MRI Thoracic Spine – Generic Standard Protocol

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

1. Executive Summary

Thoracic spine MRI is the most technically demanding of the three spinal regions. Its unique combination of anatomical features — the longest cord segment covering twelve vertebral levels, the proximity of the heart and aorta generating cardiac-transmitted motion, continuous respiratory excursion of the surrounding thorax, a small spinal cord diameter, and the transition zones between thoracolumbar and cervicothoracic anatomy — creates a distinct set of protocol challenges that are not encountered in the cervical or lumbar spine. Understanding these challenges and their solutions is the core purpose of this document.

The primary roles of thoracic spine MRI are: cord signal assessment (the spinal cord terminates at the conus at approximately T12–L1, making the thoracic spine the longest cord-bearing spinal region); myelopathy characterisation and localisation; vertebral marrow evaluation for fracture, neoplasm, and infection; discal disease assessment; and ligamentous and paraspinal tissue assessment.

Compared with CT, thoracic spine MRI offers decisive superiority for cord signal, marrow infiltration, ligamentous injury, and disc-cord relationships. CT retains advantages for cortical bone detail (fracture morphology, surgical hardware planning), dense calcification (calcified thoracic disc herniations — a clinical entity far more common in the thoracic than cervical or lumbar spine), and ossification of the posterior longitudinal ligament (OPLL). In the thoracic spine, the combination of MRI and CT is therefore more frequently required than in the other spinal regions.

Disc herniation in the thoracic spine is far less common than in the cervical or lumbar regions. When present, it has a high rate of calcification, making CT complementary to MRI for surgical planning. Thoracic disc herniation causing myelopathy is a diagnostic and surgical challenge where MRI defines cord compromise and CT defines ossification extent.

1.1 Core Strengths

• Spinal cord characterisation: Direct cord signal, morphology, and diameter assessment across the entire thoracic length — the longest cord-bearing region accessible by MRI.
• Bone marrow evaluation: Superior sensitivity for marrow infiltration (metastasis, infection, fracture oedema) compared to all other modalities.
• Multi-level coverage in a single acquisition: The entire thoracic spine (T1–T12) is covered in the standard sagittal acquisition — a critical advantage given the frequency of multi-level disease (metastases, spondyloarthropathy, osteoporotic fractures).
• No ionising radiation: Essential for younger patients and serial oncological follow-up.
• Paravertebral and posterior mediastinal tissue assessment: Paraspinal masses, epidural lesions, and thoracic outlet structures are partially evaluable within the thoracic spine FOV.

1.2 Intrinsic Limitations of the Generic Protocol

Respiratory and cardiac motion: These are the dominant technical challenges specific to the thoracic spine and have no equivalent in the cervical or lumbar regions. The heart and great vessels lie immediately anterior to the thoracic spine; respiratory excursion of the thoracic cage is continuous; the aorta runs along the left lateral aspect of the thoracic vertebral bodies. All three structures generate phase-encoding artefacts that propagate through the cord.

Long acquisition time vulnerability: A thoracic sagittal acquisition covering T1–T12 at 3–4 mm slice thickness requires 35–50 slices — significantly more than the lumbar spine. At standard TR values, this translates to longer per-sequence acquisition times that increase cumulative motion exposure.

Calcified thoracic disc herniation: Calcified disc material appears as signal void on MRI and cannot be characterised; CT is required for extent and surgical planning. MRI may underestimate the extent of calcified herniations.

Small cord diameter: The thoracic cord has a diameter of only 6–8 mm in its mid-thoracic extent — smaller than the cervical cord. At standard protocol parameters, subtle intramedullary signal changes (small MS plaques, early myelomalacia) may be below reliable detection threshold, particularly at 1.5T.

Dorsal CSF flow artefact: The dorsal thoracic subarachnoid space generates particularly prominent CSF pulsation artefacts on T2 sequences, producing a characteristic signal void that can simulate an extradural collection, cord atrophy, or dural arteriovenous fistula.

When a dedicated child protocol is required: Suspected thoracic dural arteriovenous fistula (dedicated vascular sequences, 3D TOF angiography), MS thoracic cord monitoring (dedicated thin-slice protocol), primary thoracic cord tumour workup, suspected thoracic epidural abscess (STIR + contrast + emergency coverage), spinal deformity assessment (coronal 3D), post-operative instrumented thoracic spine, and suspected thoracic disc herniation with suspected calcification (CT complementary). For sequence-level protocol optimisation, vendor terminology and artefact management, see the dedicated MRIninja page STIR Sequence.

2. Main Clinical Indications

2.1 Standard Indications

Thoracic myelopathy is the most important indication for thoracic spine MRI. Myelopathy from any cause — spondylosis, disc herniation, neoplasm, demyelination, ischaemia, AVM, infection — presents with gait disturbance, hyperreflexia, sensory level, and sphincter dysfunction. ACR Appropriateness Criteria for Myelopathy designate MRI without contrast as "usually appropriate" for the initial evaluation of chronic or progressive myelopathy [1]. The thoracic region must be included in any myelopathy workup when the clinical sensory level or cord signs suggest thoracic origin.

Thoracic back pain with clinical concern for structural pathology: Unlike lumbar back pain, where non-specific pain is common and guidelines discourage routine imaging, thoracic back pain has a higher a priori probability of significant structural or systemic disease (vertebral fracture, neoplasm, infection). The ACR and Carelon guidelines support thoracic MRI when: neurological signs are present; back pain is associated with systemic symptoms (fever, weight loss, malignancy); pain is disproportionate to clinical expectation; or radiographs are non-diagnostic [2]. The generic non-contrast protocol is appropriate for the initial structural survey.

Vertebral compression fractures — acute vs. chronic differentiation: MRI is the most sensitive and specific tool for determining whether a vertebral compression fracture is acute (bone marrow oedema: T1 dark, STIR bright) or chronic (fatty marrow replacement: T1 bright, STIR dark), and for detecting pathological fractures (metastasis, myeloma, infection) versus benign osteoporotic fractures. The standard non-contrast protocol with T1 and STIR is often sufficient for this differentiation; DWI may add diagnostic value (child page). For sequence-level protocol optimisation, vendor terminology and artefact management, see the dedicated MRIninja page Diffusion-Weighted Imaging (DWI) Sequence.

Suspected vertebral metastases or myeloma: When malignancy is known or suspected, MRI of the thoracic spine is appropriate to detect marrow infiltration and epidural extension [2]. The generic protocol detects most vertebral marrow disease; dedicated protocols with DWI and post-contrast sequences increase sensitivity for subtle infiltration.

Thoracic disc herniation: Much less common than cervical or lumbar, but clinically important because thoracic disc herniation can produce myelopathy. MRI defines disc-cord relationship, cord signal, and canal compromise. CT is frequently complementary for calcification assessment.

Spondyloarthropathy and inflammatory back pain: Thoracic involvement in ankylosing spondylitis, psoriatic arthritis, and other spondyloarthropathies is assessed by STIR for bone marrow oedema at corner erosions and enthesopathy sites. Full-spine coverage including thoracic is standard in axial SpA evaluation [3].

Post-traumatic thoracic spine assessment: In the subacute setting after confirmed or suspected thoracic trauma, MRI assesses cord signal change, ligamentous integrity, disc injury, and epidural haematoma. In acute trauma, CT is first-line for osseous assessment; ACR 2024 supports MRI for suspected ligamentous or cord injury [4].

Incidental thoracic findings on other examinations: Vertebral lesions identified on CT, PET/CT, or lumbar spine MRI frequently require thoracic MRI for characterisation and extent determination.

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 thoracic spine MRI.

3.1 Anatomy-Specific Preparation Items

Prior thoracic spine surgery and instrumentation: Posterior fixation constructs (pedicle screws, rods), interbody cages, anterior spinal plates, and vertebroplasty/kyphoplasty cement are common in the thoracic spine. Modern titanium instrumentation is generally MR-compatible but generates susceptibility artefact that degrades adjacent tissue assessment — particularly problematic in the thoracic spine where the cord is in close proximity to posterior hardware. The standard protocol is diagnostically limited in heavily instrumented thoracic spines. Vertebroplasty cement generates relatively modest susceptibility artefact compared to metallic hardware.

Cardiac implantable electronic devices (CIEDs): The thoracic spine is in closer proximity to implanted cardiac devices (pacemakers, ICDs, CRT devices) than any other spinal region. Device-specific MR conditional clearance must be obtained before scanning, and the standard protocol must be adapted to SAR limits required by the conditional approval. Many CIED protocols limit SAR to <2 W/kg whole-body — this may constrain TR, ETL, and number of slices per acquisition at 3T. Protocol must be reviewed by the supervising radiologist and, where required, cardiology coordination obtained.

Respiratory devices and monitoring: Patients with chronic respiratory disease on home oxygen or non-invasive ventilation (NIV) may require monitoring during the examination. The presence of respiratory physiotherapy devices near the chest requires screening.

Clothing: Remove all metallic clothing fasteners, underwired bras, and chest-level metallic accessories. Cardiac monitoring leads (if used for cardiac-gated protocols) require MR-compatible materials.

Pain management: Thoracic pain from vertebral fractures, metastatic disease, or disc herniation can prevent adequate immobilisation during long acquisitions. Oral or IV analgesia before the examination, or positioning with pillow support under the knees, reduces involuntary motion.

Patient history directly modifying the protocol:

• Known malignancy → whole-spine DWI and contrast (child page); extended FOV
• Suspected infection → STIR + contrast; consider whole-spine coverage
• Prior thoracic instrumentation → standard protocol limitations; consider MARS sequences if available
• Respiratory disease/dyspnoea → respiratory gating considerations
• Suspected thoracic AVM/AVF → dedicated vascular protocol (child page)
• Pregnancy → non-contrast mandatory

3.2 Patient Positioning on the MRI System

Patient position: Supine, head-first entry. Standard position for thoracic spine MRI.

Coil selection: A dedicated spine phased-array surface coil (posterior table-integrated spine coil + anterior body matrix) provides optimal SNR coverage for the thoracic spine. The posterior spine coil integrated into the table provides posterior coverage; anterior body coil elements should be applied over the anterior chest. For thoracic spine MRI, the active coil elements should span from approximately T1 to T12 — confirm activation on the console. If only a limited spine coil array is available (cervical coil or lumbar coil without thoracic elements), this represents a significant diagnostic limitation that must be documented.

Centering: The isocentre should be positioned at the mid-thoracic level (T6–T7), corresponding to approximately the level of the inferior angle of the scapula in most adults. This ensures maximal coil sensitivity across the full thoracic spine. For examinations requiring both cervical and thoracic coverage in a single protocol, the centre must be positioned to include both regions within acceptable coil sensitivity range, or two separate acquisitions at different isocentres must be planned.

Anatomical alignment: The patient's spine must align with the table axis. Thoracic scoliosis (common in the elderly, particularly in osteoporotic vertebral fracture patients) produces misalignment that requires protocol adaptation (curved planar reconstruction from 3D acquisition, or individually planned coronal slabs).

Respiratory management: The thoracic spine presents a unique challenge: the thoracic cage performs continuous respiratory excursion, and the heart and great vessels are immediate anatomical neighbours. There are several management strategies:

• Free breathing with saturation bands: Most clinical thoracic spine MRI is performed during free breathing with an anterior saturation band over the chest and heart. This suppresses the signal from respiratory-motion structures and reduces (but does not eliminate) cardiac and respiratory ghosting. This is the standard clinical approach.
• Respiratory gating: A respiratory belt or navigators trigger acquisition to the respiratory cycle. Significantly reduces respiratory ghosting at the cost of 50–100% longer acquisition time due to the variable TR from respiratory triggering. Used in selected departments for sequences most vulnerable to respiratory motion (STIR, high-SNR T2).
• Breath-hold: Applicable only to very fast sequences (TSE-VFA, single-shot variants); standard T2 TSE and STIR acquisition times preclude true breath-hold.
• Cardiac gating: ECG or peripheral pulse triggering synchronises acquisition to the cardiac cycle, eliminating cardiac-transmitted CSF pulsation and cord motion. Used in dedicated cord imaging protocols. Increases acquisition time similarly to respiratory gating.

Comfort strategies: Thoracic spine examinations are longer than targeted lumbar disc evaluations (35–45 minutes for the full standard protocol). Knee support under the lower limbs reduces lumbar discomfort and indirectly reduces gross motion. Warm blankets improve compliance in older patients. Brief verbal inter-sequence check-ins maintain cooperation.

Pre-scan technologist checks:

1. Verify all spine coil array elements active on console.
2. Centre laser at mid-thoracic level (T6–T7).
3. Confirm no metallic items on chest or upper back.
4. Verify respiratory belt position if gating is to be used.
5. Check that cardiac monitoring leads are MR-compatible if CIED protocol applies.
6. Acquire three-plane localiser and verify thoracic spine from T1 to T12 (plus cervicomedullary junction and thoracolumbar junction) within the FOV before committing to the full protocol.

4. Standard Protocol Design

4.1 Mandatory Core Sequences

4.2 Conditional Sequences

4.3 Rationale Summary Per Sequence

Sagittal T2 TSE — the primary diagnostic sequence and the planning reference for all others. Bright CSF, bright disc, dark cortical bone, intermediate cord. The thoracic cord is the primary diagnostic target as in the cervical spine — but the artefact environment is substantially worse due to cardiac-transmitted CSF pulsation, respiratory motion, and the dorsal cord pulsation flow void. The standard sagittal T2 is acquired with anterior saturation bands over the cardiac silhouette and great vessels.

What it detects well: Cord signal (T2 hyperintensity from myelopathy, oedema, demyelination, ischaemia), disc herniations (most clearly ossified herniations appear as signal void), spinal canal compromise, vertebral body alignment, CSF space.

What it misses: Subtle marrow oedema (STIR is more sensitive), calcified disc herniations (CT required), cord lesions in the setting of severe dorsal CSF flow artefact.

Critical dorsal CSF flow artefact: The dorsal thoracic subarachnoid space generates a characteristic crescentic or band-like CSF pulsation signal void on sagittal T2. This artefact is not present at all levels, appears and disappears across adjacent slices, and can simulate extradural compression, cord atrophy, arachnoid web, or even a dural arteriovenous fistula (the latter has a characteristic serpiginous appearance and should be confirmed on post-contrast imaging and vascular sequences). The technologist must acquire flow-compensated T2 sequences where available, and the radiologist must recognise this artefact pattern.

Sagittal T1 TSE — bone marrow characterisation. Identical principles to cervical and lumbar spine T1. In the thoracic spine, the combination of sagittal T1 and STIR is the primary tool for: vertebral fracture acuity characterisation (acute T1 dark/STIR bright vs. chronic T1 bright/STIR dark); metastatic marrow infiltration detection; and differentiating osteoporotic from pathological fracture (the latter often shows T1 dark incomplete fracture lines with preservation of convex endplate morphology in acute osteoporotic fractures vs. T1 diffuse dark signal in metastases).

Sagittal STIR — the bone marrow oedema and cord lesion sentinel. In the thoracic spine, STIR is even more critical than in the lumbar spine for two reasons: thoracic cord lesions (demyelination, myelitis, ischaemia) often produce T2 changes subtle enough to be missed on standard T2 but visible on STIR; and thoracic vertebral marrow disease (osteoporotic fractures, metastases, infection) is almost invariably detected on STIR at a stage when T1 or T2 alone may be equivocal. Multiple studies have demonstrated STIR superiority over T2 for thoracic cord demyelinating lesion detection [3, 4].

Critical warning: STIR duration at thoracic level is longer than at cervical or lumbar due to the requirement for full thoracic coverage (more slices). Respiratory motion accumulates over the longer acquisition time. The STIR sequence is therefore more vulnerable to respiratory artefact than T2 TSE in the thoracic region.

Axial T2 TSE — cross-sectional disc-cord-canal assessment. In the thoracic spine, two axial approaches are used depending on clinical indication:

• Targeted axial: Slices planned over the levels of cord compression or pathology identified on the sagittal T2. This is efficient and appropriate when pathology is localised.
• Full thoracic axial: Coverage from T1 to T12. Required when the clinical question involves the full thoracic cord (myelopathy with uncertain level, multi-level disease, neoplasm screening). Longer and more motion-sensitive.

Axial T2 in the thoracic spine must be planned perpendicular to the cord at each level. Respiratory and cardiac motion produce more severe artefact on axial sequences than sagittal because each individual axial slice acquisition spans a shorter time — so a single respiratory or cardiac motion event corrupts proportionally more of each slice's k-space.

4.4 Sequence Matching and Cross-Sequence Consistency

Sagittal T1, T2, and STIR must share identical geometry for direct vertebral level-by-level comparison — essential for fracture acuity characterisation (T1/STIR combination). Copy sagittal T1 and STIR geometry from the sagittal T2 reference.

Axial T2 must be planned from the acquired sagittal T2, not from the scout.

For contrast examinations, pre-contrast T1 fat-suppressed must precisely match post-contrast geometry. In the thoracic spine, motion between pre- and post-contrast acquisitions is more common than in the lumbar spine; if significant cord position change occurs between sequences, the radiologist must be alerted.

For serial follow-up (oncological monitoring, MS surveillance), identical coil configuration, isocentre, and sequence geometry are essential. In post-fracture monitoring, the key comparison is T1 and STIR signal evolution over time.

4.5 Fat Suppression in Thoracic Spine MRI

The same fat suppression principles apply as in the cervical and lumbar spine, with two important thoracic-specific modifications:

STIR remains the preferred fat suppression technique for sagittal bone marrow oedema detection. Its B0-independence is particularly valuable in the thoracic spine, where the large FOV (T1–T12), the chest wall fat-air interfaces, and the respiratory motion-related B0 fluctuations make spectral fat saturation less reliable than at the lumbar level.

Cardiac and respiratory motion affects fat suppression uniformity: Spectral fat saturation pulses applied during respiratory excursion experience B0 field shift due to thoracic cage displacement, reducing suppression efficiency. This is an inherent limitation of spectral methods in the thoracic region and supports STIR preference.

Post-contrast T1 fat-suppressed: SPAIR or Dixon are the preferred methods for the same reasons as in the cervical spine. The chest wall fat directly adjacent to the thoracic spine makes fat suppression on post-contrast T1 sequences important for lesion conspicuity.

Fat suppression is not applied to: standard sagittal T1 (bright marrow required); standard axial T2 (natural contrast); standard axial T1.

5. Optimisation Strategy

5.1 Artifact Reduction by Source

Cardiac and aortic pulsation artefact is the dominant motion artefact source unique to the thoracic spine and the most clinically consequential. The heart generates periodic motion transmitted to the thoracic cord via CSF pulsation; the thoracic aorta (which lies immediately left-lateral to the T5–T12 vertebral bodies) generates direct vascular ghosting in the phase direction. These artefacts appear as periodic ghost images of the heart and aorta displaced along the S-I phase direction, overlying or obscuring the cord and vertebral column at displaced positions. Aortic pulsation ghosts can simulate vertebral marrow signal change, cord signal change, or epidural pathology.

Reduction strategies:

• Anterior saturation band (mandatory): placed over the cardiac silhouette, mediastinum, and aorta before all sagittal and axial sequences
• S-I phase encoding for sagittal sequences: displaces ghosts cranially and caudally
• Flow compensation (gradient moment nulling): reduces velocity-dependent phase errors from CSF and vascular pulsation; accepted in exchange for slightly longer minimum TE
• Cardiac gating: eliminates pulsation artefact entirely; used in dedicated cord imaging protocols; increases acquisition time 50–100%
• Increasing NSA/NEX: partially averages non-coherent components of pulsation ghosting
• Shorter acquisition time: reduce time within which cardiac ghosts accumulate; parallel imaging R=2–3

Respiratory artefact is the second dominant challenge specific to the thoracic spine. Respiratory excursion displaces the thoracic cage, causing both bulk motion of the patient and field shift artefacts from moving air-tissue interfaces. On sagittal sequences, respiratory ghosting manifests as repeated blur copies of the thoracic wall and subcutaneous fat displaced in the S-I phase direction. On STIR sequences, B0 field fluctuation during respiration also produces regional variation in fat suppression quality.

Reduction strategies:

• Anterior saturation band (mandatory): suppresses thoracic wall fat and soft tissue signal before it generates ghosts
• S-I phase encoding: displaces respiratory ghosts cranially and caudally
• Respiratory gating: belt-triggered acquisition eliminates respiratory ghosting at the cost of ~2× acquisition time; used for STIR in departments with strict quality standards
• STIR TSE-VFA (variable flip angle): 2.3× faster than standard STIR, reducing cumulative motion exposure; non-inferior diagnostic quality demonstrated [Pos-5]
• Patient instruction: "Breathe normally and try to breathe at a regular rate"

Dorsal CSF flow artefact (signal void) is a thoracic-specific phenomenon that is not encountered to the same degree in the cervical or lumbar spine. Pulsatile CSF flow in the dorsal thoracic subarachnoid space — combined with respiratory bulk flow — generates turbulent flow that creates focal, level-variable signal voids on T2. These appear as crescentic dark areas in the dorsal subarachnoid space that may simulate cord compression, syrinx, or cord atrophy. The characteristic features that distinguish flow void from pathology:

• Present on some slices but absent on adjacent slices (level-variable, non-anatomical pattern)
• Changes with repetition or between sequences
• No corresponding STIR abnormality
• No corresponding T1 signal change
• Characteristic crescentic or band-like shape following the posterior cord surface

Reduction strategies:

• Flow compensation (gradient moment nulling) on sagittal T2
• Cardiac gating eliminates it completely
• Comparison with STIR sequence (flow void present on T2, absent on STIR = artefact; pathology visible on both sequences)

Vertebral level counting error — not a signal artefact but a systematic diagnostic error that constitutes a positioning "artefact" in the broader sense. Incorrect level counting at the thoracic spine is a clinically serious error that has led to wrong-level surgery. Prevention strategies are described in Section 4.6.

Chemical shift artefact at discovertebral junctions: Same principle as lumbar and cervical spine. At the thoracic level, wide bandwidth is particularly important because the long FOV increases the area affected by any given chemical shift displacement.

Metal artefact from instrumentation: ACDF or posterior fixation hardware at the cervicothoracic junction, thoracic pedicle screws, vertebroplasty cement, and costovertebral fusion hardware all degrade image quality. Titanium hardware produces less artefact than stainless steel. TSE is less susceptible than GRE. MARS sequences (SEMAC/WARP/MAVRIC-SL) are available for severe artefact. For sequence-level protocol optimisation, vendor terminology and artefact management, see the dedicated MRIninja page Gradient Echo (GRE/FLASH) Sequence.

5.2 Protocol Efficiency and Throughput

Routine full protocol: Sagittal T2 + T1 + STIR + targeted axial T2 = approximately 30–40 minutes at 1.5T; 25–35 minutes at 3T with parallel imaging.

Abbreviated protocol for urgent myelopathy: Sagittal T2 + STIR (15–20 minutes) identifies cord signal, marrow oedema, and gross canal compromise at all thoracic levels. Axial T2 can be targeted to the identified pathology level. This abbreviated approach is appropriate for urgent clinical scenarios.

When 3D TSE is valuable: 3D isotropic T2 TSE (SPACE/CUBE/VISTA) with 1 mm isotropic voxels enables: reconstruction perpendicular to the cord axis at any level; coronal reformats for bilateral comparison and scoliosis assessment; sagittal reformats along the curved cord axis in kyphoscoliosis. Acquisition time at 3T is 6–10 minutes with compressed sensing. Evidence supports comparable disc herniation detection to 2D with the additional advantage of flexible reformatting [ref-3D].

Whole-spine protocol: When whole-spine coverage is clinically required (metastases, myeloma, spondyloarthropathy), the thoracic spine is covered as part of a whole-spine acquisition. Sagittal STIR of the full spine (cervical + thoracic + lumbar) is the most clinically efficient approach for whole-spine marrow screening. Acquisition time for whole-spine STIR at standard parameters is approximately 9–12 minutes; TSE-VFA implementation reduces this to approximately 4–5 minutes [Pos-5].

5.3 Field Strength Considerations

Clinical recommendation: 3T is preferred for cord signal assessment, demyelinating disease, and small lesion detection. 1.5T remains adequate for standard fracture, metastasis, and disc assessment, and is mandatory for patients with 1.5T-only conditional implants.

6. Contrast Use Principles Specific to Thoracic Spine MRI

6.1 Non-Contrast Standard Protocol — Sufficient For

The non-contrast standard protocol is adequate for: routine myelopathy assessment (initial structural survey); vertebral compression fracture evaluation (acute vs. chronic differentiation); disc herniation and canal assessment; spondylosis monitoring; screening for metastatic marrow disease (STIR highly sensitive); non-specific thoracic back pain with red flags; initial assessment in acute trauma.

6.2 Gadolinium Indicated — Thoracic Spine-Specific Contexts

Suspected thoracic epidural abscess or spondylodiscitis: GBCA is required for confirmation of diagnosis, extent characterisation, and differentiation from other causes of cord compression and marrow oedema. Full spine coverage strongly recommended as infection is often multi-level.

Suspected intradural extramedullary or intramedullary cord lesion: Meningiomas, nerve sheath tumours, ependymomas, astrocytomas, and leptomeningeal disease require enhancement for diagnosis and extent characterisation. A non-contrast protocol is insufficient for this category.

Post-operative thoracic spine with new neurological deterioration: Differentiation of post-operative change (epidural fibrosis, scar) from recurrent or new cord compression or infection requires contrast.

Suspected dural arteriovenous fistula (DAF): Post-contrast T1 may show serpentine intradural flow voids and enhancing draining veins. Dedicated vascular sequences (3D TOF, CE-MRA) are typically required in addition (child page).

Active inflammatory myelopathy (demyelination, NMOSD): Post-contrast T1 detects active blood-spinal cord barrier disruption in acute demyelinating lesions. Used when disease activity will change management.

Suspected neoplastic cord involvement: Intramedullary enhancement characterises neoplastic vs. non-neoplastic cord disease. Always required when cord mass lesion is identified.

Equivocal vertebral marrow lesion: When a lesion cannot be characterised on non-contrast sequences (atypical haemangioma vs. metastasis), post-contrast fat-suppressed T1 adds specificity.

6.3 Post-Contrast Acquisition Timing

Standard timing: 3–5 minutes post-injection for most thoracic spine indications. For suspected leptomeningeal disease or epidural involvement, delayed acquisition at 10–15 minutes increases sensitivity. Document injection time in PACS. Pre-contrast fat-suppressed T1 mandatory before injection.

7. Reporting Essentials

7.1 Interpretation Framework

Cord assessment is the first priority: All thoracic MRI reports must explicitly address cord signal, morphology, and canal compromise before any other findings. Thoracic cord lesions may be present without patient awareness.

Level precision: All vertebral and disc level findings must be specified using a clearly documented counting method. "Approximately T6–T7" is not acceptable for surgical planning; "T6–T7, counted from the C1–C2 articulation" is the required standard.

Acute vs. chronic framework (same as lumbar/cervical): T1 + STIR combination drives acuity assessment.

7.2 Mandatory Reporting Checklist

Spinal cord (primary target):

• Cord signal: T2 hyperintensity location, level, extent, acuity
• Cord diameter at each level: atrophy, expansion
• Canal compromise level and severity
• Thoracolumbar junction and conus level (include in every thoracic report)
• Dorsal subarachnoid flow void: distinguish from pathology

Vertebral bodies (T1–T12, each level):

• Alignment and height
• Bone marrow signal (T1 + STIR)
• Fracture deformity, cortical integrity
• Modic endplate changes
• Vertebral level explicitly documented with counting method

Intervertebral discs (T1–T2 to T12–L1):

• Disc height and signal
• Herniation: level, direction, calcification status (note CT recommendation if calcified)
• Disc-cord relationship at each level

Spinal canal and neural structures:

• Canal AP diameter at each level
• Thecal sac compression
• Cord position (central vs. displaced)
• Neural foraminal assessment (T2 axial)

Posterior elements and costovertebral junctions:

• Facet degeneration, effusion
• Ligamentum flavum thickening
• Costovertebral joint pathology if relevant

Paravertebral and paraspinal structures:

• Paraspinal muscle bulk and signal
• Posterior mediastinal masses, lymphadenopathy
• Aortic calibre (visible on lower thoracic sequences)

Technical items:

• Motion artefact impact on cord signal assessment
• Fat suppression quality
• Level counting method documented
• Comparison with prior studies

7.3 Structured Reporting

Indication → Technique → Comparison → Findings (cord first, then systematic by level, with counting method documented) → Impression (clinically actionable) → Limitations → Critical communication.

7.4 Incidental Findings — Clinical Decision Framework

Usually benign: Typical vertebral haemangiomas (T1 bright, T2 bright, well-defined); mild disc desiccation; age-appropriate kyphosis; small Schmorl's nodes; mild facet degeneration.

Requires documentation and follow-up: Indeterminate vertebral lesion (too small or atypical for haemangioma — recommend CT or dedicated MRI with contrast); aortic dilatation ≥3 cm; incidental paraspinal or posterior mediastinal mass; unexpected nerve root sheath dilatation (Tarlov cyst-like).

Urgent/clinically important: Unexpected cord compression or cord signal change; unexpected epidural mass; unexpected vertebral metastases or aggressive lesion; large aortic aneurysm (≥5 cm) — direct communication with referring clinician; suspected epidural abscess not previously identified.

8. MRI Technologist Pearls

8.1 Sequence Order Logic

Recommended standard order:

1. Three-plane localiser — verify C6 to L1 coverage; check position of anterior saturation band
2. Sagittal T2 TSE — first diagnostic sequence; planning reference for all others; anterior saturation band applied; flow compensation if available
3. Sagittal STIR — copied from sagittal T2 geometry; anterior saturation band applied
4. Sagittal T1 TSE — copied from sagittal T2 geometry
5. Axial T2 TSE — planned from the acquired sagittal T2; targeted to pathology levels or full thoracic coverage as indicated

Rationale: Sagittal T2 and STIR provide the maximum diagnostic yield per unit time. Axial planning requires review of the sagittal T2 and must be re-planned, not copied from the scout.

8.2 Positioning Tricks

• Anterior saturation band check before every sequence: This is the single most important quality measure for thoracic spine MRI. Verify band position on the planning image before starting. A misplaced band produces either (a) persistent cardiac/aortic ghosting (band too far from heart) or (b) vertebral signal suppression (band too far posterior).
• Knee support: Same as in the lumbar spine — reduces lower back discomfort and indirectly reduces gross motion during long thoracic acquisitions.
• Regular breathing instruction: "Breathe normally and try to keep a regular, slow breathing rate" before each sequence. Irregular breathing patterns generate worse ghosting than regular rhythmic breathing.
• Count levels from the localiser: During the localiser review, identify T1 and T12 on the sagittal view and document the counting in the scan notes before proceeding with the full protocol.
• STIR quality check at completion: Before ending the STIR acquisition, immediately review the first slices on the console to confirm fat suppression is uniform — thoracic STIR has the highest failure rate of any spine STIR sequence. If fat suppression failed, repeat the sequence before the patient leaves the scanner.
• Saturation band on axial sequences: Apply the anterior saturation band on axial sequences as well as sagittal — cardiac motion degrades axial T2 at the lower thoracic levels where the heart is closest.

8.3 Fast Salvage Protocol

Core minimum (emergency myelopathy): Sagittal T2 + STIR = 9–15 minutes; identifies cord signal and marrow disease at all thoracic levels.

8.4 Common Avoidable Errors

9. Quality Control Checklist

Coverage:

• Sagittal series includes C6–C7 junction superiorly
• Sagittal series includes conus medullaris at T12–L1 inferiorly
• Both costovertebral junctions visible on lateral sagittal slices
• Axial series covers pathology level(s) with adequate cranial and caudal margin
• Full thoracic axial coverage if clinically required

Sequence completeness:

• Sagittal T2: acquired, reviewed, anterior saturation band effective (cardiac ghosting minimised)
• Sagittal STIR: acquired, fat suppression visually uniform (subcutaneous fat is dark across full thoracic FOV)
• Sagittal T1: acquired, no major motion degradation
• Axial T2: planned from sagittal T2, correctly angulated perpendicular to cord at each level

Artefact assessment:

• Cardiac/aortic pulsation ghosting not obscuring cord signal on sagittal sequences
• Respiratory ghosting not obscuring marrow signal
• Dorsal CSF flow void noted if present and distinguished from pathology
• Fat suppression quality confirmed at full thoracic extent on STIR
• No phase wrap overlying cord or vertebral column

Level identification:

• Vertebral levels confirmed by at least one validated counting technique
• Conus level identified and documented
• Any uncertainty in level counting documented in acquisition notes

Contrast (if used):

• Pre-contrast T1 fat-suppressed acquired before injection
• Injection time documented
• Post-contrast acquisition timing documented
• STIR NOT acquired post-contrast

Labelling and orientation:

• Patient identifiers correct
• Left-right orientation verified on axial images
• Level labels confirmed on key pathological levels

Critical finding communication:

• Any unexpected cord compression, cord signal change, or epidural abscess flagged for immediate radiologist review and direct clinical communication

11. Evidence Gaps and Ongoing Debate

Cardiac gating necessity in standard thoracic protocol: Cardiac gating eliminates pulsation-related cord signal artefact but doubles acquisition time. No prospective study has quantified the diagnostic benefit-to-time cost ratio of cardiac gating in routine thoracic spine MRI (vs. MS-specific protocols). Expert consensus supports its use in dedicated cord lesion protocols but not in standard degenerative/oncological workflows.

STIR vs. TSE-VFA for thoracic STIR: TSE-VFA STIR is substantially faster and shows less respiratory artefact in the thoracic region [Pos-5], but comparative diagnostic accuracy for thoracic cord lesion detection (vs. standard STIR TSE) has not been prospectively evaluated.

DWI for vertebral fracture characterisation: Evidence supports DWI in distinguishing pathological from benign vertebral fractures, but optimal b-values, ADC thresholds, and acquisition parameters for the thoracic spine specifically are not standardised.

Whole-spine vs. targeted thoracic protocols in myelopathy: The clinical value of whole-spine coverage (cervical + thoracic + lumbar in one examination) versus targeted thoracic coverage in myelopathy workup is not established by controlled prospective data. The frequency of skip lesions and multi-region involvement in MS and neoplastic disease suggests whole-spine coverage is more sensitive, but the evidence is largely expert consensus.

3D TSE for standard thoracic protocol: Evidence for 3D isotropic T2 TSE equivalence to 2D TSE for thoracic spine disc and cord assessment is available primarily from lumbar spine studies with extrapolation; dedicated thoracic comparative evidence is limited.

Field strength and cord lesion detection: While 3T is theoretically superior for small cord lesion detection, no randomised comparative study has demonstrated improved clinical outcomes (as opposed to lesion count as a surrogate endpoint) in thoracic cord myelopathy with 3T vs. 1.5T.

AI reconstruction for thoracic spine: Deep learning reconstruction has been applied to spine MRI broadly but validation in the thoracic region — particularly for cord signal and motion-affected sequences — is at early stages.

12. Evidence-Based References

A. Guidelines / Consensus / Society Recommendations

[1] Agarwal V, Shah LM, Parsons MS, et al; Expert Panel on Neurological Imaging. ACR Appropriateness Criteria® Myelopathy: 2021 Update. J Am Coll Radiol. 2021;18(5S):S73–S82. (Evidence Level: High — Guideline) Relevance: Primary ACR guideline for myelopathy imaging; defines MRI as the preferred investigation for progressive myelopathy.

[2] Carelon Medical Benefits Management. Appropriate Use Criteria: Imaging of the Spine. Version 2022. (Evidence Level: High — Guideline) Relevance: Evidence-based indications for thoracic spine MRI in back pain, myelopathy, infection, and neoplasm.

[3] Ortiz AO, Levitt A, Shah LM, et al. ACR Appropriateness Criteria® Suspected Spine Infection. J Am Coll Radiol. 2021;18(11S):S488–S501. (Evidence Level: High — Guideline) Relevance: Defines MRI with contrast as the investigation of choice for suspected spinal infection including thoracic spondylodiscitis.

[4] Expert Panel on Musculoskeletal Imaging; ACR. ACR Appropriateness Criteria® Acute Spinal Trauma: 2024 Update. J Am Coll Radiol. 2025;22(5S). PMID: 40409895. (Evidence Level: High — Guideline) Relevance: Defines CT as first-line for acute thoracic trauma with MRI for ligamentous/cord injury.

[5] ACR Appropriateness Criteria® — Inflammatory Back Pain / Axial Spondyloarthritis: 2021 Update. J Am Coll Radiol. 2021;18:S340–S360. DOI: 10.1016/j.jacr.2021.08.003. (Evidence Level: High — Guideline) Relevance: Supports STIR for whole-spine marrow assessment in axial spondyloarthropathy including thoracic involvement.

B. Systematic Reviews / Meta-analyses

[6] Brinjikji W, Luetmer PH, Comstock B, et al. Systematic literature review of imaging features of spinal degeneration in asymptomatic populations. AJNR. 2015;36(4):811–816. PMID: 25430861. (Evidence Level: High — Systematic review) Relevance: Establishes prevalence of asymptomatic thoracic disc findings; informs cautious incidental disc reporting.

C. Important Original Studies

[7] 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. 2016;37(5):970–977. PMID: 26797141. DOI: 10.3174/ajnr.A4655. (Evidence Level: Moderate — Original prospective study) Relevance: Foundational evidence for STIR superiority in thoracic cord lesion detection (STIR 93.8% vs. FSE T2 71.9% vs. PSIR 50.8%); establishes STIR as mandatory over PSIR in thoracic spine.

[8] Lauzon ML, Frayne R, Bhatt MN, et al. Clinical Utility of a Novel Ultrafast T2-Weighted Sequence for Spine Imaging. AJNR. 2018;39(8):1568–1576. PMC: 7410539. (Evidence Level: Moderate — Original prospective study) Relevance: Demonstrates that whole-spine STIR TSE (9.5 min) accumulates substantially more respiratory motion artefact in the thoracic region compared to TSE-VFA; supports VFA-based acceleration for thoracic STIR.

D. Technical MRI Papers

[9] Peh WCG, Chan JHM. Artifacts in musculoskeletal and spinal MRI: a pictorial review. Skeletal Radiol. 2001;30(4):179–191. PMID: 11398948. (Evidence Level: Technical / Foundational) Relevance: Comprehensive reference for thoracic spine artefacts including dorsal CSF flow void, cardiac pulsation, and respiratory ghosting.

[10] Noda C, Venkatesh BA, Wagner JD, et al. Primer on Commonly Occurring MRI Artifacts and How to Overcome Them. RadioGraphics. 2022;42(3). DOI: 10.1148/rg.210176. (Evidence Level: Technical) Relevance: Practical artefact identification and reduction applicable to thoracic spine.

[11] 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. (Evidence Level: Technical) Relevance: Comprehensive spinal cord imaging methodology including cardiac gating, respiratory management, and reduced FOV DWI for the thoracic cord.

[12] Del Grande F, Santini F, Herzka DA, et al. Fat-suppression techniques for 3-T MR imaging of the musculoskeletal system. RadioGraphics. 2014;34(1):217–233. PMID: 24428290. DOI: 10.1148/rg.341135130. (Evidence Level: Technical) Relevance: Fat suppression comparison including STIR vs. SPAIR vs. Dixon; supports STIR preference for large-FOV applications including thoracic spine.

E. Landmark Historical References

[13] Modic MT, Steinberg PM, Ross JS, Masaryk TJ, Carter JR. Degenerative disk disease: assessment of changes in vertebral body marrow with MR imaging. Radiology. 1988;166(1 Pt 1):193–199. PMID: 3336678. (Evidence Level: High — Landmark) Relevance: Original Modic endplate change classification applicable to thoracic as well as lumbar spine.

[14] Bydder GM, Young IR. MR imaging: clinical use of the inversion recovery sequence. J Comput Assist Tomogr. 1985;9(4):659–675. PMID: 3998939. (Evidence Level: Foundational) Relevance: Original STIR sequence description; foundational reference for fat-suppressed oedema detection in the thoracic spine and beyond.

End of document — MRI THORACIC SPINE Generic Standard Protocol — MRIninja Master Page v1.0 — April 2026

This document is designed to serve as the reference base for all child pages dedicated to specific thoracic spine pathologies, clinical indications and dedicated protocols.