Publications

Background and Aim:
Realistic S-value (Gy/Bq·s) modelling is essential for calculating absorbed dose in targeted radionuclide therapy (TRT) and establishing dose-effect relationships for cell lines during in-vitro studies. This work investigates the influence of geometry, spatial cell distribution, and Monte Carlo code selection on absorbed dose calculations for radionuclides in 96-well plates.
Methods:
Monte Carlo simulations with GEANT4 (v.11.3) and MCNP6.2 modeled radionuclide decays in 100 µL water-equivalent medium in a well (6.35 mm diameter). Two geometries were considered: a simplified monolayer, and a Gaussian or uniform distribution of 15 μm-spherical cells at the well bottom. Absorbed energy was scored using the F8 tally in MCNP and electron tracking with the Livermore physics list in GEANT4. S-values were obtained by dividing absorbed energy by target mass: whole-cell or nuclei in the individual cell model, and full monolayer or nuclear layer for the monolayer model.
Results:
For the monolayer, S-values from GEANT4 and MCNP for Lu-177 agreed within 5%, with differences attributed to cross-section data. The realistic model predicted higher absorbed doses, with monolayer S-values lower by 22% in MCNP and 8% in GEANT4. In the individual cell model, MCNP estimated S-values higher than GEANT4 by 11% (whole cell) and 8% (nucleus). Spatial arrangement effects were minor (~3%). Ongoing work will assess whether these differences reflect true absorbed dose variations or model assumptions across radionuclides.
Conclusions:
Geometry definition had the largest impact on S-values, followed by code selection, highlighting the importance of standardized models for reliable, reproducible dosimetry in preclinical TRT.

Yirong Zhang, Valerio Cosmi, Satyajit Ghosh, Ruud M. Ramakers, and Freek J. Beekman

Image uniformity is crucial in preclinical SPECT and PET imaging. This study investigates whether ultra-fine object sampling can improve image quality for ultra-sensitive PET-SPECT pinhole collimators. During resolution recovery in statistical reconstruction, artifacts may appear in images of non-sparse objects due to deconvolution effects. The VECTor PET-SPECT system (MILabs B.V., The Netherlands), equipped with an XYZ stage for precise bed movement, enables flexible object sampling. Our hypothesis that smaller sampling steps improve uniformity is based on the premise that finer sampling captures more high-resolution information from photons traveling near pinhole cone surfaces. We simulated a VECTor system with a 160-pinhole supercluster collimator (3.5 mm diameter) and triangular NaI(Tl) detector configuration. A cylindrical phantom (14 mm diameter) with uniform F-18 concentration (2 MBq/ml) was scanned in multi-planar mode (5 axial planes) with 8, 16, or 32 bed positions per plane, totaling 40, 80, and 160 positions respectively. Additional axial angular shifts (22.5°, 11.25°, 5.625°) improved inter-layer sampling. Reconstruction used Dual-Matrix Dual-Voxel OSEM (0.4 mm voxels) with 1 mm FWHM Gaussian post-filtering. Uniformity improved from 33.00% (40 positions) to 28.38% (80 positions) and 25.06% (160 positions), representing relative improvements of 14.0% and 11.7%. These results demonstrate that finer sampling significantly enhances uniformity in small-animal pinhole PET-SPECT imaging.

Index Terms—Sampling, SPECT, PET, pinhole, small-animal

I. Introduction

Resolution recovery in image reconstruction can introduce artifacts in uniform phantom images due to deconvolution effects. We investigate whether ultra-fine object sampling improves image quality for ultra-sensitive collimated PET-SPECT systems. The VECTor PET-SPECT system (MILabs B.V.) features an XYZ stage enabling precise bed movement, allowing testing of smaller sampling steps to improve uniformity. Our hypothesis posits that finer sampling captures more high-resolution information from photons traveling near pinhole cone surfaces. This study evaluates how increasing bed positions affects image uniformity.

The aim of this study is to investigate the potential of finer sampling of multi-plane trajectories (MPTs) for improving image uniformity in high-sensitivity small-animal SPECT even with large pinholes. To evaluate this, we quantitatively assessed how image uniformity changes as the number of bed positions increases in a five-plane MPT acquisition.

II. Methods

A high-sensitivity VECTor PET-SPECT system with 160-pinhole supercluster collimator (3.5 mm diameter) and 19 mm thick monolithic NaI(Tl) detectors was simulated using GATE. A cylindrical phantom (14 mm diameter, 6 mm length) contained 2 MBq/ml 18F. Scanning employed five transaxial planes with 8, 16, or 32 bed positions per plane (total 40, 80, and 160 positions respectively). Data from all positions were jointly reconstructed using Dual-Matrix Dual-Voxel OSEM, incorporating collimator/detector blur and depth-of-interaction modeling. A 3D Gaussian post-filter (1 mm FWHM) was applied.

Uniformity was assessed in 1.6 mm thick slices using: Uniformity = (max − min)/(max + min) × 100% within a circular ROI covering 75% of the phantom diameter.

III. Results

Uniformity improved from 41.23% (40 positions) to 35.60% (80 positions) and 30.62% (160 positions), representing relative improvements of 13.6% and 14.0% at 100 iterations.

Uniformity as a function of iteration number confirms that finer sampling provides improved uniformity.

References

[1] B. Vastenhouw and F. J. Beekman, “Submillimeter total-body murine imaging with U-SPECT-I,” J. Nucl. Med., vol. 48, no. 3, pp. 487–493, Mar. 2007.
[2] F. J. Beekman et al., “Positron range-free and multi-isotope tomography of positron emitters,” Phys. Med. Biol., vol. 66, no. 6, p. 065011, Mar. 2021.
[3] S. Ghosh, V. Cosmi, R. M. Ramakers, F. J. Beekman, and M. C. Goorden, “Ultra-high energy spectral prompt PET,” Phys. Med. Biol., vol. 70, no. 7, p. 075010, 2025.
[4] M. C. Goorden et al., “VECTor: a preclinical imaging system for simultaneous submillimeter SPECT and PET,” J. Nucl. Med., vol. 54, no. 2, pp. 306–312, Feb. 2013.
[5] S. Jan et al., “GATE: a simulation toolkit for PET and SPECT,” Phys. Med. Biol., vol. 49, no. 19, pp. 4543–4561, Oct. 2004.
[6] M. C. Goorden et al., “Accelerated image reconstruction by a combined dual-matrix dual-voxel approach,” Phys. Med. Biol., vol. 65, no. 10, p. 105014, May 2020.