BEST IN PHYSICS (RADIOPHARMACEUTICALS): Design and Fabrication of a Novel Pelvic-Vertebral 3D-Printed Anthropomorphic Phantom for Radiopharmaceutical Therapy Dosimetry
Abstract
Purpose
To design and fabricate a 3D-printed anthropomorphic phantom of the lumbar spine and pelvis to evaluate and improve the quantitative accuracy of SPECT/CT imaging and dosimetry for bone marrow and vertebral body lesions.
Methods
The skin, lumbar vertebrae, intervertebral discs, pelvis, and sacrum from the ICRP 145 adult male mesh-type computational phantom was used as a template for this phantom. Using the software Blender, each vertebra was modeled to accommodate radioactive spheres, and the skin surface was modified to form an outer waterproof shell. The phantom was printed on a 3D-printer using Simubone, PLA, and LW PLA. Lesions were represented by radioactive spheres constructed from 3D-printed PLA hemispherical shells to create 0.5-,1-, 2-, and 4-mL volumes. The shells were filled with epoxy resin mixed with Lu-177 to achieve an activity concentration of 700kBq/mL at the time of scanning. The phantom was scanned on a GE 870 SPECT/CT using our institution’s acquisition protocol for patients treated with Lu-177 therapeutically. Bone segmentations were created on the SPECT/CT images to compare with the 3D model volume of the phantom. Recovery coefficients (RC) were extracted and compared with standard measurements in a NEMA phantom.
Results
The total bone respective volumes of the model and segmentations was 1162.72cm3 and 1221.62cm3 (5.07% difference). The RCs found with the 3D-printed phantom were greater than those of the NEMA phantom. For 0.5 mL volumes, the 3D-printed phantom had an RC that was 2.04 times greater than the values estimated for the NEMA phantom while for 4 mL volumes, the RC was estimated to be 1.07 times greater.
Conclusion
This phantom lays the groundwork for demonstrating improved image quantification accuracy for bone marrow and lesion dosimetry. Future work for this phantom includes implementing the 3D-printed phantom's RCs into our dosimetry workflow to see absorbed dose effects.