Simple phantom fabrication for MRI‐based HDR brachytherapy applicator commissioning

Abstract A new high dose rate (HDR) brachytherapy program was initiated in a community hospital setting, with the goal of using magnetic resonance (MR) images with the implant in place during the planning process. Physics acceptance testing and commissioning was completed for key program components, including multiple applicators. To image new applicators for MRI‐based planning prior to use with patients, agar gel doped with copper sulfate was created using simple, MR‐safe household materials as a practical and inexpensive alternative to custom‐machined precision phantoms. Applicators in‐phantom were scanned in a 1.5 T MRI scanner using the same sequences developed for the brachytherapy program, then rigidly registered to high‐resolution computed tomography (CT) images to assess distortion, artifact, and geometric displacement. To date, Varian tandem and ring sets, segmented cylinders, cervical probes, endometrial applicators; and third‐party plastic needles, tandems, and vaginal guides have been imaged in phantom and are available for use clinically.


| INTRODUCTION
Currently, magnetic resonance imaging (MRI) represents the gold standard for target delineation in image-guided high dose rate (HDR) brachytherapy. 1 Implementing MRI in an HDR brachytherapy program depends on available resources, and three broad workflow categories of MRI's use in planning may be defined: MRI informed, MRI based, and MRI guided. 2 Here, MRI-informed brachytherapy describes the use of previously acquired MR images to inform optimal applicator placement at the time of implant. The MRI does not include the implant. MRI-based brachytherapy describes the use of MRI for planning after the implant is in place, with planning either based solely on the MR images, or based on MRI that is registered to a computed tomography (CT) dataset. MRI-guided brachytherapy describes the use of real-time MR image guidance to place the brachytherapy implant intraoperatively. Based on available resources and because of the potential for substantial tissue distortion due to the implant itself, MRI-based brachytherapy was considered preferable for this institution.
Careful commissioning of applicators is essential for safe and effective HDR treatment. [3][4][5] Part of this process is three-dimensional (3D) image acquisition. Previous work has described commissioning performed in preparation for MRI-based and MRI-guided HDR programs, including phantom measurements. This has been done using water phantoms: for example, Owrangi et al. 6   describe precision-machined acrylic phantoms designed to suspend needles, tandems, rings, and ovoid applicators in gel or a copper sulfate solution for MR imaging. These custom phantoms were developed and machined by those groups to include an internal stereotactic coordinate system. For this work, it was desired to fabricate simple, MR-safe, inexpensive gel phantoms for scanning in both CT and MRI without the time delay or expense of designing and manufacturing precision-machined custom acrylic phantoms.
Although simple phantoms lack the internal stereotactic coordinate system of more advanced phantom designs, rigid registration with high-resolution CT data provides a method to review rigid geometry between the imaging modalities. Inexpensive plastic storage containers (Sterilite Corporation, Townsend, MA) were obtained for the phantom bases, and a mold room electron block foam cutter was used to create custom foam inserts, cut to the shape of individual applicators and to fit snugly into the phantom bases. The foam inserts were affixed within the plastic containers, and were used to suspend the applicators within the central volume of the phantoms. Gels from agar, and agar's purified form, agarose, have been described in the literature for use in MRI phantoms for their tissue-mimicking properties. 8,11 Additives such as gadolinium chloride or nickel chloride also may be used to adjust relaxation properties of gels. 8  The agar-agar powder and water were brought to a boil on a stove top over high heat, with constant stirring. Once boiling, the mixture was removed from the heat, and the CuSO 4 solution was added.

| MATERIALS AND METHODS
The gel was cooled to approximately 48°C, stirring occasionally, to achieve a pourable and homogeneous consistency that would be cool enough not to damage the applicators. Once sufficiently cool but before the gel had set, the warm gel was poured into the prepared phantom bases around the preset applicators and allowed to solidify completely. For phantoms including needles, the gel was poured and allowed to cool completely in the phantom base, and then needles were inserted into the solid phantoms. Needles inserted multiple times resulted in extraneous voids in the gel. These voids remained visible in imaging studies, so care should be taken to insert needles only once. Photographs of one of the phantoms are included in Fig. 2.
F I G . 1. Simple supplies used for phantom construction. Dilute bleach was used to sterilize the surfaces of all components prior to fabrication, and a solution of agar-agar powder, distilled water, and copper sulfate were used for the gel.

| RESULTS AND DISCUSSION
The MR images were rigidly registered with the CT datasets, where the high-resolution CT was considered the benchmark for geometric  these applicators will be used within a workflow that includes both MRI for target delineation and CT for applicator reconstruction and organ-at-risk segmentation. With these needs in mind, the rigid registration of the datasets was considered sufficient.
A practical item to note regarding the timing between phantom construction and scanning is that agar gel has a limited shelf life once prepared. Literature suggests including toxic additives to the gel such as sodium azide may be helpful for slowing mold formation (although such additives may affect MR relaxation characteristics), 8 but for this study, shelf life was considered adequate after disinfecting phantom components and surfaces prior to fabrication and keeping the phantoms covered when not in use. Furthermore, some applicators may have a specified maximum period of use (e.g., the Varian instructions for use documentation for the universal segmented cylinder set indicates that the applicators are intended for use for less than 30 days of contact with patients).
Therefore, it is recommended to minimize the overall time between phantom fabrication, scanning, and final retrieval of applicators from within the phantom. Decreasing the time between CT and MRI scanning will also serve to minimize the likelihood of inadvertently disturbing the position of the applicators within the phantom between scans.
When using plastic wrap to protect reusable applicator components, air will be introduced between the applicator and the plastic wrap and is expected to be visible in phantom. This is noticeable in

CONFLI CT OF INTEREST
There is no relevant conflict of interest to disclose.