Long head biceps tendon instability: diagnostic performance of known and new MRI diagnostic signs

Objective: The aim of this work was to evaluate the inter-observer agreement and diagnostic performance values of 7 MRI signs (3 known and 4 new) of long head biceps tendon instability. Materials and methods: MRI of 86 patients were retrospectively evaluated. Inter-observer agreement and diagnostic performance of each diagnostic sign and of all combined signs (with the exception of detour sign) were tested for identification of biceps tendon instability, with arthroscopy as the reference standard. Results: Agreement between expert operators was moderate to good. Sensitivity, specificity, and sign accuracy showed a variation respectively of the following: humeral chondral print 72–80%, 62–68%, and 70–76%; subchondral bone edema 24–31%, 84–97%, and 50%; biceps tendon angle 60–71%, 97–100%, and 74–81%, biceps tendon-groove distance 31–47%, 90–100%, and 56–64%; long head biceps subluxation/dislocation on axial plane 49–53%, 97–100%, and 66–70%; displacement sign 74–80%, 74–100%, and 74–87%; detour sign 51–64%, 58–81%, and 62–64%; and all signs 98–100%, 32–61%, and 75–86%. Conclusion: These diagnostic signs, both known and new, individually and/or in combination, provide a valid tool in the MRI diagnosis of long head biceps tendon instability.