Imaging the Elbow

Why the elbow is hard to read. The elbow has three articulations (ulnohumeral, radiocapitellar and proximal radioulnar), the overlapping bony anatomy needs precise positioning for a true AP and a true lateral, and in the child the ossification centres appear in sequence and can mimic fractures. Subtle fractures of the radial head and coronoid are easily missed, and the neurovascular structures lie close by.

What radiographs cannot show. Ligament injuries of the UCL and LCL complex are not visible on radiographs and need MRI. The terrible triad (dislocation with radial head and coronoid fractures) needs CT for full characterisation and surgical planning.

The principle. Read every elbow radiograph systematically. Start with the fat pad signs, the most sensitive radiographic indicator of intra-articular pathology and the indirect evidence of a fracture when direct visualisation fails; then the alignment lines (anterior humeral line, radiocapitellar line); then, in a child, the ossification centres.

The minimum series. AP and true lateral in every trauma presentation, with oblique views added for specific indications. Confirm that the lateral is true before reading anything from it: the olecranon is in profile and the trochlea forms concentric circles. Fat pad signs are only reliably assessed on a true lateral, and an oblique lateral invalidates the assessment.

The anterior fat pad. Normally a small triangle. An effusion lifts it into the sail sign. A small anterior fat pad may be normal, which is why it is less specific than the posterior one.

The posterior fat pad. It lies in the olecranon fossa and is hidden there on a normal true lateral. Joint distension by effusion or haemarthrosis pushes it out of the fossa, where it appears as a small lucent line behind the distal humerus. A visible posterior fat pad is always abnormal, and after trauma it means an occult fracture until proven otherwise even when no fracture line is visible: most commonly a radial head fracture in an adult and a supracondylar fracture in a child.

The number behind "until proven otherwise". When adults with a post-traumatic effusion and normal radiographs were scanned, MRI found an occult fracture in 75%, and 87% of those were of the radial head. In children with an elevated posterior fat pad and no other radiographic abnormality, 76% went on to show periosteal new bone confirming a fracture. That paediatric figure changed practice: older series had put the risk at only 6-29%, which is why the sign was once treated as soft evidence and is now treated as a fracture.

Size does not stratify the risk. Neither the volume of the effusion nor the degree of fat pad elevation predicted whether a fracture was present. The size of the fat pad decides whether there is an effusion, not how bad the injury is, so a barely lifted posterior fat pad carries the same implication as a dramatic one, and "it was only a small effusion" is not a reason to relax follow-up. Treat the patient as having a fracture: immobilise and repeat the radiographs at 7-10 days, or obtain CT or MRI if clinical concern persists.

View larger

Ultrasound. Dynamic assessment of ligaments and tendons, effusion detection and guided injection. In lateral epicondylitis it shows a hypoechoic, thickened common extensor origin with neovascularisation and tears.

View larger

The UCL, and why arthrography is added. MRI is the gold standard for the UCL and shows the anterior bundle at its sublime tubercle insertion. Unenhanced MRI detects complete tears essentially every time, but its sensitivity across all UCL tears is only about 57% because it misses the partial undersurface tear. Intra-articular contrast lifts sensitivity to roughly 86% by showing dye tracking around the detached deep fibres while still contained by the intact superficial layer, the T-sign.

OCD of the capitellum. The physeal status on the AP radiograph is the single most decisive variable, and MRI is the gold standard for staging. A lesion is stable with an open physis, localised flattening or lucency and good motion; it is unstable with a closed physis, or fragmentation, or a loss of motion of 20 degrees or more. Stable lesions with an open physis heal with rest; once the physis is closed, surgery outperforms rest; and for unstable lesions fixation or articular reconstruction beats simple fragment excision. MR arthrography also stages osteochondral lesions and finds intra-articular loose bodies.

A positive fat pad sign or an abnormal alignment line is a starting point, not a diagnosis. The table maps the common radiographic appearances to their differentials and the discriminating feature that resolves them.

View larger

View larger

Across all major training systems, elbow imaging follows the same evidence-based, stepwise logic: plain radiographs first, CT for fracture detail, and MRI or ultrasound for soft tissue. Differences between societies are largely about access and the threshold for advanced imaging rather than the fundamental algorithm.

Global epidemiology. Radial head/neck fractures are the most common adult elbow fracture (roughly a third of elbow fractures) and the most common occult fracture behind a positive fat pad sign. Supracondylar humerus fractures are the most common paediatric elbow fracture and the leading operatively-treated fracture in children, peaking at ages 5-7 — exactly the cohort where ossification-centre interpretation is hardest. Osteochondritis dissecans of the capitellum and UCL injury cluster in adolescent and adult overhead/throwing athletes.

Registry note. Unlike hip and knee arthroplasty, elbow imaging is not tracked by joint registries; the comparative evidence base is radiology and trauma cohort literature rather than implant-survival data. Where total elbow replacement is performed (e.g. for distal humerus fractures in low-demand elderly patients), national registries such as the NJR (UK), AOANJRR (Australia) and the Norwegian Arthroplasty Register report on implant outcomes, but pre-operative characterisation still rests on CT.

High- vs limited-resource practice variation. In well-resourced settings CT and MRI are obtained promptly for complex fractures and soft-tissue diagnosis. In limited-resource settings the diagnosis frequently rests on radiographs plus disciplined clinical examination and contralateral comparison views; ultrasound — portable, inexpensive and radiation-free — is an especially valuable substitute for MRI in assessing effusion, the UCL and the common extensor/flexor origins. The systematic radiograph (fat pads, anterior humeral line, radiocapitellar line, CRITOE) is therefore the one universal, resource-independent skill every candidate must master.

Comparison views in children. Contralateral radiographs help when ossification-centre interpretation is genuinely uncertain: the centres should be symmetric, and any asymmetry suggests either a fracture or a normal variant, a normal variant being bilateral. Routine bilateral imaging doubles the radiation and rarely changes management in straightforward cases, so most guidance, including BOAST, discourages reflex comparison views in favour of disciplined CRITOE assessment and selective use only when there is real diagnostic doubt.

Does the fat pad sign mandate further imaging? A positive posterior fat pad sign after trauma predicts an occult fracture in roughly three-quarters of patients, yet MRI confirmation rarely alters acute management because most occult fractures are stable radial head injuries treated symptomatically. The pragmatic, evidence-aligned approach is to treat as a fracture and reserve CT or MRI for the cases where confirmation would change the plan: suspected displacement, a mechanical block, or an athlete who needs a prognosis.

MRI versus ultrasound for the UCL. MRI, or MR arthrography for partial tears, is the reference standard for UCL anatomy, but dynamic stress ultrasound assesses functional valgus laxity in real time and is radiation-free and repeatable. Throwing-arm thickening, hypoechoic foci and increased gapping are normal adaptive findings, so ultrasound must be interpreted against the contralateral side and the symptoms; a single study does not predict who will tear the ligament.

2D versus 3D CT for articular fracture mapping. For coronoid and radial head fragments, CT consistently reveals anatomy underestimated by radiographs, and 3D reconstruction improves appreciation of fragment size and pattern. Whether 3D meaningfully changes the operative plan over good 2D multiplanar CT remains debated, but CT in some form is now standard for terrible triad and complex distal humerus planning.