On the positioning accuracy of a novel immobilization device

Purpose: The technical feasibility of simultaneous proton beam irradiation and magnetic resonance (MR) imaging has recently been demonstrated by phantom experiments performed on a 0.22 T open MR scanner that was integrated with a static proton beam line in a clinical environment [1]. Since the scanner and Faraday cage were mounted together on a mobile transport platform, the whole setup is movable to a pencil beam scanning beam line that allows for volumetric irradiation. In perspective, this setup may be used for first in-beam MR-guided proton therapy of patients with soft-tissue sarcoma of the forearm. However, the geometry of the Faraday cage and the MR scanner limits the patient’s freedom of positioning. Patients can only be treated in a seated position with the affected arm extended sideways into the imaging and irradiation field. However, the CT scan used for treatment planning is acquired in supine position with the arm extended over the head. An in-house developed prototype CT-MR compatible positioning aid was used for reproducible positioning of the forearm in the seated position at the in-beam MR scanner as well as in supine position at the CT scanner. The aim of the present study was to assess the accuracy of arm positioning using this novel immobilization device.
Materials and Method: Six healthy volunteers (4 male, 2 female) were included in this study, that was approved by the local ethics committee. To mimic the setup at the CT scanner, they first underwent a T1-weighted gradient echo MRI scan at a clinical 3T system in supine position while being positioned on a flat tabletop overlay with their forearm fixed in the device. Secondly, the 3T MR image was digitally merged with a 0.22 T MR image of fiducial markers which have a fixed position relative to the beam in the irradiation room and are always within the field of view at the 0.22 T system. The merged image is considered the reference image prescribing the position of the forearm in the irradiation position planned for the future and is the aim of the positioning process. Thirdly, the volunteers underwent a T1-weighted gradient echo MRI scan at the 0.22 T system in seated position with their arm extended sideways in the device. In-house developed image processing software was used to rigidly co-register the locations of the markers in the latter MR image to those in the reference image. The forearm was repositioned in the device not more than twice to match the position in the reference image. The residual positioning difference of the radial bone was a measure of the positioning accuracy. Additionally, the distance of the bones to the skin surface was assessed as a measure for deformation of the arm.
Results: In 5 out of 6 the volunteers the largest positioning error in one dimension was less than 3 mm, which is comparable to that currently achieved in the clinic. The average time needed for repositioning was approximately 20 min, which is substantially longer than in routine clinical practice. Furthermore, the arm underwent a deformation that changed the skin-surface-to-bone distance mostly by approximately 3 mm.
Lastly, the device was not tolerated well by 5 volunteers due to uncomfortable arm positioning.
Conclusions: The positioning accuracy of a prototype CT-MR compatible arm holder device for in-beam MR-guided proton therapy of patients with a soft-tissue sarcoma of the forearm was evaluated. The device enables the immobilization of the forearm both in supine and seated position, with a clinically acceptable repositioning accuracy. Adjustments of the prototype holder are required to improve the patient’s comfort and to avoid the deformation of the arm.