Shoulder radiography must show the proximal humerus and glenohumeral relationship without making an acute injury worse. The ACR rates shoulder radiography as the appropriate initial examination for adult acute shoulder pain. The number and combination of projections are determined by the clinical question, order, patient condition, equipment and validated local protocol—not by a universal “four-view trauma series.”
ARRT's radiography content specifications include AP internal and external rotation, AP neutral, inferosuperior axial (Lawrence), posterior-oblique Grashey and PA-oblique scapular Y projections. That list describes examinable procedures; it does not prescribe a minimum trauma series or a fixed exposure technique.
With suspected fracture or dislocation, keep the arm in the position found unless the responsible clinician and departmental protocol direct otherwise. Do not force internal rotation, external rotation or abduction merely to reproduce a routine textbook image. Obtain the prescribed diagnostic projections by moving the patient, tube and detector; use a no-abduction alternative when needed.
The glenohumeral joint is the articulation of the humeral head and glenoid fossa. The acromion and coracoid arise from the scapula, and the lateral clavicle meets the acromion at the AC joint. Because scapular orientation varies among patients, quoted obliquity ranges are starting points: image criteria, not a memorized angle alone, determine whether the projection is correct.
For a routine AP, place the posterior shoulder against the detector, include the entire glenohumeral joint and relevant proximal humerus, and direct the central ray perpendicular at approximately 2.5 cm (1 inch) inferior to the coracoid. A 100–110 cm (40–44 inch) SID is common in US teaching, but SID and detector orientation are protocol-specific.
Positioning: When clinically safe, externally rotate the arm until the epicondyles are parallel to the detector; the palm generally faces forward. Do not obtain this forced position in acute trauma unless requested and tolerated.
Image criteria: The greater tubercle is in profile laterally, and the lesser tubercle is largely superimposed on the humeral head. Some overlap of humeral head and glenoid is expected because this is not a true AP of the scapular plane.
Positioning: When safe, flex the elbow as needed and place the posterior hand against the hip so the epicondyles are approximately perpendicular to the detector. Simply resting the palm against the thigh commonly produces neutral, not full internal, rotation.
Image criteria: The lesser tubercle is in profile medially; the greater tubercle is superimposed on the humeral head. Internal rotation may help display a Hill–Sachs impaction defect, but no single projection excludes one.
Neutral is the trauma-safe AP when rotation is contraindicated or not tolerated: leave the palm against the thigh and epicondyles about 45° to the detector. Neither tubercle is fully in profile. Label the achieved position accurately rather than calling every trauma AP “internal rotation.”
For the standard AP-oblique Grashey, place the affected posterior shoulder against the detector and rotate the torso toward that side—commonly about 35–45°—until the scapular plane is parallel to the detector. Direct a perpendicular ray to the glenohumeral joint using palpated anatomy and the departmental centering guide. Arm rotation varies with the clinical question and protocol; it is not what opens the joint.
Image criteria: The anterior and posterior glenoid rims are nearly superimposed and the glenohumeral joint space is profiled with minimal humeral-head overlap. Under- or over-rotation closes the joint. The projection can assess osseous joint-space loss and glenoid contour, but a normal radiograph does not exclude a soft-tissue labral tear.
ARRT lists Grashey as a posterior oblique and scapular Y as a PA oblique. “View” describes the finished image; “projection” describes the ray path. A patient facing the detector in RAO or LAO is not an AP oblique.
A common erect method places the affected anterior shoulder against the detector in RAO or LAO. Rotate the torso approximately 45–60°, then adjust for habitus until the scapular body is perpendicular to the detector. Direct the ray perpendicular to the scapulohumeral joint. In trauma, leave the arm supported in the position found and achieve the obliquity with the torso and equipment.
Image criteria: The thin scapular body forms the stem of the Y; the acromion and coracoid form its arms. In a reduced joint, the humeral head projects over the glenoid at the Y intersection. Failure to obtain a true lateral can mimic displacement, so use image orientation and a second diagnostic plane before assigning direction. The projection also surveys the scapular body, neck, acromion and coracoid, but dedicated scapular projections or CT may be required for fracture characterization.
An axillary projection gives the orthogonal relationship of the humeral head to the glenoid and is particularly valuable when posterior dislocation is possible. “Axillary” is a family of methods; do not combine their ray angles and detector positions.
For the nontrauma Lawrence method, position the patient supine near the table edge, raise the shoulder only enough to permit the ray path, and abduct the arm as close to 90° as the protocol requires and the patient safely tolerates. Place the detector superior to the shoulder. Direct the ray through the axilla toward the glenohumeral joint, commonly angled about 25–30° medially; adjust the angle for the actual degree of abduction and follow the department's validated method. This is the Lawrence method—not an alternative for a patient who cannot abduct.
If abduction is contraindicated, stop rather than forcing a Lawrence position. A validated Velpeau axillary projection can show the glenohumeral relationship with the arm immobilized: the seated or standing patient leans backward over the detector while a vertical or protocol-specified angled ray is centered to the joint. Geometry varies among departments, so follow the named local method. A scapular Y is another commonly used orthogonal trauma projection, although each method has limitations.
Image criteria: The glenoid is shown in profile with the humeral head's anterior/posterior relationship assessable; the coracoid is anterior. An axillary image can demonstrate glenoid-rim and humeral-head impaction fractures, but CT is often used when detailed fracture morphology is needed.
| Projection | Body / arm | Central ray | Primary image criterion | Trauma note |
|---|---|---|---|---|
| AP external | Posterior shoulder at detector; epicondyles parallel | Perpendicular, ~2.5 cm inferior to coracoid | Greater tubercle lateral profile | Do not force external rotation |
| AP internal | Back of hand on hip; epicondyles perpendicular | Same AP centering | Lesser tubercle medial profile | Not the mandatory first trauma view |
| AP neutral | Arm in position found / palm at thigh | Same AP centering | Neither tubercle fully profiled | Preferred when rotation is unsafe |
| Grashey | Affected posterior side at detector; ~35–45° toward side | Perpendicular to glenohumeral joint | Open joint; glenoid rims nearly superimposed | Not an orthogonal replacement for axillary |
| Scapular Y | PA oblique; affected anterior side at detector; ~45–60° | Perpendicular to scapulohumeral joint | Scapula lateral; head at Y if reduced | Move torso, not injured arm |
| Lawrence axillary | Supine; substantial safe abduction | Through axilla, commonly 25–30° medial | Head-to-glenoid relationship | Contraindicated if required abduction is unsafe |
| Velpeau axillary | Arm immobilized; patient leans over detector | Protocol-specific to joint | Orthogonal alignment without abduction | Named trauma alternative |
There is no ARRT statement defining a “minimum acceptable trauma shoulder series.” Follow the order and local protocol. A defensible examination assesses the joint in different planes—for example AP or Grashey plus scapular Y and/or axillary. A Grashey and routine AP are not orthogonal to one another, so a Grashey must not be presented as a substitute for an axillary image merely because abduction is painful.
Most glenohumeral dislocations are anterior. On a frontal projection the head is commonly displaced inferomedially; on a true lateral or axillary projection its direction relative to the glenoid can be determined. Posterior dislocation may show fixed internal rotation and a rounded “light-bulb” appearance on AP, but this sign is not sufficiently sensitive by itself. Confirm alignment in an orthogonal plane. A Hill–Sachs lesion is a posterolateral humeral-head impaction associated with anterior dislocation; a reverse Hill–Sachs involves the anteromedial head with posterior dislocation.
Radiographs can show proximal-humerus, scapular and glenoid-rim fractures. The classic Neer system uses four proximal-humerus segments and thresholds of more than 1 cm displacement or more than 45° angulation, but classification and management belong to the interpreting clinician. A “Bankart lesion” usually refers to an anteroinferior labroligamentous injury and is not excluded by normal radiographs; only an associated bony Bankart fragment may be directly visible.
Use a dedicated AC-joint protocol when ordered. Bilateral comparison and Zanca-type projections may be used locally. Weighted stress views remain controversial: a systematic review found no consensus on a gold-standard acute AC imaging protocol or clear routine value for weighted views. Do not diagnose separation from an oversimplified universal millimetre cutoff, and do not give patients weights unless the protocol specifically calls for them.
Radiographers evaluate positioning, anatomy coverage and image quality and must communicate urgent observations through the facility's escalation pathway. They do not independently diagnose a dislocation, assign a Rockwood or Neer grade, or choose CT/MRI outside authorization and local scope.
ACR guidance supports radiographs first for adult acute shoulder pain. If an occult fracture remains suspected after negative or indeterminate radiographs, noncontrast CT or noncontrast MRI is usually appropriate. With dislocation/instability, MRI without contrast is usually appropriate as the next study; CT without contrast or MR arthrography may be appropriate in selected cases. For suspected labral tear after negative/indeterminate radiographs, MR arthrography or noncontrast MRI is usually appropriate. For suspected rotator-cuff tear, ultrasound or noncontrast MRI is usually appropriate. These are clinician-level choices, not automatic additions by the radiographer.
Follow the facility's pregnancy-screening and justification policy. A clinically necessary shoulder radiograph should not be delayed solely because a patient is pregnant: the uterus is outside a correctly collimated shoulder beam, so fetal exposure is limited to very low internal scatter. Optimize the examination through correct collimation, technique selection and avoidance of repeats.
AAPM, with ACR endorsement, recommends discontinuing routine gonadal and fetal contact shielding. Surface shielding does not stop internal scatter and may obscure anatomy, create artifacts or increase output if it enters an AEC field. Do not routinely add gonadal, fetal or breast shields; follow current institutional policy and use any exceptional comfort shielding only when it cannot enter the field or interfere with acquisition.
Audit note: Numeric angles and distances above are common US teaching starting points and are explicitly identified as protocol-dependent. No page-specific claim is attributed to Clark's or another inaccessible textbook.
Try these ARRT-style multiple choice questions based on this article. Click an option to check your answer — correct answers turn green, wrong ones turn red.