At present, the supply of gallium-68 for medical imaging is primarily based on the [68Ge]Ge/[68Ga]Ga-generator. As these commercial generators can deliver only a limited amount of activity, and the demand for gallium-68 is increasing with the commercialization of kit-based radiopharmaceuticals, e.g. NETSPOT (DOTATATE), SomakitTOC (DOTATOC) and Illumet (PSMA-11), a future shortage of generators and gallium-68 may occur. Herein, we describe a high-yielding, automated production of this important isotope by irradiating solid targets of enriched 68Zn metal on a biomedical cyclotron. This approach may serve as an important supplement to meet future demand.
By application of silver targets with enriched zinc-68 from ARTMS on the ARTMS QIS target system (Fig. 1), gallium-68 was produced by 13 MeV proton irradiation at beam currents up to 80 µA. After irradiation, the solid target was pneumatically transferred from the cyclotron to a dissolution box in a hot cell (Fig. 2). The automated transfer comprises a very favourable and necessary feature as very high levels of radioactivity from estimated > 370 GBq gallium-68 on the target were produced at EOB. Nelson et al. recently reported on a new method for cyclotron-based gallium-68 production by irradiation of zinc-68 pellets with production yields up to 37.5 ± 1.9 GBq on target (non-purified) (6). However, their method involves manual collection of the irradiated target and subsequent transport of the target to the hot cell in a lead shield by the operator. Such a method would lead to radiation protection issues if much higher radioactivities were routinely produced, as demonstrated in the current study. Here, automated target transfer and subsequent automatic separation and radiolabelling are mandatory to avoid excessive radiation exposure to the operator.
Samples of several productions were analysed according to the draft Ph. Eur. monograph on accelerator produced gallium-68 (5). The results of the tests are summarized in Table 1. Notably, the RCP and RNP were high and only minor amounts of gallium-66/67 were present, while no other radionuclidic impurities were detected by gamma spectroscopy. Prolonged irradiation and higher beam currents expectedly improved the RNP and the AMA as judged by HPGe detection and ICP-OES, only 3.6 ng/GBq Fe and 58 ng/GBq Zn were present in the [68Ga]GaCl3 batch with the highest activity. This is more than 30 times lower than the iron contents of 0.13 ± 0.07 µg/GBq Fe found by Nelson et al. and 0.11 ± 0.07 µg/GBq Fe found by Lin et al., which was the dominating metal impurity in these studies (6, 7). The improvement in the iron content in our method, which is highly important for obtaining a high AMA in subsequent labelling of radioconjugates, is due to the introduction of the second column containing LN resin in the separation, where Fe contaminants are bound. Importantly, the RNP was ≥ 99.89% and allowed for a shelf-life of the [68Ga]Ga chloride of up to 7 hours based on RNP alone.
To estimate the purity before ICP-OES results were obtained, the produced [68Ga]Ga chloride was applied to radiolabel DOTATATE and PSMA-11. High labelling yield and AMA comparable to or higher than what is observed for the generator-produced isotope indicated a low amount of metallic impurities in the accelerator produced gallium-68. An Eckert-Ziegler synthesis module with a modified cassette and sequence directly connected to the separation module was applied to radiolabel PSMA-11 with very high levels of radioactivity (> 39.8 GBq). Several high activity productions with only 100 nmol of PSMA-11 were performed yielding up to 72.2 GBq [68Ga]Ga-PSMA-11 (uncorrected for decay; 23 min synthesis time), which was stabilized with 50 mg sodium ascorbate and 1 mL ethanol in a 12 mL volume. The radioconjugate was tested for stability up to two hours after EOS (by HPLC, colour, TLCs for free and colloid gallium) and found stable in this formulation and very high AMA. RCP was higher than 98.5% after 2 h.
In terms of patient doses, a batch of 70 GBq [68Ga]Ga-PSMA-11 would represent at least 10 doses for 50 min/scan protocol for two simultaneously running PET-scanners. If PET scanner capacity is doubled to four scanners, the batch could deliver up to 20–25 patient doses of [68Ga]Ga-PSMA-11 from a single solid target based gallium-68 production.
Otherwise, a radioconjugate (or purified [68Ga]GaCl3) of this stability and AMA could be transported to decentralized radiopharmacies possibly superseding the need for expensive generators on a continuous basis. Also, a generator carries the risk of germanium-68 breakthrough to the final product and eventually becomes long-lived radioactive waste (half-life of germanium-68: 271 days) after about 400 elutions, which needs to be disposed of. In comparison, no long-lived radioactive waste is produced by the cyclotron production route if the silver backing (containing some cadmium-109) is reused. Conversely, if only a limited amount of gallium-68 activity is needed (i.e. a few patient doses), a generator is much less technically demanding, very reliable and does not require trained personnel to perform separation and radiolabelling on synthesis modules or major capital investment in a cyclotron facility. Liquid cyclotron targets where acidic solutions containing zinc-68 salts are bombarded with protons, provide an alternative solution to get a small number of patient doses but the reported production yields of < 10 GBq at EOB or approximately 5 GBq at EOP are inferior to the solid target route (8–10). This method also carries the inherent risk of acids damaging targets and potentially cyclotron systems and require continuous target maintenance.
Other studies on solid target based [68Ga]GaCl3 productions have been published recently (11). Alnahwi et al. reported on the production of 145 GBq gallium-68, decay-corrected to EOB, by proton irradiation of pressed zinc-68 targets with 35 µA on target (12). However, the use of targets made from pressed zinc powder with expectedly reduced heat transfer compared to solid zinc limits the maximum power density that can be absorbed in the target and thus, the maximum beam current. The resulting purified activity available at EOP for radiolabelling was not stated, although the authors reported an AMA of the produced [68Ga]GaCl3 for a 20 min irradiation to be 28.3 ± 6.8 GBq/µmol. Radiolabelling of 21 nmol DOTATATE with approximately 555 MBq [68Ga]GaCl3 at time of radiolabelling resulted in a labelling efficiency of 95 ± 1.6%, which is similar to the results obtained in the current study. However, the AMA of the [68Ga]Ga-DOTATATE is not reported by the authors but can be estimated to be approximately 26 MBq/nmol at time of labelling. This is in line with the AMA we observed for [68Ga]Ga-DOTATATE using manual labelling procedures but inferior to the AMA of [68Ga]Ga-PSMA-11 (and consequently the AMA of the [68Ga]GaCl3) in our automated syntheses reaching as high as 722 MBq/nmol (range 398–722 MBq/nmol). In comparison, the AMA obtained by Nelson et al. for [68Ga]Ga chloride using DOTA-labelling was reported as 9.5 ± 1.3 GBq/µmol (6), and thus inferior to the AMA obtained in the present study.
In summary, the production method proposed here, to our knowledge, signifies a record high production of purified [68Ga]GaCl3 and of [68Ga]Ga-PSMA-11 with high RNP, RCP and very high AMA. Furthermore, the highly automated method ensures a low radiation burden to the operator, despite the multi-Curie levels of radioactivity produced. As such, this production route features a favourable supplement to generator-produced gallium-68 capable of delivering tens of patient doses for gallium-68 tracers as demonstrated with record high quantities of [68Ga]Ga-PSMA-11.