Reproducibility Audit of Directional Descriptions
A teleradiology quality reviewer audits 30 reports for directional accuracy and asks a new hire to sit in. Three excerpts are questioned: (1) "A 1.2-cm nodule is seen in the right lung, located on the left side of the displayed image." (2) "The laceration lies 3 cm proximal to the left wrist on the volar surface, extending medially." (3) "The patient was positioned prone with the head rotated to the left for the lumbar puncture; the entry site was marked 2 cm inferior to the L4 spinous process." The standard is that a second clinician must be able to reproduce the location from the words alone, without seeing the image or the patient.
1. Is excerpt (1) an error that should be sent back for correction?
Read the worked answer
No - not on the basis of the screen side alone. Send back only if the stated side conflicts with the image.
- Standard display convention for axial and coronal images viewed from the front: the left side of the screen is the patient's right. So "right lung, left side of the image" is internally consistent, not contradictory.
- The report passes the reproducibility test because it names the patient's anatomic side, which is what a clinician acts on; the screen-side clause is redundant but harmless.
- Where this breaks: sagittal images, or any image displayed without standard orientation (rare, but flagged). In those cases screen side and patient side cannot be assumed, and the reviewer should rely on the named side and require confirmation of orientation.
Rule to teach: describe the patient, not the monitor. A report that says only "left side of the image" is the one to reject - it cannot be acted on.
2. Translate excerpt (2) into an unambiguous location, and state the reference point that makes 'proximal' meaningful.
Read the worked answer
Location: palm-side (anterior) surface of the left forearm, 3 cm toward the elbow from the wrist crease, with the laceration running toward the ulnar (midline) side.
- proximal = nearer the point of attachment of the limb to the trunk. On the forearm, proximal is toward the elbow; distal is toward the fingers. So a proximal marker 3 cm from the wrist sits between the wrist and the elbow, not on the hand.
- volar = palm side; on the forearm this is the anterior surface. Its opposite here is dorsal.
- medial = toward the midline. In anatomic position the palms face forward, so the medial border of the forearm is the ulnar side - the little-finger side when the arm hangs at the side.
- Trap to avoid: medial is a body-relative term, not "toward the other arm as it lies on the table." If the arm is rotated, the words still mean the ulnar side.
Every directional term needs its reference frame (anatomic position, limb, or landmark). "Proximal to the wrist" works only because the wrist is named as the landmark.
3. For excerpt (3): what does prone-with-head-rotated-left actually look like, and is the inferior marker precise enough?
Read the worked answer
Prone = lying face down. Head rotated to the left = the left side of the face is against the table; the right ear points up. The spine is posterior/superior in this position.
- Consequence for documentation: any facial or anterior-chest finding is pressed against the table and easy to miss; the reviewer should ensure laterality statements match what was visible (right side of the face up).
- inferior = toward the feet. "2 cm inferior to the L4 spinous process" means 2 cm caudad along the midline, which is reproducible - but note that L4 is a landmark tied to the iliac crest (Tuffier's line), and in a prone patient the crest is a firmer landmark than a spinous process that may be obscured.
- Verdict: acceptable and reproducible; a stronger entry would name both the spinous process and the iliac crest level. Reject only if the note says "above/below" without a landmark, or uses left/right without saying whether that is the patient's or the observer's.