Scandium and terbium radionuclides for radiotheranostics: current state of development towards clinical application

Abstract

Currently, different radiometals are in use for imaging and therapy in nuclear medicine: ⁶⁸Ga and ¹¹¹In are examples of nuclides for positron emission tomography (PET) and single photon emission computed tomography (SPECT), respectively, while ¹⁷⁷Lu and ²²⁵Ac are used for β⁻- and α-radionuclide therapy. The application of diagnostic and therapeutic radionuclides of the same element (radioisotopes) would utilize chemically-identical radiopharmaceuticals for imaging and subsequent treatment, thereby enabling the radiotheranostic concept. There are two elements which are of particular interest in this regard: Scandium and Terbium. Scandium presents three radioisotopes for theranostic application. ⁴³Sc (T_1/2 = 3.9 h) and ⁴⁴Sc (T_1/2 = 4.0 h) can both be used for PET, while ⁴⁷Sc (T_1/2 = 3.35 d) is the therapeutic match—also suitable for SPECT. Currently, ⁴⁴Sc is most advanced in terms of production, as well as with pre-clinical investigations, and has already been employed in proof-of-concept studies in patients. Even though the production of ⁴³Sc may be more challenging, it would be advantageous due to the absence of high-energetic γ-ray emission. The development of ⁴⁷Sc is still in its infancy, however, its therapeutic potential has been demonstrated preclinically. Terbium is unique in that it represents four medically-interesting radioisotopes. ¹⁵⁵Tb (T_1/2 = 5.32 d) and ¹⁵²Tb (T_1/2 = 17.5 h) can be used for SPECT and PET, respectively. Both radioisotopes were produced and tested preclinically. ¹⁵²Tb has been the first Tb isotope that was tested (as ¹⁵²Tb-DOTATOC) in a patient. Both radionuclides may be of interest for dosimetry purposes prior to the application of radiolanthanide therapy. The decay properties of ¹⁶¹Tb (T_1/2 = 6.89 d) are similar to ¹⁷⁷Lu, but the coemission of Auger electrons make it attractive for a combined β⁻/Auger electron therapy, which was shown to be effective in preclinical experiments. ¹⁴⁹Tb (T_1/2 = 4.1 h) has been proposed for targeted α-therapy with the possibility of PET imaging. In terms of production, ¹⁶¹Tb and ¹⁵⁵Tb are most promising to be made available at the large quantities suitable for future clinical translation. This review article is dedicated to the production routes, the methods of separating the radioisotopes from the target material, preclinical investigations and clinical proof-of-concept studies of Sc and Tb radionuclides. The availability, challenges of production and first (pre)clinical application, as well as the potential of these novel radionuclides for future application in nuclear medicine, are discussed.

INTRODUCTION

In nuclear medicine, the term “theranostics” refers to the application of radiopharmaceuticals for diagnosis and therapy.¹ Radionuclides emitting γ-radiation or positrons can be used for single photon emission computed tomography (SPECT) and positron emission tomography (PET), respectively. The visualization of accumulated radioactivity and its quantification for dosimetry can provide important information towards the application of radionuclide therapy using the same targeting agent. The basic principle of radiotheranostics when using radioisotopes of the same element was applied more than 70 years ago in thyroid cancer patients.² This idea was conceptualized by Dr S. Hertz and realized by treating the first patient with radioiodine in 1941.^{3, 4} For several decades, radioiodine has been employed in clinical routine,⁵ however, the use of radiometals towards the same principle has been more challenging. Currently, somatostatin analogs are used in combination with ⁶⁸Ga for PET imaging and, in some cases, with ¹¹¹In for SPECT prior to β^¯-radionuclide therapy using ¹⁷⁷Lu T_1/2 = 6.65 d; Eβ⁻ _average = 134 keV; Eγ=113 keV, 208 keV or ⁹⁰Y T_1/2 = 2.67 d; Eβ⁻ _average = 933 keV.^{6, 7} The same concept was applied with prostate-specific membrane antigen (PSMA)-targeting ligands that have recently been introduced in clinical practice for PET imaging and therapy.^8–10 In some cases, α-emitters, such as ²²⁵Ac and ²¹³Bi, were also combined with PSMA ligands for the treatment of patients with advanced prostate cancer.^{11, 12} In these, and many other, examples, the diagnostic radiometal was always a different element to the therapeutic counterpart. Radionuclides of the same element (i.e. radioisotopes) would allow the preparation of chemically-identical radiopharmaceuticals for diagnosis and therapy enabling the concept of radiotheranostics in the truest sense. In this regard, scandium and terbium are of particular interest, as they present several radioisotopes which may be of particular value for clinical translation.

Scandium has attracted the attention of researchers and nuclear physicians alike, due to the existence of matched radionuclides for the possibility of theranostic applications (Table 1).^{13–15 43}Sc and ⁴⁴Sc are promising for PET imaging. ⁴⁷Sc is a β^–-emitter suitable for therapeutic purposes, which also produces γ-ray emission useful for SPECT imaging. The application of ⁴³Sc/⁴⁴Sc (T_1/2 = 3.9 and 4.0 h, respectively) for PET would be advantageous with regard to several aspects when comparing it to the currently-employed ⁶⁸Ga (T_1/2 = 68 min): the almost fourfold longer half-lives of ⁴³Sc/⁴⁴Sc would enable the shipment of ⁴³Sc/⁴⁴Sc-radiopharmaceuticals to distant PET centers. In addition, images could be acquired over longer periods. Finally, the stable co-ordination of Sc with 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) allows the application of the same targeting agents as will subsequently be used for therapeutic applications.^{16 43}Sc/⁴⁴Sc may, therefore, be employed for diagnosis, as well as for planning and monitoring targeted radionuclide therapy with ¹⁷⁷Lu and ⁹⁰Y. The exact matched therapeutic counterpart, ⁴⁷Sc, would be even more appealing, as it can enable the concept of using chemically-identical radiopharmaceuticals with the same kinetic properties for diagnosis and therapy.

Table 1.

Relevant decay characteristics of Sc and Tb isotopes

Isotope	Half-life	β+   Eaverage [keV] (I)	x and γ with I > 10%   E [keV] (I)	β–   Eaverage [keV] (I)	Conv. & Auger electrons(>1 keV) Eaverage [keV] (I)	α  E [keV] (I)
43Sc	3.9 h	476 (88%)	372 (23%)	–		–
44Sc	4.0 h	632 (94%)	1157 (100%)	–		–
47Sc	3.35 d	–	159 (68%)	162 (100%)		–
149Tb	4.1 h	730 (7%)	42–50 (69%), 165 (26%), 352 (29%), etc.	–	32 (85%)	3967 (17%)
152Tb	17.5 h	1140 (20%)	42–50 (65%), 344 (64%)	–	36 (69%)	–
155Tb	5.32 d	–	42–50 (108%), 87 (32%), 105 (25%)	–	19 (204%)	–
161Tb	6.89 d	–	45–53 (39%), 75 (10%)	154 (100%)	19 (227%)	–

Terbium is unique in that it represents radioisotopes for all four modalities in nuclear medicine (Table 1).^{17 155}Tb emits γ-radiation for SPECT imaging and ¹⁵²Tb decays by the emission of positrons useful for PET. The decay of ¹⁶¹Tb is characterized by the emission of low-energy β^¯-particles and γ-rays, similar to ¹⁷⁷Lu but, additionally, comprises a significant number of Auger/conversion electrons (~12 e⁻/decay). It is, therefore, a promising candidate for therapeutic purposes.^{18–20 149}Tb decays by the emission of α-particles, potentially allowing its use for α-therapy.²¹ Since terbium belongs to the group of lanthanides, stable coordination is feasible with DOTA, a macrocyclic chelator that is commonly used for chelation of ¹⁷⁷Lu. The production of Tb nuclides and its chemical (lanthanide) separation are not trivial processes, which may explain why Tb radioisotopes have not been translated to clinical routine yet.

Scandium and terbium are not only related in that they present several medically-interesting radioisotopes, but they also have similar chemical properties, allowing the coordination of the same chelator and, hence, the same targeting agent.

Herein, we report on the production routes and separation methods of Sc and Tb radionuclides and summarize the application of the respective radioisotopes in preclinical settings. The first clinical proof-of-concept studies using Sc and Tb radionuclides are also presented. The availability, challenges of production and the potential of these novel radionuclides for future application in nuclear medicine are discussed.

PRODUCTION OF SCANDIUM AND TERBIUM RADIONUCLIDES

Production of scandium radioisotopes

⁴⁴Sc

The production of ⁴⁴Sc was initially proposed by the development of a ⁴⁴Ti/⁴⁴Sc generator.²² The first such generator allowed elution of ~185 MBq ⁴⁴Sc in a volume of 20 ml.²³ A 10-min post-processing step of the ⁴⁴Sc eluate was established to reduce the volume and hydrochloric acid concentration, as well as to remove oxalate anions.^{22, 24 44}Sc obtained via this production route was applied for the labeling of DOTATOC, allowing its application in a patient for the first time.²⁵ Recently, generator-produced ⁴⁴Sc was also used for the preparation of ⁴⁴Sc-PSMA-617 and injected into prostate cancer patients.²⁶

Alternatively, the production of ⁴⁴Sc was proposed using natural calcium at a cyclotron.²⁷ Activities of up to 650 MBq ⁴⁴Sc were obtained after irradiation of the Ca target for 1 h, however, ^44mSc, ⁴⁷Sc and ⁴⁸Sc were present in the final product as radioisotopic impurities. This may not be of concern for preclinical research, however, due to the long half-lives of these radioisotopes, it would be unfavorable for clinical application.^{27, 28} Krajewski et al reported the production of ⁴⁴Sc by irradiation of enriched ⁴⁴Ca targets at a cyclotron and investigated optimal proton beam energies to obtain the best ratio between ⁴⁴Sc and undesired ^44mSc.²⁹ At the Paul Scherrer Institut (PSI), Switzerland, the production of ⁴⁴Sc from enriched ⁴⁴Ca at a cyclotron was investigated in detail, resulting in the first preclinical application of a ⁴⁴Sc-labeled biomolecule in tumor-bearing mice.³⁰ The irradiation process as well as the separation method based on extraction chromatography was optimized to enable the production of quantities of up to ~2 GBq radionuclidically pure ⁴⁴Sc when using proton beam energies of ~11 MeV.¹⁴ These developments resulted in the first application of cyclotron-produced ⁴⁴Sc, in a clinical proof-of-concept study performed at Bad Berka, Germany.³¹

⁴³Sc

The cyclotron production of ⁴³Sc was proposed by proton irradiation of natural Ca targets for 1–2 h with α-particles, in an optimum energy range of 23.9–28.1 MeV to obtain activities for medical application.^32–35 The ⁴³Sc formation was induced via the nuclear reactions ⁴⁰Ca(α,p)⁴³Sc and ⁴⁰Ca(α,n)⁴³Ti→⁴³Sc. The formed products were of high radionuclidic purity (>99%) and impurities were below 1.5*10⁻⁵% of the ⁴³Sc activity when irradiating enriched ⁴⁰Ca targets. Another option for the production of ⁴³Sc would be the irradiation of enriched ⁴²Ca targets with deuterons using the ⁴²Ca(d,n)⁴³Sc nuclear reaction, however, no experimental data were acquired to date. A drawback of these production routes is the limited availability of cyclotrons with high-current α-beams, as well as cyclotrons providing deuteron beams.^{32, 34}

Recently, the production of ⁴³Sc was investigated by comparing two different production routes based on proton irradiation of enriched ⁴⁶Ti and ⁴³Ca target material, respectively.¹⁵ The production via the ⁴⁶Ti(p,α)⁴³Sc nuclear reaction yielded a modest ~200 MBq ⁴³Sc of high radionuclidic purity (>98%) at the end of a 7 h irradiation period when using 97% enriched ⁴⁶Ti.¹⁵ The production via the ⁴³Ca(p,n)⁴³Sc nuclear reaction resulted in higher quantities of ⁴³Sc, but the product consisted of a mixture of ⁴³Sc and ⁴⁴Sc at an activity ratio of 2:1, when using 57.9% enriched ⁴³CaCO₃.¹⁵ PET images of Derenzo phantoms were acquired with ⁴⁴Sc obtained from irradiated ⁴⁴Ca targets, as well as with ⁴³Sc obtained via the ⁴⁶Ti-based production route or as a mixture with ⁴⁴Sc from ⁴³Ca irradiations. In line with the smaller positron energy of ⁴³Sc, the resolution was found to be the best for pure ⁴³Sc [full width at half maximum (FWHM) = 1.87 ± 0.14 mm], followed by the ⁴³Sc/⁴⁴Sc mixture (FWHM = 2.04 ± 0.06) and ⁴⁴Sc (FWHM = 2.12 ± 0.04) (Figure 1).³⁶

Figure 1.

Transversal sections of PET scans of Derenzo phantoms (hole diameter ranging from 0.8 to 1.3 mm in 0.1 mm increments; 1.0 and 1.3 indicate the holes of 1.0 mm and 1.3 mm diameter, respectively) filled with (a) > 99% ⁴⁴Sc, (b) 66.2% ⁴³Sc/33.3% ⁴⁴Sc and (c) 98.2% ⁴³Sc. The acquisition of the PET scans was performed within the energy window of 400–700 keV for 30 min, in order to obtain a total number of ~6 × 10⁷ coincidences. Figure adapted from Domnanich et al. EJNMMI Radiopharmacy and Chemistry 2017; 2:14.¹⁵ PET, positron emission tomography.

⁴⁷Sc

The production of ⁴⁷Sc can be achieved via a number of different nuclear reactions using a cyclotron, reactor or electron linear accelerator. Misiak et al reported on the cyclotron production of ⁴⁷Sc via the ⁴⁸Ca(p,2n)⁴⁷Sc nuclear reaction.³⁷ Using a proton energy range of 24→17 MeV would be optimal for this purpose, however, the radionuclidic purity using this route was determined to be only ~87%. Using enriched ⁴⁸Ca for irradiation with 20 MeV protons may be a feasible route for the production of GBq activity levels of ⁴⁷Sc, however, the high cost of enriched ⁴⁸Ca has been prohibitive to investigate this production route to date.³⁷ An alternative route investigated the α-particle irradiation of ⁴⁴Ca targets at a cyclotron, inducing the ⁴⁴Ca(α,p)⁴⁷Sc nuclear reaction, but it provided a comparatively lower yield and radionuclidic purity.³⁵

Domnanich et al investigated the production of ⁴⁷Sc by irradiation of ⁴⁶Ca targets in a high neutron flux reactor using the ⁴⁶Ca(n,γ)⁴⁷Ca→⁴⁷Sc nuclear reaction.³⁸ The produced ⁴⁷Ca decays to ⁴⁷Sc with a half-life of 4.5 days, hence, presenting an elegant generator-like concept enabling multiple separations of in-grown ⁴⁷Sc in an analogous process as described for the cyclotron-produced ⁴⁴Sc.¹³ The same authors also reported on the irradiation of ⁴⁷Ti at a reactor to obtain ⁴⁷Sc via the ⁴⁷Ti(n,p)⁴⁷Sc nuclear reaction. This production route yielded low activities of ⁴⁷Sc which was accompanied with a high percentage of co-produced ⁴⁶Sc, however. Both characteristics are not desirable for radiopharmaceutical applications.¹³ Similar findings were reported elsewhere.^{39, 40}

The method of ⁴⁷Sc production using electron linear accelerators (linacs) has also been proposed in the literature.^{41, 42} Mamtimin et al studied the production of ⁴⁷Sc via the ⁴⁸Ti(γ,p)⁴⁷Sc nuclear reaction by Monte Carlo simulations,⁴² while Yagi et al and Rotsch et al reported preliminary experimental results obtained with enriched ⁴⁸Ti.^{43, 44} Irradiation of natural Ti targets would result in many different Sc radioisotopes and, thus, low specific activity of ⁴⁷Sc. Employing enriched ⁴⁸Ti targets when irradiating with 22 MeV beam would maximize the ⁴⁷Sc production and, simultaneously, minimize the formation of other unwanted Sc radionuclides.⁴² Another option, reported by Starovoitova et al and Rane et al, was to use an electron linac to produce ⁴⁷Sc via the ⁴⁸Ca(γ,n)⁴⁷Ca→⁴⁷Sc nuclear reaction.^{41, 45}

Production of terbium radioisotopes

¹⁴⁹Tb, ¹⁵²Tb, ¹⁵⁵Tb

Based on calculations of excitation functions for light and heavy ion-induced reactions, Beyer et al suggested the ¹⁵²Gd(p,4n)¹⁴⁹Tb nuclear reaction as a promising route for the ¹⁴⁹Tb production.²¹ Irradiation of even highly-enriched ¹⁵²Gd targets (which are not currently available) would presumably result in a mixture of various Tb radioisotopes, however.¹⁷ Alternatively, ¹⁴⁹Tb could be produced via the indirect ¹⁴²Nd(¹²C,5n)¹⁴⁹Dy→¹⁴⁹Tb or directly via the ¹⁴¹Pr(¹²C,4n)¹⁴⁹Tb nuclear reaction, but with only limited production yields and radionuclidic purity.²¹ In initial experiments at the heavy ion cyclotron in Dubna, Russia, ¹⁴⁹Tb was produced by irradiation of a ^natNd₂O₃ target for 1.25 h utilizing the indirect route to obtain a moderate yield of 2.7 MBq ¹⁴⁹Tb.²¹ The production of ¹⁵²Tb at a cyclotron via the ¹⁵²Gd(p,n)¹⁵²Tb nuclear reaction would be problematic as well, because of the lack of highly-enriched target material and the formation of radioisotopic impurities. Proton irradiation of enriched ¹⁵⁵Gd via the ¹⁵⁵Gd(p,n)¹⁵⁵Tb reaction should be feasible to produce clinically-relevant quantities of ¹⁵⁵Tb, however.⁴⁶

¹⁴⁹Tb, ¹⁵²Tb and ¹⁵⁵Tb were successfully produced by the high-energy proton irradiation of a tantalum foil target to induce spallation.¹⁷ The spallation products were released from the target and ionized. Mass separation of the isotopes, along with their isobars and pseudoisobars, was carried out using the online mass separator at ISOLDE/CERN, Switzerland, as previously reported,¹⁷ thereby, allowing the collection of the desired radionuclides in zinc-coated gold foils.^{47, 48} After dissolution of the zinc layer of the foil, the ¹⁴⁹Tb, ¹⁵²Tb or ¹⁵⁵Tb were separated from the matrix and its isobar and pseudoisobar impurities, respectively, using cation exchange chromatography and α-hydroxyisobutyric acid as eluent. This procedure provided the final product in α-hydroxyisobutyric acid solution at pH 4.75 enabling direct radiolabeling of tumor-targeting ligands.¹⁷

¹⁶¹Tb

The production of ¹⁶¹Tb was proposed by Lehenberger et al using the ¹⁶⁰Gd(n,γ)¹⁶¹Gd→¹⁶¹Tb nuclear reaction, a similar concept to the production of no-carrier added ¹⁷⁷Lu (Figure 2).¹⁹ Highly-enriched ¹⁶⁰Gd targets were irradiated for 2–3 weeks at the spallation-induced neutron source at PSI or for 1 week at the high neutron flux nuclear reactor at Institut Laue-Langevin, France.¹⁷ The product was separated from the Gd target material using ion exchange chromatographic methods to ensure provision of the final product in a small volume of dilute hydrochloric acid, as is the case for the commercially available ¹⁷⁷Lu.^{17, 19}

Figure 2.

(a) Production route of no-carrier-added ¹⁷⁷Lu via the ¹⁷⁶Yb(n,γ)¹⁷⁷Yb→¹⁷⁷Lu nuclear reaction. (b) Analogous production route of ¹⁶¹Tb via the ¹⁶⁰Gd(n,γ)¹⁶¹Gd→¹⁶¹Tb nuclear reaction. Figure adapted from Karlsruhe nuclide chart, 8th edition, 2012 (https://www.nucleonica.com/).

PRECLINICAL APPLICATIONS

Scandium radioisotopes

In order to obtain stable conjugation of scandium radioisotopes with tumor-targeting agents, the topic of chelation had to be addressed. Various open-chain polyamino-polycarboxylate ligands and macrocycles were investigated for their suitability to coordinate Sc.^{16, 49} It was determined that open-chain chelators formed complexes with Sc much faster than DOTA.⁴⁹ On the other hand, DOTA bound Sc nuclides more efficiently, as compared to 1,4 7-triazacyclononane-1,47-triacetic acid (NOTA) and 1,4,7-triazacyclodecane-1,4,7-triacetic acid (10-ane).¹⁶ DOTA chelators with one methylphosphonic/phosphinic acid pendent arm (DO3AP, DO3AP^PrA and D3AP^ABn) were investigated as alternative chelators.⁵⁰ Another possibility for the coordination of Sc could be to use the 1,4-bis(carboxymethyl)−6-[bis(carboxymethyl)]amino-6-methylperhydro-1,4-diazepine (AAZTA) chelator, allowing radiolabeling under mild conditions (room temperature).⁵¹ Chakravatry et al labeled a CHX-A″-DTPA-functionalized antibody Fab fragment at room temperature within 45 min, resulting in >60% yield at an appreciable specific activity.⁵² CHX-A″-DTPA is known to form more stable complexes than DTPA chelators and, hence, demonstrated stable coordination of ⁴⁴Sc in mouse serum over 6 h at 37°C.⁵²

⁴⁴Sc

⁴⁴Sc was used for the radiolabeling of a variety of ligands, including somatostatin analogs,^53–55 a bombesin analog,⁵⁶ RGD peptides,⁵⁷ folate derivatives³⁰ and PSMA ligands.⁵⁸ Pre-clinical experiments, including PET imaging, were performed in mice bearing tumors that express the target of interest. Eigner et al reported on the use of generator-produced ⁴⁴Sc, in combination with a DOTA-derivatized puromycin.⁵⁹ It allowed the visualization of Walker carcinoma 256 (rat breast carcinoma) and AT1 carcinoma (subline of Dunning R3327 rat prostate carcinoma) tumors via PET imaging in rats due to protein synthesis-specific accumulation of ⁴⁴Sc-DOTA-puromycin.⁵⁹ Among the first pre-clinical experiments performed with cyclotron-produced ⁴⁴Sc was the study reported by Müller et al, where ⁴⁴Sc-folate was investigated in vitro using KB cells (human cervical cancer cells) as well as in KB tumor-bearing mice.³⁰ In this study, it was also shown that the tissue distribution profile of ⁴⁴Sc-folate was equal to that of ¹⁷⁷Lu-folate. The authors concluded that ⁴⁴Sc could serve for dosimetry purposes prior to ¹⁷⁷Lu-based radionuclide therapy in a clinical setting.³⁰ A similar pre-clinical setting was employed by the same authors using PSMA-617 for labeling with ⁴⁴Sc, ¹⁷⁷Lu and ⁶⁸Ga to compare the radionuclides’ influence on the overall in vitro and in vivo behavior.⁵⁸ It was found that ⁴⁴Sc-PSMA-617 most closely resembled the ¹⁷⁷Lu-labeled version, showing almost identical in vitro properties and biodistribution data and, hence, representing the concept of radiotheranostics. Domnanich et al demonstrated that a DOTA chelator is more suitable for ⁴⁴Sc complexation compared to NODAGA by using two pairs of peptides, DOTA/NODAGA-RGD and DOTA/NODAGA-NOC, respectively.⁵⁵ In this regard, the behavior of ⁴⁴Sc was opposite to that of ⁶⁸Ga, which exhibited an increased stability when using NODAGA-functionalized peptides.⁵⁵ While labeling of small molecules and peptides with ⁴⁴Sc was exemplified many times, only few studies were reported about the ⁴⁴Sc-labeling of proteins. Among those was the application of ⁴⁴Sc for the labeling of a CHX-A’’-DTPA-functionalized Fab fragment of Cetuximab, a chimeric human-murine IgG1 monoclonal antibody that binds specifically to the epidermal growth factor receptor.⁵² Serial PET scans over a 6 h period revealed rapid and epidermal growth factor receptor-specific accumulation in U87MG tumors (human glioblastoma cell line) in a xenograft mouse model. Renal clearance was prominent, resulting in extensive accumulation of activity in the kidneys, whereas excretion via the hepatobiliary route was also observed.⁵² Based on these promising results, the development of new ⁴⁴Sc-based radiopharmaceuticals for immunoPET imaging was encouraged by the authors.⁵² Another study investigated ⁴⁴Sc in combination with an anti-HER2 affibody molecule (DOTA-Z_HER2:2891).⁶⁰ Affibody molecules are small scaffold proteins (58 amino acids, 7 kDa), known to refold after denaturation at elevated temperatures, hence, labeling at 95°C was possible.⁶¹ The ⁴⁴Sc-DOTA-Z_HER2:2891 was investigated in vitro and in vivo, using HER2-positive SKOV-3.ip tumor cells (human ovarian cancer cells) and in mouse SKOV-3.ip xenograft models. HER2-specific tumor uptake was demonstrated in biodistribution studies and increased tumor-to-background contrast was achieved at delayed time points (6 h after injection), due to the longer half-life of ⁴⁴Sc as compared to ⁶⁸Ga.⁶⁰

⁴³Sc

Radiolabeling of DOTA and DOTA-functionalized somatostatin receptor analogs (DOTATATE and DOTANOC) using ⁴³Sc has been successfully demonstrated,^{15, 32,35} however, pre-clinical studies using this radionuclide have not been published to date. In a recent study performed at PSI by the authors of this article, ⁴³Sc was used for the labeling of PSMA-617. ⁴³Sc-PSMA-617 was injected into a mouse bearing PC-3 PIP/flu tumors (PSMA-positive and PSMA-negative prostate cancer cells, respectively) for PET/CT imaging at different time points after injection (unpublished data), as was previously reported by Umbricht et al for ⁴⁴Sc-PSMA-617.⁵⁸ The same was also done with ⁴⁴Sc- and ⁴⁷Sc-labeled PSMA-617 for PET and SPECT imaging, respectively. The obtained images of mice that received ⁴³Sc-, ⁴⁴Sc- and ⁴⁷Sc-labeled PSMA-617 are shown in Figure 3. The in vivo behavior of these chemically-identical radioligands is the same, independent of the Sc radioisotope employed. The result may be slightly improved in the case of ⁴³Sc when compared to ⁴⁴Sc, however, this was only visible in pre-clinical phantom studies.¹⁵ More importantly, there is the possibility of reducing the dose burden when using ⁴³Sc, due to the absence of high-energy γ-radiation as emitted by ⁴⁴Sc.

Figure 3.

PET/CT and SPECT/CT images as MIPs of mice 2 h after injection of (a) ⁴³Sc-PSMA-617, (b) ⁴⁴Sc-PSMA-617 and (c) ⁴⁷Sc-PSMA-617. PSMA-=PC-3 flu tumor; PSMA+=PC-3 PIP tumor, Ki = kidney, Bl = urinary bladder. The PET/CT and SPECT/CT images were prepared using VivoQuant software. The scales of PET and SPECT images were cut by 1 and 10%, respectively, to make the accumulated activity in tumors and kidneys better visible (unpublished results). BI, urinary bladder; Ki, kidney; MIPs, maximum intensity projections; PET, positron emission tomography; PSMA, prostate-specific membrane antigen; SPECT, single photon emission computed tomography.

⁴⁷Sc

The concept of theranostic application (PET/β⁻-therapy) with the matched pair of ⁴⁴Sc/⁴⁷Sc radionuclides was proposed and exemplified in a preclinical study using a DOTA-folate conjugate.^{13 44}Sc-folate resulted in distinct visualization of FR-positive KB tumors using PET. The therapeutic potential of ⁴⁷Sc-folate was successfully demonstrated by a delay in tumor growth and an increased survival time of treated mice, compared to an untreated control group.¹³

Terbium radioisotopes

Preclinical application of all four terbium radioisotopes has been reported by Müller et al in 2012, when a proof-of-concept study was performed with a DOTA-folate conjugate in a folate receptor-positive tumor mouse model.¹⁷ This study demonstrated the feasibility of using ¹⁵²Tb and ¹⁵⁵Tb for PET and SPECT imaging, respectively, as well as the potential of using ¹⁴⁹Tb and ¹⁶¹Tb for α- and β⁻-/Auger -electron therapy, respectively.

¹⁵²Tb

Due to an increased production yield achieved in recent years at ISOLDE/CERN, Switzerland, it was possible to investigate ¹⁵²Tb in a more detailed study using a mouse model of somatostatin receptor-positive AR42J rat tumor xenografts (rat pancreatic cancer cells).⁶² DOTANOC, a clinically-established somatostatin analog, was labeled at variable specific activity and used for imaging purposes using a pre-clinical PET/CT scanner.⁶² The tissue distribution profile of ¹⁵²Tb-DOTANOC was equal to that of ¹⁷⁷Lu-DOTANOC when applied to in the same animal model, showing favorable tumor-to-background ratios at higher specific activities.⁶² These results confirmed the assumption that the ¹⁵²Tb-labeled compound would show the same in vivo behavior as its ¹⁷⁷Lu-labeled match and could, therefore, be used for imaging and dosimetry prior to ¹⁷⁷Lu-based radionuclide therapy. In this regard, and in view of the application in combination with longer-circulating radioligands and antibodies, the 17.5 h half-life of ¹⁵²Tb would be considered an advantage.⁶²

¹⁵⁵Tb

The γ-ray energies of ¹⁵⁵Tb make this radionuclide well suited for SPECT imaging (Table 1). Derenzo phantom studies performed with ¹⁵⁵Tb showed an excellent spatial resolution compared to that obtained with the clinically-established SPECT nuclide ¹¹¹In (T_1/2 = 2.80 d, Eγ=171 keV and 245 keV; Figure 4).⁶³ SPECT/CT imaging was investigated using fast-clearing DOTA-peptides, including a minigastrin-based peptide and DOTATATE, as well as longer-circulating biomolecules such as an albumin-binding DOTA-folate (cm09) and the L1-cell adhesion molecule (L1-CAM)-antibody chCE7. These biomolecules were labeled with ¹⁵⁵Tb and used for preclinical imaging of tumor-bearing mice 4 h after injection of the radiolabeled peptides or 2 days and 3 days after injection of ¹⁵⁵Tb-folate and ¹⁵⁵Tb-chCE7, respectively.⁶³ The images were of high quality, enabling visualization of even small lesions in the abdominal region, as well as in the liver of an intraperitoneal SKOV-3.ip tumor mouse model, days after injection of ¹⁵⁵Tb-chCE7. ¹⁵⁵Tb was proposed as an alternative option to ¹¹¹In for dosimetry planning prior to the application of lanthanide-based radionuclide therapy.⁶³

Figure 4.

SPECT images of Derenzo phantoms filled with (a) ¹⁵⁵Tb (2.6 MBq) and (b) ¹¹¹In (4 MBq). Figure adapted from Müller et al. Nucl Med Biol 2014;41 Suppl:e58-65.3]⁶³ SPECT, single photon emission computed tomography.

¹⁴⁹Tb

The first pre-clinical application of ¹⁴⁹Tb was reported by Beyer et al.⁶⁴ A SCID mouse model of leukemia was used to demonstrate the effect of ¹⁴⁹Tb-labeled rituximab, a CD20-targeted antibody outfitted with CHX-A-DTPA chelators.⁶⁴ Mice were administered with the radioimmunoconjugate (5.5 MBq, 5 µg per mouse) 2 days after intravenous injection of 5 million Daudi cells (human Burkitt lymphoma cells) before the appearance of manifested tumors. In 89% of the treated mice, the therapy resulted in tumor-free survival of more than 4 months, a significant increase in survival time as compared to the other groups.⁶⁴ Untreated animals developed signs of Burkitt lymphoma, which required euthanasia within 37 days.⁶⁴ Animals treated with a low dose of unlabeled rituximab (5 µg  per mouse) had to be euthanized within 43 days, while the treatment with a high dose of rituximab (300 µg per mouse) increased the lifetime, resulting in a median survival of 100 days. Preliminary dose estimation for humans reported by Beyer et al revealed that a therapeutic activity of 5 GBq ¹⁴⁹Tb-rituximab would result in a bone marrow radiation dose far below the critical level.^{64 149}Tb was also used for targeted radionuclide therapy using a DOTA-folate conjugate in a proof-of-concept study with KB tumor-bearing mice.¹⁷At a later stage, the same model was used to investigate two different activity levels (2.2 MBq and 3.0 MBq per mouse, respectively) of ¹⁴⁹Tb-folate. A dose-dependent effect was observed with regard to the tumor growth delay, as well as the median survival time which was increased to 30.5 day and 43 days, respectively, as compared to a median survival of 21 days in the case of untreated control animals.⁶⁵ The therapeutic effect of ¹⁴⁹Tb-folate was also confirmed in vitro, where the radiofolate reduced the tumor cell survival in a FR-specific manner.⁶⁵ More recently, it was demonstrated for the first time in a pre-clinical setting that, based on the coemission of positrons, ¹⁴⁹Tb can be used for PET imaging.⁶⁶ An AR42J tumor-bearing mouse was injected with ¹⁴⁹Tb-DOTANOC (7 MBq) followed by PET/CT imaging 2 h later. The high-quality images allowed distinct visualization of the tumor xenografts (Figure 5). The concept of α-PET is entirely novel and would make ¹⁴⁹Tb attractive for future clinical translation, potentially allowing the visualization of applied ¹⁴⁹Tb-radioligands.

Figure 5.

MIP of PET/CT image of AR42J tumor-bearing mouse 2 h after injection of ¹⁴⁹Tb-DOTATOC (7 MBq). Figure adapted from Müller et al. EJNMMI Radiopharmacy and Chemistry 2016;1:5.⁶⁶ BI, urinary bladder; Ki, kidney; MIP, maximum intensity projection; PET, positron emission tomography; Tu, tumor.

¹⁶¹Tb

De Jong et al reported on the first preclinical study performed with ¹⁶¹Tb in 1995 using ¹⁶¹Tb-octreotide to investigate the biodistribution in normal rats.⁶⁷ It was found that ¹⁶¹Tb-octreotide cleared faster from the blood than ¹¹¹In-octreotide, however, liver uptake was higher for the ¹⁶¹Tb-labeled version.⁶⁷ The uptake in somatostatin receptor-positive tissues, such as CA20948 tumor (rat pancreatic cell line), pancreas and adrenals, was shown to be specific. Based on the results of this study, the authors concluded that ¹⁶¹Tb would be of interest for intraoperative imaging and radionuclide therapy.⁶⁷ The first therapy study was reported in 2012, with a small number of KB tumor-bearing mice using a ¹⁶¹Tb-folate for a proof-of-concept investigation, where its potential for tumor growth inhibition was demonstrated (Figure 6).¹⁷

Figure 6.

Pre-clinical therapy study: the graph represents the tumor growth relative to the average tumor size (tumor volume) determined at Day 0, which was set to 1 (indicated as relative tumor size). The study was performed with five untreated mice (controls: green/blue) and five mice treated with ¹⁶¹Tb-folate (11 MBq/mouse; red/orange). In the treated group, four of the five mice showed complete tumor remission as shown by overlapping graphs. This research was originally published in JNM, Müller et al.¹⁷ PET, positron emission tomography; SPECT, single photon emission computed tomography.

The imaging potential using pre-clinical SPECT, as well as the therapeutic effect of ¹⁶¹Tb-folate, was investigated in a one-to-one comparison with ¹⁷⁷Lu-folate and FR-positive KB and IGROV-1 tumor xenografts, respectively.⁶⁸ In line with the increased mean absorbed tumor dose after administration of ¹⁶¹Tb-folate, it was more effective to treat KB tumors than ¹⁷⁷Lu-folate applied at the same activity. As a consequence, the median survival time in ¹⁶¹Tb-folate treated mice was increased (54 days) as compared to mice treated with ¹⁷⁷Lu-folate (35 day) or control mice (31 days) in the KB tumor models. The same held true for IGROV-1 tumor mice, even though the difference in tumor growth delay and survival time was less pronounced. Both radiofolates produced high-resolution pre-clinical images of the tissue distribution profile in tumor-bearing mice. More recently, Haller et al conducted a study over 8 months in order to investigate and compare potential undesired side effects to the kidneys after ¹⁶¹Tb- and ¹⁷⁷Lu-folate therapy, respectively.⁶⁹ The evaluation of kidney function, as well as histopathological analysis of the renal tissue, revealed a dose-dependent damage with increasing scores of 1, 3 and 4 after administration of ¹⁶¹Tb-folate at 10, 20 and 30 MBq, respectively. These results were comparable to the damage (scored as 2, 3 and 4, respectively) after application of¹⁷⁷Lu-folate at the same activities. The authors concluded that Auger electrons—even though they contributed to the mean absorbed kidney dose—did not cause additional renal damage.⁶⁹ The anti-L1-CAM antibody, chCE7, was also labeled with ¹⁶¹Tb for direct comparison with the ¹⁷⁷Lu-labeled version.^{70 161}Tb-chCE7 showed higher radiotoxicity in nude mice (maximal tolerated dose: 10 MBq per mouse) as compared to ¹⁷⁷Lu-chCE7 (maximal tolerated dose: 12 MBq per mouse). At equitoxic doses, ¹⁶¹Tb-chCE7 was, however, more effective tin the growth delay of IGROV-1 tumors when compared to ¹⁷⁷Lu-chCE7.⁷⁰

CLINICAL PROOF-OF-CONCEPT STUDIES

Scandium-44

The first clinical application of generator-produced ⁴⁴Sc was performed in 2009.²² PET scans were acquired of a patient injected with ⁴⁴Sc-DOTATOC, resulting in high-quality images up to 18 h after injection.²² More recently, ⁴⁴Ti generator-produced ⁴⁴Sc was employed for a first-in-man study of using ⁴⁴Sc-PSMA-617.²⁶ Four male patients with histologically-proven prostate carcinoma, who were scheduled for ¹⁷⁷Lu-PSMA-617 therapy, were selected for the study. After the intravenous administration of ⁴⁴Sc-PSMA-617 (50.5 ± 9.3 MBq), the patients underwent a dynamic PET scan of the abdomen over a period of 30 min, followed by static PET/CT scans from skull to midthigh up to 18 h after injection (Figure 7).²⁶

Figure 7.

MIP (top) and representative sections (bottom) of PET/CT examination of a patient suffering from mCRPC, with high tumor load, using (a) ⁴⁴Sc-PSMA-617 (50 MBq, 60 min p.i.), and (b) ⁶⁸Ga-PSMA-11 (120 MBq, 60 min p.i.). (c) On the right hand side, the planar scintigraphy is shown (top) and a representative section of the post-therapy SPECT/CT scan, about 24 h after application of 6.7 GBq ¹⁷⁷Lu-PSMA-617. Figure adapted from Eppard et al. Theranostics 2017;7:4359–69.²⁶ mCRPC, metastasized castration-resistant prostate cancer; MIP, maximum intensity projection; PET, positron emission tomography; PSMA, prostate-specific membrane antigen.

Accumulation of radioactivity was found in multiple soft tissue and skeletal metastases, whereas physiological radioligand uptake was observed in the kidneys, liver, spleen, small intestine, urinary bladder and salivary glands. The kidneys received the highest absorbed radiation dose (0.354 mSv/MBq) after application of ⁴⁴Sc-PSMA-617.²⁶ The PET images obtained at 2 h after administration of ⁴⁴Sc-PSMA-617 were compared to those obtained earlier using ⁶⁸Ga-PSMA-11. Visual assessment of the PET scan acquired with ⁴⁴Sc-PSMA-617 was at least of equal quality as the one obtained previously with ⁶⁸Ga-PSMA-11 (Figure 7). Quantitative analysis revealed no differences between uptake of ⁴⁴Sc-PSMA-617 and ⁶⁸Ga-PSMA-11 in most organs, however, reduced accumulation of ⁴⁴Sc-PSMA-617 was observed in the kidneys when compared to ⁶⁸Ga-PSMA-11.²⁶

Singh et al were the first to report on a proof-of-concept imaging study in patients using ⁴⁴Sc produced at a cyclotron.³¹ Due to the high yield (~2 GBq) obtained from this production method, it was possible to ship ⁴⁴Sc over a distance of 600 km from the production site at PSI, Switzerland, to Zentralklinik Bad Berka, Germany, requiring two half-lives’ travel time. At Bad Berka, ⁴⁴Sc was used for the labeling of DOTATOC.³¹ Two male patients with well-differentiated, non-functional and functional neuroendocrine neoplasm, respectively, were selected for this study. After injection of 78 MBq and 96 MBq ⁴⁴Sc-DOTATOC, eight sets of PET/CT scans were acquired at variable time points over a period of 24 h.³¹ Distinct uptake of radioactivity was observed in malignant lesions already at early time points, with best tumor-to-background at about 4 h after injection.

The tissue distribution was somewhat different when compared with that of ⁶⁸Ga-DOTATOC (Figure 8), which may be attributed to the low specific activity of ⁴⁴Sc-DOTATOC (~1.4 MBq/µg, >55 µg). The authors stated that decreasing the amount of injected peptide to <50 µg would be the aim for future studies in this regard. ⁴⁴Sc-DOTATOC showed significant potential for PET/CT imaging and as a suitable diagnostic match to currently-used therapeutic radionuclides such as ¹⁷⁷Lu and ⁹⁰Y - as well as ⁴⁷Sc, once it becomes commercially available.³¹

Figure 8.

Comparison of serial images of the transverse section of liver, representing the lesion in segment VII (green arrow), obtained by PET/CT imaging of Patient 1 using somatostatin analogs for restaging after the third cycle of PRRT. (a) At 18 months after PRRT, the lesion was not detected on PET/CT image obtained with ⁶⁸Ga-DOTATOC; (b) at 27 months after PRRT, the lesion was detected using ⁴⁴Sc-DOTATOC; (c) a concurrent MRI performed within 24 h of the PET/CT scan obtained with ⁴⁴Sc-DOTATOC co-registered the lesion seen on the PET/CT image; (d) at 36 months after PRRT, the lesion was detected on PET/CT images obtained with ⁶⁸Ga-DOTATOC. Figure adapted from Singh et al.³¹ Cancer Biother Radiopharm 2017;32:124–32 [31] PET, positron emission tomography; PRRT, peptide receptor radionuclide therapy.

Terbium-152

The production of ¹⁵²Tb at ISOLDE/CERN, followed by the chemical separation from catcher foils and impurities at PSI, resulted in a pure product in sufficient yields for use in a first-in-man application in 2016.^{71 152}Tb was used for labeling of DOTATOC at Zentralklinik, Bad Berka, and administered to a male patient with diagnosed metastatic, well-differentiated, functional neuroendocrine neoplasm of the terminal ileum.⁷¹ PET/CT scans were performed for restaging of the disease 8 years after the sixth cycle of peptide receptor radionuclide therapy. The scans were acquired over a period of 24 h after injection of 145 MBq ¹⁵²Tb-DOTATOC.⁷¹ The resulting images allowed visualization of even small metastases, with increased tumor-to-background contrast at later time points (Figure 9). It was concluded that ¹⁵²Tb would be of particular value for dosimetry prior to radionuclide therapy, due to its much longer half-life (17.5 h) enabling delayed PET scans, as compared to the very short half-life (68 min) of ⁶⁸Ga which cannot serve this purpose.⁷¹

Figure 9.

PET/CT images of a patient with neuroendocrine neoplasm of the terminal ileum obtained at 2 h after injection of ¹⁵²Tb-DOTATOC. (a) MIP image shows the Ki, the UB, the Sp and the IS. (b/c) Transverse sections of PET/CT fusion images demonstrate radiopeptide uptake in lymph node metastases in the right costophrenic region and in the right internal mammary chain (green arrows), as well as in Segment 7 of the liver (blue arrows) and in a skeletal metastasis in the left third rib adjacent to the sternocostal junction (red arrows). Figure adapted from Baum et al. Dalton Trans 2017;46:14638–46.⁷¹ IS, injection site; Ki, kidneys; MIP, maximum intensity projection; Sp, spleen; UB, urinary bladder.

CONCLUSION

Based on the results of (pre)clinical studies, it is clear that ⁴⁴Sc holds great promise for future applications as a novel PET nuclide for diagnostic imaging and for monitoring therapy in clinics. Whether ⁴³Sc will replace ⁴⁴Sc in future is unclear and critically dependent on the possibilities of producing it in large quantities and at reasonable cost. The investigation of ⁴⁷Sc for therapy is still in its infancy and a potential utility for this novel β⁻-particle-emitting radionuclide has to be defined.

Among the Tb radionuclides, ¹⁶¹Tb is undoubtedly the most advanced in terms of production and pre-clinical investigation. The similarity of ¹⁶¹Tb to ¹⁷⁷Lu makes this novel radiolanthanide highly attractive for therapeutic application, potentially enabling a more effective therapy due to the coemission of Auger electrons while using the same targeting agents. ¹⁵⁵Tb could be produced at a cyclotron with potentially high yields; significant research efforts in this regard will be necessary, however. As a diagnostic match, ¹⁵⁵Tb may be of particular value for visualization and staging of malignancies and for dosimetry calculation prior to therapy with clinically employed and new radiolanthanides, such as ¹⁷⁷Lu and ¹⁶¹Tb, respectively.

¹⁵²Tb and, in particular, the α-particle emitting ¹⁴⁹Tb are of interest for clinical application, however, the production of these Tb radioisotopes remains challenging and will only be possible at specific production sites where a mass separator is available. Even though such facilities are currently under construction at several sites worldwide, the general availability of these nuclides may remain scarce, in particular, the short-lived ¹⁴⁹Tb which can only be used in close proximity to the production site.

The question on whether it would be beneficial or even essential, to use chemically identical radioligands for diagnosis and therapy is a controversial topic in the community and can only be addressed once the concept of “matched pair” radionuclides is realized in a clinical setting. Taking all relevant aspects into consideration, scandium provides the most promising diagnostic match whereas, in the case of terbium, the therapeutic radioisotopes are of most clinical interest. It is, therefore, possible and likely that ^43/44Sc could be used in tandem with ^149/161Tb, since both elements can be stably coordinated with a DOTA chelator. This means that using the same tumor targeting agent for Sc/Tb-based radiotheranostics would be feasible.

Both scandium and terbium remain of utmost interest for further development, with regard to routine production in large quantities for future application in nuclear medicine. Basic research and more detailed (pre)clinical investigations will be required, to enable the specific aims of individual dose plans for personalized radionuclide therapy and the option of therapy monitoring, as well as therapy regimes tailored to their specific medical situation.