ESTRO 38 Abstract book

S95 ESTRO 38

K. Singhrao 1 , J. Fu 1 , Y. Gao 2 , P. Hu 2 , Y. Yang 1 , J.H. Lewis 1 1 University of California Los Angeles, Radiation Oncology, Los Angeles, USA; 2 University of California Los Angeles, Radiology, Los Angeles, USA Purpose or Objective To develop a system of easily made and formed materials with adjustable T1 and T2 relaxation times, and x-ray attenuation properties for mimicking soft tissues and bone in both CT and MR imaging. The growing role of MR in radiotherapy has increased the demand for phantoms capable of cross-modality quality assurance or end-to-end testing. Most commercially available materials and phantoms do not mimic the T1 and T2 relaxation times, and CT numbers for different tissue types simultaneously on both modalities. Producing materials with the capability of producing tissue mimicking contrast on both modalities allows a ground truth to be established for validating a number of medical imaging and therapeutic applications such as MR-CT image registration and each component of MR-only radiotherapy workflows. Material and Methods The effects on T1 and T2 relaxation times, and CT number were measured using a range of concentrations Gd contrast (0 – 250 umol/g), agarose (0 - 4 g), glass microspheres (GMs) (0 – 5 g) and CaCO 3 (0 – 25 g) in a common carrageenan-water gelatinizer. Samples were prepared with the additives, 1.5 g carrageenan and water to weigh a total of 50 g. Samples were imaged in a tissue validation phantom. Each tissue validation phantom contained 20 samples with fixed concentration of CaCO 3 or GMs. MR images were acquired on a 3T Siemens Skyra and CT images were acquired on a Siemens Somatom. T1 and T2 relaxation time maps were generated using voxel-wise inversion recovery and spin echo signal fitting. A multivariate linear regression-based model (based on fitted coefficients) was generated to predict T1 and T2 relaxation times and CT number based on the concentrations of the 4 additives. Skeletal muscle, adipose tissue, white matter, gray matter, liver, prostate, bone marrow, glandular breast and trabecular bone tissues were mimicked to validate the fit model results and demonstrate the flexible range of values attainable with this system of materials. Results Figure 1(a,c,d) demonstrates that the carrageenan-based system of materials can span a range of T1 values from 82 ms to 2180 ms, T2 values from 12 ms to 475 ms and CT numbers from -117 HU to 914 HU. Figure 1(b,e,f) demonstrates that the addition of 10% CaCO 3 decreases the maximum T1 range by 36% and maximum T2 range by 83%. Addition of 10% glass microspheres decreases the maximum T1 range by 41% and maximum T2 range by 99%. Using the fit model, we were able to mimic the T1, T2 and CT numbers well compared to literature reported in vivo measurements (Table 1).

Conclusion We have created a system of carrageenan-based materials capable of simultaneously producing tissue-like contrast for a wide range of human tissues in 3.0T MR and CT imaging. The materials can be cast and formed in the shapes of organs to create anthropomorphic phantoms. OC-0187 Comparison of proton range predictions between Single- and Dual-Energy CT using prompt gamma imaging T. Boon-Keng 2 , Y. Xie 1 , F. O’Grady 2 , A. Lalonde 3 , J. Petzoldt 4 , J. Smeets 4 , G. Janssens 4 1 Massachusetts General Hospital, Department of Radiation Oncology, Boston, USA; 2 University of Pennsylvania, Department of Radiation Oncology, Philadelphia, USA ; 3 University of Montreal, Department of Physics, Montreal, Canada; 4 Ion Beam Applications, Research, Louvain-la-Neuve, Belgium Purpose or Objective Dual-energy CT (DECT) has been shown to be more accurate for predicting proton stopping power ratios (SPR) than traditional single-energy CT (SECT) calibrations in both surrogate and animal tissue phantoms. Prompt gamma imaging (PGI) is an in vivo imaging modality that has been used to demonstrate the feasibility of proton range verification. In this study, PGI was used to quantify the accuracy of SPR calibration from different CT modalities in a pelvic phantom. Material and Methods Three sets of CT scans were obtained with the Rando phantom using a Siemens Definition Edge CT scanner: 1) 120 kVp SECT; 2) 80/140 kVp sequential DECT scan (SQCT); and 3) 120 kVp Twin-beam scan with Ti/Au dual filter (TBCT). The HU-to-SPR calibration curves were established using the stoichiometric method on SECT and the Eigen tissue decomposition method on DECT scans. Two pencil beam scanning proton plans were created: one with two iso-energy layers and one to cover an artificially drawn target mimicking a prostate treatment. Both plans