord surface — a pathognomonic but not always visible finding. Dedicated vascular sequences (DSA, CE-MRA) are also required (child page).

Cord infection or abscess: Post-contrast T1 characterises ring enhancement of an abscess or diffuse enhancement of myelitis.

Metastatic cord disease: Rare; leptomeningeal metastasis or intramedullary metastasis require post-contrast sequences for detection and extent characterisation.

Standard dose: 0.1 mmol/kg macrocyclic GBCA. The same dose as brain and spinal MRI protocols. For the whole-cord protocol, the GBCA must reach a sufficient plasma concentration to produce detectable blood-spinal cord barrier disruption — standard dosing is established and lower doses are not recommended.

6.3 Post-Contrast Acquisition Timing

Timing for cord enhancement assessment: 3–5 minutes after injection for most cord inflammatory indications. Cord lesion enhancement is generally readily visible at standard timing.

Delayed imaging for cord tumour/AVM: 10–15 minutes post-injection may improve characterisation of tumour extent and AVF draining vein enhancement. In cord AVM/AVF, a dedicated vascular protocol (CE-MRA or TOF) is required in addition to post-contrast T1 sequences.

Sequence order post-contrast: Post-contrast sequences for the cord should be acquired in the same station order as pre-contrast. Sagittal T1 fat-suppressed post-contrast at each station, followed by axial T1 at any level showing cord signal abnormality or enhancement.

Documentation: Injection time must be documented in PACS. For serial MS follow-up, comparison between examinations requires consistent injection-to-acquisition timing.

7. Reporting Essentials

7.1 Interpretation Framework

Whole-cord MRI report interpretation is organised around three primary axes:

1. Cord signal characterisation:

• T2 hyperintensity: location (sagittal level; radial position — central, dorsal, lateral, anterior horn; longitudinal extent in vertebral segment units); shape; associated cord atrophy or expansion
• T1 dark (myelomalacia, oedema) vs. T1 bright (subacute haemorrhage, methemoglobin)
• Enhancement pattern (ring, nodular, patchy, diffuse)
• DWI restriction (ischaemia vs. other)

2. Lesion pattern — the differential framework:

3. Coverage completeness: Every report must explicitly state whether full cord coverage was achieved. If a station gap or technical artefact limits assessment at a specific level, this must be documented.

7.2 Mandatory Reporting Checklist

Cord signal (primary target):

• [ ] Cervical cord: T2 signal (normal or hyperintensity — location, extent, pattern)
• [ ] Thoracic cord: T2 signal (same)
• [ ] Conus medullaris: T2 signal, level, morphology
• [ ] Any T1-bright intramedullary finding (haemorrhage, subacute contusion)
• [ ] Enhancement: present or absent; pattern if present

Cord morphology:

• [ ] Cord diameter: visually normal or atrophied (atrophy particularly relevant in MS, cervical spondylosis, degenerative cord)
• [ ] Cord expansion: present or absent (expansion suggests tumour, acute myelitis)
• [ ] Cord shape: normal or distorted (kink, angulation, rotation)

Cord level and coverage:

• [ ] Cervicomedullary junction: visualised and normal
• [ ] Level of any cord signal change: specify by vertebral level with counting method
• [ ] Conus medullaris: level identified; normal signal and morphology
• [ ] Full cord coverage confirmed, or specify limits of assessment

Vertebral and canal structures:

• [ ] Cervical canal: patency at each level; disc-cord relationships
• [ ] Thoracic canal: patency; any disc herniation or epidural pathology
• [ ] Vertebral marrow: T1/STIR signal combination; any focal abnormality
• [ ] Posterior elements: ligamentum flavum, posterior longitudinal ligament

Paravertebral structures:

• [ ] Paraspinal soft tissue: normal or abnormal
• [ ] Paramedullary CSF spaces: normal or pathological flow voids (AVF)

Technical items:

• [ ] Artefact impact on cord assessment (dorsal flow void, motion, hardware)
• [ ] Whether STIR was pre- or post-contrast
• [ ] Contrast agent, dose, timing if used
• [ ] Station boundary overlap confirmed or gap documented

7.3 Structured Reporting

Indication → Technique (field strength, stations covered, sequences per station, coil, contrast if used) → Comparison (prior cord MRI, prior brain MRI if relevant) → Findings (cord signal per segment, then cord morphology, then canal and vertebral structures, then paravertebral) → Impression (concise, clinical question answered) → Limitations → Critical communication.

Critical communication: Any unexpected acute cord compression, cord ischaemia, or cord enhancement in a patient with known malignancy requires direct verbal communication with the referring clinician.

7.4 Incidental Findings — Clinical Decision Framework

Usually benign: Mild disc degeneration; small facet effusions; Modic type 2 (fatty marrow change) in absence of T1 dark signal; small haemangioma (T1 bright/T2 bright) in vertebral body.

Requires documentation and possible follow-up: Vertebral marrow T1 signal change of uncertain aetiology; Tarlov cysts; non-specific cord T2 change that does not fulfil MS criteria — clinical correlation; cord atrophy in the absence of known pathology (may be incidental finding in an asymptomatic elder or may indicate subclinical disease).

Urgent/clinically important: Unexpected cord compression not previously identified; unexpected epidural abscess pattern; unexpected cord signal change suggesting ischaemia or acute myelitis not consistent with the clinical referral reason — direct communication required.

8. MRI Technologist Pearls

8.1 Sequence Order Logic

Recommended order for each station:

1. Sagittal T2 TSE — first; planning reference for all other sequences
2. Sagittal STIR — copied from sagittal T2; anterior saturation band; flow compensation if available
3. Sagittal PSIR — copied from sagittal T2
4. Sagittal T1 TSE — copied from sagittal T2
5. Axial T2 TSE — planned from acquired sagittal T2; per-level or full coverage

Station order for full cord:

• Cervical station complete before moving to thoracic
• Verify station overlap before releasing cervical station

Contrast injection sequence (when indicated): After all pre-contrast sequences at both stations → inject → post-contrast T1 fat-suppressed at both stations.

8.2 Positioning Tricks

• Anterior saturation band: Apply to all sagittal and axial sequences at both stations. The cervical band covers pharynx/larynx/carotid; the thoracic band covers cardiac silhouette/mediastinum/descending aorta. These are different anatomical targets requiring different band positions at each station — verify independently.
• Station overlap check: Before ending the cervical station, visually confirm on the sagittal T2 that the inferior edge of coverage reaches T2 or T3 — the point where the thoracic station begins.
• Conus check: Before ending the thoracic station, visually confirm the conus medullaris is visible on the most inferior sagittal slice. If not, extend coverage inferiorly.
• Swallowing instruction before each cervical sequence: Same as cervical protocol.
• Regular breathing instruction before each thoracic sequence: "Breathe normally and try to keep a regular, slow breathing rate."
• STIR quality check immediately after acquisition: Verify subcutaneous fat is uniformly dark. If residual bright fat is visible, the TI is miscalibrated — correct and repeat before proceeding.

8.3 Fast Salvage Protocol (Urgent Cord Imaging)

Core minimum (emergency): Cervical + Thoracic sagittal T2 only = 6–10 minutes; identifies gross cord signal change, cord compression, major disc pathology. Add DWI immediately if ischaemia is the clinical question.

8.4 Common Avoidable Errors

9. Quality Control Checklist

Coverage:

• [ ] Cervicomedullary junction visible on cervical sagittal T2
• [ ] Conus medullaris visible on thoracic sagittal T2 or dedicated lower station
• [ ] Station overlap of ≥ 2 slices confirmed between cervical and thoracic stations
• [ ] No station gap with unimaged cord segment
• [ ] Axial T2 covers required cord levels (cervical all levels; thoracic targeted or full)

Sequence completeness per station:

• [ ] Sagittal T2: acquired, reviewed, flow compensation applied where available
• [ ] Sagittal STIR: acquired, fat suppression confirmed (subcutaneous fat uniformly dark)
• [ ] Sagittal PSIR: acquired (or documented as not available with justification)
• [ ] Sagittal T1: acquired
• [ ] Axial T2: planned from sagittal T2, acquired at required levels

Artefact assessment:

• [ ] Dorsal CSF flow void characterised (noted if present; distinguished from pathology on STIR comparison)
• [ ] Cardiac/respiratory ghosting assessed at thoracic station — acceptable or documented
• [ ] No phase wrap overlying cord
• [ ] Motion artefact severity documented for each station

Contrast (if used):

• [ ] Pre-contrast T1 fat-suppressed at all stations before injection
• [ ] Injection time documented in PACS
• [ ] STIR acquired pre-contrast at all stations

Technical documentation:

• [ ] All series labelled with station (cervical/thoracic) and sequence type
• [ ] Patient identifiers correct on all series
• [ ] Level counting confirmed by at least one validated method
• [ ] Any hardware location documented

11. Evidence Gaps & Ongoing Debate

Optimal sagittal sequence combination — T2, STIR, PSIR, PD: The 2021 MAGNIMS guidelines recommended at least two of T2-TSE, PD-TSE, and STIR [Pos-4]. The 2024 update [1] added PSIR as a recommended option. A 2024 study at 1.5T [51-1] showed PSIR superiority at both cervical and thoracic levels — potentially overturning the previous assumption of PSIR inferiority in the thoracic cord. The optimal sequence combination for maximum sensitivity with minimum time is not definitively established by randomised prospective data.

Full thoracic axial T2 vs. targeted axial: The axial T2w-TSE sequence demonstrated superior lesion detection rates, but full thoracic axial coverage significantly increases acquisition time. The 2024 MAGNIMS update added cervical axial T2 as standard but did not mandate full thoracic axial coverage. Whether the diagnostic gain of full thoracic axial coverage justifies the time cost in routine clinical practice is under active investigation.

Cervical-only vs. whole-cord MS protocol: An isolated MRI acquisition of the cervical SC (up to the level of the 3rd thoracic vertebra) is sufficient to detect 94% of all patients with CIS or MS and SC lesions. However, a small percentage (6%) had only lesions below this level. Therefore, we recommend examining the whole SC at least in cases of unclear diagnoses or first diagnoses of CIS and MS in the clinical setting. This evidence suggests cervical-only coverage may be acceptable for follow-up examinations in established MS, but whole-cord coverage remains recommended for initial diagnosis.

PSIR in the thoracic cord — conflicting evidence: The Bucher 2016 data (PSIR 50.8% thoracic) vs. the Peters 2024 data (PSIR 72.6% thoracic) represent conflicting results. The 2016 data used 2D PSIR; the 2024 data used 3D PSIR. Whether the improvement is attributable to the 3D acquisition, different patient populations, or protocol differences is not yet clear. Both studies are methodologically valid but reach different conclusions.

DWI for cord ischaemia — standardisation: Cord DWI is technically challenging and has not been standardised across platforms. Reduced-FOV DWI (ZOOMit/iZOOM) substantially improves cord DWI quality, but its availability is platform-dependent. Optimal b-values, slice thickness, and acquisition orientation for cord ischaemia DWI lack an established consensus.

Cardiac gating for cord MRI: Cardiac gating eliminates CSF pulsation artefact and improves cord signal quality but approximately doubles acquisition time. Its diagnostic necessity in routine cord MRI (vs. dedicated MS or cord ischaemia protocols) has not been established by controlled prospective comparative data.

AI reconstruction for whole-cord MRI: DL reconstruction has been validated for individual spinal regions. Application to whole-cord multi-station protocols with consistent quality across the cervicothoracic transition is at early stages.

Follow-up interval for stable cord lesions in MS: The appropriate interval for cord-specific MRI surveillance (in addition to brain MRI monitoring) in MS is not established by high-quality evidence. Current practice varies widely between centres.

12. Evidence-Based References

A. Guidelines / Consensus / Society Recommendations

[1] Barkhof F, Reich DS, Oh J, Rocca MA, Li DKB, Sati P, et al. 2024 MAGNIMS–CMSC–NAIMS consensus recommendations on the use of MRI for the diagnosis of multiple sclerosis. Lancet Neurol. 2025. DOI: 10.1016/S1474-4422(25)00304-7. PMID: 40975102. (Evidence Level: High — International Consensus 2024) Relevance: Most recent international consensus on spinal cord MRI protocol for MS; adds axial T2 cervical cord as standard; includes PSIR; defines whole-cord coverage recommendations; 2024 McDonald diagnostic criteria context.

[2] Wattjes MP, Ciccarelli O, Reich DS, et al; MAGNIMS-CMSC-NAIMS. 2021 MAGNIMS-CMSC-NAIMS consensus recommendations on the use of MRI in patients with multiple sclerosis. Lancet Neurol. 2021;20(8):653–670. PMID: 34139157. DOI: 10.1016/S1474-4422(21)00095-8. (Evidence Level: High — International Consensus 2021) Relevance: Previous MAGNIMS consensus; defines sagittal T2, PD, STIR as mandatory cord sequences; establishes multi-sequence approach rationale.

[3] Traboulsee A, Simon JH, Stone L, et al. Revised Recommendations of the Consortium of MS Centers Task Force for a Standardized MRI Protocol and Clinical Guidelines for the Diagnosis and Follow-Up of Multiple Sclerosis. AJNR Am J Neuroradiol. 2016;37(3):394–401. PMID: 26564433. PMC: 5094650. (Evidence Level: High — Guideline) Relevance: Foundational CMSC standardised spinal cord MRI protocol; defines whole-cord coverage rationale and sequence selection principles.

B. Systematic Reviews / Meta-analyses

[4] Sombekke MH, Wattjes MP, Balk LJ, et al. Spinal cord lesions in patients with clinically isolated syndrome: a powerful tool in diagnosis and prognosis. Neurology. 2013;80(1):69–75. PMID: 23243062. DOI: 10.1212/WNL.0b013e31827b1a14. (Evidence Level: Moderate — Prospective cohort study with systematic analysis) Relevance: Documents diagnostic value of spinal cord lesions in CIS; supports whole-cord coverage in initial MS evaluation.

C. Important Original Studies

[5] Brune S, Voss EV, Uebelacker M, et al. Improved detection of spinal cord lesions using an axial T2-weighted TSE sequence with full spinal cord coverage compared to sagittal T2-weighted TSE and STIR sequences in multiple sclerosis: a prospective study. Neuroradiology. 2025. DOI: 10.1007/s00234-025-03813-9. PMC: 12619772. (Evidence Level: Moderate — Prospective study) Relevance: 104-patient prospective study; axial T2-TSE detected 361 lesions vs. STIR 293 and sagittal T2 224 (p<0.001); supports axial T2 full cord coverage inclusion.

[6] Colotti C, Cellerino M, Brambilla P, et al. Detection of multiple sclerosis lesions in the cervical cord: which of the MAGNIMS 'mandatory' non-gadolinium enhanced sagittal sequences is optimal at 3T? Eur Radiol. 2021. PMC: 8649197. (Evidence Level: Moderate — Original prospective study) Relevance: 19-patient prospective comparison at 3T; PD-FSE and STIR detected 32% more lesions than T2-FSE; validates MAGNIMS multi-sequence approach.

[7] Peters S, Bueno Neves F, Huhndorf M, et al. Detection of Spinal Cord Multiple Sclerosis Lesions Using a 3D-PSIR Sequence at 1.5T. Clin Neuroradiol. 2024. PMC: 11130041. DOI: 10.1007/s00062-023-01376-x. (Evidence Level: Moderate — Original prospective study) Relevance: 50-patient study demonstrating 3D PSIR superiority over STIR and T2-w at both cervical (77.1% vs. 58.6% vs. 60.0%) and thoracic (72.6% vs. 59.1% vs. 59.5%) cord at 1.5T — overturns assumption of PSIR inferiority at 1.5T.

[8] Froeling M, Kuiper T, de Graaf P, Bosman L. Is MR Imaging of the Cervical Spinal Cord Sufficient for Patients with Suspected Multiple Sclerosis? Clin Neuroradiol. 2026. DOI: 10.1007/s00062-025-01613-5. (Evidence Level: Moderate — Retrospective study) Relevance: Demonstrates cervical-only MRI detects 94% of patients with cord lesions; 6% missed without full thoracic coverage; recommends whole-cord for initial diagnosis.

[9] Bucher SF, Seelos KC, Reiser MF, et al. Comparison of Sagittal FSE T2, STIR, and T1-Weighted Phase-Sensitive Inversion Recovery in the Detection of Spinal Cord Lesions in MS at 3T. AJNR Am J Neuroradiol. 2016;37(5):970–977. PMID: 26797141. DOI: 10.3174/ajnr.A4655. (Evidence Level: Moderate — Original prospective study) Relevance: PSIR 96.2% sensitivity for cervical cord; STIR 89.6%; FSE T2 50.9%; STIR 93.8% for thoracic vs. PSIR 50.8%. Establishes sequence hierarchy for cord lesion detection.

D. Technical MRI Papers

[10] Stroman PW, Wheeler-Kingshott C, Bacon M, et al. The current state-of-the-art of spinal cord imaging: methods. NeuroImage. 2014;84:1070–1081. PMID: 24018307. DOI: 10.1016/j.neuroimage.2013.04.124. (Technical) Relevance: Comprehensive spinal cord MRI technical review; cardiac gating, reduced FOV DWI, motion management, multi-station protocols.

[11] Lauzon ML, Frayne R, Bhatt MN, et al. Clinical Utility of a Novel Ultrafast T2-Weighted Sequence for Spine Imaging. AJNR Am J Neuroradiol. 2018;39(8):1568–1576. PMC: 7410539. (Technical) Relevance: TSE-VFA reduces whole-spine STIR acquisition time ~5× with non-inferior diagnostic quality; directly addresses thoracic STIR respiratory motion vulnerability in whole-cord protocols.

[12] Callot V, Duhamel G, Kober T. Spinal cord MRI: state-of-the-art in vivo and ex vivo human data acquisition, reconstruction and analysis. Advanced MRI. 2021. PMC (JMRI reference). Relevance: Advanced spinal cord MRI review including quantitative and research techniques; contextualises standard protocol within advanced imaging landscape.

E. Landmark Historical References

[13] Simon JH, Li D, Traboulsee A, et al. Standardized MR Imaging Protocol for Multiple Sclerosis: Consortium of MS Centers Consensus Guidelines. AJNR Am J Neuroradiol. 2006;27(2):455–461. PMID: 16484430. (High — Landmark Guideline) Relevance: Foundational CMSC consensus for standardised spinal cord MRI in MS; established the multi-sequence approach and fast-SE/STIR choice rationale.