Full-thickness supraspinatus tendon tears: correlation of findings by arthroscopy and magnetic resonance imaging.

PURPOSE: The main aim of this study was to correlate measurements of the width and retraction of isolated full-thickness supraspinatus tendon tears determined by magnetic resonance imaging (MRI) with measurements recorded by arthroscopy using a continuous millimetre scale. METHODS: A total of 53 individuals with isolated supraspinatus tears and retraction < 30 mm underwent arthroscopy at our centre. Tear width and retraction measured by arthroscopy (reference standard) and MRI (index test) on a continuous millimetre scale were compared. All measurements were performed by a single radiologist blinded to intra-operative findings and one surgeon who had previous access to MRI results. RESULTS: The average tear retraction was 12.60 ± 4.89 mm by arthroscopy and 16.81 ± 6.29 mm by MRI. Those measures exhibited moderate correlation (r = 0.643, p < 0.001) and an average difference of 4.21 mm (p < 0.001). Average tear width was 12.87 ± 4.15 mm by arthroscopy and 14.19 ± 5.20 by MRI. Those measures exhibited moderate correlation (r = 0.526, p < 0.001) and an average difference of 1.32 mm, which was not statistically significant (p = 0.109). CONCLUSION: Measures of retraction and width obtained by MRI and arthroscopy exhibited moderate correlation in small- or medium-sized supraspinatus tears.