ESTRO 37 Abstract book

S931

ESTRO 37

prostate motion, markers made of silicone are included on their surface, which are visible in MRI. Small pockets are included on the surface of the bladder and rectum to position OSLDs and dosimetric films for dose measurements. In feasibility studies, the reproducibility of organ motion was proven with accuracy of 1mm. These studies were based on the verification of the end-to-end testing, feasibility of dose measurements with OSLDs and dosimetric films.

weeks, a typical timeframe for a standard course of radiotherapy. After initial printing, half the blocks were stored in open boxes, the other half in sealed bags with desiccant. Variances in HU and density over time were evaluated for the four materials. Various clinical photon and electron beams were used to evaluate potential errors in clinical depth dose as a function of assumptions made during treatment planning. The clinical depth error was defined as the distance between the correctly calculated 90% isodose line and the 90% isodose line calculated using clinically reasonable, but simplified, assumptions. The simplified clinical assumptions we considered were: (1) that the material can be considered to be water equivalent, (2) that the CT calibration curve accurately predicts the material density, (3) that the average density for a material can be used for any block of that material, and (4), that individual block correction factors are applied to determine the correct density. Results The average HU measurements of individual blocks of PLA, ABS, NinjaFlex, and Cheetah varied by as much as 121, 30, 178, and 30 HU, respectively. Figure 1 shows the standard CT calibration curve plotted as a solid line, with each block represented as different symbols. Note that for all materials, the CT calibration curve underestimates their actual material density. The HU variation for individual blocks over 5 weeks was very small for all materials, regardless of storage environment. The magnitude of clinical depth errors depended strongly on the material, energy, and assumptions, but some were as large as 9.0 mm. Table 1 shows the errors for just PLA for each energy considered depending on the four assumptions we examined. As can be seen in Table 1, the best case is to apply individual correction factors, and the worst case is to assume water equivalence.

Conclusion With the ADAM-pelvis phantom, the methodological basis was realized to build phantoms that enable end-to-end testing in MR-guided radiotherapy. These can further be used to build even more complex functional models. The combination of 3D-printing methods, organs made of silicone and different mixtures of gels were proven to be suitable to build an anthropomorphic phantom. It was shown in various studies that patient-like motion induced uncertainties during radiotherapy are well mimicked by the phantom. Motion and dilatation can be identified by localizers. Using dosimeters, the applied dose to the bladder and rectum can be determined. In a next step the dose distribution in the prostate could be determined by using dosimetric gel. EP-1738 3D printed material uncertainty and its consequences for radiation oncology applications D. Craft 1 , S. Kry 1 , P. Balter 1 , M. Salehpour 1 , W. Woodward 2 , R. Howell 1 1 The University of Texas MD Anderson Cancer Center, Radiation Physics, Houston, USA 2 The University of Texas MD Anderson Cancer Center, Radiation Oncology, Houston, USA Purpose or Objective There has been growing interest recently in 3D printing technology in the field of radiation oncology. This is because 3D printing can simply and inexpensively fabricate patient-specific devices such as tissue compensators, boluses, and phantoms. However, most 3D printing materials have not been well characterized, including their radiologic tissue-equivalence. The purposes of this study were to (1) determine the variance in Hounsfield Units (HU) for printed objects, (2) determine if HU varies over time, and (3) calculate the clinical dose uncertainty caused by these material variations in a variety of common use cases. Material and Methods For a sample of 10 printed blocks each of PLA, NinjaFlex, ABS, and Cheetah, the average HU and physical density were tracked at initial printing and over the course of 5

Figure 1:

Table 1:

Conclusion