A Method for Partial Volume Correction of Dose Distributions Derived From PET-Based Theranostic Dosimetry
IR Marsh1*, TJ Bradshaw2, JJ Grudzinski2, R Hernandez2, E Aluicio-Sarduy2, BP Bednarz2, (1) Johns Hopkins University, Baltimore, MD, (2) University of Wisconsin-Madison, Madison, WI
Presentations
SU-F-202-2 (Sunday, 7/10/2022) 2:00 PM - 3:00 PM [Eastern Time (GMT-4)]
Room 202
Purpose: Theranostic dosimetry estimates in small lesions, commonly targeted in radiopharmaceutical therapy (RPT), are frequently underestimated due to partial volume effects inherent to quantitative PET assessment. In this work we present a methodology through which region-based partial volume correction (PVC) may be applied directly to PET-derived 3D dose distributions.
Methods: A clinical PET/CT imaging study was performed using a Jaszczak phantom containing 6 hollow spheres ranging between 0.5-16 ml (9.8-31.3 mm, diameter) and filled at a source to background ratio of 10:1 with ⁸⁶Y solution produced onsite. Samples of hot-sphere and background solutions were taken, and activity concentration was verified in a calibrated gamma counter. Theranostic dosimetry for ⁹⁰Y was performed using both true and PET image-derived activity distributions in a Monte Carlo voxel-based internal dosimetry platform. A formula for PVC based on ⁸⁶Y PET recovery coefficients (RC) and directly applicable to 3D ⁹⁰Y dose distributions was developed and assessed in a dataset of companion canine patients receiving experimental imaging and therapy with ⁸⁶Y/⁹⁰Y-NM600.
Results: Recovery of ⁸⁶Y activity in phantom spheres ranged between 0.33 and 0.69 based on contours drawn to match known sphere volumes. Following theranostic dosimetry, mean dose estimates in spheres ranged between -25% to -56% of the true dose distribution. Application of the inverse RC factor resulted in overcorrection by up to 34% while the developed dose-recovery formula accurately recovered dose across all sphere sizes. For small lesions (0.5-5.5 ml) observed in canine patients receiving ⁸⁶Y-NM600, PVC dosimetry estimates were 59.6% ± 18.6% higher than that of raw PET-based dosimetry.
Conclusion: It is feasible to directly apply PVC to PET-based 3D dose distributions for small lesions in theranostic dosimetry applications. The methodology for dose recovery presented here provides a practical framework through which to improve the quantitative accuracy of personalized RPT.
Keywords
Quantitative Imaging, Dosimetry, Nuclear Medicine
Taxonomy
IM/TH- Radiopharmaceutical Therapy: General (most aspects)
Contact Email