Endorsing Titanium-Scandium Radionuclide Generator for PET and Positronium Imaging

TL;DR

The work addresses how to sustainably supply $^{44}$Sc for PET and positronium imaging by advancing a $^{44}$Ti/$^{44}$Sc generator. It identifies the $^{45}$Sc(p,2n)$^{44}$Ti route as the most promising production path, analyzes alternative charged-particle reactions, and develops target, yield, and separation strategies to enable practical, cyclotron-supported production. Thick-target yield calculations and purification schemes (ion-exchange and solid-phase extraction) are outlined alongside concrete facility capabilities, supporting decentralised, long-term availability of $^{44}$Sc for clinical and research PET/PLI applications. The integration with J-PET and its potential to reduce diagnostic costs further enhances the method's impact, offering a pathway to broader access to advanced PET imaging and positronium lifetime imaging in diverse healthcare settings.

Abstract

The development of PET and positronium imaging techniques is strictly related to the availability of suitable radionuclides and robust radiochemistry platforms. Among the emerging candidates, $^{44}$Sc has attracted significant interest due to its favourable physical properties, including a half-life of $\sim$4 hours, a pure $β^{+}$ emission profile, and the additional prompt $γ$-emission that enables advanced triple-photon detection schemes. These characteristics make $^{44}$Sc particularly promising for highresolution imaging and novel quantitative methodologies. However, routine clinical and preclinical implementation requires a practical, sustainable, and cost-efficient production route. In this context, we propose a titanium-scandium radionuclide generator as an optimal solution. This study focuses on optimising the synthesis of the long-lived parent isotope, $^{44}$Ti ($T_{1/2}$ = 59.1 years), from which $^{44}$Sc can be selectively eluted in a chemically pure form when needed. An analysis of various production pathways was conducted, including proton and deuteron reactions on scandium, as well as $α$-particle and lithium-induced reactions on calcium, to determine the most efficient reaction parameters, target design, and expected yield. Furthermore, we identify some existing cyclotron facilities suitable for implementing this technology. Results indicate that efficient $^{44}$Ti production is achievable using proton beams in the 20-30 MeV range under extended irradiation conditions. The proposed generator system would enable routine and decentralised $^{44}$Sc supply. Its integration with the novel J-PET scanner may significantly reduce diagnostic costs and improve access to advanced PET imaging in regions with limited medical imaging infrastructure.

Figures (10)

• Figure 1: Experimental excitation function data for the $^{45}$Sc(p,2n)$^{44}$Ti reaction, including studies by McGee et al.McGee1970 (black squares), Levkovskij Levkovskij1991 (red circles, normalised by a factor of 0.8), Ejnisman et al.Ejnisman1996 (green upward triangles), Daraban et al.Daraban2009 (magenta downward triangles), and Ditrói et al.Ditroi2024 (cyan crosses). The IAEA database recommendation IAEA-Medical (solid blue line), based on a Padé fit to evaluated data, and the TENDL-2023 evaluation TENDL2023 (dotted orange line) are also presented.
• Figure 2: Experimentally measured cross-sections $\sigma$ for proton-induced nuclear reactions on $^{45}$Sc as a function of incident proton energy $E_{\mathrm{p}}$. The data show the excitation functions for the main $^{45}$Sc(p,2n)$^{44}$Ti production channel (red circles) and its primary competing reactions: $^{45}$Sc(p,n)$^{45}$Ti (orange diamonds), $^{45}$Sc(p,2np)$^{43}$Sc (blue squares), $^{45}$Sc(p,np)$^{44g}$Sc (green upward triangles), and $^{45}$Sc(p,np)$^{44m}$Sc (magenta downward triangles). Data from Daraban et al.Daraban2009 (scandium isotopes and $^{44}\text{Ti}$) and Ejnisman et al.Ejnisman1996 ($^{45}\text{Ti}$).
• Figure 3: Excitation function for the $^{45}$Sc(d,3n)$^{44}$Ti reaction. Experimental cross-section data from Hermanne et al.Hermanne2012 (red circles) and Tsoodol et al.Tsoodol2021 (green squares) are shown alongside a Padé fit (solid blue line) and the TENDL-2023 evaluation (dotted orange line) TENDL2023. The plot illustrates the relatively low cross-sections for this channel, peaking in the 30--37 MeV range.
• Figure 4: Excitation functions for deuteron-induced reactions on $^{45}$Sc, showing the main channel $^{45}$Sc(d,3n)$^{44}$Ti (red circles) alongside competing reactions: $^{45}$Sc(d,2n)$^{45}$Ti (orange diamonds), $^{45}$Sc(d,p3n)$^{43}$Sc (blue squares), $^{45}$Sc(d,p2n)$^{44m}$Sc (green upward triangles), $^{45}$Sc(d,p2n)$^{44g}$Sc (magenta downward triangles), and $^{45}$Sc(d,p)$^{46}$Sc (black crosses). Data from Hermanne et al.Hermanne2012.
• Figure 5: Theoretical excitation functions of the $^{42}$Ca($\alpha$,2n)$^{44}$Ti (solid blue line), $^{43}$Ca($\alpha$,3n)$^{44}$Ti (dashed green line), and $^{44}$Ca($\alpha$,4n)$^{44}$Ti (dotted orange line) reactions as predicted by the TENDL-2023 library TENDL2023. The experimental data point for $^{42}$Ca reaction (red circles) is from Levkovskij Levkovskij1991.