Room: AAPM ePoster Library
The accuracy of personalized dosimetry for radiopharmaceutical therapy (RPT) is limited by uncertainties related to quantitative nuclear medicine scans (i.e. PET, SPECT), which serve as input data to RPT dosimetry calculations. Partial volume corrections (PVC) in the form of recovery coefficients (RC) are often applied to address the image degradation from inherent resolution limitations of the scanner. However, measurements used to quantify RCs are often taken with idealized nuclear medicine phantoms. A 3D printed anthropomorphic phantom with replaceable, patient-specific, 3D printed inserts facilitates an easy and inexpensive way to calculate RCs to ensure accurate activity quantification for more accurate, personalized RPT dosimetry. In this work, the dosimetric impact of RCs derived from a 3D printed anthropomorphic phantom are studied using the Monte Carlo dose calculation platform RAPID.
RCs were calculated as a function of insert volume for both a Jaszczak phantom and a phantom 3D printed based on real patient anatomy that was filled with an Iodine-131 solution and imaged using SPECT/CT. The SPECT image data was corrected using RCs derived from the anthropomorphic phantom. The impact of PVC upon the resultant dose distributions is evaluated and compared against an idealized scenario.
As the size of the Jaszczak sphere increased, the recovery coefficient also increased with the 16ml sphere having a RC=0.48. The 3D printed insert (~20ml) has a larger recovery coefficient (RC=0.58) than the Jaszczak spheres. Using the SPECT imaging data in the RPT dosimetry workflow, the mean dose to the tumor is underestimated by 32% without using PVC and overestimated by 6% using PVC with the 3D printed insert RC when compared to ground truth.
This work highlights the importance of accurate RC calculations for PVC of SPECT-based Iodine-131 RPT dosimetry and offers an inexpensive way to integrate patient-specific SPECT PVC into dosimetry workflows.
Funding Support, Disclosures, and Conflict of Interest: This work is supported by NIH SPORE CA196513-01. BB and JG are co-founders of Voximetry, Inc., a nuclear medicine dosimetry company in Madison, WI.