Ultrasound in Musculoskeletal Practice

How the image forms. Ultrasound uses sound above the audible range (more than 20kHz), and MSK imaging uses 5-18MHz. Piezoelectric crystals in the transducer generate the waves, which are transmitted into tissue, reflected at tissue interfaces and received by the same transducer to construct the image.

Resolution against penetration. Higher frequency gives better axial resolution, the ability to distinguish structures along the beam axis: about 0.1mm at 15MHz. Higher-frequency waves are also absorbed more rapidly, so useful penetration at 15MHz is about 3-4cm, whereas 5MHz reaches 15-20cm.

This trade-off is the fundamental physics concept. You cannot have maximum resolution and maximum depth at the same time.

Acoustic impedance. Each tissue has a characteristic acoustic impedance (Z = density × speed of sound), and sound is reflected at interfaces between tissues of different impedance. The greater the mismatch, the stronger the reflection, which is why cortical bone, with its very high impedance, produces an extremely bright surface reflection.

Coupling gel. Air's acoustic impedance is very different from tissue's, so without gel nearly all the energy is reflected at the skin-air interface. Gel eliminates that interface and lets the sound enter the tissue efficiently.

Anisotropy is the most important artefact in MSK work. When the beam strikes a tendon at any angle other than perpendicular (90 degrees), the reflection is deflected away from the transducer, and the normally bright, fibrillar tendon appears falsely hypoechoic, mimicking tendinopathy or a tear. It is the most common cause of false-positive tendon pathology.

Where it appears and how to resolve it. It is greatest at curved surfaces: the rotator cuff insertion, the Achilles insertion and any tendon that curves around a bony prominence. Confirm a finding in both long and short axis, and gently tilt the transducer to restore perpendicularity; if the dark area resolves with tilt, it is artefact.

View larger

Acoustic shadowing. Dense structures such as bone cortex and calcification reflect or absorb nearly all the sound and leave a dark shadow behind them. The shadow helps identify calcification but obscures deeper structures.

Posterior acoustic enhancement. Fluid-filled structures (cysts, effusions) transmit sound more readily than surrounding tissue, so the area behind them is brighter. This helps confirm that an anechoic structure is fluid-filled.

Reverberation. Multiple reflections between two highly reflective parallel surfaces create repeated echoes at regular intervals deeper in the image. It is common near metal surfaces and at air-tissue interfaces.

Edge artefact (refraction). Hypoechoic lines form at the curved edges of round structures, such as cysts and tendons in cross-section, and can mimic a tear at the margin of a tendon.

Doppler imaging detects moving blood and is central to assessing vascularity in tendinopathy, synovitis and soft-tissue masses. Colour and power Doppler are not interchangeable, and the distinction is a common viva point.

Colour Doppler encodes the mean Doppler frequency shift as colour, displaying the direction of flow (towards or away from the transducer, conventionally red and blue) and a sense of velocity. It is angle-dependent and relatively insensitive to slow flow.

Power Doppler displays the integrated power (amplitude) of the signal regardless of direction. It gives no direction or velocity, but it is far more sensitive to the low-volume, slow flow that matters in MSK, which is why it is the mode of choice for the neovascularisation of tendinopathy and for low-grade synovitis.

Grading synovitis. EULAR grades synovitis on combined greyscale and power Doppler, on a semiquantitative scale from grade 0 to 3.

View larger

Settings for slow flow. To capture slow flow without artefact, use a low pulse repetition frequency (PRF) and a low wall filter, and set the colour gain just below the level that fills the image with random colour (flash artefact).

Ultrasound is the primary imaging modality for rotator cuff assessment in many centres, with sensitivity for full-thickness tears of 89-95% in experienced hands, comparable to MRI.

How it compares with MRI. Pooling 65 studies that used a surgical reference standard, there was no significant difference in sensitivity or specificity between ultrasound and MRI for either partial- or full-thickness tears. On summary ROC area, MR arthrography was most accurate at 0.935, with ultrasound 0.889 and MRI 0.878.

Ultrasound therefore sits numerically above plain MRI rather than trailing it, although the pairwise comparison of those curves found no significant difference. The meaningful step up is intra-articular contrast, rather than the switch from sound to magnetic resonance.

What that means in practice. Ordering an MRI purely because ultrasound "isn't good enough" for a cuff tear is not evidence-based; ultrasound is a valid first-line investigation. MRI earns its place for what it shows besides the cuff: muscle quality and Goutallier grade, labrum, bone marrow and occult fracture.

The caveat belongs in the same breath. These pooled figures come largely from expert centres, ultrasound performance is operator-dependent in a way MRI is not, and a service without trained MSK sonographers should not assume the numbers apply to it.

The standard shoulder protocol runs as follows:

1. Biceps tendon (long head): transverse and longitudinal views in the bicipital groove, looking for tenosynovitis (fluid around the tendon), subluxation, dislocation and tears.
2. Subscapularis: internal rotation brings the tendon to the anterior scanning window. Assess for partial- and full-thickness tears; dynamic external rotation shows the tendon rolling over the lesser tuberosity.
3. Supraspinatus: modified Crass position (hand on back pocket), in long and short axis. The appearances of partial- and full-thickness tears are set out in the hypoechoic tendon table below.
4. Infraspinatus and teres minor: external rotation with the arm adducted. They are less commonly torn but are assessed as part of the complete protocol.
5. AC joint: superior assessment for osteophytes, effusion and instability.

Dynamic impingement testing. Under real-time ultrasound the examiner sees subacromial bursal thickening and cuff compression during abduction. This is unique to ultrasound and impossible with MRI.

View larger

View larger

Ultrasound guidance improves injection accuracy and safety by showing the needle in real time, but the size of that benefit is strongly target-dependent, and for one common target it essentially disappears. The figures below pool 4 cadaveric studies (300 shoulders) and 9 human studies (514 patients).

The subacromial result is the one to remember, because it is counter-intuitive and commonly taught wrongly. The subacromial bursa is large and superficial, so an experienced clinician using landmarks hits it about as often as an ultrasound-guided needle does.

Guidance still helped there. In the same analysis, ultrasound-guided subacromial injection produced better pain and function at 6 weeks, which suggests the benefit comes from something other than simply being in the right compartment (more consistent distribution within it, or avoiding intratendinous injection).

If you justify subacromial guidance on accuracy grounds in a viva, the evidence does not support you. Most target-specific estimates rest on single studies, and whether better accuracy translates into better patient-reported pain and function is itself target-dependent and still debated; some benefit of corticosteroid may be systemic rather than local.

Other guided procedures include:

• Hip joint injection: anterior approach targeting the femoral head-neck junction. Essential for obese patients, in whom blind injection accuracy drops significantly.
• Knee joint aspiration: suprapatellar pouch approach, with ultrasound confirming the needle is within the effusion before aspiration.
• Tendon sheath injection: De Quervain stenosing tenosynovitis and trigger finger, with precise delivery into the sheath without tendon injury.
• Barbotage for calcific tendinitis: ultrasound-guided needling and aspiration of calcific deposits in the rotator cuff.

Key principles of guided injection:

1. An in-plane technique shows the full length of the needle.
2. Confirm the needle tip position before injecting. Loss of tip visualisation is a stop signal, not an invitation to advance blindly.
3. Watch the injectate spread to confirm correct placement.

View larger

Ultrasound is the gold standard for assessing the neonatal hip in developmental dysplasia of the hip (DDH), because the femoral head is predominantly cartilaginous and invisible on plain radiography until ossification begins at 3-6 months of age.

The Graf method. It uses a coronal ultrasound image of the hip in the standard section, in which the iliac bone, labrum and femoral head are all visible. Two angles are measured from this image:

• Alpha angle, between the iliac bone and the bony acetabular rim. Normal is more than 60 degrees; less than 50 degrees is dysplastic.
• Beta angle, between the iliac bone and the cartilaginous rim (labrum). Normal is less than 55 degrees.

Dynamic assessment. Barlow and Ortolani manoeuvres under real-time ultrasound can demonstrate hip instability that may not be evident on static imaging.

Reproducibility in a screening programme. Very good, but conditionally. Across 2071 scans from 798 infants, each read by the original operator and an independent one, intra- and inter-observer variability of the alpha and beta angles was not statistically significant (intraclass correlation above 0.8), and agreement on Graf class exceeded 91% with Cohen kappa above 0.8.

The condition attached. That agreement held when the standard section was correctly obtained, and the clause is the whole operator-dependence argument. The reliability is a property of the technique performed properly, not of ultrasound in general, and a poorly acquired image degrades agreement.

Because the alpha angle drives treatment decisions in a screened population of well babies, an off-axis section does not produce an uncertain result so much as a confidently wrong one. This is why Graf training and image-quality audit are part of any screening programme rather than optional refinements.

High-frequency ultrasound resolves peripheral nerves in detail and is a routine first-line test for entrapment. In short axis a nerve has a honeycomb appearance, hypoechoic fascicles set in hyperechoic connective tissue; in long axis it shows a fascicular, railroad-track pattern.

Nerve or tendon. Two features separate a nerve from an adjacent tendon: a nerve shows much less anisotropy, so it does not black out as the beam tilts, and it does not glide on muscle contraction the way a tendon does.

Carpal tunnel syndrome. Measure the median nerve's cross-sectional area (CSA) at the tunnel inlet (pisiform level). The nerve is enlarged in entrapment, a CSA above roughly 9 to 10 square millimetres supporting the diagnosis, and the wrist-to-forearm CSA ratio increases.

Measuring it reproducibly. Mark the distal wrist crease, the inlet and serial forearm levels, keep the transducer perpendicular, and trace inside the hyperechoic epineurial rim. Compare CSA along the nerve rather than relying on a single image.

View larger

Beyond the area. Look also for flattening of the nerve within the tunnel and palmar bowing of the flexor retinaculum. Echotexture alone is not a severity scale: grade severity with cross-sectional area, flattening and clinical or electrodiagnostic findings. Ultrasound complements nerve conduction studies and can reveal a structural cause, such as a ganglion, an anomalous muscle or tenosynovitis.

View larger

Cubital tunnel. The ulnar nerve is enlarged at the medial elbow. Dynamic scanning during elbow flexion can demonstrate ulnar nerve subluxation or frank dislocation over the medial epicondyle, and a snapping triceps, a real-time finding unique to ultrasound.

Other lesions seen on nerve ultrasound:

• Common peroneal nerve at the fibular neck: entrapment, ganglion
• Morton neuroma: a hypoechoic interdigital mass, with the sonographic Mulder click on transverse compression
• Traumatic and stump neuromas
• Nerve sheath tumours (schwannoma, neurofibroma): a fusiform mass entering and exiting the nerve

Ultrasound detects cortical breach directly: the hyperechoic cortical line steps, angulates or is interrupted. It is accurate enough to function as an adjunct to radiography rather than a curiosity. Pooling 30 prospective studies (n=3506):

• Paediatric forearm fractures: sensitivity 93.1% (95% CI 87.2-96.4), specificity 92.9% (95% CI 86.6-96.4)
• Adult ankle fractures: sensitivity 89.5%, specificity 94.2%
• Assessing adequacy of fracture reduction: sensitivity 94-100%

Where it changes practice is the child. Confirming a forearm fracture and checking its reduction at the bedside without ionising radiation is a direct application of the paediatric dose principle. In resource-limited settings the same logic applies where radiography is scarce or slow.

Checking the reduction. The reduction figures mean ultrasound can check the position before the patient leaves the plaster room rather than after a further radiograph. Specificity for reduction ranged from 56% to 100%, however, so ultrasound can call a reduction adequate when it is not.

The boundary. This is an adjunct, not a replacement. Accuracy varies by anatomical site and by operator, ultrasound does not characterise intra-articular extension, comminution or associated injuries the way a radiograph does, and a negative scan in a child with convincing clinical signs still warrants imaging. Use it to answer a focused question quickly, not to survey the injury.

View larger

Strong evidence supports ultrasound for rotator cuff assessment and guided interventions.

Broader applications of ultrasound continue to expand in orthopaedic practice.