Advanced imaging often is obtained in children suspected of having osteoid osteoma. We hypothesized that magnetic resonance imaging gives a falsely aggressive appearance and that computed tomography is better for identifying osteoid osteoma. This is the first prospective blinded study comparatively assessing these imaging studies in children. Twelve preoperative magnetic resonance imaging scans of confirmed cases of osteoid osteoma were collected. Three radiologists blinded to the diagnosis reviewed these images. Only a clinical history of skeletal pain was given. Lesions were classified as benign-latent, benign-aggressive, or malignant. Radiologists were asked to list their primary diagnosis (with a confidence level from 1-10). Seven of these 12 children also had computed tomography scans that subsequently were reviewed in similar manner. With computed tomography scans, lesions were accurately identified as benign-latent (15/21 readings, 71%) and as osteoid osteoma (14/21 readings) more frequently than with magnetic resonance imaging scans (7/36 readings, 19%). LEVEL OF EVIDENCE: Diagnostic study, Level III-1 (study of nonconsecutive patients--no consistently applied reference gold standard). See the Guidelines for Authors for a complete description of levels of evidence.