Zlatkin (1989) — Rotator Cuff Tears: Diagnostic Performance of MR Imaging.

Rotator Cuff Tears: Diagnostic Performance of Mr Imaging.

·Radiology·1989·265 citations·Shoulder & Elbow

This study tested how well MR imaging diagnoses suspected rotator cuff tears compared with contrast arthrography. Surgery served as the reference standard in 32 patients, with 8 asymptomatic volunteers added. A structured MR scoring system was developed to make interpretation more objective.

Study Snapshot

Design

Prospective diagnostic cohort

Blinding: Single-blind

Setting: Single academic hospital

Objective

Whether MR imaging outperforms arthrography for diagnosing rotator cuff tears using surgery as reference

Outcome(s)

MR sensitivity and specificity for rotator cuff tear vs surgery

Subjects

32 surgical patients + 8 asymptomatic volunteers

MR imaging (all 40 subjects)
Contrast arthrography (24 patients)

Inclusion

Clinical rotator cuff tendonopathy
Positive impingement test
Failed nonoperative treatment

Statistics

Sensitivity/specificityROC curve analysisLinear regressionLogistic regression

Key Findings

MR imaging outperformed arthrography for all tears: sensitivity 0.91 vs 0.71 and specificity 0.88 vs 0.71. With surgery as the reference standard, MR was the more accurate noninvasive test.
A weighted 7-point MR scoring system raised sensitivity to 1.0 and specificity to 0.92. The score combined tendon signal, tendon morphology, and the status of the subacromial-subdeltoid fat plane and bursae.
The score cleanly separated tear severity: full-thickness tears averaged 6.6 (range 5-7), partial tears 3.5 (range 3-4), and intact cuffs 1.3 (range 0-4). A cutoff score of 3 defined a tear.
All 17 full-thickness tears were correctly diagnosed on MR. The only misses were 2 partial-thickness tears of the superior surface, read as intact but abnormal tendons.
MR detected 3 small full-thickness tears (all under 1.5 cm) that were invisible on a normal arthrogram. This is the key advantage of imaging the tendon directly rather than relying on contrast leak.
MR-measured tear size matched surgical measurement almost exactly (r = .95), so MR gives the surgeon reliable preoperative sizing for planning the repair.
Arthrography carries real morbidity: 74% of patients report worse pain 24-48 hours after injection, a strong argument for a noninvasive alternative.

Board PearlMR imaging outperformed arthrography for rotator cuff tears (sensitivity 0.91 vs 0.71) and can detect small full-thickness tears missed on a normal arthrogram.

Clinical Relevance

When you see shoulder impingement pain that fails conservative care, MR is the noninvasive study that both confirms a tear and maps it for the surgeon. This paper helped move rotator cuff imaging away from invasive arthrography toward MR by showing higher sensitivity and specificity against a surgical reference standard.

Learn the grading logic. A tear requires an abnormal tendon (grade 2 or 3) PLUS loss of the subacromial-subdeltoid fat plane. Fluid in the bursa on T2 supports it. Remember the failure mode: partial tears of the superior (bursal) surface are the ones MR and arthrography both miss. Subtle tears are best seen on T1 and proton-density images, not T2.

MR sizing correlated with surgery at r = .95, which is why it guides repair choice (tendon-to-bone vs graft) and helps flag retracted, atrophic cuffs that may not be repairable. Interpret the near-perfect numbers with caution: including asymptomatic volunteers as controls inflates both sensitivity and specificity.

Rotator Cuff Tears: Diagnostic Performance of Mr Imaging.

Study Snapshot

Key Findings

Clinical Relevance

Related Articles

Rotator Cuff Tears: Diagnostic Performance of Mr Imaging.

Study Snapshot

Key Findings

Clinical Relevance

Related Articles

Rotator Cuff Tears: Diagnostic Performance of Mr Imaging.

Study Snapshot

Key Findings

Clinical Relevance

Related Articles

Rotator Cuff Tears: Diagnostic Performance of Mr Imaging.

Study Snapshot

Key Findings

Clinical Relevance

Related Articles