Coming soon
Video walkthrough
We’re producing a short video for every module — check back soon.
Exam category: Procedures → Extremity
Why it matters: The upper limb carries more testable detail than any other region — ten separate anatomic areas, each with a routine series, a part position, a centering point, and an evaluation criterion. On top of that sit the named/eponym views (Stecher, Gaynor-Hart, Coyle, Grashey, Lawrence, Neer), which are the highest-yield trap material here: each exists to solve a specific clinical problem, and the registry tests why that view is chosen, not just its name.
Work distal to proximal: fingers → hand → wrist → forearm → elbow → humerus → shoulder → clavicle/AC joints. Routine series for most joints follow the same pattern (AP or PA, lateral, oblique), and the eponym views exist specifically for the cases routine views can't handle — a bone that superimposes on itself, a joint that can't flex enough for a standard oblique, or a specific pathology that needs a specific angle to show.
The names are easiest to learn as two transverse rows of four. Starting on the lateral (thumb) side each time:
The scaphoid is the largest bone of the proximal row, articulates with the radius, and is the most frequently fractured carpal — which is why it gets its own projections. The capitate is the largest carpal overall, and the pisiform the smallest. The hamate carries the hooklike hamulus on its palmar surface.
Bending the wrist toward one side pulls apart the carpals on the opposite side. That single rule covers both deviation projections:
Routine wrist: PA, lateral, PA oblique. The special views:
Memory cue: the deviation and the carpal canal both trade one geometry for another. Deviate away from what you want to see; and whatever hyperextension the patient can't give you, add back as central-ray angle.
Routine: AP, lateral.
The forearm is not radiographed pronated. Pronating the hand pivots the radius so it crosses over the ulna in the upper forearm, superimposing the two bones. Radiographing AP with the hand supinated (palm up, anatomic position) keeps the radius and ulna parallel and side by side, which is the whole point of the projection.
The same pivot governs the elbow obliques, and the direction decides the outcome:
The elbow's three fat pads and stripes are visible only on the lateral projection; on the AP they are superimposed over bone.
The trap. A visible posterior fat pad only means something if the elbow was flexed exactly 90°. If the elbow is extended beyond 90°, the olecranon slides into the olecranon fossa and elevates the posterior fat pad on its own — the pad then appears whether the exam is negative or positive, and the sign is worthless. Before reading the sign, confirm the flexion.
Routine shoulder: AP internal/external rotation (non-trauma) or AP neutral (trauma, no rotation attempted). Rotating the arm swings the two tubercles of the proximal humerus around, and each rotation profiles a different one:
| Rotation | How it's done | Epicondyles | What ends up in profile |
|---|---|---|---|
| External | Arm abducted slightly, hand supinated | Parallel to the IR | Greater tubercle, in full profile on the lateral aspect of the humeral head; the lesser tubercle is superimposed over the head. This is the true AP of the proximal humerus |
| Internal | Arm abducted slightly, hand pronated | Perpendicular to the IR | Lesser tubercle, in full profile on the medial aspect of the head; only an outline of the greater tubercle shows over the head. This is the lateral of the proximal humerus |
| Neutral (trauma) | Arm left at the side "as is" — no rotation attempted | Roughly 45° to the IR | Neither. Both tubercles are usually superimposed by the humeral head, giving an intermediate, oblique-looking appearance |
The AP humerus uses the same external-rotation logic: abduct the arm slightly and gently supinate the hand so the epicondyles are parallel and equidistant from the IR, with the CR to the midpoint of the humerus.
Why neutral is the trauma answer. If a fracture or dislocation is suspected, do not attempt to rotate the arm at all. Neutral is not a worse version of the other two — it is the projection you deliberately take when rotating the patient's arm could do harm.
Three eponym views, each solving a different problem:
| Method | What it's for | Key positioning fact |
|---|---|---|
| Grashey (AP oblique) | Opens the glenohumeral joint space free of superimposition — fractures/dislocations, Bankart lesion, glenoid rim erosion | Body rotated 35-45° toward the affected side; CR perpendicular |
| Lawrence (inferosuperior axial) | Gives a lateral view of the proximal humerus in relation to the scapulohumeral (glenohumeral) joint — degenerative conditions; Hill-Sachs defect. The coracoid process and lesser tubercle come out in profile | Arm abducted ~90°, CR angled medially 25-30° through the axilla. Never attempt if fracture/dislocation is suspected — requires arm abduction |
| PA oblique (scapular Y lateral) | Trauma-safe view for fractures and dislocations of the proximal humerus | Body obliqued until the scapula is on end; acromion and coracoid form the two upper arms of a "Y". Humeral head over the base of the Y = not dislocated; anterior dislocation throws it below/in front, posterior behind |
| Neer (supraspinatus outlet) | Trauma-safe; specifically demonstrates the coracoacromial arch for shoulder impingement | Same anterior-oblique body position as the scapular Y, but with a 10–15° caudal CR angle — that angle is what makes it the outlet view. Arm is abducted slightly; do not rotate it |
Memory cue — same body, different beam: the scapular Y and Neer are the same anterior-oblique position; Neer adds a 10–15° caudad CR to open the supraspinatus outlet. Lawrence and Grashey need the arm rotated or abducted, so they carry trauma risk — the scapular Y and Neer don't require arm rotation, which is why they're the trauma-safe pair.
Clavicle: AP or PA, plus an axial projection to project it off the ribs/lung fields.
AC joints: AP bilateral, taken with and without weights — weighted views stress the joint to reveal subtle separations that don't show on an unweighted image. A shoulder or clavicle series should clear a fracture first.
A note on technique for AC joints. Grids generally are not used for body parts measuring 10 cm or less, and AC joints usually measure under 10 cm. So they are commonly done nongrid at a lower kVp (about 70–75) — a thin part throws little scatter, and less kVp gives adequate penetration with better contrast. (Departments vary; a grid raises dose because the exposure factors have to go up with it.)
The scapula sits behind the ribs, so both routine views are built to get it out from behind them:
Rotate the patient into an anterior oblique until an imaginary line between the superior angle of the scapula and the AC joint is perpendicular to the IR — putting the thin scapular body on end. Because patients differ, the amount of body obliquity ranges from about 45° to 60° from the frontal position. Abduct the arm slightly so the proximal humerus doesn't superimpose the ribs, but do not rotate the arm.
The acromion and coracoid processes form the two upper limbs of the Y, the scapular body the leg, and the glenoid cavity sits at the junction where they converge.
Which limb is which. The coracoid is the anterior landmark, the acromion the posterior one. So the head goes under the coracoid for an anterior dislocation and under the acromion for a posterior one — the direction of the dislocation names the process it hides under.
Everything above is the exception list. This is what the exceptions are exceptions to: the routine series for each part, how the part physically goes down on the receptor, the centering point the beam goes to, and the evaluation criteria — the specific finding on the finished image that proves you positioned it correctly. The registry tests all four directly.
| Projection | Part position |
|---|---|
| PA finger | Hand pronated, the palmar surface of the finger in contact with the IR, finger extended and aligned to the long axis of the receptor |
| Lateral finger | Hand turned into a lateral position, thumb side up, affected finger extended and the unaffected fingers flexed away so they don't superimpose it. That puts the ulnar (medial) surface of the affected digit against the receptor — a lateromedial. Exception — the index finger: turn the hand the other way, so the index is the digit touching the IR. Its radial (lateral, thumb-side) surface now rests on the receptor, which makes it a mediolateral. Definition improves with the part closest to the receptor |
| AP thumb | Internally rotate the whole hand until the posterior (dorsal) surface of the thumb rests on the IR. A PA thumb is a fallback only when the patient can't manage the AP — it loses definition to the increased OID |
| PA hand | Hand pronated, palmar surface on the IR, fingers extended and spread slightly so the digit shafts and soft tissues don't overlap |
| "Fan" lateral hand | Rotate to a lateral position, thumb side up, then spread the fingers and thumb into a fan on radiolucent supports so every digit is separated and parallel to the IR. This is the preferred lateral when the phalanges are the area of interest — the spreading is what frees each phalanx from its neighbors |
| Lateral hand in extension / flexion | Same thumb-up lateral, but with the digits superimposed rather than fanned. Offered as the alternative to the fan lateral for localizing foreign bodies in the hand and fingers; the flexed version is often less painful |
| Lateral wrist | Elbow flexed 90° with the humerus abducted so humerus and forearm lie on the same horizontal plane; rotate the arm outward so the medial (ulnar) forearm and hand rest against the IR — a thumb-up lateral |
| AP forearm | Elbow extended, hand and arm fully extended palm up (supinated), epicondyles equidistant from the IR |
| AP humerus | Arm abducted slightly, hand gently supinated so the epicondyles are parallel and equidistant from the IR |
The one rule underneath most of this row: get the part you are imaging closest to the receptor, and put the anatomy you want to see in profile. That is why the finger is PA and not AP, why the index finger reverses the lateral, and why the thumb is the one digit imaged AP.
| Part | Routine projections | CR centered to |
|---|---|---|
| Finger | PA, lateral, PA oblique | PIP joint of the finger in question |
| Thumb | AP, lateral, PA oblique | First MCP joint |
| Hand | PA, lateral, PA oblique (rotated laterally 45°, toward the thumb side) | Third MCP joint |
| Wrist | PA, lateral, PA oblique (rotated laterally 45°) | Midcarpal area |
| Forearm | AP, lateral | Mid-forearm — both the wrist and elbow joints must be on the image |
| Elbow | AP, lateral (flexed 90°), obliques | Mid-elbow joint — about ¾ inch (2 cm) distal to the midpoint of a line between the epicondyles |
| Humerus | AP, lateral | Midpoint of the humerus |
Two patterns make the centering column memorable instead of seven separate facts. For the digits and hand, the CR goes to a joint named by the part — PIP for a finger, first MCP for the thumb, third MCP for the whole hand. For the long bones, it goes to the middle of the bone — mid-forearm, mid-humerus. The wrist and elbow are the two that need their own answer: midcarpal area, and the mid-elbow joint 2 cm distal to the interepicondylar midpoint.
| Projection | Correct positioning is evidenced by |
|---|---|
| PA hand | Equal concavity on both sides of the proximal metacarpal shafts, with near-equal spacing between them |
| PA oblique hand (45°) | Metacarpal midshafts do not overlap. Excessive overlap means over-rotation; too much separation means under-rotation |
| Lateral hand | Distal radius and ulna superimposed, and the metacarpals and phalanges of digits 2–5 superimposed into a single column; the thumb appears slightly oblique and completely free of the others |
| PA wrist | Equal concavity on each side of the proximal metacarpal shafts, and the distal radius and ulna separated (minimal overlap only at the distal radioulnar joint) |
| PA wrist — no deviation | The long axes of the third metacarpal and the midforearm are aligned. That alignment is what "neutral, without radial or ulnar deviation" looks like. Radial deviation instead foreshortens the scaphoid and drops its tuberosity next to the radius; ulnar deviation decreases the foreshortening |
| Lateral wrist | The ulnar head superimposed over the distal radius, and the proximal second through fifth metacarpals aligned and superimposed |
| AP forearm | Humeral epicondyles in profile, with the radial head, neck and tuberosity only slightly superimposed by the ulna |
| AP elbow | Both epicondyles in profile |
| Lateral (external) oblique elbow | Radial head, neck and tuberosity free of superimposition by the ulna; lateral epicondyle and capitulum elongated and in profile |
| Medial (internal) oblique elbow | Coronoid process of the ulna in profile, with the radial head and neck superimposed over and centered on the proximal ulna |
| Lateral elbow (90° flexion) | The three concentric arcs appear symmetrically aligned — (1) the trochlear sulcus, (2) the double-lined outer ridges of the capitulum and trochlea, and (3) the trochlear notch of the ulna. About half the radial head is superimposed by the coronoid process, the olecranon process is in profile, and the epicondyles superimpose |
| Scapular Y | The thin body of the scapula seen on end, without rib superimposition, with the acromion and coracoid as nearly symmetric upper limbs of the Y |
Why "epicondyles in profile" keeps appearing. It is the same test in three places — AP forearm, AP elbow, AP humerus. The epicondyles are the widest, most palpable landmark of the distal humerus, so making them equidistant from the IR is what "no rotation" physically means for the whole arm. If a question asks how you confirm an AP upper-limb projection wasn't rotated, that phrase is almost always the answer.
The registry tests upper-limb pathology and terminology alongside positioning, because the requisition often names the injury rather than the projection. These are the named ones, sorted by where they happen:
| Name | What it is |
|---|---|
| Skier's thumb | A sprain or tear of the ulnar collateral ligament of the thumb near the MCP joint, from the thumb being bent back — typically a fall on an outstretched hand |
| Bennett fracture | Fracture of the base of the first metacarpal, with the fracture line entering the carpometacarpal joint, generally with posterior dislocation or subluxation |
| Boxer fracture | Transverse fracture through a metacarpal neck, most commonly the fifth — from punching something |
| Baseball (mallet) fracture | Fracture of the distal phalanx from a ball striking the end of an extended finger. The DIP joint is partially flexed and there is frequently an avulsion at the posterior base of the distal phalanx |
| Colles fracture | Fracture of the distal radius with the distal fragment displaced posteriorly. Classically a forward fall onto an outstretched arm |
| Smith (reverse Colles) fracture | Same site, opposite direction — distal radial fragment displaced anteriorly. Classically a backward fall onto an outstretched arm |
| Barton fracture | Fracture and dislocation of the posterior lip of the distal radius, involving the wrist joint |
| Hutchinson (chauffeur) fracture | Intra-articular fracture of the radial styloid process |
| Monteggia fracture | Fracture of the proximal half of the ulna together with dislocation of the radial head — often from blocking a blow with the raised forearm |
The two pairs worth separating deliberately. Colles vs. Smith differ only in which way the distal radial fragment goes — posterior for Colles, anterior for Smith, which is why Smith is called the reverse. Bennett vs. boxer are both metacarpal, but Bennett is the base of the first (into the CMC joint) and boxer is the neck of the fifth. Opposite ends of the hand, opposite ends of the alphabet of metacarpals.
A bone age study asks how mature a child's skeleton is — for growth potential, and because malnutrition and endocrine and non-endocrine disorders can delay bone growth. The most common image evaluated is a single PA of the left hand and wrist, because the hand and wrist ossify on a fairly consistent schedule with many centers to read at once.
Two methods for turning that image into an age:
How to keep them apart: Greulich-Pyle matches a picture; Tanner-Whitehouse adds up a score.
Questions test why an eponym view is chosen, not just its name — e.g. picking Lawrence for a suspected shoulder dislocation would be actively contraindicated (it requires abduction), when Neer is the trauma-safe choice for exactly that scenario.
A second family of traps runs on direction, where both options are real and only one is right for the question asked: ulnar vs. radial deviation, external vs. internal shoulder rotation, cephalad vs. caudad on the clavicle, Colles vs. Smith, coracoid vs. acromion on a scapular Y, and "more angle" vs. "less angle" when a patient can't hyperextend for the carpal canal. In each pair, work from the underlying reason rather than trying to memorize which is which — deviation opens the far side, the rotation profiles the tubercle it swings outward, and lost patient motion is made up with added central-ray angle.
Source: Bontrager, Textbook of Radiographic Positioning and Related Anatomy, 10th ed., Ch. 4-5, with the fracture terminology from Ch. 15 and the bone age methods from Ch. 16; McQuillen Martensen, Radiographic Image Analysis, 6th ed., Ch. 4-5 for the lateral-wrist part position, the third-metacarpal/midforearm neutrality criterion, and the coracoid/acromion reading of a scapular Y dislocation. Named views cross-checked against the current ARRT spec's Attachment A view list — all confirmed present. Part positions, routine CR centering points and evaluation criteria (section 6) are taken from the per-projection "Part Position", "CR" and "Evaluation Criteria" entries rather than summarized from memory.
Ready for the module check?
5 questions, freshly drawn each attempt. 85%+ to pass. Retries are unlimited and immediate.