Background MR-guided high-intensity focused ultrasound is a noninvasive treatment modality that uses focused ultrasound waves to thermally ablate tumors within the human body while minimizing side effects to surrounding healthy tissues. This technology is FDAapproved for certain tumors and has potential to be a noninvasive treatment option for extremity soft tissue tumors. Development of treatment modalities that achieve tumor control, decrease morbidity, or both might be of great benefit for patients. We wanted to assess the potential use of this technology in the treatment of extremity desmoid tumors. Questions/purposes (1) Can we use MR-guided highintensity focused ultrasound to accurately ablate a predetermined target volume within a human cadaver extremity? (2) Does MR-guided high-intensity focused ultrasound treatment stop progression and/or cause regression of extremity desmoid tumors? Methods Simulated tumor volumes in four human cadavers, created by using plastic markers, were ablated using a commercially available focused ultrasound system. Accuracy was determined in accordance with the International Organization of Standards location error by measuring the farthest distance between the ablated tissue and the plane corresponding to the target. Between 2012 and 2014, we treated nine patients with desmoid tumors using focused ultrasound ablation. Indications for this were tumor-related symptoms or failure of conventional treatment. Of those, five of them were available for MRI followup at 12 months or longer (mean, 18.2 months; range, 12-23 months). The radiographic and clinical outcomes of five patients who had desmoid tumors treated with focused ultrasound were prospectively recorded. Patients were assessed preoperatively with MRI and followed at routine intervals after treatment with MRI scans and clinical examination. Results The ablation accuracy for the four cadaver extremities was 5 mm, 3 mm, 8 mm, and 8 mm. Four patients' tumors became smaller after treatment and one patient has slight progression at the time of last followup. The mean decrease in tumor size determined by MRI measurements was 36% (95% confidence interval, 7%-66%). No patient has received additional adjuvant systemic or local treatment. Treatment-related adverse events included first-and second-degree skin burns occurring in