A complete brachial plexus MRI protocol: coil setup, STIR and T2 sequencing, contrast dosing, artifact control, and key pathologies to recognize.

Brachial Plexus MRI Protocol: 10 Essential Steps for Diagnostic-Quality Nerve Imaging

At a glance

Protocol summary

Core sequences

• Large FOV coronal STIR (screening, both plexuses)
• Small FOV axial T2 fat-suppressed (fascicular detail)
• Coronal 3D T1 SPACE (isotropic anatomical roadmap)

Contrast protocol

• 10–15 mL (0.1 mmol/kg) gadolinium-based agent
• Flow rate 2.0 mL/s
• Chaser: 100 mL saline at 2.0 mL/s

Artefact reduction

• Primary artefact: breathing / pulsation ghosting
• Respiratory gating or triggering
• Superior–inferior saturation bands

Key pitfalls

• Ghosting mimicking nerve root asymmetry
• Magic angle hyperintensity misread as neuritis
• Vessel–nerve confusion on axial T2

Introduction: why brachial plexus MRI is technically demanding

A well-executed brachial plexus MRI protocol is one of the most technically unforgiving studies in the neuromuscular imaging repertoire, and mastering it separates a department that produces confidently reportable studies from one that generates endless repeat requests. The target anatomy is small, obliquely oriented, mobile with respiration, and immediately adjacent to the subclavian vessels, lung apex, and cervical fat planes, all of which generate competing signal that can obscure the very structures the referrer needs assessed. Unlike a routine cervical spine study, this examination must simultaneously resolve millimetre-scale nerve fascicles and cover a large field of view stretching from the neural foramina to the axilla, a combination that punishes any shortcut in coil selection, patient positioning, or sequence timing.

Referrals for a brachial plexus MRI protocol typically arrive from three very different clinical doors, and each imposes its own demands on the technologist and reporting radiologist. Trauma teams request the study to assess a newborn with obstetric palsy or an adult motorcyclist for root avulsion following a traction injury. Oncology teams request it to stage a Pancoast tumor, evaluate perineural spread of a known malignancy, or distinguish post-radiation fibrosis from tumor recurrence in a previously irradiated field. General neurology, physiatry, and orthopaedic teams request it to investigate an ache, weakness, or numbness that has defied electrophysiological testing, often after months of diagnostic uncertainty. Each of these three referral pathways demands a slightly different balance of coverage, resolution, and contrast use, which is precisely why a single rigid protocol is rarely appropriate without informed operator judgement at the console.

Historically, plexus imaging was regarded as a niche, low-volume study performed only at specialist referral centres, largely because the anatomy exceeded the resolving power of early scanner hardware and coil technology. The arrival of dedicated neurovascular surface coils, three-dimensional isotropic sequences such as SPACE, and refined fat-suppression techniques has transformed this into a study that any well-equipped department can perform reliably, provided the protocol is followed with discipline.^[2] This article works through the complete brachial plexus MRI protocol used across the SATMED Health 30-day protocol mastery series, covering anatomy, sequence selection, contrast dosing, safety, pathology recognition, and the distinct pitfall patterns seen at the scanner, at the workstation, and in the referring clinic.

From a departmental training perspective, plexus imaging rewards deliberate, structured onboarding far more than most other MRI protocols, precisely because the failure modes are so predictable and so consistently repeated across different technologists and different institutions. A new technologist who understands why respiratory gating matters, and a new resident who understands why the magic angle effect exists, will each avoid the majority of the errors documented in this article before ever encountering them on a real patient. The remainder of this article is structured to support exactly that kind of structured onboarding, moving methodically from anatomy through technique, safety, pathology, and the three-tier pitfall framework that closes the loop between scanner, workstation, and clinic.

Beyond the individual patient encounter, the cumulative healthcare burden of plexus pathology is not trivial: delayed diagnosis of a traumatic root avulsion narrows the window for viable nerve grafting or transfer surgery, delayed recognition of Pancoast invasion delays oncologic staging and treatment planning, and repeated non-diagnostic studies caused by avoidable technical artefact consume scanner time that could otherwise serve other patients on an already stretched imaging waiting list. Getting this protocol right on the first attempt therefore carries value well beyond the individual report.

Anatomy of the brachial plexus

The brachial plexus is formed from the ventral rami of C5 through T1, occasionally with a contribution from C4 (prefixed plexus) or T2 (postfixed plexus), and this variability is itself a source of diagnostic confusion if the reporting radiologist is unfamiliar with normal variants. The plexus is conventionally described using the mnemonic Roots, Trunks, Divisions, Cords, Branches, and this sequence maps directly onto the anatomical compartments imaged during the study, giving the technologist and radiologist a shared vocabulary for describing the level of any abnormality.

Roots and the interscalene triangle

The five roots emerge from the neural foramina and travel through the interscalene triangle, bounded anteriorly by the anterior scalene muscle and posteriorly by the middle scalene muscle, with the first rib forming the floor of the space. This is the site most vulnerable to traction injury, and nerve root avulsion at this level produces a traumatic pseudomeningocele — a key imaging target on coronal and axial sequences angled through the foramina. Because the roots are intimately related to the exiting spinal nerve and dorsal root ganglion, careful correlation with routine cervical spine sequences is often required when a combined radiculopathy and plexopathy is suspected.

Trunks and the supraclavicular segment

C5 and C6 unite to form the upper trunk, C7 continues alone as the middle trunk, and C8 with T1 unite to form the lower trunk. The trunks lie within the posterior triangle of the neck, superficial and posterior to the subclavian artery, coursing between the scalene muscles toward the first rib before passing over it into the axilla. The upper trunk is the most commonly injured segment in traction and obstetric injury, producing the classic Erb-Duchenne palsy pattern with a waiter’s-tip posture, while the lower trunk is preferentially affected in true neurogenic thoracic outlet syndrome and in Pancoast tumor invasion, producing a Klumpke-type pattern with intrinsic hand weakness.

Divisions and the retroclavicular segment

Each trunk splits into an anterior and a posterior division as it passes behind the clavicle, a transition that reorganises the fibres destined for the flexor and extensor compartments of the upper limb. This short retroclavicular segment is anatomically compressed between the clavicle and the first rib, and it is frequently the most technically difficult portion of the plexus to resolve without dedicated thin-slice coverage, since susceptibility artefact from the adjacent bone can degrade signal precisely where clinicians most need clarity, particularly in suspected costoclavicular thoracic outlet syndrome.

Cords and the infraclavicular segment

The six divisions reorganise into three cords — lateral, posterior, and medial — named according to their relationship to the second part of the axillary artery rather than to any bony landmark. This infraclavicular segment sits within the axilla, surrounded by the pectoral muscles anteriorly and the subscapularis and latissimus dorsi posteriorly, and is best resolved on coronal three-dimensional isotropic acquisitions that can subsequently be reformatted along the long axis of the plexus to follow its oblique course from neck to axilla.

Terminal branches and extended coverage

The cords give rise to the five major terminal branches: the musculocutaneous, axillary, radial, median, and ulnar nerves, each carrying a predictable combination of fibres from specific root levels that allows localisation of injury from the clinical examination alone in experienced hands. Imaging of terminal branch pathology, for example isolated axillary neuropathy following shoulder dislocation, typically requires an extended field of view into the proximal arm beyond the standard plexus coverage, and this should be explicitly requested by the referring clinician or anticipated by the reporting radiologist reviewing the clinical history before the study begins. A useful practical rule is that any referral mentioning a specific terminal branch by name, rather than the plexus generally, should prompt a check of whether standard coverage will actually reach the segment in question, since the default field of view for a screening plexus study is deliberately centred more proximally and may not extend far enough distally to capture an isolated terminal branch lesion.

Embryological development

The ventral rami that form the brachial plexus develop from the cervical and upper thoracic somites during the fifth and sixth weeks of gestation, migrating alongside the developing upper limb bud as it extends laterally from the body wall. This shared developmental origin explains why congenital plexus anomalies, including rare duplications or anomalous root fusions, are occasionally encountered alongside other upper limb malformations, and why the segmental root pattern established in early embryogenesis persists as the anatomical basis for the dermatomal and myotomal localisation still used in clinical examination today.

Vascular relations and fascial anatomy

The subclavian artery separates the roots and trunks of the plexus from the subclavian vein, running immediately anterior and inferior to the trunks as they cross the first rib, and this close vascular relationship is exactly why pulsation ghosting so reliably contaminates plexus imaging unless it is deliberately suppressed at the console. The axillary artery continues this relationship distally, running centrally between the three cords that are themselves named for their position relative to it. The prevertebral fascia envelops the roots and trunks as they exit the interscalene triangle, and thickening or asymmetric enhancement of this fascial envelope can itself be an early sign of inflammatory plexitis or perineural tumor spread, a finding easily overlooked without dedicated fat-saturated sequences carried through the full length of the plexus.

Anatomical variants

Beyond the prefixed and postfixed variants already described, radiographers and radiologists should remain alert to variable trunk and cord configurations, an occasional accessory phrenic nerve contribution from the C5 root, and physiological asymmetry in cord length between the two sides. Awareness of these variants prevents both under-recognition of true pathology and, just as importantly, over-calling a normal variant as an abnormal finding, a distinction that matters considerably when a surgical colleague is planning an operative approach based directly on the anatomy described in the report.

MR tissue relaxation values

Understanding the relaxation behaviour of the tissues surrounding the plexus is what allows a radiographer to predict, and troubleshoot, image contrast before a single sequence is run. Peripheral nerve has intermediate T1 and mildly prolonged T2 relative to skeletal muscle, and this modest difference is the physical basis for the nerve-to-fat contrast that every fat-suppressed sequence in this protocol is built to exploit.

Because nerve T2 is only modestly longer than muscle T2, fat suppression is essential to make this narrow contrast window visible: without it, the small T2 difference between nerve and adjacent fat is invisible against the far brighter background fat signal that dominates the image. This is the physical rationale behind both the STIR sequence, which nulls fat by inversion irrespective of field strength, and fat-saturated fast spin-echo T2, which suppresses fat by chemical-shift-selective pulses and is more sensitive to field inhomogeneity.^[3] At 3T, the increased chemical shift between fat and water protons widens the frequency separation exploited by spectral fat saturation, which can improve suppression uniformity in a well-shimmed field but also increases vulnerability to fat-saturation failure at the periphery of the coil, where field homogeneity is poorest, precisely the region occupied by the axillary and infraclavicular segments of the plexus.

Diffusion tensor imaging of the plexus, while not part of the routine ten-step protocol described here, exploits a further relaxation-adjacent property of nerve tissue: the fractional anisotropy of water diffusion along intact, organised fascicles. Where fascicular architecture is disrupted by tumor infiltration or severe traction injury, fractional anisotropy falls and mean diffusivity rises, providing a quantitative adjunct to the qualitative T2 signal changes described throughout this article. Departments building an advanced plexus service may wish to add this sequence for oncologic staging or complex traction injury cases, though it adds meaningfully to total scan time and post-processing burden.^[3]

The relaxation values summarised above also explain why a technologist reviewing images at the console should never judge fat suppression quality from a single slice in isolation. Field inhomogeneity across the wide coverage this protocol demands means that fat suppression can be excellent centrally, over the trunks and divisions, while failing at the periphery of the coil, over the more peripheral cords and terminal branches, purely as a consequence of the physics summarised in this table rather than any error in sequence prescription. Reviewing the full slice stack, rather than a single representative image, is therefore an essential quality step before the patient leaves the scanner.

Scanning technique: 10-step protocol

The ten steps below are presented in the order they should typically be executed at the console, moving from broad screening coverage toward progressively higher-resolution, more focused acquisitions. This sequencing is deliberate: it allows the technologist to identify gross asymmetry or discontinuity early, on the wide-coverage STIR sequence, before committing scan time to the smaller field-of-view acquisitions that will ultimately answer the specific clinical question posed by the referral.

1. Patient setup: position the patient supine, arms adducted symmetrically at the sides with the shoulders relaxed and depressed to a matched degree bilaterally; asymmetrical arm position is a common and entirely avoidable source of side-to-side comparison error that can mimic true pathology on the contralateral side.
2. Coil selection: use a dedicated neurovascular or flexible phased-array surface coil combination spanning from the skull base to the mid-humerus, rather than relying on the body coil alone, since local coil signal is what ultimately permits fascicular-level resolution.
3. Localiser and centring: centre the field of view over the C6 vertebral body to ensure symmetric bilateral coverage of both plexuses whenever a comparative assessment is clinically required, as it is in the majority of non-oncologic referrals.
4. Large FOV coronal STIR: acquire this first as the screening sequence, covering root, trunk, and cord segments bilaterally to detect asymmetry, oedema, or frank discontinuity before committing to smaller, more focused acquisitions.
5. Small FOV axial T2 fat-suppressed: centre this sequence on the abnormality identified on the screening STIR, or in acute trauma on the interscalene and supraclavicular segments by default, to resolve individual fascicles at high in-plane resolution.
6. Coronal 3D T1 SPACE (isotropic): acquire for a high-resolution anatomical roadmap and subsequent multiplanar reformation along the oblique course of the plexus, providing the surgical and oncologic teams with reformats they can interrogate independently.
7. Phase-encoding direction optimisation: set the phase-encoding direction superior–inferior or anterior–posterior as required by the dominant artefact source, to displace respiratory and vascular ghosting away from the plexus itself rather than superimposing it directly.
8. Respiratory gating or triggering: apply a respiratory bellows trigger or navigator echo, particularly when imaging the lower trunk and infraclavicular segment adjacent to the moving lung apex and diaphragm.
9. Superior–inferior saturation bands: place spatial saturation bands to suppress inflowing vascular signal and pulsatile ghosting arising from the subclavian and axillary vessels running immediately alongside the plexus.
10. Post-contrast sequences and reformatting: where clinically indicated, acquire a post-contrast fat-saturated 3D T1 sequence matched in geometry to the pre-contrast acquisition, then generate maximum intensity projection and curved planar reformats for the referring surgical or oncology team.

Common technical troubleshooting during acquisition

Even with careful upfront planning, individual patients occasionally present technical challenges that require real-time adjustment rather than a fixed protocol applied unchanged. A patient unable to lie fully supine due to shoulder pain may need the contralateral arm elevated slightly to maintain symmetry, with this deviation explicitly documented for the reporting radiologist. A patient with an irregular respiratory pattern, common in acute pain states, may render bellows-based triggering unreliable, in which case switching to a navigator-based technique or accepting a modest increase in averages to average out the resulting ghosting can preserve diagnostic quality without abandoning the study.

Metal artefact from a prior surgical clip or an implanted device near the shoulder is another frequent real-time challenge; where feasible, increasing receiver bandwidth and adjusting the frequency-encoding direction to run parallel to any linear metallic structure will reduce the resulting distortion considerably more effectively than attempting to compensate after the fact during image review.

Protocol customisation by referral pathway

Although the ten steps above form the backbone of every brachial plexus study, experienced departments adjust emphasis according to the referral pathway rather than applying an identical protocol to every patient regardless of clinical context. In the trauma pathway, priority is given to thin-slice axial coverage through the interscalene triangle and neural foramina, since the clinical question centres on root continuity and pseudomeningocele formation, and contrast is frequently omitted entirely.

In the oncology pathway, coverage is extended inferiorly to include the lung apex and chest wall when a Pancoast tumor or nodal disease is suspected, and post-contrast fat-saturated 3D T1 imaging becomes a routine rather than optional component of the study. In the general neurology and physiatry pathway, where inflammatory or hereditary plexopathy is the leading differential, symmetric bilateral coverage takes priority over unilateral high resolution, since the diagnostic question frequently hinges on demonstrating asymmetry or multifocality between the two sides rather than resolving fine detail on one side alone.

A further variation applies to the paediatric obstetric pathway, where the referral typically concerns a neonate or infant with suspected obstetric brachial plexus palsy. Smaller anatomy demands a proportionally smaller field of view and thinner slices to maintain adequate spatial resolution across correspondingly smaller nerve roots, while sedation or feed-and-wrap technique, rather than adult-style verbal coaching, becomes the primary strategy for motion control. Departments imaging this population regularly should maintain a dedicated paediatric plexus protocol card distinct from the adult trauma protocol described in this article.

In practice, most departments will settle on 3T as the field strength of choice for brachial plexus imaging whenever it is available, given the substantial SNR advantage that can be converted directly into either finer spatial resolution or shorter, more motion-robust acquisitions. However, 1.5T remains entirely acceptable, and in patients with implants or metalwork near the shoulder, it may actually be preferable owing to reduced susceptibility artefact and greater SAR headroom for the long echo trains this protocol depends upon.^[4]

From a workflow perspective, the full ten-step protocol described above typically occupies a single appointment slot of 30 to 40 minutes once patient positioning, coil setup, and any required contrast administration are included. Departments running a high volume of plexus referrals often find it worthwhile to build a standing, vendor-saved protocol card that pre-loads coil selection, field of view, saturation band placement, and respiratory triggering settings, reducing variability between technologists and shortening room turnover time without compromising diagnostic quality.

Contrast media protocol

Contrast is not required for every brachial plexus study; pure post-traumatic traction injury assessment can often be performed entirely non-contrast, since the diagnostic question in that setting is one of anatomical discontinuity and fluid signal rather than tissue enhancement, and this distinction matters for both patient throughput and resource use across a busy imaging department. Contrast becomes essential, however, whenever tumor, inflammatory plexitis, or infection sits on the differential diagnosis, because the pattern and intensity of enhancement is frequently what distinguishes reactive oedema from true neoplastic infiltration or an abscess collection requiring urgent drainage.

In the pure post-traumatic setting, where the clinical question is whether a nerve root has been avulsed and whether a pseudomeningocele has formed, non-contrast STIR and T2 fat-saturated sequences alone are generally sufficient, because the diagnostic signal of interest is CSF-like fluid signal and structural discontinuity rather than tissue enhancement. Omitting contrast in this scenario spares the patient unnecessary gadolinium exposure, shortens the total appointment time, and avoids the additional venous access step entirely, all without compromising diagnostic yield for the traumatic indication being investigated.

Many departments standardise the mechanical delivery of this protocol using an integrated power-injector extension line system such as SATLine, paired with pre-filled contrast delivery devices such as SATSyringe, to ensure the injection rate and chaser volume specified above are reproduced consistently across every technologist and every shift, reducing the variability that otherwise creeps into manually prepared injections. Consistent venous access technique, including confirmation of a secure cannula and a test flush before the main bolus, remains the single most effective way of preventing extravasation in a region where the injection site is often deliberately distant from the plexus itself, in the contralateral or ipsilateral antecubital fossa.

Current European and North American contrast society guidance continues to recommend using the lowest gadolinium dose consistent with diagnostic quality, reflecting ongoing scrutiny of gadolinium retention in body tissues following repeated administrations over a patient’s lifetime. For the brachial plexus protocol, the 10 to 15 mL dose specified here already represents a standard, weight-based single dose rather than a double or triple dose regimen, and departments should resist requests to increase volume beyond this in the absence of a specific clinical justification such as suspected leptomeningeal or perineural spread requiring maximal contrast-to-noise ratio.^[28]

In the paediatric population, the same 0.1 mmol/kg weight-based dosing principle applies, scaled down proportionally for smaller body weight, with injection rate adjusted accordingly to maintain an equivalent bolus profile; a fixed adult-volume dose should never be administered to a child regardless of how the request form is phrased, and departments imaging children regularly should maintain a clearly displayed paediatric dosing chart at the injector station.

Specific absorption rate (SAR)

The brachial plexus protocol relies heavily on long fast spin-echo echo trains for both the T2 fat-saturated and STIR sequences, and both of these are RF-intensive by design.

SAR management is therefore a genuine operational constraint on this protocol, particularly at 3T and in patients carrying implants near the shoulder or chest wall where local tissue heating is a legitimate safety concern rather than a theoretical one. Because the protocol requires bilateral, large field-of-view coverage for the screening STIR sequence followed by a further high-resolution axial acquisition and a lengthy 3D SPACE sequence, the cumulative RF burden across the full examination is considerably higher than for many other regional MRI protocols of comparable duration, and this should be factored into scanner-side monitoring throughout the appointment rather than checked only once at the end.

The limits summarised in the table below draw directly on three internationally recognised frameworks: the International Electrotechnical Commission’s device-level safety standard, the American Association of Physicists in Medicine’s quality assurance guidance, and the European Commission’s Radiation Protection 185 guidelines on operator and patient RF exposure thresholds. Aligning departmental SAR monitoring practice with all three, rather than a single national standard, ensures the protocol remains compliant regardless of the regulatory jurisdiction in which a given department operates.

Five SAR-reduction strategies

• Reduce echo train length in the fat-saturated T2 turbo spin-echo sequences to lower cumulative RF deposition per repetition, accepting a modest increase in overall scan time as the trade-off; this is often the first and simplest adjustment made when the console flags an approaching SAR limit mid-examination.
• Lower the refocusing flip angle using variable flip angle, hyperecho-type turbo spin-echo schemes, which can cut SAR by 30–50 percent with minimal penalty to image contrast, and are supported on most modern platforms as a vendor-provided option that can often be enabled without a full protocol rebuild.
• Increase repetition time modestly where the scan-time budget allows, spreading the same total RF energy over a longer interval and reducing the time-averaged SAR without materially changing image contrast, though this should be weighed against the total appointment slot length available in a busy department.
• Favour STIR over spectral fat-saturation in patients approaching SAR limits, since STIR is generally less RF-demanding per unit of fat suppression achieved in inhomogeneous regions such as the neck and shoulder, where spectral saturation often needs repeated or higher-power pulses to compensate for poor shimming across the wide coverage this protocol requires.
• Use parallel imaging acceleration to reduce the number of refocusing pulses required per unit acquisition time, directly lowering time-averaged SAR while also shortening total scan time and reducing motion vulnerability, though as discussed later in this article, acceleration must be increased incrementally and validated locally rather than pushed to the maximum available factor by default.^[9]

In practice, most departments combine two or three of these strategies rather than relying on any single adjustment, since the SAR budget for the full ten-step protocol accumulates across sequences rather than resetting between them. A sensible default is to pair a variable flip angle turbo spin-echo scheme with a modest parallel imaging factor, reserving further TR extension or echo train reduction for individual patients who approach the limit despite these baseline settings.

It is worth noting that SAR headroom is not static across an examination: as body temperature rises marginally over a long appointment, or as the scanner’s internal SAR model accumulates energy across successive sequences, a patient who comfortably tolerated the screening STIR sequence may approach the limit by the time the 3D SPACE acquisition begins. Monitoring the console’s cumulative SAR indicator throughout the appointment, rather than relying solely on the pre-scan estimate, allows the technologist to intervene proactively before a sequence is aborted mid-acquisition.

Top 10 pathologies

The ten conditions below span the trauma, oncology, and general neurology referral pathways described in the introduction, and together they account for the great majority of clinically actionable findings encountered in routine brachial plexus MRI practice. Each entry lists the characteristic T1 and T2 signal behaviour, a brief note on epidemiology or differential diagnosis, and the specific way the finding shapes protocol choice and reporting emphasis.

Taken together, these ten diagnoses illustrate why a single fixed protocol cannot serve every brachial plexus referral equally well: trauma-pattern injuries reward thin-slice coverage through the foramina, oncologic referrals reward contrast administration and extended inferior coverage to the lung apex, and inflammatory or hereditary conditions reward symmetric, bilateral high-resolution coverage capable of demonstrating subtle, patchy, or multifocal signal change that a unilateral, narrowly targeted study would simply never show.

When two or more of these ten diagnoses remain genuinely plausible after the standard protocol has been completed, additional targeted sequences rather than a repeat of the entire examination are usually the most efficient next step. Adding diffusion-weighted imaging to clarify a solid mass, extending post-contrast coverage inferiorly when nodal disease is suspected, or repeating STIR imaging at a later date to assess interval change in an inflammatory plexopathy each answer a specific residual question without subjecting the patient to a full repeat appointment.

It is also worth emphasising that these ten conditions are not mutually exclusive within a single patient. A patient with neurofibromatosis type 1 and a known plexiform neurofibroma may later develop a malignant peripheral nerve sheath tumor arising within that same lesion, and a patient previously treated with radiotherapy for breast cancer may develop both radiation fibrosis and, years later, a genuinely recurrent metastatic deposit within the same irradiated field. Reporting radiologists should therefore resist the temptation to settle on the first plausible diagnosis from this list and instead actively consider whether the imaging findings are fully explained by a single process or whether two coexisting processes better account for the full pattern of signal change observed.

Pitfalls — radiographers

The primary scanning pitfall documented for this protocol is breathing and pulsation ghosting, arising from the immediate proximity of the lower trunk and cords to the moving lung apex and the pulsatile subclavian vessels. This single artefact is responsible for the majority of avoidable repeat scans in brachial plexus imaging, and it deserves deliberate, protocol-level mitigation rather than an ad hoc fix applied only after a study has already failed quality review. Because the lower trunk, divisions, and cords sit directly against the moving chest wall and the pulsatile subclavian vessels, ghosting from these two independent motion sources can overlap and compound one another, producing an artefact pattern that is considerably harder to resolve retrospectively than either source alone.

A useful departmental habit is to review the screening coronal STIR at the console before proceeding to the small FOV axial acquisition, specifically checking for residual ghosting across the lower trunk. If ghosting is visible at this early stage, it is far more time-efficient to adjust phase-encoding direction or add a saturation band immediately than to discover the same artefact degrading the definitive axial sequence twenty minutes later.

Pitfalls — radiologists

The primary interpretation pitfall for this protocol is the magic angle effect: nerve fascicles oriented at approximately 55 degrees to the main magnetic field can display artefactual T1 and short-echo-time hyperintensity that closely mimics true neuritis or inflammatory oedema, and this effect is particularly troublesome in the brachial plexus given its inherently curved, obliquely running course, which guarantees that some segment of every study will pass through the critical angle range regardless of patient positioning. Unlike a straight peripheral nerve such as the median or ulnar nerve in the forearm, the plexus cannot be reliably repositioned out of the magic angle range, making familiarity with this artefact, rather than its avoidance, the more realistic goal for the reporting radiologist.

None of these four interpretation pitfalls is exotic or rare; each is encountered routinely in general radiology practice, and each is fully preventable provided the reporting radiologist maintains a low threshold for reviewing the long-TE T2 series, the isotropic 3D dataset, and the contralateral side before finalising an impression that describes a focal abnormality.

Pitfalls — non-radiology physicians

Each of these clinical pitfalls shares a common thread: they arise not from any deficiency in the imaging itself, but from an incomplete feedback loop between the referring team and the radiology department. A brief, specific clinical question on the request form, combined with a willingness to repeat imaging when the clinical picture and an initial normal study genuinely disagree, resolves the great majority of these pitfalls without requiring any change to the scanning or interpretation protocol described elsewhere in this article.

Pitfall comparison summary

Departments seeking to reduce the incidence of these three pitfall categories over time benefit from a structured, periodic audit process: reviewing a sample of recent plexus studies specifically for evidence of respiratory or vascular ghosting, cross-checking a sample of reports against subsequent clinical or surgical follow-up where available, and holding a brief joint case-review session with referring clinical teams at intervals throughout the year. This kind of structured feedback loop, rather than any single technical fix, is what sustains protocol quality over time as staff rotate through the department and as scanner software and coil hardware are periodically upgraded. A department that treats this article as a one-time training exercise rather than a living reference will inevitably see pitfall rates drift back upward as institutional memory fades; revisiting the anatomy, technique, and pitfall sections periodically as part of routine continuing education helps preserve the gains made from initial implementation.

AI and automation

Automated nerve-segmentation and lesion-detection tools are emerging for peripheral nerve MRI, though the evidence base for artificial intelligence applied specifically to the brachial plexus remains earlier-stage than for more established applications such as brain or cardiac imaging. Published work demonstrates deep-learning segmentation of peripheral nerves from MR neurography datasets with encouraging concordance to expert manual tracing, supporting a plausible future role in quantitative, fascicle-level assessment of nerve calibre and signal change over serial follow-up studies.^[10]

More recent work has specifically explored deep-learning image reconstruction applied to three-dimensional brachial plexus neurography, demonstrating that accelerated acquisitions reconstructed with deep-learning methods can preserve diagnostic image quality while meaningfully shortening scan time, a genuinely practical benefit given how motion- and time-sensitive this examination is.^[11] Several broader, FDA-cleared and CE-marked platforms now offer general musculoskeletal and neuro lesion-detection assistance that can be applied to plexus studies as part of a wider worklist triage strategy, though radiologist confirmation of every flagged finding remains mandatory and non-negotiable.^[12] Departments adopting these tools should validate performance locally on their own case mix before relying on automated flagging for plexus pathology specifically, since published performance figures rarely generalise perfectly across scanner vendors, coil configurations, and patient populations.

A further active research direction combines deep-learning reconstruction with diffusion tensor tractography of the plexus, aiming to produce automated, quantitative fascicle-level maps that could eventually support objective, serial monitoring of nerve recovery after injury or surgical repair, rather than relying solely on qualitative signal assessment by an individual reader. This work remains largely confined to academic centres at present, and routine clinical deployment across general radiology practice is not yet established; departments should treat it as a promising direction to watch rather than a tool ready for immediate adoption into the standard ten-step protocol described in this article.

A further practical limitation worth stating plainly is that no AI tool currently carries a specific regulatory clearance for brachial plexus pathology detection as a standalone indication; existing clearances generally cover broader musculoskeletal or general neuro-oncologic applications that happen to be applicable to plexus studies as a secondary use case. Radiologists should therefore treat any AI-generated finding on a plexus study as a second opinion to be weighed alongside their own independent reading, not as a substitute for the structured, segment-by-segment interpretation approach described throughout this article.

As acquisition speed continues to improve through deep-learning reconstruction, the practical bottleneck in plexus imaging is increasingly shifting from scan time toward interpretation time, since a technically excellent, artefact-free study still requires a systematic, segment-by-segment review to extract its full diagnostic value. Structured reporting templates, built around the roots-trunks-divisions-cords-branches framework introduced earlier in this article, remain the most reliable way to ensure that acquisition speed gains translate into genuine interpretation quality rather than simply faster throughput of studies reviewed less thoroughly.

Further reading

1. 7 Proven Strategies for Optimizing MRI Sequences in 2026
2. 2026 Contrast Media Guidelines: eGFR Thresholds & Safe Administration Protocol
3. Gadolinium-Enhanced MRI in Brain Metastases: Enhancement Patterns, Protocols & AI Radiomics
4. Vision-Language Models & Generative AI 2026: The Future of Radiology Reporting
5. Reducing Medical Waste Toxic Emissions: Complete Guide 2026

These five resources extend directly on themes introduced throughout this article: sequence optimisation principles applicable well beyond the plexus, current contrast safety thresholds relevant to every gadolinium-enhanced study performed in the department, enhancement pattern analysis in a different but conceptually related neuro-oncologic context, the broader trajectory of AI-assisted reporting referenced in the automation section above, and departmental sustainability considerations that increasingly shape purchasing decisions around consumables such as contrast delivery systems and coil hygiene products.

Reducing artefacts with patients and parameters

The most critical scanning parameters that impact image quality in a brachial plexus MRI protocol include spatial resolution, signal-to-noise ratio, image contrast, and artefact control, and each of these four domains interacts with the others in ways that the technologist must actively balance rather than optimise in isolation.

1. Spatial resolution

Spatial resolution defines the ability to distinguish small details in an image, and in this protocol that means distinguishing individual nerve fascicles from adjacent fat and vessel. Matrix size: increasing the matrix size, in both the frequency and phase directions, increases spatial resolution but decreases SNR because the resulting voxel size becomes smaller. Field of view (FOV): reducing the FOV also increases spatial resolution, but a smaller FOV likewise results in smaller voxels and a corresponding reduction in SNR. Slice thickness: thinner slices provide higher spatial resolution and reduce partial volume averaging across the small, obliquely running plexus fascicles, but they significantly decrease SNR and must be balanced against the total available scan time.

2. Signal-to-noise ratio (SNR)

SNR represents the strength of the diagnostic signal relative to the inherent background noise in the image; a high SNR produces crisp, clear images, whereas a low SNR produces a grainy, diagnostically limited result. Number of averages (NEX/NSA): increasing the number of averages acquires the same data multiple times, which improves SNR, though doubling the number of averages roughly doubles the total scan time. Receiver bandwidth: decreasing the bandwidth limits the amount of noise recorded per unit signal and boosts SNR, but a lower bandwidth increases both scan time and chemical shift artefact, a particularly relevant trade-off at the fat-nerve interface central to this protocol. Coil selection: using dedicated, localised surface coils rather than the whole-body coil captures a far stronger local signal and heavily improves SNR across the entire course of the plexus.

3. Image contrast

Contrast determines how different tissues are distinguished from one another, for example distinguishing nerve from muscle, fat, or fluid. Repetition time (TR): TR is the time between consecutive RF pulses; a short TR maximises T1 tissue contrast, while a long TR minimises it. Echo time (TE): TE is the time between the RF pulse and the peak of the resulting echo signal; a short TE minimises T2 effects, while a long TE maximises T2 weighting, making fluid-filled and oedematous areas appear very bright, exactly the contrast this protocol relies upon for pathology detection. Flip angle: the flip angle controls the excitation of protons, and adjusting it changes tissue contrast, a consideration that is especially critical in any gradient-echo based angiographic add-on sequence used to clarify vascular anatomy.

4. Artefact control

Artefacts are visual distortions or ghosting that degrade image quality, and in this protocol they are dominated by respiratory and vascular motion. Phase-encoding direction: swapping the phase and frequency axes can shift motion-induced artefacts, such as breathing or blood flow ghosting, away from the primary region of interest rather than through it. Flow compensation and gating: this utilises physiological triggers, for example a respiratory bellows, to minimise blurring and ghosting caused by pulsatile or respiratory motion, and is the direct technical implementation of the remedy specified for this protocol’s primary artefact. Parallel imaging: this utilises multiple coil elements simultaneously to reduce the number of phase-encoding steps required, significantly cutting down scan time and reducing the window of vulnerability to motion artefact overall.

A practical console checklist, applied consistently before every acquisition in this protocol, closes the loop between these four parameter domains: confirm phase-encoding direction matches the dominant expected artefact source, confirm respiratory gating or triggering is active for any sequence covering the lower trunk or infraclavicular segment, confirm saturation bands are placed to suppress vascular pulsation, and confirm the selected parallel imaging factor is within the department’s locally validated range for the coil in use. Working through these four checks takes under a minute at the console but accounts for the majority of avoidable repeat scans documented in departmental quality reviews.

It is worth stating explicitly that these four parameter domains, spatial resolution, SNR, image contrast, and artefact control, cannot be optimised independently of one another, since almost every adjustment that improves one domain degrades at least one of the others. Increasing matrix size to improve spatial resolution reduces SNR; decreasing bandwidth to recover that SNR increases chemical shift artefact; and adding averages to compensate for the resulting noise extends scan time and reintroduces vulnerability to the very respiratory and vascular motion this protocol works so hard to control. Recognising this web of trade-offs, rather than treating any single parameter as freely adjustable in isolation, is what separates a genuinely optimised protocol from one that has simply been tweaked without full consideration of the downstream consequences.

Parallel imaging protocols and parameters

Turbo, or echo-train, factor selection directly trades off scan time, spatial resolution, and image blurring in the fat-saturated T2 and STIR sequences that form the diagnostic core of this protocol. Higher turbo factors shorten scan time considerably but increase T2 blurring along the phase-encoding direction, an effect that is particularly undesirable when the clinical goal is resolving thin, closely spaced nerve fascicles.

As a practical rule, increasing the acceleration factor beyond the SNR reserve of the coil array reintroduces precisely the artefacts, residual aliasing and ghost overlay, that the protocol is otherwise designed to eliminate. Acceleration should therefore be increased incrementally and validated against a phantom or a known reference case before being deployed routinely on patients, rather than adopted wholesale from a vendor default without local verification.^[13]

In practical terms, most major vendor platforms now offer some form of simultaneous multi-slice or compressed-sensing acceleration in addition to conventional parallel imaging, and these newer techniques can further reduce scan time for the 3D SPACE sequence in particular, since it is the most acquisition-intensive component of the full protocol. As with conventional parallel imaging, any newly introduced acceleration technique should be validated locally against the department’s existing reference images before it replaces the standard protocol across the full patient population.

A brief note on reporting standards is worth adding here: a structured report for this protocol should explicitly state the plexus segment involved using the roots, trunks, divisions, cords, and branches terminology introduced earlier in this article, describe signal behaviour on both the STIR and long-TE T2 sequences rather than one alone, and state explicitly whether post-contrast imaging was performed and, if so, the enhancement pattern observed. This level of structured detail is what allows a referring surgeon or oncologist to act on the report directly, without needing to review the images themselves before planning further management.

Conclusion

A diagnostically reliable brachial plexus MRI protocol depends on disciplined execution of three core sequences, the large FOV coronal STIR, the small FOV axial fat-saturated T2, and the coronal 3D T1 SPACE, supported throughout by respiratory gating and vascular saturation to control the protocol’s dominant artefact of breathing and pulsation ghosting, consistently and without exception across every patient scanned under this protocol. Contrast, when clinically indicated, follows a standard 10 to 15 mL, 2.0 mL/s protocol with a matched saline chaser, and should be reserved for cases where tumor, plexitis, or infection genuinely enters the differential diagnosis.

Recognising the ten high-yield pathologies covered in this article, from root avulsion and Parsonage-Turner syndrome through to schwannoma, Pancoast invasion, and radiation plexopathy, requires careful correlation of T1 and T2 signal behaviour with enhancement pattern and clinical timeline. Equally important is understanding the three-tier pitfall framework that spans the scanner, the reporting workstation, and the referring clinic: each professional group is vulnerable to a distinct and predictable failure mode, and only genuine cross-disciplinary awareness closes the diagnostic loop from request to actionable report.

Departments that build this awareness into structured onboarding, standing protocol cards, and periodic case-review audits will consistently produce brachial plexus studies that answer the clinical question on the first attempt, sparing patients repeat appointments and sparing referring clinicians the diagnostic delay that an inadequate first study inevitably introduces. As three-dimensional isotropic imaging, deep-learning reconstruction, and quantitative fascicle-level analysis continue to mature, the fundamentals covered in this article, disciplined coil selection, deliberate artefact control, and careful correlation between signal behaviour and clinical context, will remain the foundation on which every future advance in plexus imaging is built.

For radiographers, the single highest-value habit to take from this article is reviewing the screening coronal STIR at the console for residual ghosting before proceeding further. For radiologists, it is maintaining a reflexive habit of correlating any short-TE hyperintensity against a long-TE sequence before attributing it to pathology. For referring physicians, it is providing a specific clinical question on every request form rather than a generic indication. None of these three habits requires additional equipment or training time measured in more than a few minutes, yet together they resolve the great majority of the pitfalls documented throughout this article. Departments that embed all three into routine practice, rather than treating them as optional refinements, will find that brachial plexus MRI stops being the unpredictable, technically fraught study it is sometimes reputed to be, and becomes instead a routine, reliably diagnostic examination like any other in the departmental repertoire.

References