MRI Lumbar Spine – Generic Standard Protocol

MRIninja Knowledge Base | Master / General Page Version 2.0 — April 2026 | Evidence review through April 2026 Audience: Radiologists · Neuroradiologists · MRI Technologists · Advanced Students

1. Executive Summary

Magnetic resonance imaging is the gold-standard modality for the evaluation of lumbar spine pathology. Its fundamental advantages over other imaging modalities — superior soft tissue contrast, absence of ionising radiation, multiplanar capability, and simultaneous visualisation of osseous, discal, ligamentous, neural and vascular structures — make it the first-line investigation for the vast majority of clinical indications in this anatomical district [1, 2].

The generic standard lumbar spine protocol is designed as a diagnostically broad acquisition: it must detect or exclude the most prevalent pathological categories including intervertebral disc degeneration and herniation, central canal and foraminal stenosis, facet joint disease, ligamentous hypertrophy, vertebral body pathology, and epidural disease. It is not designed to maximise sensitivity for a single, specific pathological entity. This distinction is fundamental: the generic protocol is a triage tool of high breadth but variable depth.

Compared with computed tomography (CT), lumbar spine MRI offers decisive superiority for soft tissue discrimination, direct visualisation of the spinal cord and nerve roots, assessment of disc hydration and annular integrity, bone marrow characterisation, and detection of epidural and paravertebral soft tissue processes without radiation exposure [3]. CT retains advantages for acute fracture cortical detail, spondylolisthesis with dynamic imaging, complex osseous anatomy pre-surgery, and situations where MRI is contraindicated or unavailable. Plain radiography provides only indirect structural information and is insufficient for neural compression, disc pathology or soft tissue disease assessment. Ultrasound has no meaningful role in lumbar spine diagnostic imaging.

1.1 Core Strengths

• Nerve root and thecal sac assessment: Direct neural structure visualisation under compression without radiation or intrathecal contrast [4].
• Disc characterisation: Assessment of nucleus pulposus hydration, annular tears, disc height loss, endplate changes, and herniation morphology with unmatched detail [5].
• Bone marrow evaluation: Sensitive detection of marrow oedema, infiltration, and fatty replacement — enabling early diagnosis of vertebral metastases, osteoporotic fractures, and inflammatory endplate changes [6].
• Multiplanar acquisition: Sagittal, axial, and coronal planes without repositioning.
• No ionising radiation: Critical advantage for serial follow-up and younger populations.
• Incidental finding detection: Paraspinal, retroperitoneal and pelvic findings identifiable within the field of view.

1.2 Intrinsic Limitations of the Generic Protocol

The generic standard protocol represents an intentional compromise between diagnostic breadth, acquisition time, field of view, and SNR. These trade-offs must be explicitly understood.

Temporal resolution: Single static acquisition only. Dynamic stenosis and position-dependent symptoms are not assessed.

Spatial resolution trade-offs: Routine 2D TSE acquisitions optimise SNR within acceptable acquisition times but sacrifice the isotropic resolution of 3D acquisitions. Small annular tears, micro-foraminal compromise, and subtle endplate lesions may be at or below reliable detectability. For sequence-level protocol optimisation, vendor terminology and artefact management, see the dedicated MRIninja page Turbo Spin Echo (TSE/FSE) Sequence.

Field of view constraints: The standard protocol focuses on L1–S1. Pathology at the thoracolumbar junction, sacrum, or sacroiliac joints may be incompletely assessed unless FOV is deliberately extended.

Post-operative limitations: The non-contrast protocol cannot differentiate recurrent disc herniation from epidural fibrosis — gadolinium is required (see dedicated child page).

Inflammatory and neoplastic disease: The standard protocol detects gross disease but lacks the sensitivity of dedicated sequences (STIR, fat-saturated T1, DWI, post-contrast) optimised for bone marrow infiltration or epidural extension. For sequence-level protocol optimisation, vendor terminology and artefact management, see the dedicated MRIninja page STIR Sequence.

Metallic implants: Susceptibility artefacts from surgical hardware significantly degrade image quality and require dedicated MARS/SEMAC/MAVRIC-SL sequences not included in the standard protocol.

No functional assessment: Neural function, cord perfusion, and axonal integrity are not assessable.

When a dedicated child protocol is required: post-operative evaluation with gadolinium, suspected infection or inflammatory spondyloarthropathy, primary or metastatic spinal neoplasm, vascular malformations, detailed sacroiliac joint assessment, high-resolution nerve root imaging, and any case with metallic implants.

2. Main Clinical Indications

2.1 Standard Indications

Low back pain with radicular symptoms (lumbosacral radiculopathy): The most frequent indication globally. MRI is indicated when conservative management has failed (typically after 4–6 weeks), neurological deficit is present, or red flags are identified [7]. The standard protocol reliably detects disc herniation, foraminal and lateral recess stenosis, and nerve root compression. Generally sufficient for initial diagnostic workup in uncomplicated radiculopathy.

Neurogenic claudication and lumbar spinal stenosis (LSS): MRI is the preferred modality for pre-operative planning and confirming diagnosis in suspected degenerative LSS. The standard protocol quantifies central canal dimensions, lateral recess narrowing, and multi-level disease extent. Standing MRI may offer incremental value in patients with discordant clinical and imaging findings, but is not universally available [8].

Non-specific low back pain with clinical suspicion of significant structural pathology: Guidelines do not support routine imaging for uncomplicated non-specific LBP without red flags [7, 9]. When clinical assessment identifies features warranting investigation, the standard protocol provides a broad structural survey.

Pre-operative planning for degenerative disc disease: The standard protocol informs surgical decision-making regarding level selection, foraminal dimensions, and adjacent segment status. CT provides complementary osseous anatomy detail in complex cases.

Post-traumatic evaluation (non-acute): In the sub-acute and chronic post-traumatic setting, MRI assesses ligamentous integrity, disc injury, vertebral marrow oedema, epidural haematoma resolution, and neural compromise.

Suspected spondylolisthesis workup: MRI demonstrates neural compression, disc disease, and posterior element status. Dynamic assessment typically requires supplementary radiography or upright MRI.

Degenerative disc disease monitoring and serial follow-up: Serial MRI monitors disease progression, treatment response, and is used in clinical trial settings. Standardised positioning and sequence parameters are essential for valid comparison.

Scoliosis evaluation (selected cases): Indicated in atypical scoliosis presentations, rapid progression, or suspected intraspinal anomaly. Coronal 3D acquisitions may be required for full deformity characterisation.

Suspected Modic endplate changes: Bone marrow oedema at the discovertebral junction is well characterised by the standard protocol. Modic types 1–3 are reliably differentiated on standard sagittal T1 and T2 sequences [10].

Lumbar spine protocol — practice patterns in the United States: A 2024 survey of 193 musculoskeletal radiologists found that the most common protocol combines sagittal T1, T2, and STIR sequences (77.7% of respondents), with axial T1+T2 pair (54.5%) or T2 alone (34.2%) for axial imaging [A-survey]. This data confirms wide consensus around the core sagittal triad with departmental variation in axial and conditional sequences.

2.2 Urgent Red Flags Requiring Expedited or Emergency Imaging

Note: For acute traumatic cord injury, CT remains first-line for osseous evaluation in most trauma protocols, with MRI added for ligamentous, disc, and cord parenchymal assessment.

3. Preparation Reference

3.1 Anatomy-Specific Preparation Items

Metallic implants and prior surgery: Pedicle screws, interbody cages, dynamic stabilisation devices, interspinous spacers, spinal cord stimulators, and intrathecal pump catheters must be identified before imaging. These may contraindicate MRI, require conditional clearance, or necessitate a dedicated MARS protocol. The standard generic protocol is diagnostically inadequate in patients with extensive posterior instrumentation.

Clothing and belt buckles: Metal clasps, underwired bras, and belt buckles near the lumbar region introduce susceptibility artefacts. Patients should receive examination gowns. Trouser zips at lumbar level produce focal artefact.

Pain management considerations: Acute lumbar pain causes motion degradation. If clinically appropriate, analgesia before the examination reduces motion artefact and improves diagnostic yield. Severe mobility limitations must be communicated to the technologist before positioning.

Bladder and bowel preparation: Patients should void before examination. Full bladder and bowel gas cause motion-related artefact on longer acquisitions and may distort FOV positioning for lower lumbar sequences. Bowel preparation is not routinely required, but excessive peristalsis contributes to phase-encoding artefact in coronal acquisitions.

Patient history affecting protocol design:

• Prior lumbar surgery → contrast protocol (child page)
• Active or suspected infection → STIR and contrast (child page)
• Known malignancy → extended coverage, DWI, contrast (child page)
• Pregnancy → non-contrast protocol mandatory
• Severe renal insufficiency (eGFR < 30 mL/min/1.73m²) → contrast risk stratification per local GBCA protocol
• Scoliosis or significant deformity → coronal acquisitions, extended FOV

Claustrophobia: Patients tolerate the lumbar spine position better than brain or cardiac MRI due to head near the bore entrance. Reassure patients that the head often remains outside. Anxiolytic protocol may be required in severe cases.

3.2 Patient Positioning on the MRI System

Patient position: Supine, feet-first entry. Universal and accepted standard for lumbar spine MRI.

Coil selection: A dedicated spine phased-array surface coil is mandatory. Modern systems use a posterior spine coil integrated within the table combined with an anterior surface element over the abdomen/pelvis. Verify correct coil element activation on the console before scanning.

Centering: The isocentre should be positioned at the iliac crest level, corresponding approximately to L4–L5. This places the critical mid-lumbar and lumbosacral levels within maximum coil sensitivity. For tall patients or protocols requiring T12–S2 coverage, repositioning or two separate acquisitions may be needed.

Anatomical alignment: The spine should align with the table axis in the coronal plane. Lateral rotation introduces asymmetry in axial slice planning and misrepresents foraminal dimensions. A foam cushion under the knees (approximately 15–20 cm) reduces lumbar lordosis, improves comfort, and reduces posterior element overlap on axial images.

Immobilisation: Abdominal compression straps are not used for lumbar spine MRI (unlike abdominal MRI). Patients should remain still, breathe normally, and avoid coughing. Foam padding stabilises the knees and prevents pelvic rotation.

Pre-scan technologist checks:

1. Confirm correct coil activation on console.
2. Verify alignment laser crosses at iliac crest level.
3. Ensure no metallic items remain within the field of view (belt area, trouser zip).
4. Confirm patient comfort and ability to remain still.
5. Document mobility limitations or inability to fully extend lower limbs.
6. Acquire localiser immediately and verify lumbar spine is within FOV before proceeding.

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. Bright nucleus pulposus in hydrated discs; bright CSF; dark cortical bone. Directly depicts disc degeneration (Pfirrmann grading [11]), disc herniation, canal stenosis, conus morphology, and spinal alignment. Must be acquired first: it is used as the planning reference for all other sequences. Quality of this sequence defines the diagnostic value of the entire examination.

What it detects well: disc height and hydration, herniation morphology, canal compromise, conus and cauda equina, vertebral body height, gross bone marrow signal change, ligamentum flavum thickness (T2 dark), CSF space.

What it misses: subtle bone marrow oedema (requires STIR), fat-isointense marrow disease, epidural fat obliteration (better on T1), soft tissue characterisation without fat suppression.

Technologist note: Gibbs ringing artefact at the cord surface can simulate a syrinx; verify with wider matrix or dedicated thin-slice acquisition. CSF pulsation ghosting in the AP direction is the dominant artefact; verify phase direction is AP before acquisition.

Sagittal T1 TSE — bone marrow characterisation. Bright fat-containing normal marrow on T1 is the baseline against which any bone marrow pathology is identified: T1 signal loss in the marrow (dark on T1) indicates replacement of fat by infiltrate, oedema, or sclerosis. This is the critical sequence for vertebral metastasis screening and for characterising Modic endplate changes.

What it detects well: bone marrow T1 signal — T1 dark + STIR bright = acute oedema/Modic 1; T1 bright (fat) = chronic Modic 2; T1 dark + STIR dark = sclerosis/Modic 3. Subacute haemorrhage (methemoglobin T1-bright). Epidural fat outline. Post-gadolinium T1 enhancement (when fat-suppressed).

What it misses: disc dehydration less conspicuous than on T2; neural compression less well shown; soft tissue oedema may be masked by background fat signal.

Sagittal STIR — the bone marrow oedema and inflammation sentinel sequence. STIR nulls fat T1 (regardless of frequency — B0 independent), leaving only tissue with elevated water content (oedema, inflammation, neoplastic infiltration, infection). The combination of T1 suppression and additive T1+T2 contrast enhancement makes STIR uniquely sensitive to bone marrow oedema.

What it detects well: vertebral marrow oedema (acute fracture, Modic 1, infection, metastasis), disc inflammation, paraspinal soft tissue oedema, nerve root oedema, ligamentous tears (increased water content).

What it misses: chronic changes with little oedema (Modic 2, stable sclerosis); lower SNR than T2 — cannot replace T2 as the primary structural overview.

Critical pitfall: STIR must never be acquired after gadolinium. Gadolinium shortens T1 of enhancing tissues to values overlapping fat T1; these tissues will be suppressed rather than highlighted, producing false-negative STIR images. STIR must always precede contrast administration.

Axial T2 TSE (disc-level series) — the nerve root compression and foraminal stenosis sequence. Planned individually per disc level, parallel to each intervertebral disc. Provides cross-sectional anatomy of the disc, thecal sac, lateral recesses, neural foramina, and facet joints at each level. Cannot be replaced by the sagittal series alone: foraminal compromise and far-lateral herniations are only reliably assessed in the axial plane.

What it detects well: disc herniation direction and type (central, paracentral, foraminal, far-lateral), lateral recess stenosis, facet degeneration and synovial cysts, epidural soft tissue, nerve root morphology.

What it misses: disc hydration less conspicuous than sagittal T2; conus assessment not possible.

Technologist note: Each disc level must be individually angulated parallel to the disc space — this is non-negotiable. Slices planned only from the scout (not from the sagittal T2) produce incorrect angulations. Failure to include the subarticular recess (above and below each disc) is the most common coverage error.

Axial T1 TSE (disc-level series) — complementary to axial T2. Bright epidural fat on T1 provides natural contrast delineating thecal sac and nerve roots. Loss of epidural fat signal is an indirect sign of space-occupying pathology. Particularly useful for foraminal fat obliteration (indicating foraminal compromise) and for differentiating disc material from blood products.

Departmental practice: In the United States, 54.5% of lumbar spine protocols include axial T1+T2 pairing; 34.2% use axial T2 alone [A-survey]. Axial T1 is omittable in time-limited or abbreviated protocols for uncomplicated radiculopathy when bone marrow pathology is not suspected.

4.4 Sequence Matching and Cross-Sequence Consistency

Sagittal series geometric matching: Sagittal T1, T2, and STIR must be acquired with identical or closely matched slice thickness, gap, FOV, number of slices, and centring. These three sequences are the primary comparison set: T1 vs STIR signal pattern at each vertebral body defines the Modic change type and fracture acuity. Any geometric mismatch renders this comparison unreliable. On modern scanners, copy the slice geometry from the sagittal T2 to the sagittal T1 and STIR using the console copy-geometry function.

Axial series matching: Axial T2 and T1 disc-level series must share identical angulation at each level. Copy geometry from axial T2 to axial T1.

Pre- and post-contrast matching: In contrast examinations, pre-contrast T1 fat-suppressed must precisely match post-contrast. Geometry mismatch invalidates enhancement comparison. Motion between acquisitions must be flagged. Automated power injector is strongly recommended to maintain patient stillness.

Serial follow-up reproducibility: For treatment monitoring and clinical trial imaging, identical slice geometry, coil configuration, and field strength must be maintained across examinations. Archive protocol parameter and geometry files per patient.

4.5 Fat Suppression — Regional Principles

Fat suppression in lumbar spine MRI is sequence-dependent, indication-driven, and field-strength sensitive.

STIR (Short-TI Inversion Recovery): The preferred fat suppression method for sagittal bone marrow oedema detection. Nulls fat by its T1 (TI ≈ 150–175 ms at 1.5T; ≈ 180–220 ms at 3T), independent of spectral frequency. Robust to B0 inhomogeneity across the large FOV of the lumbar spine. Cannot be used post-gadolinium.

Spectral fat saturation (SPIR/SPAIR/ChemSat): Higher SNR than STIR; better suited for post-contrast T1 imaging. Susceptible to B0 inhomogeneity at the large lumbar spine FOV, particularly at the lumbosacral junction and body edges. At 3T, inhomogeneous fat saturation is more frequent than at 1.5T. SPAIR (spectral attenuated inversion recovery) provides better B0 robustness than simple CHESS while maintaining higher SNR than full STIR [12].

Dixon technique: Multi-echo acquisition producing water-only, fat-only, in-phase, and opposed-phase images. Most robust fat suppression available — independent of B0 inhomogeneity. Increasingly preferred at 3T for both pre- and post-contrast T1 sequences. Evidence shows Dixon T2 and T1 images score higher for fat suppression uniformity than STIR and CHESS respectively in lumbar spine imaging [13]. Fat-only images additionally characterise fat-containing lesions (haemangioma, lipoma). Fat-water swapping artefacts may occur in regions of extreme B0 heterogeneity.

When fat suppression is not used:

• Standard sagittal T1 (bright marrow fat is the diagnostic signal)
• Standard axial T2 (epidural fat provides natural contrast for disc/root differentiation)
• Standard axial T1 (same reason)

5. Optimisation Strategy

5.1 Artifact Reduction by Source

Motion artefact is the most frequent cause of non-diagnostic lumbar spine MRI. Voluntary repositioning, coughing, breathing, and involuntary bowel peristalsis transmitted through body mass all degrade image quality. Motion produces ghosting along the phase-encoding direction, blurring of disc margins, and loss of nerve root definition. Artefacts may simulate epidural pathology, disc prolapse, or cord signal change. Reduction strategies: phase direction optimisation (Section 4.6); pre-examination analgesia in severe pain; brief patient communication before each sequence; shorter acquisition times where SNR permits; anti-peristaltic agents (glucagon, hyoscine butylbromide) are not routinely used but may be considered for severe bowel motion cases. Any motion artefact obscuring the disc margin, thecal sac contour, or nerve root at a clinically relevant level requires repeat of that sequence before the patient leaves.

CSF/vascular pulsation artefact: Lumbar CSF pulsates with the cardiac cycle; epidural venous plexus adds a secondary source. Ghost images of the thecal sac are displaced along the phase direction, potentially simulating disc herniation, epidural mass, or cord lesion. AP phase direction on sagittal sequences displaces ghosts anteriorly away from posterior elements. Flow compensation (gradient moment nulling) can reduce vascular ghosts but increases minimum TE and reduces slice count per TR. Cardiac gating is rarely used for routine lumbar MRI due to time overhead.

Chemical shift artefact (first order): Displacement at fat-water interfaces (endplate-disc junction, nerve root-foramen fat) creates bright and dark bands at boundaries in the frequency-encoding direction. Dark band can simulate narrow disc height or cortical erosion. At 3T, chemical shift is twice as prominent as at 1.5T at equivalent bandwidth. Increasing receive bandwidth reduces displacement; this is particularly important at 3T.

Susceptibility artefact: Ferromagnetic implants cause local B0 distortion producing geometric distortion, signal void, and signal pile-up. TSE sequences are less susceptible than GRE sequences. SEMAC/MAVRIC-SL/WARP sequences dramatically reduce hardware artefact but require dedicated protocol design and significantly increased acquisition time. For sequence-level protocol optimisation, vendor terminology and artefact management, see the dedicated MRIninja page Gradient Echo (GRE/FLASH) Sequence.

Aliasing (wrap-around): Patient anatomy extending beyond the FOV in the phase direction wraps onto central structures. Most common: anterior abdominal structures overlying the spine (AP phase direction, sagittal). Prevention: increase FOV; use no-phase-wrap/phase oversampling.

Gibbs ringing: Truncation artefact from finite k-space sampling creates oscillating bands at high-contrast interfaces (CSF-cord, disc-thecal sac). Can simulate syrinx at cord surface. Prevention: adequate matrix size; caution with partial Fourier; zero-filling.

5.2 Protocol Efficiency and Throughput

Routine full protocol: Sagittal T2 + T1 + STIR + axial T2 + axial T1 = approximately 25–35 minutes at 1.5T; 20–28 minutes at 3T with parallel imaging.

Abbreviated/abbreviated protocol for low-complexity indications: Sagittal T2 + axial T2 (12–15 minutes) is adequate for disc herniation and canal compromise assessment. A 2023 study showed limited T2-only protocol has acceptable diagnostic performance for degenerative disc disease in pain intervention clinics [54]. Sagittal STIR should be added whenever acute fracture, infection, or neoplasm is a clinical possibility.

When 3D is worth the time: Isotropic 3D T2 TSE (SPACE/CUBE/VISTA) enables multiplanar reformatting from a single acquisition. Comparable sensitivity to standard 2D for disc herniation and canal compromise has been demonstrated at 1.5T with significantly shorter scan time (8.5 vs 18.7 minutes) [38]. 3D acquisitions particularly benefit complex anatomy, scoliosis, and high-resolution foraminal assessment.

When 2D is more robust: 2D acquisitions are less affected by patient motion (motion corrupts only the affected slice); offer better bone marrow characterisation with dedicated T1 and STIR sequences; have more predictable contrast behaviour. For standard clinical indication in cooperative patients, 2D remains the reference standard.

5.3 Field Strength Considerations

At 3T, the longer tissue T1 relaxation times require TR increases in T1 sequences (TR 550–800 ms vs 450–700 ms at 1.5T) to maintain comparable T1 contrast. SAR constraints may limit ETL at 3T, particularly for STIR sequences; variable flip angle TSE (TSE-VFA) approaches reduce SAR while maintaining comparable T2 contrast [45,46].

6. Contrast Use Principles Specific to Lumbar Spine MRI

6.1 Non-Contrast Standard Protocol — Sufficient For

The non-contrast standard protocol is diagnostically adequate for: uncomplicated radiculopathy and disc herniation evaluation; central canal and foraminal stenosis assessment; degenerative disc disease grading; Modic endplate change characterisation; pre-operative planning in degenerative disease; acute vertebral fracture evaluation; spondylolisthesis assessment; screening for incidental pathology.

6.2 Gadolinium Indicated — Lumbar Spine-Specific Contexts

Post-operative lumbar spine (most critical indication): Gadolinium is required to differentiate recurrent or residual disc herniation (does not enhance) from epidural fibrosis (enhances uniformly), which cannot be reliably distinguished on non-contrast imaging alone [15, 16]. Pre-contrast T1 fat-suppressed must always precede injection.

Suspected spondylodiscitis (infection): Contrast enhancement characterises the activity of infection, differentiates infectious from Modic degenerative changes, and demonstrates paraspinal and epidural abscess extent. Gadolinium is required for diagnosis confirmation and surgical planning.

Suspected primary or metastatic spinal neoplasm: Enhancement characterises lesion vascularity, delineates epidural extent, and distinguishes benign from aggressive lesions. Post-contrast sequences with fat suppression maximise lesion-to-background contrast.

Leptomeningeal disease / nerve root enhancement: Intradural nerve root enhancement suggests inflammatory, infectious, or neoplastic involvement of nerve roots or meninges and is only detectable post-contrast.

Inflammatory spondyloarthropathy (selected cases): Gadolinium may be added when STIR findings are equivocal and active sacroiliitis or vertebral inflammation characterisation will change management. ASAS/EULAR guidelines note that gadolinium is not routinely required if STIR is positive [17, 18].

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

6.3 Post-Contrast Acquisition Timing

Standard timing: Post-contrast T1 fat-suppressed sequences are acquired within 3–10 minutes of injection for most lumbar spine indications. This window optimises enhancement of granulation tissue (post-operative fibrosis), inflammatory tissue, and vascular lesions while intravascular gadolinium concentration has partially equilibrated.

Post-operative disc vs scar: The well-established clinical standard requires immediate post-contrast imaging (within 5 minutes) to show enhancement of fibrosis; delayed imaging (> 30 minutes) risks non-specific disc herniation enhancement by diffusion, potentially confounding the diagnosis. The pre-contrast T1 fat-suppressed acquired immediately before injection is the critical comparison.

Injection time documentation: The exact time of injection must be documented in PACS and the report for all post-contrast examinations. Interpretation of enhancement patterns depends on timing.

7. Reporting Essentials

7.1 Interpretation Framework

Lumbar spine MRI interpretation requires systematic assessment of all anatomical compartments. The primary diagnostic axes are:

Degenerative vs non-degenerative: The vast majority of lumbar spine MRI findings are degenerative. The interpreter must actively identify features inconsistent with simple degeneration: aggressive bone marrow changes, multiple vertebral levels with signal abnormality, paraspinal tissue changes, fever history.

Acute vs chronic: T1 dark + STIR bright = acute oedema; T1 bright (fat) = chronic fatty change/Modic 2; T1 dark + STIR dark = sclerosis. This combination distinguishes acute from chronic vertebral changes without contrast.

Focal vs diffuse: Diffuse bone marrow signal change (multiple levels, uniform) is more likely neoplastic or inflammatory; focal single-level involvement may be degenerative, metastatic, or traumatic.

Neural compromise assessment: Always address the conus, thecal sac, cauda equina, and each nerve root at each level systematically. Do not assume the clinician has reviewed the axial images.

7.2 Mandatory Reporting Checklist

Vertebral bodies (each level): alignment and height; bone marrow signal (T1 + STIR); fracture deformity; Modic endplate changes (type and level).

Intervertebral discs (L1–L2 to L5–S1): disc height; T2 signal/Pfirrmann grade; disc herniation (type, location, level, nerve root compression); Schmorl's nodes.

Spinal canal and neural structures: conus medullaris (level, signal, morphology); thecal sac compression and CSF space; cauda equina nerve root appearance; central canal dimensions at each level; lateral recess stenosis; neural foraminal stenosis (each level, each side).

Posterior elements: facet joints (degeneration, effusion, synovial cysts); ligamentum flavum (thickness, hypertrophy, buckled appearance); pars interarticularis.

Paravertebral structures: paraspinal muscle bulk and signal; retroperitoneal structures visible in FOV (aorta, iliac vessels, kidneys).

Technical items: motion artefact impact; fat suppression uniformity; FOV adequacy; coil performance; comparison with prior studies; contrast agent, dose, injection time, and post-contrast acquisition time if used.

7.3 Structured Reporting

Reports should follow a structured format: Indication → Technique (field strength, sequences, coil, contrast) → Comparison → Findings (systematic by compartment) → Impression (concise, clinically actionable) → Limitations → Critical communication if required. Structured reporting improves inter-reader agreement and reduces clinically significant omissions [19].

7.4 Incidental Findings — Clinical Decision Framework

Usually benign, no action required: typical vertebral haemangiomas (T1 bright/T2 bright, well-defined); small non-aggressive Schmorl's nodes; mild facet degeneration; disc desiccation in older patients; moderate epidural lipomatosis without compression; small non-obstructing renal cysts (Bosniak I); normal aortic calibre.

Requires documentation and follow-up: indeterminate vertebral lesion (too small or atypical — consider dedicated MRI with contrast or CT); aortic dilatation ≥3 cm; incidental renal lesion Bosniak ≥IIF; significant paraspinal mass; unexpected nerve root enhancement.

Urgent / clinically important: unexpected vertebral metastasis or aggressive lesion; incidental aortic aneurysm ≥5 cm; large retroperitoneal or paraspinal mass; epidural haematoma or abscess not previously suspected — immediate clinical communication required.

8. MRI Technologist Pearls

8.1 Sequence Order Logic

Recommended standard order:

1. Three-plane localiser — immediately review for correct coverage
2. Sagittal T2 TSE — highest diagnostic priority; used as reference for all subsequent planning
3. Sagittal STIR — planned from sagittal T2 geometry
4. Sagittal T1 TSE — planned from sagittal T2 geometry; short and robust
5. Axial T2 TSE (disc-level series) — most motion-sensitive; planned from sagittal T2
6. Axial T1 TSE (disc-level series) — planned from axial T2 geometry

Rationale: If the patient cannot complete the examination, the sagittal T2 and STIR provide maximum diagnostic yield per sequence. The axial series requires active re-planning from the sagittal T2 and should be acquired after the sagittal series is confirmed adequate. Contrast sequences (pre and post-contrast fat-suppressed T1) are acquired at the appropriate position in the protocol for the clinical indication.

8.2 Positioning Tricks

• Knee cushion: foam wedge 15–20 cm under the knees reduces lumbar lordosis; significantly improves comfort for patients with acute pain; reduces posterior element overlap on axial images.
• Feet position: slight leg abduction or foot support reduces involuntary pelvic rotation from external foot rotation in supine position.
• Communication: brief verbal reminder before each sequence — "Please stay still, breathe normally, and try not to cough" — measurably reduces motion on longer acquisitions.
• Coil positioning check: physically verify that the spine coil is centred at iliac crest level. Coil displaced superiorly reduces sensitivity at the lumbosacral junction — the most clinically critical region.
• Transitional anatomy identification: if lumbarisation of S1 or sacralisation of L5 is suspected on the localiser, document and alert the radiologist. Ensure adequate coverage for accurate level counting.

8.3 Fast Salvage Protocol

Core minimum (two-sequence protocol for radiculopathy): Sagittal T2 + Axial T2 = 9–14 minutes, covering disc herniation and neural compression at all levels.

8.4 Common Avoidable Errors

9. Quality Control Checklist

Coverage:

• [ ] Sagittal series includes conus (T12–L1) superiorly
• [ ] Sagittal series includes S1–S2 inferiorly
• [ ] Both foramina visible on lateral sagittal slices
• [ ] Axial series covers all five disc levels (L1–L2 through L5–S1)
• [ ] Each axial block covers pedicle-to-pedicle at each level

Sequence completeness:

• [ ] Sagittal T2: acquired, reviewed, no major motion degradation
• [ ] Sagittal STIR: acquired, fat suppression visually uniform (subcutaneous fat nulled)
• [ ] Sagittal T1: acquired, no major motion degradation
• [ ] Axial T2: all five levels planned from sagittal T2 (not scout), correctly angulated
• [ ] Axial T1: acquired if indicated per local protocol

Image quality:

• [ ] No diagnostic-grade motion artefact at clinically critical levels
• [ ] STIR: uniform fat nulling across vertebral column (check subcutaneous fat)
• [ ] No phase wrap-around overlying spinal canal
• [ ] No significant Gibbs ringing mimicking syrinx
• [ ] Susceptibility artefact (if hardware present): documented, extent characterised

Contrast (if used):

• [ ] Pre-contrast T1 fat-suppressed acquired before injection
• [ ] Injection time documented in PACS and report
• [ ] Post-contrast acquisition timing documented
• [ ] STIR was NOT acquired post-contrast

Labelling and orientation:

• [ ] Patient identifiers correct on all series
• [ ] Left-right orientation verified on axial images
• [ ] Series correctly labelled in PACS
• [ ] Transitional anatomy flagged if present

Critical finding communication:

• [ ] Any cauda equina compression, epidural abscess, cord signal change, unexpected metastasis or haemorrhage flagged for immediate radiologist review and clinical communication

11. Evidence Gaps and Ongoing Debate

Optimal protocol abbreviation: No consensus on minimum sequence set for specific low-complexity indications. Available studies show comparable accuracy for disc herniation with abbreviated T2-only protocols [54] and rapid 3D SPACE protocols [38], but impact on non-disc pathology detection is understudied. The clinical implication of omitting STIR for acute fracture or early metastasis detection in abbreviated protocols is unknown.

2D vs. 3D TSE equivalence: Comparable accuracy for disc herniation and stenosis has been demonstrated [38, 41], with potential 3D advantage for foraminal and nerve root assessment [44]. Most studies are retrospective, single-centre, and use expert readers; generalisation to routine clinical practice requires validation.

Axial T1 necessity: Survey data shows 45.5% of US radiologists use axial T2 alone without axial T1 [A-survey]. No prospective study has defined the diagnostic cost of omitting axial T1 for specific clinical presentations.

Optimal field strength: No prospective randomised comparison of 1.5T vs. 3T diagnostic accuracy for specific lumbar spine clinical outcomes exists. Expert consensus favours 3T for complex or high-resolution indications; 1.5T remains adequate for standard protocols.

DWI role in routine protocols: Emerging evidence supports DWI for vertebral fracture characterisation and metastasis screening, but optimal b-values, ADC cutoffs, and acquisition parameters for lumbar spine remain areas of active investigation without definitive consensus.

Contrast necessity in borderline indications: The boundary between cases requiring contrast (post-operative, infection, neoplasm) and those where it is of marginal value is supported more by expert consensus than controlled prospective data. Over-use of GBCA in non-indicated lumbar spine cases remains a practice problem.

AI-assisted reconstruction: Deep learning reconstruction (Siemens Deep Resolve, Philips SmartSpeed, GE AIR Recon DL) enables acquisition time reduction or resolution improvement. Clinical validation for lumbar spine across pathologies is at early stages. For sequence-level protocol optimisation, vendor terminology and artefact management, see the dedicated MRIninja page Spin Echo DWI / Non-EPI DWI Sequence.

Variable flip angle TSE (TSE-VFA): Shows approximately 2× speed improvement over conventional TSE with comparable contrast and non-inferior diagnostic quality [45, 46]; SAR reduction is particularly relevant at 3T. Adoption is growing but not yet standard.