Evaluation of tube size variations to improve resolution in tube-of-response based image reconstruction for Ga68 PET

J Maus; P Nikulin; F Hofheinz; J Kotzerke; J van den Hoff

doi:10.1055/s-0040-1708144

Nuklearmedizin - NuclearMedicine, Table of Contents

Nuklearmedizin 2020; 59(02): 97
DOI: 10.1055/s-0040-1708144

Wissenschaftliche Vorträge

Radiomics

Evaluation of tube size variations to improve resolution in tube-of-response based image reconstruction for Ga68 PET

Authors

J Maus

¹Helmholtz-Zentrum Dresden-Rossendorf, Positronen-Emissions-Tomographie, Institut für radiopharmazeutische Krebsforschung, Dresden
P Nikulin

¹Helmholtz-Zentrum Dresden-Rossendorf, Positronen-Emissions-Tomographie, Institut für radiopharmazeutische Krebsforschung, Dresden
F Hofheinz

¹Helmholtz-Zentrum Dresden-Rossendorf, Positronen-Emissions-Tomographie, Institut für radiopharmazeutische Krebsforschung, Dresden
J Kotzerke

²Klinik und Poliklinik für Nuklearmedizin, University Hospital Carl Gustav Carus, Technische Universität Dresden, Dresden
J van den Hoff

¹Helmholtz-Zentrum Dresden-Rossendorf, Positronen-Emissions-Tomographie, Institut für radiopharmazeutische Krebsforschung, Dresden

Abstract

Full Text

Ziel/Aim Ga68-labeled tracers like Ga68-PSMA or Ga68-DOTATOC play a relevant role in clinical PET, but the substantial positron range of Ga68 (mean/max range in water: 2.7/8 mm) deteriorates spatial resolution. Integration of isotope-specific resolution recovery into iterative reconstruction should therefore be considered. In our in-house PET image reconstruction – THOR – this is possible by adjusting the tube diameter in the used tube-of-response projector. We have, therefore, investigated the potential of tube diameter optimization on achievable image resolution for Ga68.

Methodik/Methods We performed measurements with a resolution phantom at three different target-to-background ratios (20:1, 10:1, 5:1) for F18 and Ga68, respectively (Philips Ingenuity-TF PET/MR). All measurements were reconstructed with the vendor reconstruction as well as with THOR using different tube diameters and voxel sizes (2mm, 4mm). Image resolution, noise level, and magnitude of Gibbs artifacts (GA) were derived. The optimal tube diameters for F18 and Ga68 were determined respectively and applied to exemplary clinical Ga68 PSMA data.

Ergebnisse/Results The vendor reconstruction yields a resolution (FWHM) of [6.7±0.2] mm for Ga68 which is ~10 % worse than for F18 while THOR yielded [5.0±0.4] mm using the default – F18-optimized – tube size which is ~14 % worse than for F18. Increasing the tube diameter in THOR by up to 30 % is possible for Ga68 (but not for F18) without causing notable GAs. This improves the achieved resolution by ~8 % to ~4.6 mm. This is only slightly worse than the THOR result for F18 and much better than the resolution achievable for Ga68 with the vendor reconstruction. First clinical examples also demonstrate the beneficial effects of an isotope-specific tube size, exhibiting visually superior image quality.

Schlussfolgerungen/Conclusions Our preliminary results indicate that an isotope-dependent tube diameter does improve image resolution for challenging isotopes like Ga-68 without introducing notable GAs that often plaque resolution recovery attempts.