Evaluation of PC-ISO for customized , 3 D printed , gynecologic 192 Ir HDR brachytherapy applicators

San Francisco, CA 94115, USA; phone: (415) 353 7031; email: Adam.Cunha@ucsf.edu Evaluation of PC-ISO for customized, 3D printed, gynecologic 192Ir HDR brachytherapy applicators J. Adam M. Cunha,1a Katherine Mellis,1 Rajni Sethi,1 Timmy Siauw,1 Atchar Sudhyadhom,1 Animesh Garg,2 Ken Goldberg,1,2 I-Chow Hsu,1 and Jean Pouliot1 Radiation Oncology,1 University of California, San Francisco, CA; Electrical Engineering and Computer Science,2 University of California, Berkeley, CA, USA Adam.Cunha@ucsf.edu


I. INTRODUCTION
25 Gynecologic (GYN) brachytherapy applicators come in a variety of shapes and sizes to accommodate different patient scenarios. However, there is little opportunity to customize the shape of these applicators and their internal structure to the needs of each patient. As 30 a consequence, a fixed applicator might fit too loosely, which allows movement between scanning and treatment, and therefore increases dose uncertainty; it might fit too tightly, which can cause patient discomfort; or it might require extra interstitial catheters to ensure that dose 35 objectives can be met.
Rapid prototyping, or 3D printing, has the potential to address the customization limitation of current GYN brachytherapy applicators. With 3D printing, it is possible to construct conformal applicators with customized 40 channels [1]. There is currently a wide range of printing materials available for this purpose. However, to be suitable for clinical use, the material must be compatible with the brachytherapy workflow. Specifically, it must be biocompatible, sterilizable, free of CT scanning ar-45 tifacts, and have similar dose attenuation properties as water (the medium assumed by brachytherapy planning systems using the AAPM Task Group 43 formalism).
The purpose of this study is to evaluate PC-ISO (Stratasys, Eden Prairie, MN), a biocompatible, ther-50 moplastic, 3D printing material, for use in printing custom, single-use GYN brachytherapy applicators. Previous studies have shown that PC-ISO can be sterilized in multiple ways [2], including STERRAD (Advanced Sterilization Products, Irvine, CA), which is the preferred 55 sterilization method at our institution.
FIG. 1. 3D printing allows physicists and physicians to customize the size and shape of brachytherapy applicators to improve treatment. Shown above is a 3D printed applicator made of PC-ISO (white, top). This applicator was designed to fit a patient with a very wide vaginal canal, which was too large for the largest commercial applicator of the same type at our clinic (yellow, bottom).
This study evaluates the radiation properties of PC-ISO as a material for customized, single-use, GYN brachytherapy applicators. The evaluation is made using comparisons of CT scans, dose-depth curves for PC-ISO 60 and water, and using geometry that is within the scope of a typical clinical procedure. Although this study focuses on evaluating PC-ISO, the same tests can be used to evaluate other materials for brachytherapy. Figure 1 shows an example of a customized GYN cylinder appli-FIG. 2. A set of testing apparati were designed and 3D printed for this study to measure the depth-dose curve for 192 Ir. The apparati held a piece of radiochromic film and an endobronchial brachytherapy catheter. The left picture shows the testing apparatus printed in PC-ISO, and the picture on the right shows control apparatus used to suspend a piece of radiochromic film in water. The apparati were scanned in a helical CT to compute the Hounsfield unit distribution.
cator printed in PC-ISO next to a commercial applicator of the same type.

75
Fortus 400mc (Stratasys) used in this study has a resolution for PC-ISO of 0.178 mm.
Manufacturers have supported medical interests in 3D printing by introducing printing materials that pass the 85 International Standard ISO-10993 as well as the United States Pharmacopeia standards for biocompatibility [26]. PC-ISO is both USP Class VI approved and ISO-10993-1 rated. The material is also sterilizable and has high flexural and tensile strength properties that have made it 90 a common choice for many medical applications [27][28][29]. For example, PC-ISO has been explored for use in anklefoot orthoses [14], lumbar cages [30], and bone screw linking devices [31].

95
To evaluate PC-ISO, a custom testing apparatus was designed in CAD. This apparatus is shown in Figure 2 (left). The apparatus consisted of a pair of identical Lshaped blocks designed to snap together. Each block contained a single, straight channel 0.2 cm in diame-100 ter, which tightly held a 6F endobronchial brachytherapy catheter (Nucletron, Sunnyvale, CA). When snapped together, the blocks held a 3 cm by 6 cm radiochromic film segment in a 6 cm long shallow gap between the blocks. The blocks were 1 cm thick on each side of the film to 105 provide side scatter on the scale of a typical brachytherapy applicator radius. The 6 cm side of the film was radial to the channel, and 3 cm side of the film was 0.25 cm away from the central axis of the channel.
The choice of a 3 cm diameter was used because it is a 110 characteristic dimension of the typical cylinder applicator used in the clinic. The geometry of the apparatus was designed to be relevant to the geometry of applicators used in a clinical setting since the focus of this study was on the validation of PC-ISO for use in a brachytherapy 115 clinical setting. A nearly identical control apparatus was designed to leave most of the surface area of the film exposed. This apparatus was used to perform a control experiment in water. This apparatus is also shown in Figure 2 (right).

120
The testing apparatus was printed in PC-ISO using a Fortus 400mc (Stratasys), and the control apparatus was printed in ABS plastic using a uPrint Plus (3D Systems, San Francisco, CA). The stereolithography (STL) files for the testing and control blocks are available from the 125 authors upon request.
A size 6 French endobronchial brachytherapy catheter was placed in the testing apparatus. The opposite channel was left empty. There was 3 cm of channel length inside the apparatus, which allowed for 13 dwell positions 130 spaced 0.25 mm apart. Figure 3 shows the experimental setup. A dose plan with a cylindrical dose distribution was designed with equal dwell time at each of the 13 positions. The time was normalized to deliver 1500 cGy at 1 cm from the center of the channel in water. Figure 4   135 shows the isodose lines in Oncentra (Nucletron, Sunnyvale, CA) within the apparatus as would result from the designed dose plan with a 192 Ir source.
A 3 cm × 6 cm radiochromic film segment was placed between the two halves of the apparatus and the halves 140 were snapped together. The entire apparatus was submerged in water, and the dose plan was delivered to the film using a microSelectron V2 digital afterloader. The same test was repeated on the control apparatus directly afterwards. The films were allowed to develop for 24 145 hours after exposure. Then they were processed to find the dose-depth curve.
A helical CT was used to scan the PC-ISO testing apparatus, a Nucletron cylinder applicator routinely used in the clinic, and a cup of water. The distribution of 150 Hounsfield units (HU) was extracted for each from the FIG. 3. The test apparatus with a size 6 French endobronchial brachytherapy catheter was inserted into one of the end channels. The PC-ISO and control apparati were suspended in water before the dose was delivered.
FIG. 4. A dose plan was to deliver 1500 cGy at 1 cm from the center of the catheter channel in water. The dwell positions and radial dose distribution for the radiochromic film study are represented using the TG-43 dose calculation formalism in water. The films were developed for 24 hours after exposure before they were scanned to find the dose-depth curve.

DICOM RT files.
FIG. 5. Shown above is the distribution of Hounsfield units (HU) inside the PC-ISO apparatus (red), a Nucletron cylinder applicator (green), and a cup of water (blue). The mean was −1 HU for water, −10 HU for PC-ISO, and +524 for the Nucletron applicator. The mean HU value for the PC-ISO testing apparatus is closer to water than air (−1000 HU) or bone (+1000 HU) and more water equivalent than the Nucletron cylinder.

IV. RESULTS
There were no visible CT artifacts inside the testing apparatus. The distribution of the Hounsfield units (HU) 155 inside the apparatus is shown in Figure 5. The mean Hounsfield unit was −1 HU for water and −10 HU for PC-ISO. This mean HU value is closer to water than air (−1000 HU) or bone (+1000 HU). The mean HU for the testing apparatus was more equivalent to water than the 160 mean HU for the Nucletron cylinder applicator, which had a mean of +524 HU.
The percent dose depth for the testing apparatus (PC-ISO) and the control apparatus (water) is shown in Figure 6. The two curves are within 1% of each other be-165 tween 1 cm and 6 cm from the channel. Doses closer than 1 cm were excluded because that region of the film was over-saturated.

V. DISCUSSION
To be compatible with current dose planning sys-170 tems, PC-ISO should be radiologically equivalent to water within the energy range of interest, which for 192 Ir is approximately 10 2 keV with an average gamma emission energy of 380 keV. The results showed a 1% difference in dose attenuation over the range of interest, which for 175 brachytherapy is not a significant source of error compared to other sources of error such as catheter movement and contouring uncertainty. The dose attenuation results are corroborated by the HU distribution, which did not show any very-high density regions in the printed 180 medium that could effect the dose attenuation in an un-FIG. 6. The percent dose depth from the radiochromic film test for the PC-ISO testing apparatus (red) and the control apparatus (blue), which is water equivalent. The two curves were within 1% of each other between 1 cm and 6 cm, showing that the TG-43 planning system, which assumes a water medium, can be used as normal.
expected way. The spread seen for PC-ISO ( Figure 5) is due to the honeycomb internal structure characteristic of 3D printing, which creates small regular-patterned regions of higher (material) and lower (air) density. The evaluation of PC-ISO's radiation properties along with previous studies of its mechanical properties and sterilization [14,[26][27][28][29][30][31] made us confident that it was suitable for clinical use.
Since the conclusion of these tests, we have offered cus-190 tomized PC-ISO to patients in cases where the physician felt it would improve their treatment. We printed a 2.75 cm diameter segmented cylinder, which is between two standard sizes (2.5 cm and 3.0 cm) from our vendor. We printed a 2.75 and a 3.5 cm diameter solid cylindri-195 cal applicator with internal channels. These applicators were custom built for each patient from measurements taken during examination. All PC-ISO applicators were sterilized using the STERRAD procedure before implantation. A CT scan of a custom-built PC-ISO cylinder 200 applicator implanted in a patient prior to treatment is shown in Figure 7. The PC-ISO applicator is contoured on the scan but is difficult to see because of its near water-equivalency. The Nucletron tandem used during the procedure inserted in the center of the printed cylin-205 der is clearly visible at the center of the applicator.
PC-ISO applicators are virtually tissue-equivalent under CT scan -more tissue-equivalent than commercial applicators at our clinic. This level of tissue equivalence can make it difficult to find the boundary of the appli-210 cator during segmentation, especially at the tip of the applicator where the surface is curved. To address this issue, it may be possible to cover the applicators in a radio-opaque dye and condom before insertion. FIG. 7. After the conclusion of the tests outlined in this study, we printed several GYN applicators in PC-ISO when a custom built applicator would improve treatment options. The applicators were designed specifically for each patient from measurements taken during examination and were sterilized using the STERRAD procedure before implantation and treatment. A segmented PC-ISO cylinder applicator with a custom 2.75 cm diameter is shown implanted and contoured in the patient during the dose planning process in axial (top left), sagittal (top right), and coronal (bottom left) view. The Nucletron tandem (bright white) is approximately 1/3 the diameter of the full applicator. Bottom right is a 3D representation of the contoured organs, applicator (blue), and tandem (purple) with the catheter visible.

215
PC-ISO is a readily available material for 3D printing that is FDA approved for temporary implants. In this study it was evaluated for use in a brachytherapy environment. It was shown that PC-ISO has sufficiently equivalent dose attenuation properties to water at 192 Ir 220 energies to be compatible with the brachytherapy planning system and workflow. It also does not produce CT artifacts. Given these results, we printed several customized cylinders and used these cylinders on patients when it would improve their treatment. While 3D print-225 ers with the capability to print in FDA-approved materials are currently on the order of $100,000, clinics wishing to implement this technique can outsource the printing of their customized designs to printing vendors.

230
We thank the Qualcomm Undergraduate Experiences in Science and Technology (QUEST) program for providing funding and resources for this project. We also thank Serena Scott, Ph. D, for her help designing the applicators.