Objectives. To measure the national practice variations in imaging studies performed for men newly diagnosed with clinically localized prostate cancer.

Methods. We created an analytic file from 1991 to 1996 Medicare claims data using files for a random sample of 5% of all Medicare beneficiaries. Among men with newly diagnosed clinically localized prostate cancer, we identified those undergoing staging bone scans, staging computed tomography (CT), or staging magnetic resonance imaging (MRI) at the time of diagnosis. We conducted univariate and multivariate analyses adjusting for Charlson index score, age group, race, geographic region, and year of diagnosis.

Results. In all geographic regions, men receiving radiation therapy (RT) were more likely than those receiving radical prostatectomy (RP) to undergo CT. In the South, RT patients were significantly more likely than RP patients to undergo MRI and bone scans. In the West, RT patients were significantly more likely than RP patients to have bone scans. In multivariate analyses that controlled for all significant univariate findings, treatment with RT significantly predicted for the use of bone scans (odds ratio 1.24, 95% confidence interval 1.17 to 1.31), CT scans (odds ratio 3.26, 95% confidence interval 3.18 to 3.34), and MRI scans (odds ratio 1.47, 95% confidence interval 1.23 to 1.72). Regional differences in the use of imaging technologies for staging persisted in the multivariate analysis.

Conclusions. Patients undergoing RT for clinically localized prostate cancer undergo more bone, CT, and MRI scans than do patients undergoing RP, regardless of comorbidity, age, or race. In addition, a significant geographic variation was found in the use of these diagnostic tests. These variations suggest that evidence-based staging guidelines have not been met with broad physician acceptance.