Improving detection of Hill-Sachs fractures on radiographs.

OBJECTIVE

To investigate factors which affect radiographic diagnosis of Hill-Sachs fractures, and find criteria which improve detection.

MATERIALS AND

METHODS

Retrospective search was made for the term "Hill Sachs" within MRI reports in our local PACS system, and cases with post-reduction radiographs were included in the study. Prospective diagnoses and subspecialty MSK training of the interpreting radiologist of record were recorded. Images were then retrospectively reviewed by two observers and statistical analysis was performed.

RESULTS

Our retrospective study included 181 cases, of which 35% had prospective radiographic diagnosis of Hill-Sachs fracture. Retrospective review found that 73% of the radiograph series had at least 1 sign of a Hill-Sachs fracture. The internal rotation view showed a Hill-Sachs lesion in 59% of cases, but did not detect it in 14% of cases, where the lesion was instead visible on axillary, external rotation, and/or scapular Y view. Odds ratio of prospective Hill-Sachs detection on radiographs was 2.68 for musculoskeletal fellowship-trained radiologists versus non-musculoskeletal-trained radiologists.

CONCLUSION

Hill-Sachs fractures are often not recognized on post-reduction radiographs. Diagnosis of Hill-Sachs lesion can be significantly increased if radiologists are aware that the internal rotation view may fail to show the injury, and if all 4 views of a shoulder series are scrutinized.