Thyroid Ultrasound: Change of Inter-observer Variability and Diagnostic Performance after Training

ABSTRACT

PURPOSE: To investigate and compare inter-observer variability and diagnostic performance of thyroid ultrasound (US) between a faculty member and observing residents. MATERIALS AND METHODS: From October 2007 to June 2009, 18 residents underwent training in thyroid US section. Group 1 included 8 residents that were trained for the first time and group 2 included 10 residents that were trained for the second time. US features of nodules were recoded according to the composition, echogenicity, margin, calcifications, shape, and final assessment by a faculty member and residents, respectively. Following a discussion, a faculty member performed fine needle aspiration. Then, the inter-observer variability and diagnostic performance between a faculty member and residents were investigated and compared for US. RESULTS: In group 1, agreement for composition in resident 1, calcification for residents 5 and 6, and shape for resident 4 were slight, moderate, moderate, and moderate, respectively. In group 2, agreement for composition in residents 1 and 10 were moderate. Substantial or greater agreement was observed more frequently in group 2 than 1. The diagnostic performances for both the faculty and residents were high and not statistically different. CONCLUSION: Agreement for US features between a faculty and residents as well as diagnostic performance were high. Moreover, diagnostic performance of residents that underwent training a second time of training was higher than residents that underwent training only once.