Reconstructing Radiopharmaceutical Distributions Using Coded Aperture Tomography with Photon Counting Detector 📝

Author: David P. Adam, Xun Jia, Youfang Lai, Yuncheng Zhong 👨‍🔬

Affiliation: Department of Radiation Oncology and Molecular Radiation Sciences, Johns Hopkins University, Johns Hopkins University 🌍

Abstract:

Purpose: Radiopharmaceutical therapy is experiencing a resurgence in interest due to its potential of treating widespread metastatic disease in a patient-specific manner. Accurately measuring and describing pharmacokinetic behavior in patients is crucial for reconstructing absorbed doses and assessing outcomes. Current conventional approaches like SPECT/CT are hindered by logistical challenges for measurement at multiple time points over multiple days. This study proposes a novel approach using coded aperture tomography with photon counting detectors (PCDs) to reconstruct radiopharmaceutical distributions.
Methods: The system measures signals from two orthogonal angles with a coded aperture in front of a PCD, capable of detecting low photon signals with high sensitivity. We assumed the activity distribution to be reconstructed is similar to the prior image acquired at initial SPECT imaging. Using this prior information, a reconstruction model was formulated as an optimization problem with respect to decay factors for each individual activity group, e.g. tumor. We performed a proof-of-concept simulation using public SPECT and CT data for a patient treated with 177Lu-DOTATATE. Measurement data at the coded-aperture detectors for days 4-8 were computed based on ground truth activity distributions for a certain data acquisition time via raytracing through the patient for 208 keV photon and adding detector counting noise. Activity distributions were reconstructed using computed projection data. Relative root mean square error to the maximum activity value (rRMSE) was used to evaluate reconstruction results.
Results: We investigated dependence of rRMSE on data acquisition time and determined 10 min (5 min per projection) to achieve ~1% rRMSE in a benchmark study. We studied four scans from two patients with real SPECT data. rRMSE ranged from 3.9% to 7.4%. Maximum intensity projection demonstrated good correlation between reconstructed source position and ground truth.
Conclusion: The novel activity reconstruction system demonstrates potential for activity quantification in radionuclide therapy.

Back to List