Expanding the PET radioisotope universe utilizing solid targets on small medical cyclotrons.

Molecular imaging with medical radioisotopes enables the minimally-invasive monitoring of aberrant biochemical, cellular and tissue-level processes in living subjects. The approach requires the administration of radiotracers composed of radioisotopes attached to bioactive molecules, the pairing of which considers several aspects of the radioisotope in addition to the biological behavior of the targeting molecule to which it is attached. With the advent of modern cellular and biochemical techniques, there has been a virtual explosion in potential disease recognition antigens as well as targeting moieties, which has subsequently opened new applications for a host of emerging radioisotopes with well-matched properties. Additionally, the global radioisotope production landscape has changed rapidly, with reactor-based production and its long-defined, large-scale centralized manufacturing and distribution paradigm shifting to include the manufacture and distribution of many radioisotopes via a worldwide fleet of cyclotrons now in operation. Cyclotron-based radioisotope production has become more prevalent given the commercial availability of instruments, coupled with the introduction of new target hardware, process automation and target manufacturing methods. These advances enable sustained, higher-power irradiation of solid targets that allow hospital-based radiopharmacies to produce a suite of radioisotopes that drive research, clinical trials, and ultimately clinical care. Over the years, several different radioisotopes have been investigated and/or selected for radiolabeling due to favorable decay characteristics (i.e. a suitable half-life, high probability of positron decay, etc.), well-elucidated chemistry, and a feasible production framework. However, longer-lived radioisotopes have surged in popularity given recent regulatory approvals and incorporation of radiopharmaceuticals into patient management within the medical community. This review focuses on the applications, nuclear properties, and production and purification methods for some of the most frequently used/emerging positron-emitting, solid-target-produced radioisotopes that can be manufactured using small-to-medium size cyclotrons (≤24 MeV). This journal is © The Royal Society of Chemistry.

Introduction

Positron Emission Tomography (PET) is one of several diagnostic imaging tools used to non-invasively assess metabolic activity in body tissue. PET scanning is sensitive enough to quantify changes in biological processes, making it an ideal candidate for assessing the efficacy of an early treatment plan or newly developed drugs.[1] PET makes use of radiotracers which permit the visual isolation of metabolic processes within specific tissues and cells. These radiotracers are composed of a radioisotope attached to a bioactive molecule (e.g. a small molecule, peptide, antibody, etc.), as demonstrated in Fig. 1. The choice of bioactive targeting molecule mainly depends on the process to be visualized but also on the molecule's ability to incorporate the radioisotope into the chemical structure of the pharmaceutical compound. However, choosing the radioisotope requires consideration of biocompatibility in addition to the assessment of physical and biological half-life suitability for the metabolic process under investigation.[2-4] Over the years, the suitability of several PET radioisotopes for use as tracers has been advanced, along with their unique methods of production.[2,5] There are three major methods that can be used to produce radiometals – parent/progeny generators, nuclear reactors and cyclotrons.[2] For radioisotope production utilizing cyclotrons, radioisotopes can be produced using gas, liquid, or solid phase targets.[2,6] This review will focus on recent advancements in the use of solid targets.

Fig. 1

Overview of PET. Schematic depicting positron annihilation in a patient undergoing a PET scan.

The design of a solid target requires several considerations, and these factors influence its configuration and performance. Targets may be composed entirely of enriched and natural elements and alloys, and may be prepared as foils, disks or composites (through the electroplating, sputtering, sedimenting, or pressing of alloys and elements onto supporting metal substrates (target backing) such as gold, silver, tantalum, titanium, rhodium or niobium).[7-10] The specific design selected depends on factors such as substrate bond, heat transfer efficiency, ease of product isolation from the target material after irradiation, ease of target material recovery for enriched-material targets, product sensitivity to impurities and nuclear cross-section data.[8] An ideal target should be stable, allow a high rate of heat transfer, and possess favorable melting properties[6] since proton-induced irradiation generates significant amounts of heat in a small target area. When heat transfer and melting point are not suitable, innovations such as an oblique target arrangement as well as target alloying practices have been implemented to pacify these limitations.[6] The separation chemistry employed should be fast (i.e. less than one half-life of the product radionuclide) and efficient at separating the product from both the target material and co-produced impurities.[11] Commonly used separation techniques include are ion exchange chromatography, solid phase extraction (SPE) and thermochromatographic dry distillation, the schematics for which can be seen in Fig. 2, Fig. 3 and Fig. 4 respectively.

Fig. 2

Ion exchange chromatography. Schematic depicting the separation of dissolved PET radioisotopes from dissolved, unreacted target material after passage through an ion exchange column. The resin retains the PET radioisotope, and an eluent is used for its elution.

Fig. 3

Solid phase extraction. Schematic depicting the separation of dissolved PET radioisotopes from dissolved, unreacted target material after passage through a solid phase extraction column. The column retains the PET radioisotope and any impurities. Impurities are washed-off before elution is performed to obtain the PET radioisotope.

Fig. 4

Thermochromatographic dry distillation. Schematic depicting the separation of volatile PET radioisotopes from non-volatile target material after heating is performed in a furnace. The PET radioisotope condenses on a removable, cooler surface within the furnace. The radioisotope is removed from the surface using a rinse.

The radioisotopes covered in this review (see Table 1) can be produced on solid targets using small-to-medium sized, proton-accelerating cyclotrons with a maximum particle acceleration energy of 24 MeV. Production using this class of cyclotron was selected since these machines are the most prevalent in medical and research institutions. While the list of radioisotopes reviewed is a comprehensive one, the authors would like to highlight other PET radioisotopes such as 110mIn[12,13], 152Tb,[13,14]51Mn,[15,16]62Cu[16-18] and 132La[19,20] which are also available for production on the aforementioned cyclotrons, but were excluded from this review due to limitations. This review briefly discusses the biomedical application of each radionuclide while focusing mainly on the specifics of its production and purification procedures. Several recent reviews can provide more details on the applications of established and emerging PET radionuclides discussed herein ref. 21–23. A summary of the major nuclear and production properties of each radioisotope is provided in Table 1 for convenience. The nuclear information presented in this review was retrieved from the National Nuclear Data Center website by Brookhaven National Laboratory (https://www.bnl.gov).

Properties of emerging and commonly used PET radionuclides

Element	Isotope	Nuclear properties					Production conditions
		Half-life	Mode of decay (%)	Avg. β+ energy, keV	β+ endpoint energy, keV (%)	Principal γ energies, keV (Abs. %)	Major nuclear reactions	Beam energy, MeV	Target	Max. reported yield, MBq (μA h)−1
Arsenic (As)	72	26.0 h	β+ (87.8)	1117	2500 (64.2) 3334 (16.3)	511 (176.4), 630 (8.1), 834 (81.0), 1464 (1.13)	72Ge(p,n)72As	18–8	Natural or enriched GeO2 on Cu or Nb, Ge foil	90
			EC (12.2)	1528
Bromine (Br)	75	96.7 min	β+ (75.0)	773, 708, 904	1753 (53.0), 1612 (4.9), 2040 (4.0)	511 (149.1),141 (6.6), 287 (88.0), 428 (4.4)	76Se(p,2n)75Br	24–21.5	Enriched Se, enriched selenides (i.e. Ag2Se, CuAgSe, Cu2Se, PbSe)	1480
			EC (25.0)
	76	16.2 h	β+ (55.0)	1532, 375, 1800	3382 (25.8), 871 (6.3), 3941 (6.0)	511 (109), 559 (74.0), 657 (15.9), 1854 (147)	76Se(p,n)76Br	16–8	Enriched Se, enriched selenides (i.e. Cu2Se)	70
			EC (45.0)
Copper (Cu)	60	23.7 min	β+ (93.0)	872, 1325, 840	1982 (49.0), 2947 (15.0), 1912 (11.6)	511 (185.0), 826 (21.7), 1333 (88.0), 1792 (45.4)	60Ni(p,n)60Cu	14.7	Natural or enriched Ni on Au	2146
			EC (7.0)
	61	3.3 h	β+ (61.0)	524, 399	1216 (51.0), 933 (5.8)	511 (123.0), 67 (4.2), 282 (12.2), 656 (10.8)	61Ni(p,n)61Cu	14.7–9	Enriched Ni on Au	281
			EC (39.0)				64Zn(p,α)61Cu	17.6–11.7	Natural or enriched Zn	13
	64	12.7 h	β+ (17.6)	278	653 (17.6)	511 (35.2), 1346 (0.5)	64Ni(p,n)64Cu	15–10	Enriched Ni on Au or Rh	5883
			β− (38.5)
			EC (43.9)
Cobalt (Co)	55	17.5 h	β+ (76.0)	649, 436	1499 (46.0), 1021 (25.6)	511 (152.0), 477 (20.2), 931 (75.0), 1409 (16.9)	56Fe(p,2n)55Co	23.5–16	Enriched Fe	236
			EC (24.0)				58Ni(p,α)55Co	16–15	Enriched Ni, enriched Ni on Ag or Au
Gallium (Ga)	66	9.5 h	β+ (57.0)	1904, 397	4153 (51.0), 924 (3.7)	511 (114.0), 834 (5.9), 1039 (37.0), 2752 (22.7)	66Zn(p,n)66Ga	15–6	Natural or enriched Zn on Au or Ag	700
			EC (43.0)
	68	67.7 min	β+ (88.9)	836, 353	1899 (87.7), 822 (1.2)	511 (177.8), 1077 (3.2), 1261 (0.1), 1883 (0.1)	68Zn(p,n)68Ga	14–11	Enriched Zn, enriched Zn on Al, Au or Ag	1380
			EC (11.1)
Iodine(i)	124	4.2 d	β+ (22.7)	687, 975	1535 (11.7), 2138 (10.7)	511 (45.0), 603 (62.9), 723 (10.4), 1691 (11.2)	124,125Te(p,xn)124I	22–5	Natural or enriched Te, natural or enriched TeO2	111
			EC (77.3)
Manganese (Mn)	52g	5.6 d	β+ (29.4)	242	575 (29.4)	511 (58.8), 744 (90.0), 936 (94.5), 1434 (100.0)	52Cr(p,n)52gMn	20–10	Natural Cr, natural Cr on Cu or Ag	9.6
			EC (70.6)
Niobium (Nb)	90	14.6 h	β+ (53.0)	662, 726	1500 (51.0), 1641 (2.0)	511 (102.0), 142 (66.8), 1129 (92.7), 2319 (82.0)	90,91Zr(p,xn)90Nb	19–7	Natural Zr, enriched ZrO2 on Cu	596
			EC (47.0)
Scandium (Sc)	44g	4.0 h	β+ (94.3)	632	1474 (94.3)	511 (188.5), 1157 (99.9), 1499 (0.9)	44Ca(p,n)44g,mSc	18–6	Natural or enriched Ca, enriched CaO3	50
			EC (5.7)
Technetium (Tc)	94m	52 min	β+ (70.2)	1094, 639, 405	2439 (67.6), 1446 (0.9), 917 (0.9)	511 (140.3), 871 (94.2)	94Mo(p,n)94mTc	13–6	Natural Mo, enriched MoO3	111
			EC (29.8)
Titanium (Ti)	45	184.8 min	β+ (84.8)	439	1040 (84.8)	511 (169.6), 720 (0.2), 1408 (0.1)	45Sc(p,n)45Ti	16–8	Natural Sc	422
			EC (15.2)
Yttrium (Y)	86g	14.7 h	β+ (31.9)	535, 681, 883	1221 (11.9), 1545 (5.6), 1988 (3.6)	511 (64.0), 628 (32.6), 1077 (82.5), 1153 (30.5)	86Sr(p,n)86gY	15.1–11	Enriched SrCO3 or SrO	166
			EC (68.2)
Zinc (Zn)	63	38.5 min	β+ (92.7)	1042, 733, 600	2345 (80.3), 1675 (7.0), 1382 (4.9)	511 (185.5), 670 (8.2), 962 (6.5), 1412 (0.8)	63Cu(p,n)63Zn	16–6	Natural or enriched Cu	2470
			EC (7.3)
Zirconium (Zr)	89	78.4 h	β+ (22.7)	396	902 (22.7)	511 (45.5), 909 (99.0), 1713 (0.7), 1745 (0.1)	89Y(p,n)89Zr	14–9	Natural Y, natural Y2O3 on Cu	58
			EC (77.3)

List of PET radioisotopes

Halogens

Bromine: 75Br, 76Br

Applications

Since the early days of nuclear medicine, several bromine isotopes have been considered for therapy (auger emitting 77Br), SPECT (80mBr), and for PET imaging (75Br and 76Br). With intermediary physicochemical properties between fluorine and iodine, radiobromines allow for fine-tuning molecular mass, volume, polarizability, and lipophilicity as alternatives to fluorine-18 (18F; t1/2: 109.8 min, 100% β+) or radioiodines (see next section). Of the two positron emitting bromine isotopes, 75Br (t1/2: 96.7 min, 75% β+, 25% EC), has the more favorable characteristics for PET applications with a larger yield of lower energy positrons and a lower number of other gamma-generating decay modes. While the long half-life daughter 75Se (t1/2: 119.8 d, 100.0% EC) warrants careful dose estimation, several 75Br-labeled agents have undergone preclinical and clinical studies including cardiac fatty acid metabolism,[24] neuroreceptor mapping,[25-28] and glucose analogues.[29]

The other positron-emitting isotope is 76Br (t1/2: 16.2 h, 55% β+, 45% EC). It has seen wider adoption primarily because it decays to stable 76Se[1,30,31] and possesses a half-life well matched to protein pharmacokinetics as illustrated through 76Br-labeled antibodies and their fragments for preclinical tumor imaging.[1] Beyond antibody radiolabeling, there are examples of 76Br-labeled neuroimaging tracers,[32-34] peptides,[35] reporter gene studies,[36] and agents for studying sympathetic innervation.[37] Despite this there are limitations to using 76Br-labelled tracers for PET imaging applications. Similar to radioiodine, it is prone to dehalogenation in vivo where it generally remains in blood circulation and slowly accumulates in the gastric mucosa.[38] The high levels of blood retention negatively affects image quality, particularly in blood rich organs (i.e. the heart and lungs) hours after the tracer has been administered.[1,38] Furthermore, bromide has a long biological half-life in human plasma (about 10 days) which makes elimination of labile [76Br]Br− post-imaging problematic.[39] These concerns will surely be addressed as new 76Br-labelled tracers are developed, given an expanded role of this isotope with new β+ γ coincidence PET methodologies.[40,41] A recent review of the chemistry and applications of radiobromines[42] provides more insight into their applications.

Production

The primary nuclear reaction used to produce 75Br on solid-target medical cyclotrons is the 76Se(p,2n)75Br nuclear reaction.[30,43-46] Although enriched elemental selenium has been used to produce solid targets for proton bombardment,[30,44-46] its low melting point and low thermal conductivity necessitates current restrictions to maintain target integrity during bombardment. Alloyed selenides such as NiSe, Ag2Se, CuAgSe, Cu2Se and PbSe[30,44,47] have been developed as suitable alternatives which permit irradiation at higher currents, significantly increasing yields. One drawback of selenium alloy targets is poor adherence to heat-conducting target backings.[44] The beam energy used for the production of 75Br ranges from 24 to 21.5 MeV.[43,44] For this energy range, the theoretical 75Br yields (assuming 76Se-enriched targets), ranged from 1184 to 1480 MBq (μA h)−1 (i.e. 32–40 mCi (μA h)−1),[44] although production at higher energies (i.e. 28 to 22 MeV) provided yields up to three times greater.[30] The major impurity produced was 76Br[43-45] although trace amounts of 72As (t1/2: 26.0 h, 87.8% β+, 12.2% EC), 73As (t1/2: 80.3 d, 100.0% EC), 75Se (t1/2: 119.8 d, 100.0% EC) and 77Br (t1/2: 57.0 h, 0.7% β+, 99.3% EC) have been reported.[44]

There are several reactions that can be used to produce 76Br, however few can be carried out using lower-energy medical cyclotrons.[46] Nevertheless, a feasible production route is the 76Se(p,n)76Br nuclear reaction, which occurs in the 16 to 8 MeV beam energy range.[48] Like 75Br, the target was composed of an enriched selenium alloy, specifically Cu2Se[48,49] or NiSe.[47] Bombarding a Cu2Se target with beam energies ranging from 16 to 8 MeV, the yield of 76Br ranged from 65 to 70 MBq (μA h)−1 (i.e. 1.8 to 2.1 mCi (μA h)−1) at end of bombardment (EOB).[48] The primary impurity produced for this production route was 77Br (2%).[30,46]

Product purification and target recovery

The primary method used to isolate bromine from irradiated selenium was thermochromatographic dry distillation.[30,43,44,46,48] This procedure required heating the selenide targets in an enclosed, oxygen-free apparatus such as a furnace or dedicated distillation vessel. The distillation temperature depended on the type of target being processed, with 300 °C being ideal for distillation of bromine from elemental selenium targets[43,44] and temperatures ranging from 1090 to 1100 °C being ideal for the selenide targets.[30,48] To ensure an oxygen-free environment, an inert gas such as argon was continuously passed through the system. In addition to reducing oxygen exposure, the argon gas mobilizes vaporized bromine and selenium gases generated from the distillation out of the furnace and into a cold trap for collection. The vaporized selenium was the first to condense along the walls of the tubing at temperatures between 120 and 215 °C,[44] while the bromine began to condense closer to room temperature. Any uncondensed bromine can be collected by bubbling the gas stream through water[44] or a dilute NaOH solution.[48] The condensed bromine can be removed by washing with pure water or absolute ethanol.[30,44] Once cooled, the heated targets can be recycled and repeatedly irradiated, typically up to 20 times with <1% target material loss per run.[30,48] The product separation yield for the process ranged from 65–75%.[48]

Summary

Table 2 below summarizes key production and purification parameters for 75Br and 76Br not hitherto mentioned in Table 1.

Production and purification parameters for 75Br and 76Br

Radioisotope	75Br, 76Br
Target material	Elemental selenium, selenides (NiSe, Ag2Se, CuAgSe, Cu2Se, PbSe)
Product purity	—
Major impurities	76Br, 72As, 73As, 75Se, 77Br
Separation method	Thermochromatographic dry distillation
	(300–1100 °C in Ar; water or ethanol for precipitate dissolution; NaOH trap)
Product separation yield	65–75%
Target recovery yield	(100 − x)%, where ‘x’ is no. of times reused

Iodine: 124I

Apart from 99mTc, radioiodines are the most commonly used radionuclides in nuclear medicine. Complementary to the near ideal SPECT isotope 123I (t1/2: 13.2 h, 100.0% EC), the positron emitting 124I (t1/2: 4.2 d, 22.7% β+, 77.3% EC) has been used extensively in immunoPET imaging and other protein-based PET applications due to its well-understood radiochemistry and long half-life.[1,50] Thus far, 124I has been used for cancer diagnosis, monitoring and treatment planning, as well as mechanistic, and pharmacokinetic studies.[51-54] Beyond antibodies, 124I labelled peptides and small molecules targeting neurotransmitters systems, tumor hypoxia, prostate specific membrane antigen (PSMA), and neuroinflammation have also been investigated in humans.[55-57] However, the use of 124I is not without its limitations, not the least of which is its comparatively low positron yield at a high end-point energy. This results in low spatial resolution[58] which may be addressed through hybrid imaging with PET/MR instruments possessing strong magnetic fields.[59] Another drawback is the large number of gammas released from electron capture, increasing patient radiation dose, and further limiting image resolution. These can be corrected for with special software[1,3] and may be turned into an advantage using the aforementioned β+/γ coincidence tomography.[41] To avoid high doses to the thyroid gland from the high energy positron and gammas, ideally the attachment of 124I to biomolecules is kinetically inert.[60] Readers are referred to recent reviews on the radiochemistry and application of radioiodines for more information.[23,42] Taking advantage of increased interest in theranostic radiopharmaceuticals, 124I may see a surge in upcoming years as a companion to 131I (t1/2: 8.02 d, β−: 971 keV) or used for therapy itself (70% absorbed dose of 131I.[61]

While 124I can be produced through a number of different nuclear reaction pathways, however the most feasible production via low-energy proton cyclotron is through the 124,125Te(p,xn)124I nuclear reactions. The radioisotope can be produced through the irradiation of natural tellurium (isotopic composition: 0.09% 120Te, 2.55% 122Te, 0.89% 123Te, 4.74% 124Te, 7.07% 125Te, 18.84% 126Te, 31.74% 128Te, 34.08% 130Te) metalloid targets, >98% enriched 124,125Te metalloid targets, and enriched tellurium oxides electro-deposited or melted onto platinum or tantalum substrates. The optimum beam energy range used for irradiation can range from 14 to 5 MeV for 124Te-based targets, and 22 to 4 MeV for 125Te-based targets. The primary impurities produced during bombardment using 124Te-based targets were 123I (t1/2: 13.2 h, 100.0% EC), 125I (t1/2: 59.4 d, 100.0% EC), 126I (t1/2: 12.9 d, 1.0% β+, 47.3% β−, 51.7% EC) and 130I (t1/2: 12.4 h, 100.0% β−), while the primary impurities for 125Te-based targets are 123I and 125I.[9,46,54,62,63] For the 124Te(p,n)124I nuclear reaction within the 14 to 7 MeV energy range, reported yields ranged from 5 to 21 MBq (μA h)−1 (i.e. 0.1 to 0.6 mCi (μA h)−1) with major impurities equaling less than 1%. For the 125Te(p,2n)124I nuclear reaction within the 22 to 4 MeV energy range, reported yields ranged from 43 to 111 MBq (μA h)−1 (i.e. 1.2 to 3.0 mCi (μA h)−1) with impurities comprising as much as 13% of the final product.[54,62-64]

The preferred method for separating the radioiodine from the post-bombarded tellurium target material is via dry (thermochromatographic) distillation,[46,54,63] which is facilitated by the favorable melting and solidification properties of TeO2. Conventional dry distillation was carried out using a ventilated quartz tube where the irradiated TeO2 was heated to temperatures between 670 to 820 °C for periods ranging from 5 to 20 min in carrier gases of either air, argon, helium, or oxygen. This approach aimed to extract as much of the iodine product using a NaOH trap, however, any prematurely condensed iodine can be removed via dissolution using a weak buffer solution. To minimize tellurium evaporation, Al2O3/quartz wool was integrated into the distillation setup.[63] The separation efficiency for the dry distillation stage ranged from 80% to 95%.[54] Wet chemistry methods were also used to chemically isolate iodine from the tellurium targets, but these methods were less preferred due to product dilution in the final solution. The process was carried out in two stages, with the first being a target dissolution stage using an oxidizing alkali, and the second a reduction stage facilitated by the addition of aluminum powder.[9]

After the iodine has been extracted, the spent tellurium targets can be recycled. Metallic tellurium targets can be dissolved in a mixture of hydrogen peroxide and hydrochloric acid, and the resulting solution can subsequently be reduced using either hydrogen bromide or a mixture of hydrazine and sodium sulfite. Reduction via the former recovered tellurium in the form of tellurite, while the latter facilitated the recovery of metallic tellurium.[9] For tellurium oxide targets, a multistage process involving vacuum distillation, acid dissolution and chemical precipitation was used to recover metallic tellurium.[9] In general, the recovery yields for tellurium-based targets range from 60 to 90%.[54] However, despite the innovations in tellurium recovery, 124I production is still negatively affected by high target material costs.[62]

Table 3 summarizes key production and purification parameters for 124I not previously mentioned in Table 1.

Production and purification parameters for 124I

Radioisotope	124I
Target material	Natural and enriched tellurium, enriched tellurium oxide
Product purity	87–>99%
Major impurities	123I, 125I, 126I, 130I
Primary separation method	Thermochromatographic dry distillation
	(670–820 °C in air, Ar, He or O2; weak buffer for precipitated dissolution; NaOH trap)
Product separation yield	80–95%
Target recovery yield	60–90%

Transition metals

Copper: 60Cu, 61Cu and 64Cu

Copper has many positron-producing radionuclides (i.e.60Cu (t1/2: 23.7 min, 93.0% β+, 7.0% EC), 61Cu (t1/2: 3.3 h, 61.0% β+, 39.0% EC) and 64Cu (t1/2: 12.7 h, 17.6% β+, 38.5% β-, 43.9% EC)) that are finding use in PET applications primarily due to their well-established coordination chemistry.[10] Although less frequently utilized, 60Cu and 61Cu have been used to label various bioactive molecules for hypoxia and blood flow studies.[17,21,65]

The most commonly used copper radioisotope is 64Cu, with the first-in-human use as [64Cu]ASTM for tumor hypoxia.[66] There are several characteristics that make 64Cu well suited for PET imaging. Besides its well elucidated chemistry, its favorable half-life allows it to be used for radiolabeling various targeting vectors like peptides, antibodies, and nanoparticles.[1,10] Additionally, the positron energy for 64Cu (mean: 278 keV, max: 653 keV) is comparable to 18F (mean: 250 keV, max: 634 keV), leading to similar positron range and excellent spatial resolution.[67] With respect to use in immunoPET, there exists some dissonance between radiolabeling conditions and antibody stability. Moreover, some 64Cu radiotracers were found to be susceptible to metabolism in the liver, which permits transchelation of the radioisotope with other copper-binding proteins such as Cu/Zn superoxide dismutase (Cu/Zn SOD) in the liver and metallothionein in the blood.[1,3,68] These limitations may be overcome with improved chelator technology.[69,70] Lastly, it warrants mentioning the potential theranostic pairing with 67Cu (t1/2: 2.58 d, β−: 100%, avg. Eβ: 141 keV) which makes it ideal for radioimmunotherapy monitoring.[71]

As 60Cu is typically produced via the 60Ni(p,n)60Cu nuclear reaction, occurring with natural nickel (isotopic composition: 68.08% 58Ni, 26.22% 60Ni, 1.14% 61Ni, 3.63% 62Ni, 0.93% 64Ni) and enriched 60Ni targets when bombarded with a beam energy of 14.7 MeV.[65] For the 14.7 MeV bombardment of >99% enriched 60Ni electroplated on a gold target backing, the yield at EOB was 2146 MBq (μA h)−1 (i.e. 58 mCi (μA h)−1) with the primary impurities produced to include 61Cu (t1/2: 3.3 h, 61.0% β+, 39.0% EC) (0.05%) and 57Co (t1/2: 271.7 d, 100.0% EC) (0.025%).[65] When bombardment was carried out using natural nickel targets for similar conditions in the same study, EOB yields of 291 MBq (μA h)−1 (i.e. 7.9 mCi (μA h)−1) and 385 MBq (μA h)−1 (i.e. 10.4 mCi (μA h)−1) were reported along with the presence of 55Co (t1/2: 17.5 h, 76.0% β+, 24.0% EC) (0.65%) and 61Cu (0.60%) as the major impurities.[65]

Both the 61Ni(p,n)61Cu[65,72] and 64Zn(p,α)61Cu[17,72] nuclear reactions have been reported for the production of 61Cu. Production via61Ni(p,n)61Cu required proton optimum beam energies ranging from 14.7 to 9 MeV and enriched 61Ni targets. For the 14.7 MeV bombardment of >99% enriched 61Ni electroplated onto a gold target substrate with a thickness of 118 μm and diameter of 5 mm, the 61Cu yield at EOB was 281 MBq (μA h)−1 (i.e. 7.6 mCi (μA h)−1) while the primary impurity produced was 58Co (t1/2: 70.9 d, 14.9% β+, 85.1% EC) (0.04%).[65] In the same study, 60Ni(d,n)61Cu reactions were investigated but produced lower yields at higher per bombardment costs. Alternatively, 61Cu can be produced using the 64Zn(p,α)61Cu nuclear reaction. The reaction has been investigated with natural zinc (isotopic composition: 49.17% 64Zn, 27.73% 66Zn, 4.04% 67Zn, 18.45% 68Zn, 0.61% 70Zn) foils and enriched 64Zn electroplated on to gold, silver or aluminum substrates with beam energies ranging from 17.6 to 11.7 MeV (maximum production occurs circa 14.5 MeV).[17] When >99% enriched 64Zn was bombarded with 14.5 MeV protons, the yield of 61Cu was 13 MBq (μA h)−1 (i.e. 0.4 mCi (μA h)−1) with a purity exceeding 95%.[17] Despite having completed similar experiments using natural zinc foil targets in the same study, the results were too variable to draw valid conclusions. The primary impurities produced during the bombardment of zinc targets were 66Ga (t1/2: 9.5 h, 57.0% β+, 43.0% EC), 67Ga (t1/2: 3.3 d, 100.0% EC) and 68Ga (t1/2: 67.7 min, 88.9% β+, 11.1% EC).[17] Based on the data generated, 61Cu production from the bombardment of nickel was found to be more efficient than production using zinc. However, the former production route can be expensive due to the high cost associated with enriched 61Ni. Since enriched 64Zn is significantly cheaper than enriched 61Ni, the production of 61Cu from zinc targets has found some consideration as an alternative route.[17]

64Cu is primarily produced via the 64Ni(p,n)64Cu nuclear reaction.[9,62] The solid target used for the production is typically enriched 64Ni that has been electroplated onto a gold[62,73-77] or rhodium[10] target backing substrate. Although the threshold for the reaction is 2.5 MeV, production of 64Cu has been reported for beam energies ranging from 15 to 10 MeV since this range offers the highest yields.[62] Numerous yields for the production of 64Cu have been reported in the literature. McCarthy et al.[76] reported yields ranging from 85 to 185 MBq (μA h)−1 (i.e. 2.3 to 5 mCi (μA h)−1) when gold substrates electroplated with >95% 64Ni (thicknesses ranging from 132 to 311 μm) were bombarded with a 15.5 MeV proton beam, while McCarthy et al.[65] reported yields of 296 MBq (μA h)−1 (i.e. 8 mCi (μA h)−1) for greater purity targets of similar design (i.e. >99% 64Ni electroplated onto a gold substrate with a thickness of 215 μm and diameter of 5–6 mm) irradiated at 14.7 MeV. Obata et al.[77] reported yields ranging from 22 to >111 MBq (μA h)−1 (i.e. 0.6 to >3 mCi (μA h)−1) when they bombarded gold substrates that were electroplated with 94.8% 64Ni with a 12 MeV proton beam. More recently, a saturation yield of 5883 MBq (μA h)−1 (i.e. 159 mCi (μA h)−1) has been reported by Avila-Rodriguez et al. (2007) when they bombarded gold substrates electroplated with >95% enriched 64Ni with an 11.4 MeV proton beam. The primary impurity produced for this reaction is 55Co (t1/2: 17.5 h, 76.0% β+, 24.0% EC) (0.012%).[65] However, other relevant impurities include 57Ni (t1/2: 35.6 h, 43.6% β+, 56.4% EC), 56Co (t1/2: 77.2 d, 19.7% β+, 80.3% EC), 57Co (t1/2: 271.7 d, 100.0% EC) and 58Co (t1/2: 70.9 d, 14.9% β+, 85.1% EC).[73]

Isolation of the copper product from the target material begins with target dissolution in strong acid (i.e. 6 M HNO3 or 6 N HCl at temperatures ranging from 90 to 100 °C (Jeffery et al. 2012). The acidic solution is passed through an anion exchange column filled with AG1-X8 resin which retains the copper product along with the nickel target material and any cobalt impurities present. The nickel target material is eluted first using 6 N HCl[75,76] for HCl-dissolved targets, or 0.2 M HCl in 96% methanol[74] for HNO3-dissolved targets. The cobalt impurity fraction retained by the column can be eluted with 4 M HCl[75] for HCl-dissolved targets, or 0.3 M HCl in 72% ethanol[74] for HNO3-dissolved targets. Finally, the copper product can be eluted using dilute acid (i.e. 0.1–0.5 M) or distilled water[75,76] for HCl-dissolved targets, or 0.3 M HCl in 40% ethanol[74] for HNO3-dissolved targets. Separation yields exceeding 95% have been reported for the process.[75]

Implementation of a dedicated recycling scheme for 60Ni recovery is not typical due to its low material cost,[65] which stems from the large natural abundance of the isotope. However, recycling of enriched 64Ni is necessary to keep production costs down. Several methods have been developed for 64Ni recovery. The method described by Obata et al.[77] made use of two heating stages conducted in series. In the first heating stage, the solution containing the eluted nickel was evaporated to dryness at 150 °C before it was rehydrated with distilled water. In the second heating stage, the rehydrated solution was heated at 900 °C for 24 h. This resulted in the formation of 64NiO which was then used for the preparation of future targets. The minimum recycling efficiency for this process was 94%. A second method was discussed by McCarthy et al.,[76] and required several evaporation and acid dissolution steps be carried out in series. The first two acid dissolution steps were performed with HNO3 and the final acid dissolution step was performed with H2SO4. After H2SO4 dissolution, the solution was diluted and the pH was adjusted to 9 with the addition of concentrated NH4OH. The solution was available for electroplating after the addition of ammonium sulfate and a final dilution with deionized water. The overall efficiency of this recovery process exceeded 90%.

For the case of 61Cu production from zinc targets, Rowshanfarzad et al.[72] and Asad et al.[17] outlined a method which made use of multiple column separation stages in series. The zinc portion of the targets were dissolved in 10 M HCl at room temperature before passing through a cation exchange column filled with AG 50W-X8 resin (200–400 mesh). The column retained the copper product, the gallium impurities, and the unreacted zinc target material. The copper product and zinc target materials were eluted with 10 M HCl and the eluant was passed through an AG 1-X8 anion exchange column (200–400 mesh). Both the copper product and zinc target material were retained, and the copper product was selectively eluted with 2 M HCl. In order to recycle the zinc target material (as in the case of enriched targets), the column can be eluted with 0.5 M HCl.[17] This solution was then evaporated to dryness before 6 M H2SO4 was added to digest any resin residues from the prior chromatography stages. The evaporation/digestion stage was repeated before the recovered zinc was finally reconstituted as an electroplating solution. The overall yield for the 64Zn recovery process was about 90%.[17]

Table 4 summarizes key production and purification parameters for 60Cu, 61Cu and 64Cu not mentioned in Table 1.

Production and purification parameters for 60Cu, 61Cu and 64Cu

Radioisotope		60Cu, 61Cu, 64Cu
Target material		Natural and enriched nickel (all), natural and enriched zinc (for 61Cu only)
Product purity	60Cu	∼98.8 to ∼99.9%
Major impurities		61Cu, 55Co, 57Co
Product purity	61Cu	∼99.6% for 61Ni(p,n)61Cu nuclear reaction; >95% for 64Zn(p,α)61Cu nuclear reaction
Major impurities		58Co for 61Ni(p,n)61Cu nuclear reaction; 66Ga, 67Ga, 68Ga for 64Zn(p,α)61Cu nuclear reaction
Product purity	64Cu	∼95 to >99%
Major impurities		55Co, 56Co, 57Co, 58Co, 57Ni
Separation method	Ni	Anion exchange chromatography (AG 1-X8 resin)
		- Target dissolution: (a) HNO3 or (b) strong HCl
		- Ni target elution: (a) 0.2 N HCl in 96% methanol or (b) 6N HCl
		- Co impurity elution: (a) 0.3% HCl in 72% ethanol or (b) 4 N HCl
		- Copper elution: (a) 0.3 N HCl in 40% ethanol or (b) water
Product separation yield		95%
Target recovery yield		90%
Separation method	Zn	Ion exchange chromatography in series (AG 50W-X8 resin (cation) and AG 1-X8 resin (anion))
		- Target dissolution: 10 N HCl
		- (Cation exchange) Cu and Zn elution: 10 N HCl
		- (Anion exchange) Cu elution: 2 N HCl
		- (Anion exchange) Zn target elution: 0.5 N HCl
Product separation yield		—
Target recovery yield		90%

Cobalt: 55Co

Cobalt is a human micronutrient and component of the essential vitamin B12 (cobalamin). As a potential competitor radionuclide to 64Cu, 55Co (t1/2: 17.5 h, 76.0% β+, 24.0% EC) has been shown to mimic calcium intake in infarcted brain tissue. This property makes it useful in the imaging of recent cerebral brain damage resulting from cerebral tumors, stroke and other brain injuries.[78-81] Recent reports have demonstrated high resolution PET images obtained with peptide and affibody chelates of 55Co.[82,83] Ionic 55Co species are residualizing radiotracers. Significant amounts of 55Co can accumulate in organs such as the bladder, liver and kidney[84] with the majority eliminated or decayed three days post-injection. 55Co has also shown great promise in the field of immunoPET imaging and has been used to label an anti-EGFR affibody[82] as well as several peptides.[83,85] Unfortunately, widespread clinical use of this isotope may be hampered as 55Co decays into 55Fe (t1/2: 2.73 years) while releasing large amounts of high energy gamma rays from electron capture in the process.[1] Nonetheless, it remains a valuable tool for preclinical PET investigations.

There are two major nuclear reactions available to produce 55Co on small medical cyclotrons by utilizing either the 56Fe(p,2n)55Co[81,86,87] or the 58Ni(p,α)55Co.[81,88,89] Production via56Fe(p,2n)55Co involved irradiating 56Fe-enriched iron foils[87] with a minimum proton beam energy of about 16 MeV,[86,87] although the reaction peaks at an energy near 23.5 MeV.[87] However, this production route was undesirable since significant amounts of the chemically-inseparable 56Co (t1/2: 77.2 d, 19.7% β+, 80.3% EC) impurity was also co-produced.[81] The 58Ni(p,α)55Co production route can be carried out through the bombardment of natural nickel (isotopic composition: 68.08% 58Ni, 26.22% 60Ni, 1.14% 61Ni, 3.63% 62Ni, 0.93% 64Ni) foils[89] or enriched 58Ni-electroplated onto silver[88] or gold[81] disk substrates. Irradiations have been carried out using proton beams with energies of about 16 MeV (ref. 88 and 90) to 15 MeV.[81] The primary impurities produced during irradiation were 57Co (t1/2: 271.7 d, 100.0% EC), 58Co (t1/2: 70.9 d, 14.9% β+, 85.1% EC) and 57Ni (t1/2: 35.6 h, 43.6% β+, 56.4% EC), particularly when natural nickel targets were used for the production.[89] The saturation yield of 55Co when natural nickel foils were bombarded with a beam energy of 16 MeV was 236 MBq (μA h)−1 (i.e. 6.3 mCi (μA h)−1) at EOB with purity exceeding 97%.[89]55Co yields from 16 MeV proton beam bombardments of >99% 58Ni-electroplated gold disk substrates resulted in 9.3 MBq (μA h)−1 (i.e. 0.3 mCi (μA h)−1) with a purity exceeding 92%.[88]

Column-based separation techniques have been used to isolate cobalt from the nickel targets after heated dissolution with concentrated HCl.[81,88,89] The first example used an AG1-X8 anion exchange resin[81,89] which retained the cobalt product when the acidic target solution was passed through. Dilute HCl with concentrations ranging from 0.4–0.5 M (ref. 81 and 89) could be used to elute the cobalt product from the resin. Another stationary phase, DGA branched resin,[88] was demonstrated to selectively release 55Co upon elution with 3 M HCl. The overall separation efficiency for the process generally exceeded 90%.[88,89] If the method used for target preparation was the electroplating of enriched 58Ni on a metal disk substrate, 58Ni was recovered with a 94% efficiency from the spent target solution using a series of drying and chemical cleaning stages before finally being reconstituted as an electroplating solution.[88]

Table 5 summarizes key production and purification parameters for 55Co not mentioned in Table 1.

Production and purification parameters for 55Co

Radioisotope	55Co
Target material	Natural and enriched nickel
Product purity	92 to >97%
Major impurities	57Co, 58Co, 57Ni
Separation method	Anion exchange chromatography (AG1-X8 resin)
	- Target dissolution: concentrated HCl
	- Co elution: 0.4–0.5 N HCl
Product separation yield	>90%
Target recovery yield	94%

Manganese: 52gMn

A considerable amount of interest in the use of 52gMn (t1/2: 5.6 d, 29.4% β+, 70.6% EC) in PET applications has developed since its first use as a myocardial perfusion PET agent in 1985.[91] The radionuclide has been used in several preclinical imaging studies, including murine 4T1 xenografts, functional β-cells in type 1 and type 2 diabetes, and neural imaging in primates.[22,92] The interest in 52gMn stems from its decay characteristics, ease of production, and its potential to act as a surrogate to non-invasively quantify MR contrast agent.[93] The 52gMn radioisotope has a long half-life which facilitates late time-point studies up to 2 to 3 weeks following tracer administration, making it suitable for ImmunoPET applications.[1,94] Additionally, positron decay occurs at a low positron energy (β+Emax: 575 keV), promoting high resolution imaging that is comparable to 18F.[93] Despite its numerous advantages, caution is exercised when using 52gMn for PET applications since it has several high energy gamma emissions (i.e. 744, 935, and 1434 keV) coupled with prolonged residence in critical organs which require careful dosimetry studies prior to translation.

52gMn is primarily produced using the 52Cr(p,n)52gMn nuclear reaction.[92,94-96] The reaction takes place within the 20 to 10 MeV beam energy range.[93] The main target material used is natural chromium (isotopic composition: 4.35% 50Cr, 83.79% 52Cr, 9.50% 53Cr, 2.37% 54Cr) which can be loaded on to metal target substrates such as copper[95] and silver[94]via electroplating or pressing.[92,94,95] The natural chromium targets may also be a metal disk[92,93] or foil.[96] The primary impurity produced during the reaction was 54Mn (t1/2: 312.2 d, 100.0% EC) via the 54Cr(p,n)54Mn nuclear reaction,[92,94] however this impurity can be greatly reduced if 52Cr-enriched targets are used. 52mMn (t1/2: 21.1 min, 96.6% β+, 1.6% EC, 1.8% IT) was also produced[97] from the irradiation, however the radionuclide decays into 52gMn relatively quickly after its production. When chromium metal disks were bombarded with 16 MeV proton beams, a yield of 9.6 MBq (μA h)−1 (i.e. 0.26 mCi (μA h)−1) was reported at a purity level of 99.55%.[93] For the composite, chromium/silver pressed targets bombarded at the same beam energy, a yield of 5 MBq (μA h)−1 (i.e. 0.1 mCi (μA h)−1) at a purity level of about 95% was reported.[94] When natural chromium foils were irradiated with 12.5 MeV proton beams, a yield of 4.5 MBq (μA h)−1 (i.e. 0.12 mCi (μA h)−1) and a purity exceeding 99.0% was reported by Topping et al. (2013).

Product purification

The major method used to isolate manganese from the chromium targets is ion exchange chromatography. There are several different resins that can be used to carry out the isolation procedure. The most frequently used resin was the AG 1-X8 which required dissolution of the bombarded target in 11 to 12 M HCl.[93,96-98] After dissolution, the target solution was passed through the resin which selectively retained the manganese product, and elution was carried out using 0.1 M HCl. The separation efficiency of the process exceeded 97%.[97] Apart from the AG 1-X8 resin, the 50W-X8 resin from Dowex or AG was also used to perform the separation, as discussed by Chaple and Lapi[92] and Wooten et al.[95] Methods for recovering the chromium target material have not been discussed, likely because recovery schemes tend not to be feasible when the large natural abundance of 52Cr is considered.

Table 6 summarizes key production and purification parameters for 52gMn not mentioned in Table 1.

Production and purification parameters for 52gMn

Radioisotope	52gMn
Target material	Natural chromium
Product purity	95 to 99.55%
Major impurities	54Mn, 52mMn
Separation method	Ion exchange chromatography (AG1-X8 resin or other resins)
	- Target dissolution: 11 N HCl
	- Mn elution: 0.1 N HCl
Product separation yield	97%
Target recovery yield	n/a

Niobium: 90Nb

A relative newcomer PET radionuclide, 90Nb (t1/2: 14.6 h, 53.0% β+, 47.0% EC) is finding increasing consideration as a labelling agent in immunoPET applications due to close agreement between its half-life and the pharmacokinetics of peptides, antibody fragments and antibodies.[99,100] Additionally, 90Nb has a high probability of positron decay at moderate emission energies (β+Emax: 1.5 MeV), which facilitates high-resolution PET imaging.[101] Protein labelling studies have been carried out as a preliminary step, including 90Nb-labelled monoclonal antibodies rituximab and bevacizumab in high yields.[102-104]90Nb has also been used to label the peptide derivative DFO-succinyl-(d)Phe1-octreotide.[105] Although easily produced using low-energy cyclotrons, efficient separation methods ensuring quick isolation of 90Nb from bombarded target materials have not yet been established.[99,105]

90Nb is primarily produced via the 90Zr(p,n)90Nb nuclear reaction within the 19 to 7 MeV beam energy range.[99,105] However, 90Nb could also be produced via the 91Zr(p,2n)90Nb nuclear reaction when the incident beam energy exceeded 15 MeV.[105] Maximum 90Nb production occurred at a beam energy of 13.5 MeV.[105] The target material for 90Nb production was typically natural zirconium foils (isotopic composition: 51.45% 90Zr, 11.22% 91Zr, 17.15% 92Zr, 17.38% 94Zr, 2.80% 96Zr),[99,101-103,105] although the use of this target produced a significant number of impurities, making it unsuitable for clinical trial applications.[105] The primary impurities produced during the bombardment of natural zirconium were 89g/mNb (t1/2: 2.0 h, 75.0% β+, 25.0% EC/66.0 min, 81.0% β+, 19.0% EC), 91mNb (t1/2: 60.9 d, 96.6% IT, 3.4% EC), 92mNb (t1/2: 10.2 d, 0.1% β+, 99.9% EC), 95g/mNb (t1/2: 35.0 d, 100.0% β-/t1/2: 3.6 d, 5.6% β-, 94.4% IT), 96Nb (t1/2: 23.4 h, 100.0% β-), 87Y (t1/2: 79.8 h, 0.2% β+, 99.8% EC) and 89Zr (t1/2: 78.4 h, 22.7.0% β+, 77.3% EC).[102,105] An alternative production route using ZrO2 from enriched 90Zr has been proposed by ref. 105, in which case the target preparation involved sedimentation of enriched ZrO2 powder onto a copper substrate. Production using the 90ZrO2 targets had up to two orders of magnitude lower impurity levels. When three, 0.25 mm thick natural zirconium foils were irradiated with a 17.5 MeV proton beam, a 90Nb yield of 145 MBq (μA h)−1 (i.e. 3.9 mCi (μA h)−1) was reported with a purity of 97% at EOB.[102] A yield of 596 MBq (μA h)−1 (i.e. 16.1 mCi (μA h)−1) with purity exceeding 95% was reported following the bombardment of three 90ZrO2-on-copper foil targets with an incident proton beam energy of 11.9 MeV, 4 hours after EOB.[105]

There are two established methods that have been used to isolate the niobium product from post-bombarded zirconium targets. The first method was described by Radchenko et al.[100,103] and required serial ion exchange chromatography separation. Irradiated zirconium targets were first dissolved in 28 M hydrofluoric acid and diluted to 21 M before loading on to the first ion exchange column. Cation exchange was performed by passing the solution through a DOWEX 50 × 8 resin (200–400 mesh) to retain trace metal ion impurities (i.e. copper and iron), colloids and undissolved particulate, while allowing the dissolved target material and product to pass through. After washing with concentrated HF, the emerging solution from the cation exchange stage was transferred to the second ion exchange stage. Anion exchange was performed using an AG 1 × 8 resin (200–400 mesh) resin to retain the niobium product. To elute the niobium, a mixture of 1% H2O2 in 6 M HCl[101] or 0.1 M oxalic acid[103] was used. If the former was used for elution, the H2O2 should be removed by heating the eluate for 5 min at 120 °C. The overall efficiency of the extraction procedure ranged from 93 to 95%.

A second method was described by Radchenko et al.,[101,102] which made use of a liquid–liquid extraction procedure. After placing into a small ice-water bath, the bombarded zirconium targets were dissolved in 48% HF (28 M). After complete target dissolution, 10 M HCl and saturated boric acid were added. More than 99% of the niobium product was then partitioned into chloroform using 0.02 M N-benzoyl-N-phenylhydroxylamine as a phase transfer catalyst. Liquid–liquid extraction was then performed with vigorous shaking for 20 min. After washing the organic phase with a mixture of 9 M HCl and 0.001 M HF and then 9 M HCl, respectively, the niobium product was re-extracted from the organic phase using aqua regia with an extraction rate ranging from 90–95%. To increase the purity of the obtained 90Nb product, the aqueous layer was subsequently passed through an anion exchange resin. The niobium containing solution was evaporated to dryness and the resulting residue was dissolved in a mixture of 0.25 M HCl and 0.1 M oxalic acid. Anion exchange was performed on the residue using a small Aminex A27 column, and the niobium product was eluted using a mixture of 6 M HCl and 0.01 M oxalic acid. The overall efficiency of the extraction procedure ranged from 76 to 81%.

In general, methods for recovering the zirconium target material have not been discussed, likely due to the lack of feasibility of any proposed recovery scheme given the relatively large natural abundance of 90Zr.

Table 7 summarizes key production and purification parameters for 90Nb not mentioned in Table 1.

Production and purification parameters for 90Nb

Radioisotope		90Nb
Target material		Natural and enriched zirconium oxide
Product purity		>95 to 97%
Major impurities		89g/mNb, 91mNb, 92mNb, 95g/mNb, 96Nb, 87Y, 89Zr
Separation method	1	Serial ion exchange chromatography (DOWEX 50 × 8 resin (cation), AG 1 × 8 resin (anion))
		- Target dissolution: 28 M HF
		- (Cation exchange) 1st Nb elution: concentrated HF
		- (Anion exchange) 2nd Nb elution: 1% H2O2 in 6 M HCl
Product separation yield		93 to 95%
Separation method	2	Liquid–liquid extraction
		- Target dissolution: 48% HF
		- Extraction solvents: 10 M HCl + saturated boric acid, 0.02 M N-benzoyl-N-phenylhydroxylamine in CHCl3, aqua regia
Product separation yield		76 to 81%
Target recovery yield		n/a

Scandium: 44gSc

Favorable decay characteristics and the potential availability of a 44Ti/44Sc generator has brought 44gSc (t1/2: 4.0 h, 94.3% β+, 5.7% EC) into clinical focus over the past decade. Scandium-44 is a trivalent radionuclide that is readily conjugated to peptides and proteins using common macrocyclic chelators such as DOTA and NOTA. This facilitates the growing number of preclinical and clinical PET studies reported,[106-110] and was exemplified by 44gSc-labelling of PSMA-617 (i.e. peptidomimetic inhibitors) as an alternative for 68Ga-labelling.[111] Due to the four-fold longer half-life, dosimetry measurements could be obtained 19 h post-injection to better match the pharmacokinetics of the 177Lu-labelled therapeutic. Additionally, 44Sc forms a true theranostic pairing with 47Sc.[112]

There are many methods that can be used to produce 44gSc, ranging from the 44Ti/44gSc generator[62,111] to small medical cyclotrons.[62,110,113,114] With respect to cyclotron production via solid targetry, one of the most common production routes is through the 44Ca(p,n)44g,mSc nuclear reaction. The process involves the use of solid targets formed from natural calcium (isotopic composition: 96.94% 40Ca, 0.65% 42Ca, 0.14% 43Ca, 2.09% 44Ca, >0.01% 46Ca, 0.19% 48Ca) and enriched 44Ca metal,[113] as well as enriched [44Ca]CaCO3 powders.[62,114] The nuclear reaction takes place over beam energies ranging from 18 to 6 MeV, with the reaction cross-section maximums between 13 and 10 MeV.[114] The yield of 44gSc produced from the bombardment of >94% enriched [44Ca]CaCO3 powders at a beam energy of 11 MeV was 50 MBq (μA h)−1 (i.e. 1.4 mCi (μA h)−1),[114] while that obtained from the bombardment of natural calcium at a beam energy of 15.6 MeV was 32 MBq (μA h)−1 (i.e. 0.9 mCi (μA h)−1) with a purity exceeding 95%.[110] It should be pointed out that targets with high concentrations of 44Ca perform poorly as solid targets due to their low electrical and heat conductivity properties.[110] The primary impurity produced during bombardment was 44mSc (t1/2: 58.6 h, 98.8% IT, 1.2% EC) which has a negligible effect on PET studies when produced in small amounts.[110]44mSc has found new use as an in vivo generator in various pharmacokinetic studies[62,113] due to its longer half-life and eventual decay into 44gSc. Other impurities produced during the proton irradiation of 44Ca include 47Sc (t1/2: 3.3 d, 100.0% β-) and 48Sc (t1/2: 43.7 h, 100.0% β-), which were formed from trace amounts of 48Ca via the 48Ca(p,2n)47Sc and 48Ca(p,n)48Sc nuclear reactions, respectively.[110,114] Small amounts of 43Sc (t1/2: 4.0 h, 88.1% β+, 11.9% EC)[110] and 46Sc (t1/2: 83.8 d, 100.0% β-)[62] have also been reported.

After acid dissolution, the scandium product can be isolated from irradiated calcium targets through column chromatography. However, the specific column and resin used varied with the type of target bombarded. With 44Ca metal, the target was first dissolved in concentrated HCl and then a UTEVA extraction resin was used to retain both the scandium product and the target material.[110] The bulk target material was eluted with the 10 M HCl, while the scandium product was eluted subsequently with water, recovering ∼80% of the 44gSc activity at EOB. When irradiating [44Ca]CaCO3, a chelating ion exchange resin (i.e. Chelex 100) was used[114] after target dissolution in 0.1 M HCl. The resin retains both the scandium product and the target material, with the latter eluted first using 0.01 M HCl. The scandium product was removed using 1 M HCl with a separation efficiency exceeding 70%. A target-material recovery stage was further described in the work published by Krajewski et al.[114] Briefly, the target-containing effluent was concentrated with ammonia. The new solution was subsequently heated to 170 °C to decompose NH4Cl formed from the ammonia addition, and to evaporate any remaining water. 44Ca was recovered as 44CaO with a recovery efficiency of 60%, although the authors believed the small batch sizes processed strongly affected the recovery efficiency. Regardless, the addition of a recovery stage was worthwhile given the cost of 44Ca.[62]

Table 8 summarizes key production and purification parameters for 44gSc not mentioned in Table 1.

Production and purification parameters for 44gSc

Radioisotope		44gSc
Target material		Natural and enriched calcium, enriched calcium carbonate
Product purity		>95%
Major impurities		44mSc, 47Sc, 48Sc, 43Sc, 46Sc
Separation method	1	Ion exchange chromatography (UTEVA extraction resin)
		- Target dissolution: concentrated HCl
		- Ca elution: 10 M HCl
		- Sc elution: water
Product separation yield		80%
Target recovery yield		—
Separation method	2	Ion exchange chromatography (Chelex 100 resin)
		- Target dissolution: 0.1 M HCl
		- Sc elution: 0.01 M HCl
		- Ca elution: 1 M HCl
Product separation yield		70%
Target recovery yield		60%

Technetium: 94mTc

The workhorse of nuclear medicine is the SPECT isotope 99mTc.[115] As a PET surrogate, 94mTc (t1/2: 52 min, 70.2% β+, 29.8% EC) has seen limited use. Efflux transporter activity has been measured with 94mTc-sestamibi. The tracer was used in a preclinical multi-drug resistance study involving P-glycoprotein knock-out and wild-type mice, and showed delayed tracer clearance in the liver and kidney of the former group.[116] Other examples include examining cardiac function and estrogen receptor expression.[117,118] A major application of 94mTc remains the quantitative assessment of novel 99mTc radiopharmaceuticals through in vivo PET studies.[119] The limited adoption of 94mTc in PET can be attributed to several factors, with the most limiting being the prevalence of the high-energy gamma ray that is released during its decay.[120] Clinical applications using 94mTc for testing 99mTc radiopharmaceuticals are also limited by the relatively short half-life of 94mTc,[120] as well as the poor dosimetry of 94mTc which is seven-fold less favorable than that of 99mTc.[121]

94mTc can be produced using the 94Mo(p,n)94mTc nuclear reaction on small medical cyclotrons.[11,119-121] The reaction takes place within the 13 to 6 MeV energy range,[11,119] and has been carried out using both natural molybdenum foils (isotopic composition: 14.53% 92Mo, 9.15% 94Mo, 15.84% 95Mo, 16.67% 96Mo, 9.60% 97Mo, 24.39% 98Mo, 9.82% 100Mo) and 94Mo-enriched molybdenum oxide (i.e. MoO3) pellets and powders.[120-122] When using 94Mo-enriched targets for production, the major impurity produced was 94gTc (t1/2: 293 min, 10.5% β+, 89.5% EC),[119] especially when bombardment was conducted with beam energies equaling or exceeding 13 MeV.[120] When natural molybdenum was used for production, impurities such as 95g/mTc (t1/2: 61.0 d, 0.4% β+, 96.1% EC, 3.9 IT/t1/2: 20.0 h, 100.0% EC), 96g/mTc (t1/2: 4.3 d, 100.0% EC/t1/2: 51.5 min, 98.0% IT, 2.0% EC) and 99mTc (t1/2: 6.0 h, 100.0% IT) were present in addition to 94gTc.[120-122] Despite the myriad of impurities formed, 94mTc production from natural molybdenum has been considered suitable for human radiopharmaceutical studies.[120] However, special attention should be given to 96mTc (which does not emit any imageable gamma rays)[120] and 95Tc,[123] which are both produced in significant amounts and have been shown to increase patient dose. When natural molybdenum foils were bombarded with an 11 MeV proton beam, a saturation yield of 111 MBq (μA h)−1 (i.e. 3 mCi (μA h)−1) was reported.[121] However, a theoretical yield of 2000 MBq (μA h)−1 (i.e. 54 mCi (μA h)−1) has been calculated for thick targets of enriched 94Mo bombarded with beam energies ranging from 13 to 7 MeV.[120,122]

The primary method used to separate the technetium product from the metallic or oxide molybdenum target material was thermochromatographic dry distillation.[120,121] This requires heating of the irradiated targets in moist air,[120] oxygen or helium.[121] For molybdenum oxide targets heated to temperatures below 800 °C, volatile technetium species such as [94mTc]HTcO4 and [94mTc]Tc2O7 sublime and were transported by the carrier gas until condensation temperatures (i.e. 250 to 350 °C) were reached. When heating was carried out at temperatures above 800 °C (i.e. 1090 °C (ref. 120)), both the technetium product and molybdenum target material sublimed. The molybdenum condenses first in the 600 to 800 °C temperature range, while the technetium product condenses at the cooler temperatures mentioned above. The condensed technetium product can be dissolved using hot NaOH and purified using a minimized alumina column. The entire separation process required approximately 25 minutes, and had an 80 to 85% product separation efficiency. The molybdenum target material was well isolated and could be easily recovered for immediate reuse with a recovery efficiency exceeding 95%.

Table 9 summarizes key production and purification parameters for 94mTc not mentioned in Table 1.

Production and purification parameters for 94mTc

Radioisotope	94mTc
Target material	Natural molybdenum, enriched molybdenum oxides
Product purity	—
Major impurities	94Tc, 95g/mTc, 96g/mTc, 99mTc
Separation method	Thermochromatographic dry distillation
	(500 to 1090 °C in moist air, O2 or He; hot NaOH for precipitate dissolution)
Product separation yield	80 to 85%
Target recovery yield	>95%

Titanium: 45Ti

Favorable decay characteristics has garnered increase consideration of 45Ti (t1/2: 3.08 h, 84.8% β+, 15.2% EC) as a PET imaging agent. These include a reasonable half-life and decaying to a stable daughter isotope, both of which are desirable features for PET radionuclides.[124] Additionally, 45Ti has a high probability of positron decay at a moderate maximum positron energy (β+Emax: 1040).[92,124] Most positron decay events proceed directly to the ground state, making positron annihilation the primary source of gamma rays in the system.[124,125] These decay characteristics yield images that possess a high resolution.[92] In addition to its favorable decay properties, 45Ti is also available to produce as discussed below. Despite its appeal, research on 45Ti radiolabeled tracers is limited as a result attributed to titanium's complicated chemistry and its high reactivity with oxygen.[92,126] Regardless, a few studies using 45Ti-radiolabelled molecules exist. 45Ti has been used to label human serum albumin, phytate, DTPA and citrate for pre-clinical studies.[125] A few bioconjugate studies have been completed including protein and peptide targeting vectors. Biodistributions studies in mice using [45Ti]transferrin observed high 45Ti uptake in the spleen and liver[126] indicative of transchelation. Another recent example attached 45Ti to a PSMA-targeted peptidomimetic.[126] In an interesting attempt to develop dual modality probes, chelate free mesoporous silica nanoparticles were labeled with 45Ti for non-invasive in vivo quantification.[127]

45Ti is produced via the 45Sc(p,n)45Ti nuclear reaction.[128] The reaction primarily takes place within the 16 to 8 MeV energy range, with a threshold energy of 2.9 MeV.[128] Regardless, a range restricted to 15 to 8 MeV was recommended to reduce the production of isotopic and non-isotopic impurities during bombardment.[124] The primary impurities produced during bombardment were 44Ti (t1/2: 59.1 years, 100.0% EC) from the 45Sc(p,2n)44Ti nuclear reaction, and 44gSc (t1/2: 4.0 h, 94.3% β+, 5.7% EC) from the 45Sc(p,pn)44gSc and 45Sc(p,d)44gSc nuclear reactions.[128] The target material was natural scandium (isotopic composition: 100.00% 45Sc) foils.[92,124-126,128] A few values for 45Ti yields have been reported in the literature. Vavere and Welch[126] reported yields of 422 MBq (μA h)−1 (i.e. 11.4 mCi (μA h)−1) after bombarding 0.25 mm thick natural scandium foils with a 14.7 MeV proton beam. However, yields as low as 230 MBq (μA h)−1 (i.e. 6.2 mCi (μA h)−1) for bombardments at 11.8 MeV have been reported.[127]

Cation-exchange chromatography has been used to isolate the titanium product from the scandium targets, and the method has been extensively described by Vavere et al.[125] and Vavere and Welch.[126] Post-bombarded targets were first dissolved in 6 N HCl, and the final solution was added to a cation exchange column filled with AG 50W-X8 resin (100–200 mesh). 45Ti was then eluted with 6 N HCl, and this eluate was heated in the presence of a nitrogen stream to evaporate the HCl from the final product. Greater than 92% of the eluted 45Ti product was obtained with a radionuclide purity of 99.8%. Methods for recovering the scandium target material have not been discussed as such methods would likely not be feasible due to the naturally high abundance of 45Sc.

Table 10 summarizes key production and purification parameters for 45Ti not mentioned in Table 1.

Production and purification parameters for 45Ti

Radioisotope	45Ti
Target material	Natural scandium
Product purity	>99%
Major impurities	44Ti, 44gSc
Separation method	Cation exchange chromatography (AG 50W-X8 resin)
	- Target dissolution: 6 M HCl
	- Ti elution: 6 M HCl
Product separation yield	>92%
Target recovery yield	n/a

Yttrium: 86gY

As part of the original theranostic pairing,[129]86gY (t1/2: 14.7 h, 31.9% β+, 68.2% EC) is a residualizing radioisotope that has promising applications in radioimmunotherapy as an imaging compliment to its chemically-identical therapeutic counterpart, 90Y. Owing to its well-elucidated chelation chemistry, 86gY has been used to label several different antibodies, although the produced radiotracers have been limited to use in preclinical studies[1] due to a few limitations. 86gY decays primarily by electron capture, generating several gamma emissions in the process. This negatively affects imaging quality and necessitates the use of corrections.[130] A second limitation is the accumulation of unbounded 86gY (due to conjugate dissociation or transchelation) in bone marrow, resulting in acute bone marrow toxicity. However, it should be mentioned that the risk of transchelation for ligands such as DOTA is low given the high thermodynamic and kinetic inertness of Y-DOTA chelates.[131] Various peptides[132,133] and antibodies[134-137] have been radiolabeled with the intention of monitoring radioimmunotherapy. With respect to its utility as a theranostic pair, the half-life of 86gY is short compared to 90Y and may limit the flexibility of radioimmunotherapy applications.[1,138] Beyond immunoPET, 86Y has potential for use as a surrogate for gadolinium for non-invasive quantification of MR contrast agents.[139]

86gY is primarily produced via the 86Sr(p,n)86gY nuclear reaction, which also produces significant amounts of 86mY (t1/2: 47.4 min, 99.3% IT, 0.4% β+, 0.3% EC), a shorter half-life impurity which quickly decays into 86gY.[62]86gY has been produced from the bombardment of solid targets made from pressed, enriched 86SrCO3 (ref. 140–142) or 86SrO[141] powders using proton beams of energies ranging from 15.1 to 11 MeV.[140,142] Naturally, the yield of 86gY varied with the solid target material used as well as the specific irradiation conditions employed. When bombardments were carried out using 95.6% enriched 86SrCO3 solid target using a 15.1 MeV proton beam, an average 86gY yield of 48 MBq (μA h)−1 (i.e. 1.3 mCi (μA h)−1) with a purity exceeding 99% was obtained.[62,140] Alternatively, bombardments carried out on 96.4% enriched 86SrO targets using a 14.5 MeV proton beam produced much higher 86gY yields of 166 MBq (μA h)−1 (i.e. 4.5 mCi (μA h)−1), with 86mY (220%), 87g/mY (t1/2: 79.8 h, 0.2% β+, 99.8% EC/t1/2: 13.3 h, 98.4% IT, 0.8 β+, 0.8% EC) (0.27%/0.43%) and 88Y (t1/2: 106.6 d, 0.2% β+, 99.8% EC) (0.024%) being the primary impurities.[62,141]

Multistage-electrolysis and filtration were two primary methods used to extract yttrium from post-bombardment strontium solid targets. The multistage-electrolysis procedure has been reported by Reischl et al.[140] and Yoo et al.,[141] and first required dissolution of the target material in 2.8 M HNO3. To displace any oxygen and carbon dioxide dissolved in solution, the solution is sparged with an inert gas (i.e. argon) for 10 to 15 min before electrolysis began and continued through two electrolysis stages. The first electrolysis stage was carried out using a platinum plate cathode and anode. Yttrium was deposited on the cathode, followed by transfer to a platinum wire during a second electrolysis stage, where it was later removed and used for downstream labelling procedures. A simpler filtration method required dissolution of the targets in 6 M HCl, followed by the subsequent addition of 1 M NH4OH.[142] The resulting solution was then filtered under vacuum and washed with water to remove any residual strontium-containing liquid absorbed into the filter paper. Although the filtration-based separation technique was considerably faster than the multistage-electrolysis technique, filtration-based separation suffered from a lower efficiency which affected overall yttrium yields. Multistage-electrolysis boasts separation efficiencies of about 97%, while the separation yield for filtration was about 88%.[62] The strontium remaining in solution after electrolysis and filtration was available to be recovered for reuse. Depending on the chemical form of the strontium, a solution of (NH4)2CO3 (ref. 140 and 142) or NH4OH followed by the addition of saturated ammonium carbonate solution[141] was used to precipitate the strontium in the form of 86SrCO3 with a recovery efficiency exceeding 90%. The 86SrCO3 can be converted to 86SrO through thermal decomposition if desired.[141]

Table 11 summarizes key production and purification parameters for 86gY not mentioned in Table 1.

Production and purification parameters for 86gY

Radioisotope		86gY
Target material		Strontium oxides
Product purity		>99%
Major impurities		86mY, 87g/mY, 88Y
Separation method	1	Multistage electrolysis
		- Target dissolution: 2.8 M or 4% HNO3
		- Platinum plate cathode and anode
		- Inert gas environment
Product separation yield		∼97%
Separation method	2	Filtration
		- Target dissolution: 6 M HCl
		- Y precipitator: 1 M NH4OH
Product separation yield		∼88%
Target recovery yield	>90%

Zinc: 63Zn

Zinc is the second most abundant transition metal in the human body being essential to numerous somatic processes, ranging from enzymatic stability to DNA regulation.[143-145] Consequently, zinc dysregulation has been associated with many diseases, such as diabetes, Alzheimer's disease, and various cancers including breast, pancreatic and prostate cancer.[143,145,146] The critical role of zinc homeostasis in addition to its association to numerous diseases makes the development of non-invasive techniques for monitoring fluctuations in somatic zinc levels of great interest.[143,144] Of the many zinc radioisotopes, only three are positron-emitters, with 63Zn (t1/2: 38.5 min, 92.7% β+, 7.3% EC) showing the greatest promise due to its high probability for positron decay and ease of production.[143] For conjugation, zinc only has one oxidation state which makes designing inert chelates much simpler. Despite its attractive properties, 63Zn is limited by its comparatively short half-life, which restricts its viable application to about 2 hours.[143] Regardless, 63Zn radiotracers are thought to be suitable for a variety of short investigative studies such as blood/tissue zinc transport, pancreas and prostate transport kinetics, and blood–brain barrier investigations.[143] A few clinical and preclinical PET studies have used a [63Zn]zinc citrate radiotracer to monitor zinc transport.[144] One preclinical study found that the radiotracer mainly accumulated in the gastrointestinal tract of mice, although a less significant accumulation in the brain of the animals was also reported. A clinical study comparing [63Zn]citrate uptake in brains of Alzheimer's disease patients and age-matched controls found no observed difference in uptake although the clearance of radioactivity was significantly slower in the patient group.[147]

Although several nuclear reactions can be used to produce 63Zn, the only reaction suitable for production using small medical cyclotrons is the 63Cu(p,n)63Zn nuclear reaction.[144] This reaction takes place within the 16 to 5 MeV range using enriched 63Cu as well as natural copper (isotopic composition: 69.17% 63Cu, 30.83% 65Cu). In both cases, 63Zn production was maximized in the 13 to 12 MeV energy range. The primary target material for the reactions were natural copper foils[148] or natural copper that had been electroplated onto a suitable metal substrate such as gold.[146] When a natural copper foil was bombarded with a 16 MeV proton beam, the reported 63Zn yield was 2470 MBq (μA h)−1 (i.e. 66.8 mCi (μA h)−1) with a purity exceeding 99.9%.[148] The primary impurities generated during bombardment were 64Cu (t1/2: 12.7 h, 38.5% β−, 10.8% β+, 50.7% EC), 62Zn (t1/2: 9.2 h, 8.2% β+, 91.8% EC), and 65Zn (t1/2: 243.9 d, 1.4% β+, 98.6% EC).[144,148] The presence of the 65Zn isotopic impurity is problematic for PET applications due to its long half-life. To limit the production of this impurity, bombardments at beam energies exceeding 11 MeV have been recommended.[148]

A method for isolating zinc from the irradiated natural copper targets is described by ref. 148. The natural copper target was first dissolved in concentrated HNO3 before undergoing cation exchange with an AG 50W-X8 resin. The resin retained the zinc product, which can be eluted using a mixture of 0.05 N HCl and 85% acetone. The eluted zinc was obtained with a separation efficiency of 88.0%. Methods for recovering the copper target material have not been discussed as such methods would likely be impracticable due to the naturally high abundance of 63Cu.

Table 12 summarizes key production and purification parameters for 63Zn not mentioned in Table 1.

Production and purification parameters for 63Zn

Radioisotope	63Zn
Target material	Natural copper
Product purity	>99.9%
Major impurities	64Cu, 62Zn, 63Zn, 65Zn
Separation method	Cation exchange chromatography (AG 50-X8 resin)
	- Target dissolution: concentrated HNO3
	- Zn elution: 0.05 N HCl and 85% acetone
Product separation yield	88%
Target recovery yield	n/a

Zirconium: 89Zr

The well matched half-life of 89Zr (t1/2: 78.4 h, 22.7% β+, 77.3% EC) with the pharmacokinetics of antibodies has led to more than 90 clinical studies, highlighted by the success of the ZEPHIR trial.[4] These 89Zr-labelled antibodies are used to plan and monitor cancer treatments, visualize in vivo biodistribution, determine target expression, and conduct radiotherapy dosimetry. With the advancements in its production and labelling technology, 89Zr has been used to label numerous antibodies (i.e. HER1-targeted cetuximab, HER2-targeted trastuzumab, VEGF-targeted bevacizumab, CD20-targeted ibritumomab tiuxetan, transforming growth factor-β (TGF-β)-targeted fresolimumab, CD105-targeted TRC105 and PSMA-targeted 7E11) for both clinical and pre-clinical studies.[1] Most of these examples employed desferoxime (DFO) chelator which is less than ideal for zirconium[149] with significant concentrations of unbound 89Zr found with certain radiotracers such as 89Zr-fresolimumab and 89Zr-bevacizumab. This represents a major drawback of 89Zr as free radiometal will accumulate in the bone and may negatively affect bone marrow health.[150,151] Ongoing efforts to develop improved 89Zr chelates may address many of these concerns.[152-154]

Although several nuclear reactions can be used to produce 89Zr, only the 89Y(p,n)89Zr reaction can be feasibly carried out on small medical cyclotrons. Natural yttrium (isotopic composition: 100.00% 89Y) targets can be bombarded with beam energies between 14 to 9 MeV, although the actual range for reasonable, high purity production is somewhat narrower.[62] Bombardments with beam energies exceeding 11.6 MeV lead to the production of the long half-life contaminant, 88Y (t1/2: 106.6 d, 0.2% β+, 99.8% EC) formed via the 89Y(p,pn)88Y reaction,[155] while beam energies greater than 13 MeV resulted in the production of a second long half-life contaminant, 88Zr (t1/2: 83.4 d, 100.0% EC) which is inseparable from the final product and is formed via the 89Y(p,2n)88Zr reaction.[62] Bombardments with beam energies below 10 MeV are not sufficient for 89Zr production.[62] Many different forms of yttrium targets have been investigated to produce 89Zr. Foil and pressed pellet targets were produced entirely from natural yttrium. Alternatively, targets can be composites composed of yttrium or an yttrium oxide (Y2O3) layered onto a suitable metal substrate such as copper using sputtering and deposition techniques.[156,157] A high purity yield of 58 MBq (μA h)−1 (i.e. 1.6 mCi (μA h)−1)[64,158] has been reported in the literature for 89Zr production on yttrium metal targets in the aforementioned beam energy range. While the natural abundance of 89Y is monoisotopic it contains trace impurities of iron, titanium, and gadolinium. Proton interactions with those trace elements during bombardment form radionuclide impurities such as 48V (t1/2: 16.0 d, 49.9% β+, 50.1% EC), 156Tb (t1/2: 5.3 d, 100.0% EC), 65Zn (t1/2: 243.9 d, 1.4% β+, 98.6% EC) and 56Co (t1/2: 77.2 d, 19.7% β+, 80.3% EC) which negatively affect 89Zr antibody labelling and patient dosimetry.[156]

To ensure high zirconium purity, column-based separation methods are generally utilized, with the most common being weak cation exchange.[158,159] The irradiated yttrium targets were dissolved in strong (1 to 3 M) hydrochloric acid, and the solution was passed through a hydroxamate resin column. The hydroxamate column absorbed the zirconium ions in solution and later eluted using 1 M oxalic acid with a separation efficiency exceeding 70%. Depending on the application, a second column-based separation and washing stage could be used to eliminate the oxalic acid from the complex it forms with zirconium since oxalic acid is highly toxic to the human body. When performed, the result was an zirconium product with a final purity exceeding 99.9%.[156-158] Methods for recovering the yttrium target material have not been discussed because recovery of natural Y targets is not practiced due to its economic characteristics vs. the purchase of new target material.

Table 13 summarizes key production and purification parameters for 89Zr not mentioned in Table 1.

Production and purification parameters for 89Zr

Radioisotope	89Zr
Target material	Natural yttrium and its oxide
Product purity	>99.9%
Major impurities	88Y, 88Zr, 48V, 156Tb, 65Zn, 56Co
Separation method	Cation exchange chromatography (hydroxamate resin)
	- Target dissolution: 1 to 3 M HCl
	- Zr elution: 1 M oxalic acid
Product separation yield	70%
Target recovery yield	n/a

Main group elements

Gallium: 66Ga, 68Ga

Due to its comparatively long half-life, 66Ga (t1/2: 9.5 h, 57.0% β+, 43.0% EC) has received some consideration in the field of PET for longer tracking of radiogallium. However only a few examples of 66Ga PET radiotracers have been reported as the high energy emissions from its decay raises both patient and healthcare personnel safety concerns.[1] In one study, 66Ga was used to label monoclonal antibodies (mAbs) to investigate tumor angiogenesis in animals. The study found that the radiotracer did not clear quickly from the bloodstream of the animals, resulting in moderate amounts of the radiotracer being found in the lungs, liver, heart, kidneys and spleen.[160] A second animal study demonstrated that unbound 66Ga accumulated in the bone and liver, pointing to metabolism or transchelation.[161] On the other hand, 68Ga (t1/2: 67.7 min, 88.9% β+, 11.1% EC) has found extensive use as a small-molecule imaging tool within the last decade.[162-164]68Ga is particularly advantageous for the labelling of such compounds since its half-life is generally comparable to their pharmacokinetics. Another group of targeting molecules that can be radiolabeled using 68Ga are peptides. These radiotracers are easy and inexpensive to produce and show quick blood clearance and tissue diffusivity. Most of the peptides that have been labelled are analogous to somatostatin, a regulatory protein primarily produced in the nervous and digestive systems. These analogues are able to interact with somatostatin receptors found in normal human tissues (i.e. the gastrointestinal tract, kidney, thyroid, spleen, brain, pancreas) and in a variety of malignancies (i.e. neuroendocrine tumors, malignant lymphomas, renal cell carcinomas, prostate cancer, breast cancer and small cell lung cancer), thus facilitating PET investigations. 68Ga has also been used to label melanocortin analogues for investigating melanocortin-receptor containing tissues (i.e. fibroblasts, leukocytes, keratinocytes, melanocytes and endothelial and antigen-presenting cells), and bombesin analogues for investigating bombesin-receptor containing tissue found in tumors (i.e. breast and prostate cancers). It is evident that 68Ga peptide radiotracers have been used in numerous preclinical and clinical studies,[165] including nearly 400 clinical investigations. The promising potential of 68Ga radiopharmaceuticals have found clinical utility through successful imaging of PSMA, somatostatin receptors, and FAPI.[163,164,166]

66Ga can be produced from the bombardment of both natural zinc (isotopic composition: 49.17% 64Zn, 27.73% 66Zn, 4.04% 67Zn, 18.45% 68Zn, 0.61% 70Zn) and 66Zn-enriched foils,[167] or after electroplating onto gold or silver substrates.[160] The radionuclide is primarily produced via the 66Zn(p,n)66Ga nuclear reaction at a minimum beam energy of 6 MeV,[161] but irradiations are typically carried out at much higher energies (i.e. such as 9.8 MeV,[167] 13 MeV,[160] 14.5 MeV (ref. 167) and 15 MeV (ref. 161)). Yields of 511 MBq (μA h)−1 (i.e. 13.8 mCi (μA h)−1) with a purity of 99.96% have been achieved from the 14.5 MeV bombardment of 98.9% enriched 66Zn targets[167] although higher yields of 700 MBq (μA h)−1 (i.e. 18.9 mCi (μA h)−1) have been reported elsewhere.[64] The primary impurity produced during irradiation was 67Ga (t1/2: 3.3 d, 100.0% EC).[161,167]

While 68Ga has traditionally been produced with 68Ge/68Ga generators producing 1850 MBq and 3700 MBq of 68Ga, cyclotrons have alternatively been used for its production via the proton bombardment of 68Zn-enriched targets. Recent progress using liquid targets has demonstrated modest production yields above 68Ge/68Ga generators. However solid targets show significant potential for much higher capacity production.[168] Solid targets may be foils or zinc-coated metal backings (i.e. aluminum, gold or silver) that are generally prepared via electroplating or sputtering.[169-171] The primary reaction governing the production of 68Ga is the 68Zn(p,n)68Ga nuclear reaction, which takes place within the 14 to 11 MeV beam energy range.[62] Researchers have reported approximate yields of 1240 MBq (μA h)−1 (i.e. 33.5 mCi (μA h)−1) from the bombardment of natural zinc targets at a beam energy of 12.5 MeV,[171] and yields of 1380 MBq (μA h)−1 (i.e. 37.3 mCi (μA h)−1) for the bombardment of enriched >99% 68Zn electroplated silver targets with a beam energy of 12.5 MeV.[169] A novel enriched zinc target design (backed with silver) irradiated at 13 MeV demonstrated a final yield of 1213 MBq (μA h)−1 (i.e. 32.8 mCi (μA h)−1) at end of separation (EOS) with a purity exceeding 99%.[172] The major impurities produced from the reaction were 66Ga (t1/2: 9.5 h, 57.0% β+, 43.0% EC) and 67Ga (t1/2: 3.3 d, 100.0% EC), and they are inseparable from the desired isotope.[62] Measures can be taken to reduce 66/67Ga production during target irradiation including: (1) reduce the amounts of 66Zn and 67Zn in the targets by using highly enriched 68Zn to limit the 66Zn(p,n)66Ga and 67Zn(p,n)67Ga nuclear reactions, and (2) limit the maximum beam energy to 12 MeV to reduce the 68Zn(p,2n)67Ga nuclear reaction[62] respectively.[169]

Gallium can be extracted from the zinc targets through two major methods. The first method made use of cation exchange chromatography where irradiated targets were dissolved in 10 to 12 M HCl and passed through a cation exchange resin, such as the AG 50W-X8 or AG 50W-X2, which retained both the zinc target material and gallium product. The zinc target material was then eluted using 10 M HCl, and was recovered with efficiencies ranging from 76 to 80%. The gallium product was subsequently eluted with 4 M HCl with an overall separation efficiency as high as 90%.[160,167,170] To increase the purity of the obtained gallium, a washing stage using 7 N HCl was carried out immediately before eluting the gallium product.[170] Alternatively, this procedure was carried out in a fully automated way using commercially available modules with separation efficiencies up to 80%.[169,171-173] Unreacted zinc target material was recovered from the 10 M HCl solution through electrolysis using a voltage of 5 V and a current of 1 A with efficiencies exceeding 76%.[169]

The second method for isolating gallium is liquid–liquid extraction. Preparation steps for this procedure involved dissolving the irradiated zinc targets in 12 M acid and diluting the resulting solution to 7 M with water. The dissolution was then subjected to a series of extraction and back-extraction stages using various solvents (i.e. diisopropyl ether, 7 M HCl and water) until gallium was obtained with the desired purity.[167] Maximum separation efficiency for the gallium product was 79% and the recovery efficiency for the zinc target material exceeded 99%.

Table 14 summarizes key production and purification parameters for 66Ga and 68Ga not mentioned in Table 1.

Production and purification parameters for 66Ga and 68Ga

Radioisotope		66Ga, 68Ga
Target material	66Ga	Natural and enriched zinc
Product purity		>99.9%
Major impurities		67Ga
Target material	68Ga	Enriched zinc
Product purity		>99%
Major impurities		66Ga, 67Ga
Separation method	1	Cation exchange chromatography (AG 50W-X8 and AG 50W-X2 resins)
		- Target dissolution: 10 to 12 M HCl
		- Zn elution: 10 M HCl
		- Ga elution: 4 M HCl
Product separation yield		90%
Target recovery yield		76 to 80%
Separation method	2	Liquid–liquid extraction
		- Target dissolution: 12 M HCl
		- Extraction solvents: Diisopropyl ether, 7 M HCl and water
Product separation yield		79%
Target recovery yield		>99%

Arsenic: 72As

72As (t1/2: 26.0 h, 87.8% β+, 12.2% EC) is considered a promising PET radionuclide for several reasons including its long nuclear half-life which permits the radiolabeling of biomolecules with longer biological half-lives.[174,175] The resulting radiotracers can be used in several extended biochemical and physiological process studies such as immunoimaging and receptor mapping. A second favorable characteristic arises from the unique soft Lewis acid properties of arsenic to covalently bond to thiol (or sulfhydryl) groups.[1,174,175] Theoretically, this allows arsenic's radioisotopes to directly label numerous biomolecules with little to no chemical modification.[1,175] Several radioarsenic labeled antibodies and nanoparticles have been reported, although none have entered clinical use.[176-178] Finally, the positron emitting 72As provides a matched pair with 77As (t1/2: 38.8 h, 100% β−) which will behave identically in vivo for imaging and radiotherapy, respectively.[179] The theranostic advantage of the pairing makes up for the poor image resolution resulting from arsenic's large positron range (i.e. 3.6 mm in soft tissue).[58] Despite its favorable qualities however, research on the PET applications of 72As has been slow as a result of the limited supply of isotopically-enriched germanium, the primary material for the production of various arsenic radioisotopes,[1,175] although selenium has been used on occasion.[180] The limited supply of isotopically-enriched germanium increases target costs which translates into an overall increase in 72As production costs.[175]

72As can be produced using a 72Se/72As generator, or using a solid target cyclotron.[1] With respect to the latter, 72As can be produced using a variety of nuclear reactions such as 72Ge(p,n)72As, 73Ge(p,2n)72As, 74Ge(p,3n)72As and 76Se (p,x)72As.[180,181] Of the listed reactions, the most commonly used was the 72Ge(p,n)72As nuclear reaction which took place within the 18 to 8 MeV range,[174] noting that 72As production at beam energies exceeding 15 MeV contained a greater number of impurities in the final product.[174] The targets used for 72As production were typically 72Ge-enriched germanium disks[178] and 72Ge-enriched germanium oxide that had been electroplated onto substrates made of copper[174] and niobium.[175] Additionally, natural germanium oxide[174] or natural germanium (isotopic composition: 20.57% 70Ge, 27.45% 72Ge, 7.75% 73Ge, 36.50% 74Ge, 7.73% 76Ge) foil[182] have also been used to produce 72As in other studies. The primary impurities produced during bombardment were 71As (t1/2: 65.3 h, 28.3% β+, 71.7% EC), 74As (t1/2: 17.8 d, 34.0% β−, 29.0% β+, 37.09% EC), 76As (t1/2: 26.2 h, 100.0% β−),69Ge(t1/2: 39.1 h, 24.0% β+, 76.0% EC) and 67Ga (t1/2: 3.3 d, 100.0% EC) (175,178,182). The 72As production yield for >96% enriched germanium oxide electroplated onto a niobium substrate and bombarded with 16.1 MeV protons was 90 MBq (μA h)−1 (i.e. 2.4 mCi (μA h)−1), with purity exceeding 99%.[178] The theoretical 72As yield for natural germanium oxide electroplated onto a copper substrate and bombarded with a proton beam energy of 15 MeV to 8 MeV was 60 MBq (μA h)−1 (i.e. 1.6 mCi (μA h)−1).[174]

Methods for isolating arsenic from the germanium targets have been described extensively in Enferadi et al.,[174] Ellison et al.[175,178] and Guin et al.[183] In general, the methods required dissolution of the germanium targets using solutions of 6 M NaOH or hot aqua regia, followed by the addition of 6 to 10 N HCl. The addition of the HCl was necessary to convert the germanium metal to GeCl4, which is significantly more volatile than the metallic arsenic product. Thirty percent hydrogen peroxide was then added to oxidize the arsenic to the As(v) state. The final dissolution was evaporated to near dryness, after which distillation was performed on the resulting residue. About 97% of the 72As was left behind as residue from the distillation, and then further purified through ion exchange chromatography after dissolution in 10 M HCl. Anion exchange was carried out by passing the arsenic solution through a Bio-Rad AG 1 × 8 resin (200–400 mesh), which retained the arsenic product as well as any residual germanium and gallium impurities. The arsenic could be collected by eluting with 10 M HCl, and the procedure had a separation efficiency of 93%, resulting in an overall As(v) separation efficiency of about 90%. However, if it is desirous to have the arsenic in the As(iii) state instead of its As(v) state, SPE can be carried out subsequently. The SPE procedure had an efficiency of 57% and after its application, the overall separation efficiency of the As(iii) product was 51%. Recovery of the germanium target material could be achieved by treating the acidic condensate from the dissolution stage with 6 M NaOH. This neutralized the condensate, and converted GeCl4 to GeO which could be easily reduced to metallic germanium when sufficiently heated. The overall recovery efficiency reported by Ellison et al.[178] was about 60%, however it should be emphasized that higher recovery efficiencies could be achieved if the recovery process was optimized.

Table 15 summarizes key production and purification parameters for 72As not mentioned in Table 1.

Production and purification parameters for 72As

Radioisotope	72As
Target material	Natural and enriched germanium and their oxides
Product purity	>99%
Major impurities	71As, 74As, 76As, 67Ga, 69Ge
Separation method	Distillation, Anion exchange chromatography and SPE
	- (Distillation) target dissolution: 6 M NaOH or hot aqua regia + 6 to 10 N HCl + 30% H2O2
	- (Chromatography) target distillation: 10 M HCl
	- As elution: 10 M HCl
Product separation yield	>90% As(v), ∼51% As(iii)
Target recovery yield	∼60%

Conclusion

Given the surge in clinical use and interest in several metallic radioisotopes, solid target systems for small (≤24 MeV) medical cyclotrons are rapidly evolving to enable large-scale production with consistent yield, composition and quality. With this, many factors need to be considered when choosing to develop in-house radiometal production capabilities, including (i) the variability in medical cyclotron design (i.e. beam energy, current and machine configuration); (ii) target isotope availability; (iii) quality/quantity variations when compared to alternative, non-cyclotron production routes, and (iv) the intended application of the radioisotope. While many of the metallic radioisotopes discussed herein are being developed as imaging agents unto themselves, several have the potential to be theragnostic partners to a range of chemically diverse therapeutic radioisotopes, and require considerations in chemical compatibility and physical/biological half-life. Of those diagnostic radioisotopes with theragnostic applications, the community will likely see near-term adoption of cyclotron-produced 68Ga and 89Zr, and a resurgence of interest in 64Cu. Radioisotopes with element-identical theragnostic application such as 86Y, 44Sc and 124I may encounter increases in use as long as their production and/or recycling methodologies can be sufficiently optimized. Yet others, such as cyclotron-produced 99mTc, may emerge as important enablers for those regions in which supply via non-cyclotron-based alternatives remain problematic.

Alpha particle

Beta emission

Positron emission

Days

Electron capture

The endpoint energy (i.e. maximum energy) of an emitted positron

End of bombardment

Hours

Isomeric transition

Kiloelectron volts

Megabecquerels

Megaelectron volts

Microamperes

Millicuries

Minutes

Molar (i.e. mols of acid per liter)

Neutron

Normal (i.e. moles of dissociated H+ ions per liter)

Degrees celsius

Positron emission tomography

Positron

Single photon computed tomography

Solid phase extraction

Half-life

Gamma ray

Conflicts of interest

There are no conflicts of interest to declare.