Background 3D (three-dimensional) US (ultrasound) of the shoulder seems to be as accurate as 2D (two-dimensional) US when compared with MRI and surgical findings for the diagnosis of full- and partial-thickness rotator cuff tears. However, 3D ultrasound transducer acquired images have poorer inter- and intra-rater reliability for the assessment of supraspinate tendon partial tears compared to 2D images.

Objectives The aim of this study was to determine the intra- and inter-reader reliability in the analysis of conventional 2D and of 3D acquired ultrasound images in the detection of different tear types of the supraspinate tendon performed by rheumatologists with different experience level in musculoskeletal ultrasound.

Methods Non-interventional prospective observational pilot study of 2309 images (210 3D image sets) of 127 adult patients suffering from unilateral shoulder pain. 2D and 3D images were scored by three readers (one fellow sonographer EULAR intermediat level = reader B, one advanced sonographer EULAR Teacher/EFSUMB Level 2 = reader C and one advanced sonographer EULAR Teacher/EFSUMB Level 3 = reader A) independently as partial-thickness teras (bursal sided partial-thickness tear, intrasubstance partial-thickness tear, articular sided partial-thickness tear), full-thickness tear and no tear. The intra-and inter-rater reliability were calculated.

Results There was an excellent intra-rater agreement of reader A in the overall classification of supraspinate tendon tears (2D vs 3D images κ =0.892, pairwise agreement 93.81%, 3D scoring round 1 vs 3D scoring round 2 κ =0.875, pairwise agreement 92.857%). The inter-rater agreement was only moderate compared to reader C on 3D images (κ =0.497, pairwise agreement 70.95%) and fair compared to reader B (κ =0.238, pairwise agreement 42.38%). There was however excellent overall agreement in the classification of full thickness tears between the 2D and 3D images (κ =0.810) and overall good agreement in the classification of partial thickness tears (κ =0.667).

Conclusions The use of 3D ultrasound in the assessment of supraspinate tendon tears has an excellent intra-rater agreement for an advanced sonographer and an excellent inter-rater agreement in the detection of full thickness tears. We found a good inter-rater agreement for an advanced sonographer and a fair to moderate inter-rater agreement for a less experienced sonographer. The main reason for discordant results were articular sided partial thickness tears at the footprint where anisotropy makes interpretation difficult. Therefore we emphasise to consider the anisotropy at this special anatomical level and special care in probe positioning while taking the 3D scans. Nevertheless 2D evaluation has the advantage of additional information (e.g. sonopalpation, dynamic examination) in the differentiation of articular sided partial thickness tears versus tendinosis or anisotropy.

References

1. Hayter CL et al, J Ultrasound Med. 2012 Mar;31(3):449-53.

2. Co S et al, Can Assoc Radiol J. 2012 Aug;63(3):170-6.

3. Kang CH et al, Skeletal Radiol. 2009 Nov;38(11):1063-9

Disclosure of Interest None declared