The relationship between the humeral head and the anterolateral acromion was compared for subjects with and without full-thickness rotator cuff tears, as indicated by magnetic resonance imaging, using a case-control design. "Uncoverage" of the humeral head by the anterolateral acromion was defined as the percentage of the humeral head width not covered superiorly by the anterolateral acromion, as measured digitally on magnetic resonance imaging scans of the shoulder in the oblique coronal plane. The uncoverage of the humeral head by the anterolateral acromion was 32.0% for men > or = 40 years of age with normal rotator cuffs (n = 39) and 38.2% for men > or = 40 years of age with full-thickness rotator cuff tears (n = 43). The difference in uncoverage of 6.2% was statistically significant (p < 0.001). The uncoverage of the humeral head by the anterolateral acromion was 36.1% for women > or = 40 years of age with normal rotator cuffs (n = 10) and 41.8% for women > or = 40 years of age with full-thickness rotator cuff tears (n = 17). The difference in uncoverage of 5.72% was not statistically significant with the number of subjects available (p = 0.108). We concluded that increased uncoverage of the humeral head by the anterolateral acromion may be a factor in the pathogenesis of full-thickness rotator cuff tears.