Guidelines for humeral subluxation cutoff values: a comparative study between conventional, reoriented, and three-dimensional computed tomography scans of healthy shoulders.

Abstract

The humeral subluxation index (HSI) is frequently assessed on computed tomography (CT) scans in conditions of the shoulder characterized by humeral displacement. An arbitrarily set HSI cutoff value of 45% for anterior subluxation and 55% for posterior subluxation has been widely accepted. We studied whether mean values and thresholds of humeral subluxation, in relation to the glenoid and scapula, were influenced by different imaging modalities. The HSIs referenced to the scapula (SHSI) and glenoid (GHSI) were compared between conventional CT scans, CT scans reoriented into the corresponding reference plane (ie, scapular plane for the SHSI and glenoid center plane for the GHSI), and 3-dimensional (3D) CT reconstructions of 120 healthy shoulders. The 95% normal range determined the cutoff values of humeral subluxation. The SHSI thresholds for conventional, reoriented, and 3D CT scans were 33%-61%, 44%-68%, and 49%-61%, respectively. A different mean SHSI was found for each imaging modality (conventional, 47%; reoriented, 56%; 3D, 55%; P ≤ .014), with the conventional SHSI showing an underestimation in 89% of the cases. GHSI thresholds for conventional, reoriented, and 3D CT scans were 40%-61%, 44%-56%, and 46%-54%, respectively. The mean GHSI did not differ between each imaging modality (conventional, 51%; reoriented, 50%; 3D, 50%; P = .146). The SHSI and GHSI are susceptible to different imaging modalities with consequently different cutoff values. The redefined HSI cutoff values guide physicians in the evaluation of humeral subluxation in conditions characterized by humeral displacement, depending on the available image data.