Abstract
577
Objectives: Bremsstrahlung photons utilized in SPECT imaging of Y-90, an almost pure beta-emitter, have two origins: external bremsstrahlung (EB) produced as beta particles traverse tissue; and internal bremsstrahlung (IB), which arises during the beta decay process itself. However, until recently [Eur J Nucl Med Mol Imag 2017 44 (Suppl) S119-S956] the latter effect has not been considered when modeling the bremsstrahlung energy spectrum in SPECT simulations. The goal of this work is to generate and validate a Y-90 bremsstrahlung emission spectrum including both EB and IB, and to demonstrate that this pre-calculated spectrum can be used for accurate Monte Carlo simulation of bremsstrahlung SPECT/CT imaging.
Methods: The energy spectrum from a Y-90 source was measured with a high purity germanium detector (HPGe). This same source/detector geometry was then simulated using the pencyl program of the PENELOPE (version 2014) Monte Carlo electron/photon transport package. The geometry model of the HPGe detector and lead shield were based on manufacturer specifications. The beta emission spectrum of the source was taken from the BetaShape program, and the Y-90 IB emission spectrum was taken from the theoretical work of Cengiz [Cengiz & Almaz, Rad Phy & Chem 2004;661-668]. Energy deposition by photons striking the HPGe were tallied in 1 keV bins. The simulated and measured detected energy spectra were then compared. After this validation, PENELOPE was used to generate a Y-90 EB emission spectrum by simulating a point source in infinite tissue (water), tallying the bremsstrahlung spectrum at the point of generation, before any potential self-absorption. This EB spectrum (in absolute units of photons/eV/decay), was then combined with the theoretical absolute IB spectrum to produce a total Y-90 bremsstrahlung emission spectrum. This pre-calculated EB+IB photon spectrum was then coupled with the SIMIND Monte Carlo code, which includes detailed photon transport in the object and SPECT camera. Gamma-camera measurements for four Y-90 source geometries (line source, a hot sphere in air, hot sphere in a coldwater phantom and a liver/lung torso phantom with activity distribution mimicking microsphere radioembolization) were performed with a high-energy collimator to validate SPECT simulations with the EB+IB emission spectrum. Results: With detailed modeling of the HPGe system there was excellent agreement between the measured energy spectrum and PENELOPE simulation when both EB+IB was included (total detected counts agreed to within 3%). In the EB+IB emission spectrum, the total yield (per decay) of photons > 30 keV was 4.7% (EB 3.4% and IB 1.3%). The IB contribution to the total was 16% at 30 keV, 26% at 100 keV, 49% at 500 keV, 67% at 1 MeV and 77% at 1.5 MeV. Coupling the EB+IB emission spectrum with SIMIND also gave good agreement between energy spectra measured by the SPECT camera and simulation for the 3 phantom geometries (with total counts agreeing to within 5%). There was also excellent agreement between measured and simulated line profiles (FWHM agreed to within 5% and FWTM agreed to within 1%) and SPECT projection images acquired with a 105 to 195 keV acquisition window. Within this acquisition window the counts originating from IB, including higher energy IB photons that downscatter in to the window, were as high as 48% of the total counts. Conclusion: IB is a significant contributor to the Y-90 bremsstrahlung yield dominating over (tissue) EB at energies above ~ 500 keV. For accurate SPECT simulations, the IB component available from theoretical calculations can be added to the EB component from electron/photon transport. In a typical bremsstrahlung SPECT acquisition window, approximately half the counts originate from IB, hence neglecting this component in simulation studies will lead to gross underestimation of the total counts. Research Support: NIH R01 EB022075