3D-printed Magnetic Resonance (MR)-based gynecological phantom for image-guided brachytherapy training

PURPOSE/OBJECTIVES: There is a clinical need to develop anatomic phantoms for simulation-based learning in gynecological brachytherapy. Here, we provide a step-by-step approach to build a life-sized gynecological training phantom based on magnetic resonance imaging (MRI) of an individual patient. Our hypothesis is that this phantom can generate convincing ultrasound (US) images that are similar to patient scans. METHODS: Organs-at-risk were manually segmented using patient scans (MRI). The gynecological phantom was constructed using positive molds from 3D printing and polyvinyl chloride (PVC) plastisol. Tissue texture/acoustic properties were simulated using different plastic softener/hardener ratios and microbead densities. Nine readers (residents) were asked to evaluate 10 cases (1 ultrasound image per case) and categorize each as a "patient" or "phantom" image. To evaluate whether the phantom and patient images were equivalent, we used a multireader, multicase equivalence study design with two composite null hypotheses with proportion (pr) at H-01 : p(r) <= 0.35 and H-02: p(r) <= 0.65. Readerswere alsoaskedto review US videos and identify the insertion of an interstitial needle into the pelvic phantom. Computed Tomography (CT) and magnetic resonance (MR) images of the phantom were acquired for a feasibility study. RESULTS: Readers correctly classified "patient" and "phantom" scans at p(r) = 53.3% +/- 6.2% (p values 0.013 for H-01 and 0.054 for H-02, df = 5.96). Readers reviewed US videos and identified the interstitial needle 100% of the time in transabdominal view, and 78% in transrectal view. The phantom was CT and MR safe. CONCLUSIONS: We have outlined a manufacturing process to create a life-sized, gynecological phantom that is compatible with multi-modality imaging and can be used to simulate clinical scenarios in image-guided brachytherapy procedures. (c) 2022 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved.