To assess the safety and preliminary efficacy of MR guided focused ultrasound (MRgFUS) as a non-invasive treatment modality for vascular malformations that are technically difficult to treat by conventional treatment modalities. All lesions were slow-flow vascular malformations that were difficult to localize on physical examination and lacked sonographically apparent blood vessels or cystic spaces, precluding percutaneous sclerosis/ethanol ablation or surgical resection. After review by an orthopedic surgeon and an interventional radiologist for treatment, cases were referred for consideration of treatment using MRgFUS. Tumors were selected for MRgFUS based on accessibility of the tumor to the focused ultrasound beam based on assessment of the pre-operative MRI; tumors also had to maintain a distance of 1cm from the skin and major nerves and have a volume less than 20mL. The volume limitation was chosen based on experience such that the treatment could be completed within 3 hours. Written informed consent was obtained from all subjects. Ablations were performed using the InSightec ExAblate 2100 MRgFUS device; this was an off-label application of the system. Sonication planning was performed manually by the treating physician (PG, with 4 years of experience using the MRgFUS system). Thermal dose maps were monitored to confirm heating of the lesion, and energy was adjusted to avoid tissue cavitation. Patients were evaluated prior to the procedure, on the day of treatment, and up to 12 months post-procedure. Pre-procedure and post-procedure MRIs were also obtained, as well as clinical assessment of pain and functional status before and after the procedure. Four patients were treated, three males and one female, with a median age of 33 years (range 18-54). Three lesions were located in the thigh and one in the calf. The median maximal lesion dimension was 1.4cm (range 1.1-4cm), with patients demonstrating a T2 hyperintense, heterogeneously enhancing mass on pre-procedure imaging. The ExAblate 2100 transducer was operated at 0.95-1MHz with an with an average of 1521 ± 578J median energy per sonication for each lesion (range 658-1878J) . Mean number of sonications was 42 (range 14-56). Average duration of individual sonications was 12.2sec (range 7.7-20sec). The averages of the values of temperatures achieved in the targeted tissue during individual sonications were 50°C (average) and 54°C (maximum). The average treatment time was 142min (range 64-202min). There was a median follow-up period of 7 months (range 4-12). No serious adverse events occurred. There was a significant reduction in maximum daily pain on a ten-point scale (8.8±1.5 to 2±2.3,P=0.020) with a mean reduction of 76% in maximal pain. Tumor volumes prior to treatment were 4.9±5mL (range 0.4-10.1mL). After treatment, the mean non-perfused volume (NPV) was 6.8±1.9mL (range 5.1-8.4mL). The mean ratio of the initial tumor volume to the non-perfused volume (NPVR) after ablation was 6±6.2 (range 0.5-13) There was a significant reduction in mean maximal lesion dimension after FUS (2.0±1.4cm to 0.5±1.1cm,P=0.003) with complete resolution of lesional enhancement in three patients and partial reduction in one patient. Two patients had complete resolution of their symptoms and of lesional enhancement (Fig.1-2). Two patients had partial pain relief. One of these patients had no residual tumor on MRI (Fig.3), but had muscle atrophy due to long-term disuse secondary to tumor-related pain. This pain is slowly improving with physical therapy to strengthen the leg. The second patient a symptom-free period that lasted 4 months, and then developed recurrent pain. MRI showed residual lesional enhancement, likely due to proximity of part of the tumor to neurovascular structures at the time of first treatment (Fig.4). This patient had an NPVR of 0.5; all other patients had an NPVR greater than 1. A second treatment is planned in this case to allow access to the remaining tumor by changing patient position. Soft tissue vascular malformations (VM) are the most common pediatric soft tissue tumors¹. These tumors are classified based on flow dynamics. For low-flow VMs, percutaneous sclerotherapy or surgery are standard treatment options, but each have an approximately 10% recurrence rate. Percutaneous sclerotherapy is limited to lesions that are sonographically visible. By combining the ability of MRI to visualize these lesions and of FUS to ablate them, MRgFUS appears to be a promising non-invasive treatment modality to treat vascular malformations^2,3. This pilot series selected tumors that were not amenable to other treatments, and is the largest demonstrating the potential of MRgFUS to treat vascular malformations. A larger sample with longer follow up is required to establish the safety and efficacy of this modality and to determine the specific types of lesions that should be selected for treatment with MRgFUS