Technical Note: Transconvolution based equalization of positron energy effects for the use of ⁶⁸ Ge/⁶⁸ Ga phantoms in determining ¹⁸ F PET recovery

Abstract

Purpose: Avoiding measurement variability from 18 F phantom preparation by using 68 Ge/ 68 Ga phantoms for the determination of 18 F recovery curves (RC) in clinical quality assurance measurements and for PET/CT site qualification in multicentre clinical trials. Methods: RCs were obtained from PET/CT measurements of seven differently sized phantom spheres filled either with 18 F or with 68 Ga. RCs for the respective other isotope were then determined by two different methods: In the first method, images were convolved with positron range transconvolution functions derived from positron annihilation distributions found in literature. This method generated recasted images matching images using the respective other isotope. In the second method, the PET/CT system's isotope independent (intrinsic) point spread function was determined from said phantom measurements and convolved with numerical representations simulating hot spheres filled with the respective other isotope. These simulations included the isotope specific positron annihilation distributions. Recovered activity concentrations were compared between recasted images, simulated images, and the originally acquired images. Results:18 F and 68 Ga recovery was successfully determined from image acquisitions of the respective opposite isotope as well as from the simulations.68 Ga RCs derived from 18 F data had a normalized root-mean-square deviation (NRMSD) from real 68 Ga measurements of 0.019% when using the first method and of 0.008% when using the second method.18 F RCs derived from 68 Ga data had a NRMSD from real 18 F measurements of 0.036% when using the first method and of 0.038% when using the second method. Conclusions: Applying the principles of transconvolution, 18 F RCs can be recalculated from 68 Ga phantom measurements with excellent accuracy. The maximal additionally introduced error was below 0.4% of the error currently accepted for RCs in the site qualification of multicentre clinical trials by the EARL program of the European Association of Nuclear Medicine (EANM). Therefore, our methods legitimately allow for the use of long-lived solid state 68 Ge/ 68 Ga phantoms instead of manually prepared 18 F phantoms to characterize comparability of 18 F measurements across different imaging sites or of longitudinal 18 F measurements at a single PET/CT system.

Standardisation of conventional and advanced iterative reconstruction methods for Gallium-68 multi-centre PET-CT trials

“…As more and more isotopes find their way into the clinic though, this might prove logistically difficult as it could result in multiple standards that need to be regularly reviewed and updated. A method to move this field forward might be to establish a model for defining the limits based on the isotopes’ characteristics [ 37 ]. An alternative approach could be to utilise newly developed deep learning approaches to optimise the images [ 38 ].…”

“…For the investigation of the performance of Q.CLEAR and its harmonisation with PSF reconstruction, the NEMA IEC PET Body Phantom™ (Data Spectrum) was scanned on the GE Discovery 710 PET-CT scanner. Gallium-68 was inserted in all six spheres (37,28,22,17,13, and 10 mm) and the background region, achieving radioactivity concentrations of 19.9 kBq/ml and 2.4 kBq/ml, respectively. The lung insert packed with styrofoam beads was also included in the background.…”

3D printing of radioactive phantoms for nuclear medicine imaging

“…While technetium-99m is an interesting radionuclide for certain SPECT-applications, we expect that future developments in the field will focus on solid-state phantoms containing radionuclides with longer half-lives, in particular positron emitters for PET, which can be used for site qualification in multicenter clinical trials and for longitudinal measurements on a single PET/CT scanner [27]. Building on our experience with 3Dprinted 99m Tc-phantoms, we are currently developing customized procedures for 3D printing of phantoms containing radionuclides with longer half-lives (t ½ ), such as germanium-68 (t ½ = 271 days) for PET-and cobalt-57 (t ½~2 72 days) for SPECT applications.…”

Towards guidelines to harmonize textural features in PET: Haralick textural features vary with image noise, but exposure-invariant domains enable comparable PET radiomics

“…Because positron trajectories are de-correlated from the MTF, differences in positron range are irrelevant between isotopes. Additionally, the transconvolution method allows for recasting images acquired with different positron emitters [52].…”