Quality Assurance Tests for 3D Printed Bolus Used in Radiation Therapy

Abstract

3D printing continues to present innovative applications in radiotherapy (RT). One such application is the creation of 3D printed bolus for RT treatments. However, the increased accuracy provided by these 3D prints requires additional scrutiny with respect to their quality and fit. To date, there exists no widely available set of quality assurance (QA) tests for these 3D printed devices, which are seeing increased use in RT. Therefore, this study looks at formulating a comprehensive set of tests for the quality of 3D printed bolus, thus ensuring they perform as intended in RT treatments. QA was performed on seven 3D printed boluses used for various treatment sites. The shape, bulk density, and defect characteristics were used to assess 3D print quality. For each bolus, a high resolution (0.8 mm slice thickness) CT scan was acquired. The scan was imported into the planning system, where an external contour was defined. For the bulk density test, the mean Hounsfield Unit (HU) of a volume of interest (VOI) was compared with the expected HU (HUexpected). This VOI was cropped 1 mm inwards from the external contour to avoid any partial volume effects. The max and min HUs from the VOI were used to identify print defects, paying particular attention to min HU < 0 and max HU > HUexpected. The shape of the print was assessed by comparing the external contour with the original mesh used to create the 3D print. After registering the two structure sets, the congruence in shape was quantified through calculating Hausdorff distances between structure sets, and evaluating the resulting median, min, and max errors. For bulk density, the set of 3D printed boluses yielded a mean HU of 127 which, on average, was 7 HUs higher than the HUexpectedof 120. For 3D print defects, the (max, min) HUs from the set of 3D prints was (211, -300). The max HU > HUexpected is most likely due to the presence of noise. HU values < 0 (e.g. min HU) were typically observed along the edge of the 3D print, where the extruder is changing between fill patterns, causing small air gaps. For the shape of the 3D prints, the Hausdorff distances yielded average (median, max, min) errors of (0.21, 2.26, 0.00) mm. The larger errors (e.g. max error) were observed in regions where a noticeable change in 3D print shape was observed (e.g. around the nose and eyes). This information may be used to map out potentially troublesome areas. A comprehensive set of QA tests has been presented for 3D printed bolus used in RT. The robustness of these tests was demonstrated through application on a variety of 3D prints. This QA will aid in standardizing 3D print quality in RT. As this QA set is the first of its kind, it can be adopted by programs commissioning and performing regularly scheduled QA for 3D printed bolus.