High-yield overview
Technique, normal anatomy and pathological patterns for the orthopaedic exam
18.8 vs 12.1Median nerve mm² at pisiform, CTS vs control
3TPreferred field strength
<1 mmIn-plane resolution required
<3 mmSlice thickness
FascicularNormal pattern, lost in pathology
Seddon Classification
Neurapraxia: Conduction block, nerve intact
Axonotmesis: Axon disrupted, endoneurium intact
Neurotmesis: Complete nerve disruption
Key: MRI can show continuity but not differentiate axonotmesis from neurotmesis reliably
Critical Must-Knows
- Normal nerve: intermediate T1, slightly high T2, fascicular pattern
- Pathology: nerve enlargement, T2 hyperintensity, loss of fascicular pattern
- 3T MRI preferred for nerve imaging (higher SNR)
- PD or T2 fat-sat sequences best for nerve visualisation
- Denervation oedema in muscle indicates upstream nerve injury
Clinical Pearls
- “Carpal tunnel: median nerve enlarged at the pisiform (18.8 vs 12.1 mm² in controls) — no validated MRI cut-off; the 10mm² figure is an ultrasound threshold
- “Cubital tunnel: ulnar nerve enlargement proximal to tunnel
- “Brachial plexus: roots, trunks, divisions, cords, branches
- “Nerve tumour: fusiform swelling, target or fascicular sign
- “Muscle denervation: T2 high acutely, fatty replacement chronically
Clinical Warning
MRI neurography is a specialised technique. Know the normal nerve signal characteristics, signs of compression (enlargement, T2 hyperintensity), and the muscle denervation pattern that indicates nerve pathology. Carpal tunnel and cubital tunnel are common clinical applications.
Mnemonic
SCALE-TReading the Abnormal Nerve
S
S = Size (cross-sectional area enlargement)
C
C = Calibre change at an entrapment point
A
A = Architecture (loss of fascicular / honeycomb pattern)
L
L = Localisation to a known fibro-osseous tunnel
E
E = Enhancement (suggests tumour or active inflammation)
T
T = T2 signal hyperintensity (oedema) plus muscle denervation
Hook:Signal change alone is non-specific; combine size, calibre, architecture and muscle denervation before calling pathology.
Overview & Imaging Principles
What MR neurography adds
MR neurography is high-resolution, fat-suppressed MRI optimised to display peripheral nerves and detect pathology that electrodiagnostics localise but cannot anatomically map. It excels at deep or proximal nerves and the brachial and lumbosacral plexus, characterises mass lesions, and demonstrates secondary muscle denervation. It complements - rather than replaces - nerve conduction studies and high-resolution ultrasound.
Typical carpal/cubital tunnel
- Best initial test
- Clinical + nerve conduction studies
- Role of MR neurography
- Reserved for atypical or secondary causes
Superficial focal neuropathy
- Best initial test
- High-resolution ultrasound
- Role of MR neurography
- Problem-solving, deep extension
Plexus / proximal / deep nerve
- Best initial test
- MR neurography
- Role of MR neurography
- First-line imaging
Suspected nerve tumour
- Best initial test
- MRI with contrast
- Role of MR neurography
- Characterisation and surgical planning
Functional deficit, localising level
- Best initial test
- MRI muscle denervation map
- Role of MR neurography
- Identifies level and chronicity
Key Findings: Pathological Nerve Patterns
Nerve enlargement
- Description
- Increased cross-sectional area
- Significance
- Proximal to compression site
T2 hyperintensity
- Description
- Increased signal (brighter than normal)
- Significance
- Oedema, inflammation
Calibre change
- Description
- Abrupt narrowing at compression point
- Significance
- Indicates entrapment location
Notching
- Description
- Flattening at compression site
- Significance
- External compression
Loss of fascicular pattern
- Description
- Homogeneous signal
- Significance
- Fibrosis, chronic compression
Carpal Tunnel Syndrome
Median nerve cross-sectional area at the pisiform is the most discriminating measurement, but there is NO validated MRI cut-off: in the study cited on this page the mean was 18.8 mm² in CTS against 12.1 mm² in controls, so the familiar 10 mm² figure — which comes from ULTRASOUND — would call the average normal wrist abnormal. Compare with the other side and read area alongside the qualitative signs: palmar bowing of the flexor retinaculum, nerve flattening within the tunnel, T2 hyperintensity, proximal nerve enlargement, and sometimes thenar denervation.
Cubital Tunnel Syndrome
Ulnar nerve enlargement proximal to cubital tunnel (behind medial epicondyle). Nerve may sublux over epicondyle with flexion. T2 hyperintensity indicates inflammation. May see FCU or intrinsic muscle denervation.
Systematic Approach: Normal Nerve Appearance
T1 signal
- Appearance
- Intermediate (isointense to muscle)
- Significance
- Anatomic localisation
T2/PD signal
- Appearance
- Mildly hyperintense to muscle
- Significance
- Not as bright as fluid
Fascicular pattern
- Appearance
- Honeycomb appearance on axial
- Significance
- Intact nerve architecture
Size
- Appearance
- Consistent along course
- Significance
- Enlargement indicates pathology
Enhancement
- Appearance
- Minimal or none with Gd
- Significance
- Enhancement suggests pathology
Course
- Appearance
- Smooth, no deviation
- Significance
- Mass effect causes displacement
Fascicular Pattern
On high-resolution axial images, the fascicular pattern appears as multiple round/oval low signal dots (perineurium) surrounding intermediate signal fascicles. This 'honeycomb' pattern is lost in nerve injury, tumour, and inflammation.
Specific Nerve Imaging
Roots (C5-T1)
- Location
- Exit neural foramina
- Key Pathologies
- Avulsion (pseudomeningocele), stretch
Trunks
- Location
- Supraclavicular
- Key Pathologies
- Trauma, tumour, TOS
Divisions
- Location
- Behind clavicle
- Key Pathologies
- Less commonly seen
Cords
- Location
- Infraclavicular
- Key Pathologies
- Tumour, aneurysm compression
Branches
- Location
- Axilla, arm
- Key Pathologies
- Specific nerve injuries
Root Avulsion Signs
Traumatic root avulsion: pseudomeningocele (CSF collection at root level), absent nerve root on T2 myelography, denervation of paraspinal muscles (specific for preganglionic injury). MRI can assess root avulsion which has poor prognosis without grafting.
Grading Nerve Injury (Seddon / Sunderland) and What MRI Can Add
The summary flags that MRI shows continuity but cannot reliably separate axonotmesis from neurotmesis — a point worth grounding in the injury grading, because the grade drives prognosis and the wait-versus-operate decision. (The full classification and the biology of degeneration/regeneration are developed in the nerve-injury-regeneration topic; here the focus is the imaging correlate.)
Neurapraxia
- Sunderland
- Grade I
- Lesion
- Focal myelin block, axon intact
- Recovery
- Full, days to weeks
- MRI / imaging correlate
- Often normal or only mild focal T2 signal; nerve continuous; little or no muscle denervation
Axonotmesis
- Sunderland
- Grades II to IV
- Lesion
- Axon disrupted, connective sheaths progressively involved (II endoneurium intact to IV only epineurium intact)
- Recovery
- Variable - good in II, poor in IV; axon regrows about 1 mm/day
- MRI / imaging correlate
- Nerve continuous but enlarged and T2-hyperintense with muscle denervation; MRI cannot reliably separate II from III from IV
Neurotmesis
- Sunderland
- Grade V
- Lesion
- Complete transection, all layers
- Recovery
- None without surgery
- MRI / imaging correlate
- Discontinuity, gap, stump / terminal neuroma; muscle denervation
Why imaging cannot replace time and electrodiagnostics
MRI reliably shows nerve continuity, calibre, signal and secondary muscle denervation, but it cannot distinguish a recovering axonotmesis from a neurotmesis early because both show a swollen, T2-bright, denervating nerve. Sunderland grades II to IV (and the mixed grade VI of Mackinnon) all keep the nerve in continuity, so a neuroma-in-continuity on MRI does not by itself indicate surgery. Serial clinical examination, electrodiagnostic evidence of re-innervation, and the passage of time (no recovery by the expected re-innervation distance, advancing Tinel sign) remain essential; intra-operative nerve action potentials across a neuroma-in-continuity are the definitive test of whether to resect and graft.
Muscle Denervation
Acute
- Timeframe
- Less than 1 month
- T1 Signal
- Normal
- T2/STIR Signal
- High (oedema)
- Reversibility
- Fully reversible
Subacute
- Timeframe
- 1-6 months
- T1 Signal
- Normal to slightly high
- T2/STIR Signal
- High
- Reversibility
- Largely reversible
Chronic
- Timeframe
- Greater than 6 months
- T1 Signal
- High (fatty)
- T2/STIR Signal
- Variable
- Reversibility
- Irreversible fatty infiltration
Mnemonic
OEDEMA Then FATDenervation Pattern
O
O = Oedema-like signal (high T2) in acute phase
E
E = Early changes potentially reversible
F
F = Fatty replacement in chronic phase
A
A = Atrophy accompanies fatty change
T
T = Time determines reversibility
Hook:If you see muscle oedema in a specific nerve distribution, look for the nerve pathology upstream. Chronic denervation (fatty infiltration) indicates poor recovery potential.
Differential Diagnosis on Nerve MRI
Entrapment neuropathy
- Key MRI Pattern
- Proximal enlargement, T2 hyperintensity, calibre change at tunnel
- Discriminating Feature
- Abnormality localises to a known fibro-osseous tunnel
- Pitfall
- Mild T2 signal can be normal at tunnels (magic-angle, partial volume)
Traumatic neuropathy
- Key MRI Pattern
- Discontinuity, neuroma-in-continuity, stump neuroma
- Discriminating Feature
- History of trauma plus focal fusiform mass at injury level
- Pitfall
- Perineurial fibrosis may mimic enlargement
Schwannoma
- Key MRI Pattern
- Fusiform, eccentric to nerve, split-fat and target signs
- Discriminating Feature
- Separable from parent fascicles; nerve passes at edge
- Pitfall
- Cystic/ancient change can look aggressive
Neurofibroma
- Key MRI Pattern
- Fusiform, central within nerve, target sign
- Discriminating Feature
- Nerve enters and exits centrally; cannot be separated
- Pitfall
- Plexiform type in NF1 may harbour MPNST
MPNST
- Key MRI Pattern
- Over 5cm, ill-defined, heterogeneous, perilesional oedema, loss of split-fat
- Discriminating Feature
- Absent split-fat sign, low ADC on DWI, rapid growth, NF1
- Pitfall
- Target sign does NOT exclude malignancy
Intraneural ganglion
- Key MRI Pattern
- Cystic T2-bright tubular lesion tracking along nerve
- Discriminating Feature
- Connection to adjacent joint (e.g. superior tibiofibular)
- Pitfall
- Mistaken for solid tumour
Lipomatosis of nerve (fibrolipomatous hamartoma)
- Key MRI Pattern
- Coaxial-cable / spaghetti appearance, fat between thickened fascicles
- Discriminating Feature
- Macroscopic fat following fat signal on all sequences
- Pitfall
- May coexist with macrodactyly
Inflammatory / hereditary neuropathy
- Key MRI Pattern
- Diffuse symmetric multi-nerve thickening and T2 signal
- Discriminating Feature
- Bilateral, non-focal, no tunnel localisation
- Pitfall
- Misread as multifocal entrapment
Split-Fat Sign
A rim of fat partly surrounding a fusiform intraneural mass, reflecting the neurovascular bundle's fat being displaced rather than destroyed. Its presence favours a benign nerve sheath tumour; its absence is one of the most useful pointers to MPNST.
Magic-Angle Artifact: The Great Nerve-Signal Pitfall
The differential table and the controversies both warn that mild nerve T2 signal can be normal and cite the "magic-angle effect" — so it deserves to be explained, because it is the single commonest reason a normal nerve is over-called as neuropathic. (The general MRI physics is developed in the mri-imaging-principles topic; here is the nerve-specific pitfall.)
Highly ordered collagen — in nerve fascicles, tendons and ligaments — normally has a very short T2 because of dipolar interactions between water protons, so these structures are dark on most sequences. When the fibres lie at about 55 degrees to the main magnetic field (B0) those dipolar interactions are minimised, T2 lengthens, and the structure becomes artifactually bright on short-echo-time sequences (T1, proton density, PD fat-sat, short-TE gradient echo). A nerve curving around a tunnel or angling through the brachial plexus can pass through this orientation and light up — mimicking oedema.
Echo-time behaviour
- Magic-angle artifact
- Bright on short-TE sequences; fades as TE lengthens
- True neuropathy
- Bright on T2 / STIR and persists at long TE
Location
- Magic-angle artifact
- Only where the nerve lies near 55 degrees to B0 (curved/oblique segments)
- True neuropathy
- At the site of pathology (tunnel, mass, injury), any orientation
Nerve calibre
- Magic-angle artifact
- Normal
- True neuropathy
- Often enlarged with a calibre change
Fascicular architecture
- Magic-angle artifact
- Preserved
- True neuropathy
- Often lost
Muscle denervation
- Magic-angle artifact
- Absent
- True neuropathy
- May be present
How not to be fooled
If a nerve is bright only on short-echo-time sequences, only where it bends toward 55 degrees to the main field, and is of normal calibre with preserved fascicles and no muscle denervation, suspect magic angle rather than disease. Confirm by checking a long-echo-time sequence (artifact fades, true oedema persists) and the contralateral side, and never call neuropathy on signal alone - require corroborating calibre change, architecture loss or denervation.
Guidelines, Registries & Global Practice
AAOS (US, CTS CPG)
- Position on MRI / Neurography
- Routine imaging not required to diagnose idiopathic CTS; electrodiagnostics support diagnosis
- Practical Implication
- Reserve MRI for atypical or secondary causes
ACR Appropriateness Criteria (US)
- Position on MRI / Neurography
- MR neurography / high-resolution US usually appropriate for focal neuropathy and plexopathy
- Practical Implication
- Modality choice guided by site and local expertise
NICE / BOA (UK)
- Position on MRI / Neurography
- Clinical and electrodiagnostic diagnosis first; imaging for atypical features or masses
- Practical Implication
- MRI targeted, not screening
EFORT / European MSK radiology consensus
- Position on MRI / Neurography
- MR neurography and US complementary; 3T and dedicated coils recommended where available
- Practical Implication
- Protocol and expertise dependent
ESSR (European Soc MSK Radiology)
- Position on MRI / Neurography
- Endorses high-resolution US and MR neurography with standardised nerve protocols
- Practical Implication
- Drives reporting consistency
Global epidemiology context
Carpal tunnel syndrome is the commonest entrapment neuropathy worldwide (general-population prevalence roughly 3 to 6 percent of adults, higher in women and manual workers). Cubital tunnel is the second commonest. Most are diagnosed clinically and electrodiagnostically; imaging volumes vary enormously with access.
High- vs limited-resource practice
In well-resourced centres, 3T MR neurography with metal-artifact-reduction and 3D isotropic sequences allows multiplanar nerve and plexus assessment. Where MRI access or expertise is limited, high-resolution ultrasound is the pragmatic first-line nerve imaging test, and management often proceeds on clinical and electrodiagnostic grounds alone.
Registry / quality note
There is no implant registry analogue for nerve imaging. Quality is governed by acquisition standards (field strength, dedicated coils, fat suppression, in-plane resolution) and reporting expertise rather than device-survival data. Diagnostic performance figures in the literature should be read as expertise- and protocol-dependent.
Related pages: Nerve Root vs Peripheral Nerve for localising the lesion before requesting the scan, and Nerve Repairs for what the imaging findings commit you to. For the entrapments imaged most often see Carpal Tunnel, Cubital Tunnel Syndrome, Tarsal Tunnel Syndrome, Meralgia Paraesthetica and Morton's Neuroma, where MR neurography and high-resolution ultrasound are complementary. For the masses, Schwannoma, Neurofibroma and Malignant Peripheral Nerve Sheath Tumour — remembering that no imaging sign reliably separates benign from malignant. Plexus anatomy is in Brachial Plexus Anatomy and Lumbosacral Plexus Anatomy, with Adult Brachial Plexus Injury for the traumatic case.
Controversies & Areas of Uncertainty
MRI vs ultrasound vs nerve conduction studies
There is no consensus first-line imaging test for entrapment. High-resolution ultrasound is cheaper, dynamic and operator-available; MR neurography offers deep/proximal coverage (plexus, sciatic) and muscle denervation mapping. Reported diagnostic accuracy largely reflects local expertise and scanner availability rather than an inherent superiority of one modality. Electrodiagnostics remain the functional reference standard; imaging is complementary, not a replacement.
Is routine MRI needed in carpal tunnel syndrome?
Most guidelines reserve MRI for atypical presentations, suspected secondary/structural cause (mass, anomalous muscle, persistent median artery, bifid nerve), or revision surgery. Cross-sectional-area thresholds vary by level measured and overlap with controls, so MRI is not recommended to confirm routine idiopathic CTS.
Quantitative neurography (DTI, T2 mapping)
Diffusion tensor imaging-derived fractional anisotropy and apparent diffusion coefficient, plus quantitative T2 mapping, are promising biomarkers of nerve injury and recovery but remain largely research tools. Lack of standardised acquisition, normative values and reproducibility across vendors limits routine clinical use.
Significance of nerve T2 hyperintensity
Mild fascicular T2 hyperintensity occurs in asymptomatic volunteers and from magic-angle effect, so signal change alone is non-specific. Findings should be interpreted alongside calibre change, fascicular architecture, muscle denervation and the clinical/electrodiagnostic picture.
Clinical Imaging Technique
Field strength
- Recommendation
- 3T preferred over 1.5T
- Rationale
- Higher SNR for small structures
Coil
- Recommendation
- Dedicated surface coil
- Rationale
- Maximises signal-to-noise
Slice thickness
- Recommendation
- Less than 3mm
- Rationale
- Resolves fascicular detail
In-plane resolution
- Recommendation
- Less than 1mm
- Rationale
- Visualises nerve architecture
Sequences
- Recommendation
- PD fat-sat, T2 fat-sat, T1
- Rationale
- Nerve-fluid contrast
Planes
- Recommendation
- Axial perpendicular to nerve + along nerve
- Rationale
- Cross-section and longitudinal
Nerve Visualisation Principles
Nerves are best visualised on fat-suppressed T2 or PD sequences where they appear slightly hyperintense against suppressed fat. The fascicular pattern (honeycomb appearance) should be visible on high-resolution imaging. Nerve is intermediate on T1, allowing anatomic localisation.
Clinical Decision Scenarios
Practise clinical reasoning and management decisions out loud
Viva scenarioStandard
Clinical prompt
“A 45-year-old presents with hand numbness and thenar weakness. Nerve conduction studies confirm carpal tunnel syndrome. The hand surgeon requests MRI.”
Viva scenarioChallenging
Clinical prompt
“A patient presents after a motorbike accident with a flail arm. Clinical examination suggests brachial plexus injury. What would you expect on MRI?”
Viva scenarioChallenging
Clinical prompt
“A 30-year-old presents with a slowly enlarging painless mass in the forearm. Ultrasound shows a fusiform mass along the course of a nerve.”
Exam day cheat sheet
MRI Neurography Quick Reference
Normal Nerve Appearance
- T1: Intermediate (like muscle)
- T2: Mildly hyperintense to muscle
- Fascicular pattern (honeycomb)
- Size consistent along course
Compression Signs
- Nerve enlargement proximal to compression
- T2 hyperintensity (oedema)
- Calibre change at compression point
- Loss of fascicular pattern
Carpal Tunnel Criteria
- Median nerve enlarged at pisiform (18.8 vs 12.1 mm²); no validated MRI cut-off
- Palmar bowing of retinaculum
- Nerve flattening within tunnel
- Thenar denervation (late)
Denervation Pattern
- Acute: High T2, normal T1 (oedema)
- Chronic: High T1 (fatty infiltration)
- Distribution follows nerve supply
- Identifies level of injury
Evidence
Evidence
Median nerve cross-sectional area at the pisiform is enlarged in carpal tunnel syndrome
Park, Won, Oh, Kim, Hwang, Yoo • Asian J Surg (2020)
Key Findings:
- Retrospective comparison of 164 wrists: 67 with clinically and electrodiagnostically confirmed CTS versus 97 controls.
- Mean median nerve cross-sectional area at the pisiform level was 18.8 mm² in CTS versus 12.1 mm² in controls (p less than 0.05).
- Cross-sectional area at the hook of hamate did not differ significantly between groups.
Clinical implication: The pisiform level is the most discriminating site for measuring median nerve cross-sectional area on MRI in suspected CTS.
Limitation: A RETROSPECTIVE STUDY OF WORKERS THAT REPORTS GROUP MEANS AND NO DIAGNOSTIC THRESHOLD - and its control mean is the number that matters most for this page: 12.1 mm² at the pisiform in people WITHOUT carpal tunnel syndrome. Any cut-off set below that will call the average normal wrist abnormal, so the ultrasound-derived figure of 10 mm² must not be carried across to MRI, on which cross-sectional areas measure larger. The paper derives no cut-off, reports no sensitivity or specificity and gives no measure of spread around either mean, so the separation between 18.8 and 12.1 mm² cannot be converted into a test. Cases were defined by examination plus nerve conduction with anyone negative on either test excluded, which sharpens the contrast between groups and further inflates apparent separation. Single centre, one occupational population, 127 of 164 female.
Evidence
3T MRI nerve cross-sectional area correlates with electrodiagnostic severity in CTS
Ikeda, Okada, Toyama, Uemura, Takamatsu, Nakamura • Orthopedics (2017)
Key Findings:
- 70 wrists of 35 patients with unilateral idiopathic CTS imaged at 3T with nerve conduction studies.
- Median nerve cross-sectional area was greatest at the scaphoid body level and positively correlated with distal motor latency.
- Cross-sectional area at the distal radioulnar joint and hamate hook did not correlate with severity.
Clinical implication: Level of measurement matters; 3T MRI nerve area can reflect electrophysiological severity but thresholds are level-dependent.
Limitation: THIRTY-FIVE PATIENTS, AND THE CORRELATION RUNS IN OPPOSITE DIRECTIONS AT DIFFERENT LEVELS - at the scaphoid body the affected nerve was larger and area correlated POSITIVELY with distal motor latency, but at the scaphoid tubercle the affected nerve was SMALLER and the correlation was NEGATIVE, while at the distal radioulnar joint and hamate hook there was no difference and no correlation at all. So 'cross-sectional area reflects severity' holds at one of four levels and reverses at another; the tubercle finding reflects the nerve being flattened rather than swollen at the point of maximal compression. Unilateral disease was used so the unaffected hand serves as control, which is elegant but assumes the other wrist is truly normal. No diagnostic threshold is derived and none should be inferred.