Multiple production routes for ¹⁹¹Pt have been explored using either osmium or iridium targets bombarded with protons, deuterons, or α-particles highlighted in Table 1. Bonardi et al. (21) produced no-carrier-added (n.c.a.) ¹⁹¹Pt—which complemented earlier work from Parent et al. (22) and Sharma and Smith (23)—while achieving 170 MBq/μg with decontamination factors of >10⁶ via two optimized radiochemical separations (Sn(II)/ether vs. NH₂OH/dithizone extraction). Obata et al. (20) measured excitation functions, finding peak cross sections of ∼623–635 mb for ¹⁹¹Pt at ∼26–32 MeV, with theoretical thick-target yields of ∼108–192 MBq/μA-h for both proton and deuteron irradiation using ^natIr or ¹⁹³Ir targets. Furthermore, they noted ∼25 MeV protons as the optimal energy, though advanced target dissolution methods were needed due to iridium exhibiting superior resistance to acid (20). Obata et al. (24) addressed this by using an alkali-fused Ir target and in situ HCl digestion, followed by solvent extraction and anion exchange, yielding 17.4 ± 1.1 MBq/μA-h at EOB (7.1 ± 0.4 MBq/μA-h post separation) with >99% radionuclidic purity.

Areberg et al. (25) demonstrated the first use of [¹⁹¹Pt]cisplatin (Figure 1A) for tumor imaging. Fourteen patients received [¹⁹¹Pt]cisplatin (≥99% radionuclidic purity)—synthesis based on the work reported by Hoeschele et al. (26)—and showed clear gamma-camera visualization of tumors in multiple anatomical sites (25). Building on this, the same group (27) reported organ-specific absorbed and effective doses for [¹⁹¹Pt]cisplatin (and ^193mPt/^195mPt analogs)—advancing beyond earlier whole-body mean dose calculations by Lange et al. (28).

Recent work has leveraged the auger electrons emitted from ¹⁹¹Pt towards targeted therapy. Obata et al. (29) developed a resin-based method to isolate n.c.a. [^188,189,191Pt]Pt(II)Cl₄²⁻, and a one-pot radiosynthesis of [*Pt]cisplatin, yielding 30%–40% without intermediate evaporation. Using tracer-level [^189,191Pt]cisplatin, Obata et al. (30) showed only 0.6% overall cell uptake in cells, yet ∼20% of internalized platinum localized to the nucleus and ∼2% bound covalently to DNA (0.28 ± 0.02% ID/mg) (30). Single-cell assays confirmed that auger electrons caused direct DNA double-strand breaks, validating [^189,191Pt]cisplatin as an extremely localized therapeutic with minimal systemic toxicity (30). Obata et al. (31) compared ¹⁹¹Pt coordination to Cys, DTPA, EDDA (Figures 1B–D) to evaluate the in vitro behavior to analogous ¹¹¹In-labeled (t_1/2 = 2.8 d, 100% EC) agents (31–34). They demonstrate that free ¹⁹¹PtCl₄²⁻ undergoes rapid thiol coordination with Cys, significantly reducing protein binding at 60°C (∼10%) compared to 45°C (∼42%) (31). In contrast, labeling with DTPA and EDDA resulted in moderate radiochemical yields (70%–80%) and reduced protein binding only to ∼42% and ∼30%, respectively (31). Furthermore, ¹⁹¹Pt was complexed with the DNA-targeting dye Hoechst33258 via DTPA ([¹⁹¹Pt]Pt-DTPA-Hoechst33258; >95% radiochemical purity) and Cys ([¹⁹¹Pt]Pt-Cys-Hoechst33258; ∼90% radiochemical purity) to compare with [¹¹¹In]In-DTPA-Hoechst33258 (>95% radiochemical purity) and found both ¹⁹¹Pt-based complexes displayed one order of magnitude greater DNA-binding than the ¹¹¹In analog (31). Notably, [¹⁹¹Pt]Pt-Cys-Hoechst33258 induced DNA damage more effectively than its DTPA counterpart, suggesting that thiol-based ¹⁹¹Pt labeling enhances delivery to DNA and elevates therapeutic potential (31).

Obata et al. (35) conjugated ¹⁹¹Pt to a oncogene MYCN-specific pyrrole-imidazole polyamide (PIP) scaffold (¹⁹¹Pt-MYCN-PIP) bearing Cys, tri-arginine (R3) for cellular penetration (36), and a fluorescent compound coumarin (GCC-Cys-R3-coumarin control, ¹⁹¹Pt-GCC-PIP). The MYCN gene is a transcription factor that is amplified in human neuroblastoma and is related to the patient's prognosis (35). Noted in Obata et al. (35), targeting cancer-related genes with PIPs have been utilized in preclinical studies with mice and marmosets (37, 38), along with developments of MYCN-targeting PIP in Yoda et al. (39) showed promising specific targeting ability and therapeutic effects. With 50%–70% radiochemical yield and >95% radiochemical purity, ¹⁹¹Pt-MYCN-PIP achieved ∼10-fold higher uptake and DNA-binding in MYCN-amplified vs. non-amplified neuroblastoma cells, and reduced MYCN expression in vitro (35). Omokawa et al. (40) synthesized a sugar-conjugated platinum complex, FGC-Pt (cis-dichloro[(2-fluoro-α-_D-glucopyranosidyl)propane-1,3-diamino-2-propyl]platinum) (41) and labeled it with n.c.a. ¹⁹¹Pt by either direct activation (61.7% radiochemical purity) or post-labeling of neutron-activated [¹⁹¹Pt]K₂PtCl₄ (14.5 ± 7.3% radiochemical yield; 93.8% radiochemical purity), with the latter method providing significantly higher yield and purity. In healthy mice, both [¹⁹¹Pt]FGC-Pt preparations showed almost identical biodistribution at 24 h—and γ-counting correlated with ICP-MS measurements (r = 0.92, p < 0.05), confirming their utility for quantitative imaging (40). Most recently, Obata et al. (42) developed a PSMA-targeting ¹⁹¹Pt-trithiol complex showing a 46-fold uptake advantage in PSMA⁺ vs. PSMA⁻ tumors (in vitro), outperforming the Cys-based analog (2.2 ± 0.3).

The production routes to obtain ^193mPt are shown in Table 1. Uddin et al. (44) measured the experimental excitation function for the ¹⁹²Os(α,3n)^193mPt reaction—building on previous work by Hilgers et al. (45)—reporting a peak cross section of 1.47 ± 0.19 b (66.63 keV x-ray) and 1.53 ± 0.21 b (135.5 keV γ-ray), both at 36.4 ± 0.2 MeV. As the authors noted, several methods for the dissolution of osmium had been reported (45–47). An optimized electrolytic technique was carried out to prepare highly enriched ¹⁹²Os, where the authors noted, low electrodeposition yields were minimal to this point (44). Jones et al. (48) reported a maximum deposition of 9.5% at pH 13—which encouraged the authors to focus on this effort. Chakrabarty et al. (47) on the other hand, reported a high yield of ∼80% for an isotopically enriched osmium sample, where efforts by Uddin et al. (44) were devoted to optimizing the electrolytic deposition process. By using their electrolyte, a maximum electrodeposition yield of ∼75% at pH 12.8 was achieved for the enriched osmium, with 15% lower for natural osmium. Adopting radiochemical separation techniques from Bonardi et al. (21) and Hilgers et al. (45), Uddin et al. (44) oxidized the osmium sample with the Ni backing in concentrated nitric acid and evaporated out the liquid. The OsO₄ was distilled and trapped in 4.7 N KOH, while the residual Pt was dissolved in 3 N HCl, and reduced from Pt(IV) to Pt(II) with SnCl₂. The [Pt(SnCl₃)₅]³⁻ anion was extracted into the ether phase, achieving a radiochemical yield of 80%–96% across 20 individual osmium samples (44). Compared to Hilgers et al. (45) and predictions from nuclear model codes [TALYS (49) and STAPRE (50)], the measured excitation functions from (44) showed excellent agreement across the energy range. Integral yield calculations 1 μA for 1 h yielded ∼10 MBq/μA-h of ^193mPt and ∼0.06 MBq/μA-h of ^195mPt within the optimal energy window of 40→30 MeV, establishing ¹⁹²Os(α,3n)^193mPt as the most effective cyclotron-based route for producing clinically relevant quantities of ^193mPt (44).

Uddin et al. (51) demonstrated a small-scale, cyclotron-based production of ^193mPt via ¹⁹²Os(α,3n)^193mPt reaction, achieving 99% radionuclidic purity and a specific activity of 1 GBq/μg ^193mPt, effectively overcoming the limitations of low specific activity associated with reactor-based (n,γ) production on ¹⁹²Pt targets as highlighted by Azure et al. (52). Target dissolution and OsO₄ distillation, previously reported in Hilgers et al. (45) and Uddin et al. (44), combined with a SnCl₂-ether extraction sequence developed by Ahmed and Koch (53) and Koch and Yates (54), enabled 85% recovery of enriched Os and 90% radiochemical yield of ^193mPt (51). The experimental batch yield at EOB was ∼10 MBq using a 1.6 μA beam for 3 h, corresponding to ∼40% of the theoretical value predicted from the excitation function of the ¹⁹²Os(α,3n)^193mPt reaction (44, 51). In contrast, (n,γ) production using 5 mg of 57% enriched ¹⁹²Pt (ϕ = 4 × 10¹⁴ n cm⁻² s⁻¹; 7 d) yielded 3 GBq with a specific activity of only 0.6 MBq/μg ^193mPt (51, 52). Moreover, α-induced production results in minimal ^195mPt impurity (0.5%) compared to the (n,γ) route (∼12%), emphasizing its suitability for scalable, high-purity Auger-electron radionuclide production (51, 52).

Lange et al. (55) performed the radiosynthesis of cisplatin labeled with ^193mPt and subsequent biodistribution on rabbits and mice. From their findings following intravenous injection, most of the activity accumulated in the kidneys, urine, and liver, with rapid excretion of the radiolabeled complex (79% eliminated by 24 h) (55). A year later, the same group (28), performed distribution studies and dose calculations for ^193mPt and ^195mPt and reported similar biodistribution results from the prior study, along with similar behavior with the ^195mPt-labeled analog (28). Azure et al. (52) performed the first microscale synthesis of carboplatin labeled with ^193mPt, reporting [^193mPt]carboplatin (Figure 2) uptake had saturated by 2–3 in V79 cells, and similar findings to [^195mPt]transplatin in Howell shown in Figure 3A (56). Notably, ∼70% of internalized ^193mPt was in the nucleus, with ∼60% of that bound to DNA (52)—substantially higher targeting than observed with ^195mPt (25% cellular radioactivity in nucleus, 42% bound to the DNA) (56).

High specific activity ^195mPt is best obtained via indirect reactor routes or enriched target irradiation, and all its production routes are shown in Table 1. Knapp et al. (59) produced n.c.a. ^195mPt by irradiating enriched ¹⁹³Ir to produce ^195mIr (t_1/2 = 3.67 h)—via ¹⁹³Ir(n,γ)¹⁹⁴Ir(n,γ)^195mIr—which then decays (β⁻) to ^195mPt while taking advantage of the high thermal flux of the High-Flux Isotope Reactor (HFIR) at Oak Ridge National Laboratory (ORNL) to surpass the specific activities achievable by direction ¹⁹⁴Pt(n,γ) or ¹⁹⁵Pt(n,n') routes (26, 59, 60). ^195mPt was separated from bulk Ir via thiourea-HCl elution on cation resin—where methods were previously reported by Siegfried et al. (61) and Berg and Senn Jr (62).—yielding high purity of ^195mPt (59). Hilgers et al. (45) measured the ¹⁹²Os(α,n)^195mPt reaction, reporting a maximum cross section of 4.4 ± 0.7 mb at 22.1 ± 0.7 MeV, and projected ∼0.09 GBq yield—about an order of magnitude lower than reactor methods (63). Vosoughi et al. (64) irradiated ^natPt in a reactor (3 × 10¹³ n cm⁻² s⁻¹, 30 h, 5 MW power), obtaining 16.28 MBq of ^195mPt. The product was allowed to decay for 48-h due to short-lived Au/Pt impurities and solvent extraction separation was performed following an established method by Vimalnath et al. (65), they obtained radiochemical yield and separation efficiency of ≥99% and 99.4%, respectively (64). However, specific activity was only ∼0.8 MBq/mg, much lower compared to the reported <37 MBq/mg (59) and 15.9 MBq/mg (66) that were achieved with enriched ¹⁹⁴Pt targets at ORNL and SAFARI-1 reactors, respectively (64). Bodnar et al. (67) aimed to develop a method of preparation of ^195mPt with high specific activity via a photonuclear reaction. Obtaining ^195mPt via the ¹⁹⁷Au(γ,np)^195mPt reaction, they implemented a novel technique for gold extraction and produced high specific activity ^195mPt >1 Ci/mg (67). Wawrowicz and Bilewicz (57) tested the double-neutron capture approach but proved it to be impractical due to an unknown second-step cross section and difficult target dissolution, yielding <10% recovery. Therefore, until nuclear data and chemical processing improves, double-capture routes offer no advantage (57).

Leveraging reactor-produced n.c.a. ^195mPt, Zeevart et al. (66) prepared [^195mPt]cisplatin for a Phase 0 clinical trial on healthy patients (66). Using an optimized synthesis—building on work by Smith (68)—[^195mPt]cisplatin was obtained in >95% radiochemical yield (^195mPt and ¹⁹⁷Pt combined), with co-produced impurities (¹⁹²Ir, ¹⁹¹Pt, Au isotopes) below detection (66). Sathekge et al. (69) obtained whole-body planar scans and SPECT/CT images up to 144 h post-[^195mPt]cisplatin injection in five volunteers. Bodnar et al. (67) also optimized the radiosynthesis of ^195mPt-cisplatin from earlier works of Chernyaev (70) and Dykiy et al. (71) for in vitro and in vivo evaluation. They confirmed induced necrosis and apoptosis in vitro at mass doses over five orders of magnitude lower than conventional cisplatin doses (67, 70, 71). In mice with Ehrlich tumors, a single [^195mPt]cisplatin dose achieved 65% tumor growth inhibition—and 100% animal survival—vs. 35% inhibition by conventional cisplatin (67).

Apart from cisplatin analogs, Aalbersberg et al. (72) conducted a preclinical evaluation of ^195mPt SPECT using NanoSPECT/CT and U-SPECT⁺/CT scanners following thermal neutron irradiation of ¹⁹⁴Pt in the High Flux Reactor (HFR) in Petten, the Netherlands. They achieved sub-millimeter resolution and linear quantification over a wide activity range (0.035–4.36 MBq), confirming accurate in vivo Pt distribution measurements (72). SPECT-based quantification, calibrated using a ^195mPt dilution series, correlated strongly with ex vivo gamma-counting and graphite-furnace atomic absorption spectroscopy (GF-AAS), validating accurate in vivo quantification of platinum biodistribution (72). Although the study validated the feasibility of ^195mPt SPECT in small animals, the authors noted limitations including low specific activity 3–4 MBq per injection, small sample size, and the need to improve purification methods to extend imaging with radiolabeled cisplatin (72). Muns et al. (73) characterized a metal-organic linker, [ethylenediamineplatinum(II)]²⁺ (called Lx) with antibody-drug conjugates (ADCs) for in vivo stability and tumor targeting using ^195mPt and ⁸⁹Zr (t_1/2 = 78.36 h). Nearly identical ^195mPt and ⁸⁹Zr biodistributions in tumor-bearing mice confirmed the in vivo stability of the Pt(II)-histidine coordinative bond within Lx (73). However, the amounts of platinum incorporated into Lx-based ADCs and the specific activity of ^195mPt were too low to support preclinical or clinical SPECT imaging studies (73).

Nadar et al. (74) synthesized a n.c.a ^195mPt-BP complex, shown in Figure 3B, to achieve bone-targeting Auger-electron therapy. This complex was introduced previously by Margiotta et al. (75). In healthy C57BL/6N mice (2.5 mM Pt, 24 h), ICP-MS showed a 4.5-fold higher uptake in hard tissue (12.18 ± 0.56%ID/g) vs. its bisphosphonate-free precursor Pt(NO₃)₂(en) (2.69 ± 0.26%ID/g), and accomplished reducing off-target retention in many organs including the kidney (5.70 ± 0.15 vs. 3.38 ± 0.28%ID/g) (74). Pt-BP also induced minimal Pt-DNA adduct formation (<0.5% of total Pt in most tissues; kidney: 2.8%, spleen: 1.4%) compared to the precursor (kidney: 4.8%, spleen: 9.8%), confirming that bisphosphonate conjugation both enhances bone selectively and spares healthy tissues for DNA damage (74). In micro-SPECT/CT studies, ^195mPt-BP rapidly localized to growth plates, whereas ^195mPt(NO₃)₂(en) accumulated specifically in soft tissues (74). Laser ablation ICP-MS (LA-ICP-MS) further validated 73.5% co-localization of ^195mPt-BP, showing almost a four-fold increase accumulation of Pt in bone compared to the precursor—highlighting its specific bone-binding mechanism (74). In a subsequent study, Nadar et al. (76) treated mice with intratibial bone tumors using ^195mPt-BP and [^195mPt]cisplatin. ^195mPt-BP exhibited significantly higher and sustained accumulation in metastatic lesions with 2.8–3.3-fold higher uptake than the contralateral tibia, indicating selective targeting (76). In contrast, ^195mPt-cisplatin exhibited lower uptake (≤3.7%ID/g) with no evidence of lesion selectivity at any time point (76). Therapeutic efficacy was assessed via γ-H2AX staining—a biomarker specific for double-strand DNA breaks—revealing that ^195mPt-BP induced a 4.6-fold greater fraction of γ-H2AX-positive tumor cells (1.66 ± 0.4%) compared to ^195mPt-cisplatin (0.36 ± 0.1%) and an 11-fold increase over non-radioactive Pt-BP (0.15 ± 0.1%) (76). These results confirm that bone-targeted ^195mPt-BP delivers Auger radiation directly to tumor-associated bone lesions with superior efficacy compared to [^195mPt]cisplatin (76).

Most recently, de Roest et al. (77) explained [^195mPt]cisplatin uptake in cisplatin-sensitive and -resistant head-and-neck cancer models. They found that cisplatin-resistant HNSCC cell line (VU-SCC-OE) accumulated more [^195mPt]cisplatin in DNA and exhibited greater capacity to repair cisplatin-induced crosslinks compared to the cisplatin-sensitive HNSCC cell line (VU-SCC-1131), with a DNA retention ratio of 3.4 vs. 1.45 (77). The authors concluded that [^195mPt]cisplatin imaging is not predictive of tumor sensitivity to cisplatin but may serve as a tool for assessing cisplatin-related off-target toxicity (77).

A variety of production methods exist for ¹⁰³Pd, including reactor- and accelerator-based routes which is described in Table 2. Sudar et al. (80) reported a maximum cross-section of 505 ± 26 mb at 10.05 ± 0.19 MeV (via x-ray measurements) and identified the optimal energy range for maximizing specific cross-sections (300–500 mb) and yields to be between 8 and 12 MeV. The authors compared between neutron-counting studies—including those by Albert (81), Johnson et al. (82), and Hansen and Albert et al. (83)—and activation measurements—Blaser et al. (84), Harper et al. (85), Treytl and Caretto (86), Mukhammededov and Vasidov (87), and Hermanne et al. (88)—from energies 2.8–400 MeV, confirming good agreement across studies, with discrepancies at lower energies mainly attributing to systematic uncertainties and differences in target preparation (80). Building on this, Hussain et al. (79) provided a comprehensive evaluation of all accelerator-based production routes for n.c.a. ¹⁰³Pd, integrating six reaction channels (89–95) reported in Table 2 using EXFOR data and key literature sources, and by normalizing the raw measurements with three nuclear-reaction codes (STAPRE (50), TALYS (49), and EMPIRE (96)) to produce recommended excitation functions with 95% confidence limits. Furthermore, they investigated another indirect precursor of ^natPd(p,x)¹⁰³Ag→¹⁰³Pd (97, 98) that can form up to 70% of total ¹⁰³Pd via ¹⁰³Ag decay but suffers from long-lived impurities and complex chemistry, limiting their large-scale clinical applicability (79).

Manenti et al. (99) optimized n.c.a. ¹⁰³Pd production via the ¹⁰³Rh(d,2n) reaction using a stacked-foil activations method at deuteron energies from 5 to 33 MeV on the JRC-Ispra and ARRONAX cyclotrons (beam currents 100–170 nA, 1 h irradiations). Experimental cross-sections rose steadily above the 3.62 MeV threshold, peaking at 1,261 ± 71 mb at 15.0 ± 0.4 MeV, and then declined gradually at higher energies (99). Comparison with prior data and models showed good agreement with Hermanne et al.'s (92) γ-ray measurements and close agreement with the recommended values of Hussain et al. (79), while Ditroi et al. (100) reported cross-sections up to 15% lower (99). Furthermore, EMPIRE-II and EMPIRE-3.2.2 (96) both reproduced the experimental curve within uncertainty, whereas TENDL-2015 (49) underestimated cross-sections above 10 MeV (99). Thick-target yields (TTYs) were computed from integrated thin-foil data, reporting that up to ∼12 MeV, deuteron-induced TTYs matched those of the ¹⁰³Rh(p,n) route (99). Above 12 MeV, deuteron yields exceed proton yields by up to a factor of two—reflecting the higher (d,2n) cross-section at medium energies and marking deuteron beams as especially attractive for high-throughput production (99). Radionuclidic purity within the 5–33 MeV window is excellent as authors noted only ¹⁰¹Pd (t_1/2 = 8.47 h) co-produces above its 22 MeV threshold, greatly simplifying post-irradiation separation (99). The higher stopping power of 13.3 MeV deuterons also reduces target mass, with a 188 μm Rh foil suffices for full absorption vs. 214 μm for 10.5 MeV protons, marginally easing radiochemical separation (99). Despite these advantages, high-energy deuteron cyclotrons remain scarce, which may constrain routine clinical-scale ¹⁰³Pd production (99).

Ohya et al. (101) demonstrated an efficient method for producing no-carrier-added ¹⁰³Pd, followed by radiochemical separation and target material recycling. The radiochemical separation incorporated a Bi-Rh alloying pretreatment at 500°C, enabling high-yield dissolution of the Rh target and achieving a 93 ± 4% dissolution efficiency (101). Following co-precipitation to remove Bi and palladium radionuclides—including ¹⁰⁰Pd and ¹⁰³Pd— a dimethylglyoxime (DMG)-based extraction, achieved 99 ± 1% yield (101). The radiopalladium was subsequently back-extracted from chloroform using aqueous ammonia, yielding 97 ± 2% of [¹⁰³Pd(NH₃)₄]²⁺ (101). The entire process was completed within 3.5 h, yielding a ¹⁰³Pd radiochemical yield of 87% and >99% radionuclidic purity (101). During the recycling process, 91 ± 3% of the Rh target was efficiently recovered with minimal Bi contamination (9 μg per 50 mg Rh) through cation exchange purification; therefore, providing a framework for clinical-scale ¹⁰³Pd radionuclide production (101).

Krol et al. (102) presented the first feasibility study on the production of ¹⁰³Pd via the ¹⁰³Rh(p,n)¹⁰³Pd reaction using cyclotron irradiation of a liquid target. By achieving an EOB activity of 1.03 ± 0.05 MBq (20.06 ± 0.97 MBq/μA) under optimized conditions (30 ± 0.5 μA, 1 h irradiation, 200 psi top up pressure, and 16.4 mg/ml metal-salt concentration), they demonstrated that liquid targets can reliably yield research-scale quantities of ¹⁰³Pd suitable for radiochemistry (102). Furthermore, an anion-exchange separation using Dowex 1 × 8 resin with 1 M HNO₃ for rhodium elution achieved a 90.1 ± 2.1% recovery from the irradiated target solution, while a 1:1 mixture of 0.5 M NH₃ + NH₄Cl for palladium elution resulted in a 103.8 ± 2.3% recovery—achieving a rhodium reduction factor of ∼10⁶ (102). More recently, Laouameria et al. (103) addressed previous limitations by developing a diffusion-driven extraction to separate ¹⁰³Pd from its stable ¹⁰³Rh target, relying on the metals' differing vapor pressures. Using their radionuclide separation equipment (RSE), they achieved an overall separation of 17 ± 2% and deposition yields of 77 ± 2% on Nb foil and 49 ± 2% on ZnO/W discs, respectively (103). Furthermore, using the ZnO/W disc substrate, the method produced 31.9 MBq EOB with a specific activity of 8.1 GBq/g, representing a streamlined alternative to traditional wet-chemistry approaches for Auger-emitter production (103).

Blasko et al. (104) conducted a study on a cohort of 230 men with clinically T1-T2 prostate cancer treated exclusively with ¹⁰³Pd brachytherapy. The study found an overall 9-year biochemical control rate of 83.5%, with PSA-only progression observed in just 4.3% of patients (104). The findings validated ¹⁰³Pd brachytherapy as an effective and durable treatment option cross a range of risk groups, achieving high biochemical and clinical outcomes in patients with organ-confined prostate cancer (104).

Li et al. (105) developed an electroless plating method to fabricate ¹⁰³Pd brachytherapy seeds by directly depositing ¹⁰³Pd onto carbon bar substrates, thereby eliminating the metallic pre-coatings and the complex pellet assemblies required from prior reports. Under hydrazine-based bath conditions optimized in Li et al. (106), this method achieves a 98% deposition efficiency and a ¹⁰³Pd utilization rate of 51%, which is more than double (∼25%) seen with traditional silver bars (105). By streamlining the plating process and cutting material losses, the approach reduces both fabrication cost and complexity, paving the way for more economical, high-performance ¹⁰³Pd seed production and broader clinical adaptation (105).

Researchers have also explored ¹⁰³Pd in nanoparticle-based brachytherapy. Laprise-Pelletier et al. (107) evaluated the therapeutic efficacy, biodistribution, and tolerability of two formulations of ¹⁰³Pd-doped Pd@Au nanoparticles (NPs) in a prostate cancer xenograft model. Like Djoumessi et al. (108), the Pd NP synthesis achieved a high encapsulation efficiency of 87% for all ¹⁰³Pd atoms incorporated into the 10–14 nm cores (107). Comparing to Moeendarbari et al. (109), who reported 80% tumor inhibition after a 1.5 mCi implant given in 40 μl, the present study achieved similar therapeutic effects using a tenfold smaller volume (4 μl at 1.6–1.7 mCi) (107). Fach et al. (110) formulated ¹⁰³Pd within gold-palladium (AuPd) alloy nanoparticles, intrinsically radiolabeled with ¹⁰³Pd, capable of forming biodegradable gel-like implants upon injection. Therapeutic efficacy of ¹⁰³Pd-nanogels in a tumor-bearing mouse model indicated doses of 25 MBq [¹⁰³Pd]AuPd-nanogel produced a robust tumor-growth delay and double median survival compared to controls, with no systemic toxicity (110). Building on this, Sporer et al. (111) compared injectable ¹⁰³Pd-brachytherapy seeds that form biodegradable LOIB-based solids in situ, using either intrinsically radiolabeled PdAuNPs or a novel SSIB-[16]aneS₄ chelator. The [¹⁰³Pd]PdAuNPs were synthesized by co-reduction of [¹⁰³Pd]PdH₂Cl₄ and AuHCl₄ surface-functionalized with a lipophilic coating and dispersed in LOIB:EtOH to achieve an overall radiochemical yield of 83% or via conjugation of the [16]aneS₄ chelator shown in Figure 4A to a lipophilic sucrose septaisobutyrate (SSIB), followed by complexation with [¹⁰³Pd]PdH₂Cl₄ in 99% yield (111). While both formulations reached activities of 1–1.5 GBq/ml with negligible release (<1%) of radioactivity over 30 days, the chelator strategy deems to be favorable as it avoids non-degradable gold and offers a versatile platform for other radiometals (111).

Hindie et al. (112) used the Monte Carlo track-structure code CELLDOSE (113) (for electrons) in conjunction with PHITS (114) (for photons) to quantify energy deposition from ¹⁰³Pd/^103mRh at the cell surface, within the cytoplasm, and in the nucleus enabling normalized comparison against ¹⁶¹Tb and ¹⁷⁷Lu. In the single-cell model, ¹⁰³Pd delivered 7- to 10-fold higher nuclear absorbed dose and 9- to 25-fold higher membrane dose than ¹⁷⁷Lu—driven primarily by Auger and conversion electrons—with ¹⁶¹Tb showing intermediate dose profiles (112). Annamalaisamy et al. (115) reported the first radiosynthesis and evaluation of ¹⁰³Pd₂(bpy)₂ale (Figure 4B), designed as an in vivo ¹⁰³Pd/^103mRh generator for bone-targeted Auger-electron therapy—extending prior work by Cipriani et al. (116) and Fathy et al. (117). At pH of 7 and 60°C, the radiosynthesis achieved >85% radiochemical yield by iTLC, and preparative HPLC confirmed radioactive and non-radioactive complexes were identical (115). Notably, iTLC showed complete retention of parent ¹⁰³Pd and daughter ^103mRh—significantly improving upon macrocyclic ¹⁰³Pd/^103mRh generators reported by Jensen et al. (118), which exhibited ∼7% ^103mRh release—related to the electron-donating bipyridyl ligand quenching “Coulomb explosion” effect discussed in Nath et al. (115, 119). The result from the work of Jensen et al. (118) can be explained by works of van Rooyen et al. (120) and Szucs et al. (121) who conducted detailed recoil energy calculations associated with the emission of Auger electrons, photons, and neutrinos (115). Finally, ¹⁰³Pd₂(bpy)₂ale exhibited potent multimodal toxicity via Auger electrons and demonstration chemotoxicity comparable to cisplatin by works of Zhao et al. (122), highlighting its theragnostic potential (115).

Highlighted in Table 2, ¹⁰⁹Pd is produced using an enriched ¹⁰⁸Pd (98%) metal target, which was performed by Chakraborty et al. (123), obtaining a specific activity of ∼1.85 GBq/mg (50 mCi/mg) at a thermal neutron flux of 3 × 10¹³ n cm⁻² s⁻¹ for 3 days (78). In the review by Boros and Packard (78), a dissolution method is carried out in heated aqua regia and is subsequently evaporated and heated to dryness with 12 N HCl to form H₂PdCl₄. Silver-111 (¹¹¹Ag) is co-produced and can be removed by coprecipitation with small amount of AgNO₃ (78). The supernatant containing ¹⁰⁹Pd is later dissolved in dimethylsulfoxide (DMSO) to produce ¹⁰⁹Pd(DMSO)₂Cl₂ for subsequent syntheses (78). Hien et al. (126) reported thermal neutron capture cross-section (σ₀) and resonance integral (I₀) of the ¹⁰⁸Pd(n,γ)¹⁰⁹Pd, backing previous work of thermal neutron capture cross sections (127–134) and resonance integral data (131, 135) for this reaction.

Porphyrin derivatives are well known to preferentially accumulate in malignant tumors via photodynamic mechanisms (136–139), and early efforts to radiolabel these macrocycles with therapeutic radionuclides—such as ¹⁰⁹Pd-hematoporphyrin (140) and ¹⁰⁹Pd-antimelanoma antibodies (141)—demonstrated targeting potential but lacked tumor retention. To expand upon this potential, Das et al. (142) radiolabeled a porphyrin derivative (DHBEP) with n.c.a. ¹⁰⁹Pd to create a highly stable, rapidly tumor-localizing radiopharmaceutical. The novel ligand DHBEP was synthesized via a two-step sequence and complexed with ¹⁰⁹PdDMSO₂Cl₂ at 80°C for 1 h, achieving >98% radiochemical purity. The ¹⁰⁹Pd-DHBEP complex remained stable at >97% after 48 h (∼4 half-lives of ¹⁰⁹Pd) at room temperature in saline (142). Biodistributions studies with Swiss mice bearing fibrosarcoma tumors revealed high tumor uptake at 30 min p.i. [(5.28 ± 1.46%IA/g)] and activity was cleared via the renal pathway (142).

Pineau et al. (125) evaluated TE1PA, shown in Figure 5, to demonstrate its suitability for complexation with both natural and radioactive palladium towards radiopharmaceutical development. Under all conditions and comparing TE1PA to cyclam, TE1Bn (benzyl cyclam), TE1Py (pyridylmethyl cyclam), they reported significant improvement in inertness of [¹⁰⁹Pd][Pd(TE1PA)]⁺ over [¹⁰⁹Pd][Pd(cyclam)]²⁺ at room temperature over a 24-h period, highlighting the enhances properties of the picolinate derivative (125).

Gharibkandi et al. (143) developed ¹⁰⁹Pd-coated gold nanoparticles (Au@¹⁰⁹PdNPs) functionalized with polyethylene glycol (PEG) conjugated to trastuzumab for targeted therapy of HER2-positive cancers. The resulting Au@Pd-PEG-trastuzumab radiobioconjugate averaged 9.5 antibodies per nanoparticle and demonstrated high HER2-specific uptake in SKOV-3 cells, achieving >99% internalization within 1 h, consistent with findings reported by Gaweda et al. (143, 144). The authors compared the cytotoxicity of radiobioconjugates labeled with the Auger emitter ¹²⁵I (t_1/2 = 59.49 d; Au@Pd¹²⁵I-trastuzumab), β⁻ emitter ¹⁹⁸Au (t_1/2 = 2.69 d; ¹⁹⁸Au-trastuzumab), and the ¹⁰⁹Pd/^109mAg in vivo generator (Au@¹⁰⁹Pd-trastuzumab) (143). With consistent activity concentrations of 20 MBq/ml, the ¹⁰⁹Pd/^109mAg-based conjugate demonstrated significantly higher cytotoxicity than those conjugates radiolabeled with either ¹²⁵I or ¹⁹⁸Au, highlighting the therapeutic advantage of simultaneous emission of both radiation types from this generator design (143). A subsequent study in 2024 (145) improved ¹⁰⁹Pd production using ¹⁰⁸Pd, achieving >500 MBq/mg from the natural palladium target and >2 GBq/mg from the enriched palladium target (78). Their findings indicated that Pd NPs labeled with ¹⁰⁹Pd were significantly more cytotoxic at similar activities than those labeled with either ¹³¹I or ¹²⁵I (145). Analogous to ¹⁰³Pd, Annamalaisamy et al. (115) also reported the radiosynthesis and evaluation of ¹⁰⁹Pd/^109mAg in situ generator bound to a mixed bipyridyl-bisphosphonate scaffold, ¹⁰⁹Pd₂(bpy)₂ale, for bone-targeted radionuclide therapy. in vitro, the conjugate significantly reduced metabolic viability in prostate and ovarian cancer cells, with cytotoxicity depending on both activity concentration and exposure time (115).

Shown in Table 3, Salek et al. (149) irradiated isotopically enriched osmium (¹⁹⁰Os, 97.8%) in the 5 MW Tehran Research Reactor (ϕ = 4 × 10¹³ n cm⁻² s⁻¹) for 15 days with subsequent fusion in a mixture of KOH-KNO₃, reporting a specific activity of ∼325 mCi/mg. The dissolution method for osmium reported by Brihaye et al. (150) has been established in subsequent steps to form K₂OsCl₆; and carried out for all reported studies in this review. Additionally, osmium by-products ¹⁸⁵Os (t_1/2 = 15.4 d) via ¹⁸⁴Os(n,γ)¹⁸⁵Os reaction and ¹⁹³Os (t_1/2 = 30.2 h) via ¹⁹²Os(n,γ)¹⁹³Os reaction are produced only in trace amounts (149). These are neglected as ¹⁸⁵Os decays to stable ¹⁸⁵Re, and ¹⁹³Os decays quickly (149). However, an unavoidable longer-lived impurity is ¹⁹²Ir (t_1/2 = 73.8 d), which is produced when stable ¹⁹¹Ir—the stable decay product of ¹⁹¹Os—undergoes a ¹⁹¹Ir(n,γ)¹⁹²Ir reaction during irradiation (146, 149, 151). Brihaye et al. (150) demonstrated two separation methods—distillation and solvent extraction—between ¹⁹¹Os and ¹⁹²Ir. Using these methods, they achieved a separation efficiency of 100% by distillation and 99.9% efficiency by solvent extraction (150). Salek et al. (149) modified the extraction method and yielded a 98.8 ± 0.48% ¹⁹¹Os recovery, while completing the procedure in 30 min.

In a study performed by Jamre et al. (148), BLM (Figure 6A) was radiolabeled with ¹⁹¹Os by reacting it with K₂OsCl₆. The total labeling and formulation of ¹⁹¹Os-BLM took approximately 24 h, resulting a >95% radiochemical yield and >97% radiochemical purity, with <3% free ¹⁹¹Os- K₂OsCl₆ detected by radio-TLC (148). They reported the ¹⁹¹Os-BLM complex remained stable in aqueous solution for ∼72 h. Biodistribution studies (4 h, 24 h, 48 h, 72 h, and 14d p.i.) for ¹⁹¹Os-BLM demonstrated high uptake in the lungs and moderate accumulation in the liver and spleen, all remaining >1% ID/g throughout the study (148). In vivo imaging at 24, 48, and 72 h confirmed these retention patterns as well (148).

Labeling APMTS (Figure 6B) with ¹⁹¹Os, Moghaddam et al. (152) achieved >95% radiochemical yield in a 12 h synthesis with a specific activity of 21.5 GBq/mmol, while the complex remained >95% stable for at least 48 h (152). In the biodistribution studies (4, 24, 48, and 72 h p.i.) using the ¹⁹¹Os-APMTS complex, liver uptake and kidney uptake peaked by 48 h (5.2%–6.7% ID/g), while there was low blood, heart, bone retention by 24 h and negligible by 72 h (<0.5% ID/g) (152). A follow-up study by Moghaddam-Banaem et al. (151), demonstrated the preparation of ¹⁹¹Os-phyate complex shown in Figure 6C that could be used for radiosynovectomy applications. Using 10 mg of sodium phytate, the complex forms in ∼24 h with a labeling yield >98% detected by radio-chromatography, while remaining stable in an aqueous solution for at least 72 h (151). Biodistribution studies (0.5, 4, 24, 72 h p.i.) showed most of the injected dose remained in the joint with minimal uptake in the kidney, and other organs considered negligible (<0.5% ID/g) (151).

Due to its widespread use, ¹⁹²Ir is routinely produced in nuclear reactors via the ¹⁹¹Ir(n,γ)¹⁹²Ir reaction, using either Na₂IrCl₆ targets—described by Ananthakrishnan (157)—or iridium wire, as applied in clinical settings by Schaeken et al. (153, 158). All production routes are shown in Table 4. Irradiating Na₂IrCl₆ under standard conditions —10 mg; ϕ = 1.5 × 10¹³ n cm⁻² s⁻¹; 7 days—can yield 12 GBq of ¹⁹²Ir, with specific activity >185 GBq per gram Ir (157). After irradiation, the targets are dissolved in 10 ml of 0.1 N HCl, yielding radiochemical solutions with concentrations ranging from 74 to 370 MBq/ml and >99% radionuclidic purity (157). This method remains the benchmark for high-activity, high-purity ¹⁹²Ir production for clinical brachytherapy (157). As reactor-produced ¹⁹²Ir is carrier-added, accelerator routes have been explored to produce n.c.a. ¹⁹²Ir with potentially higher specific activity.

Via the ¹⁹²Os(p,n)¹⁹²Ir reaction, Hilgers et al. (153) measured a peak cross-section of 68 ± 8 mb at 9.1 ± 0.5 MeV, while identifying an optimal production window of 8–16 MeV (∼0.16 MBq/μA-h ¹⁹²Ir). The authors confirmed their experimental data with nuclear model codes [EMPIRE-II (96) and ALICE-IPPE (159)] and pointed out that though a cyclotron approach yields lower activity than those achieve via reactor-based production, the specific activity could be much higher (153). They estimated under realistic irradiation conditions (30 h, ϕ = 3.74 × 10¹⁵ p/s), projected batch yields could reach ∼5.6 GBq—serving as a complementary approach and broadening access to high specific activity ¹⁹²Ir brachytherapy sources (153). Langille et al. (155) demonstrated that a 12.8 MeV proton beam on naturally abundant, electroplated osmium targets yields ¹⁹²Ir with an average measured cross section of 46.4 ± 6.2 mb, which compared well with literature values of Hilgers et al. (153) and Szelecsenyi et al. (160). Targets underwent oxidative dissolution (H₂O₂/HCl) and anion-exchange chromatography on Dowex 1 × 8, with the process delivering an overall radiochemical efficiency of ∼80% and radionuclidic purity of 100% (155). Building on established microwave-assisted syntheses of non-radioactive complexes [(ppy)₂Ir(μ-Cl)₂Ir(ppy)₂] and Ir(ppy)₂(bpy)—reported earlier by Alam et al. (161), Bura et al. (162), and Wu et al. (163)—the authors performed the first radiosynthesis of an iridium cyclometallation reaction by adding n.c.a. [IrCl₆]³⁻ to the microwave reaction (155). They achieved up to 68% radiochemical purity of Ir(ppy)₂(bpy) with a maximum specific activity of 0.54 ± 0.14 Ci μmol⁻¹ (20 ± 5.2 GBq μmol⁻¹) (155).

Tarkanyi et al. (164) reported the first experimental cross sections for the ¹⁹²Os(d,2n)¹⁹²Ir reactions up to 21 MeV, employing a stacked-foil technique with 84.5% enriched ¹⁹²Os targets electrodeposited on 25 μm thick Ni foils, thereby observing a cross sectional peak of 370 ± 46 mb at 12.1 ± 0.8 MeV. Although reactor-based ¹⁹²Ir production yields remain higher, the deuteron route results in a n.c.a. product of ¹⁹²Ir with significantly higher specific activity (153, 164). Compared with the earlier ¹⁹²Os(p,n) process via Hilgers et al. (153), the (d,2n) channel delivers higher cross sections and thick-target yields in the same energy window; however, due to smaller and higher-current proton cyclotrons being more readily available, the choice for the ¹⁹²Os(p,n)-reaction is preferred (164). Dovbnya et al. (165) reported the first experimental demonstration of photonuclear ¹⁹³Ir(γ,n)¹⁹²Ir on natural iridium using a tantalum bremsstrahlung converter integrated within a neutron moderator, which enhanced ¹⁹²Ir yields by ∼50% via the ¹⁹¹Ir(n,γ)¹⁹²Ir reaction and delivered up to ∼900 MBq/h under 40 MeV, 4 μA beam conditions. Computational simulations with PENELOPE-2008 software (166)—supplemented by evaluated photonuclear cross sections—accurately reproduced experimental yields for ¹⁹²Ir as well as co-produces isotopes (¹⁹⁰Ir, ⁹⁰Mo, ⁹⁹Mo), validating the mixed γ- and n-flux model (165). Compared to traditional reactor-based ¹⁹¹Ir(n,γ)¹⁹²Ir production (74 MBq/h; >1,000 MBq/h-g) and cyclotron-based ¹⁹²Os(p,n)¹⁹²Ir production (>185 MBq/h; without carrier), this electron-accelerator approach offers competitive batch yields and modular flexibility (153, 157, 165). While the specific activity is low, the authors suggest that optimizing activation-cooling regimes and employing enriched ¹⁹³Ir targets could enable scalable, reactor-free ¹⁹²Ir production suitable for medical and industrial applications (165).

¹⁹²Ir has been significantly utilized in high dose rate (HDR) brachytherapy, offering a steep dose gradient that concentrates therapeutic radiation within tumors while minimizing damage to the surrounding normal tissue (167). Jayakody et al. (167) reviewed a suite of independent verification methods—including radiochromic films, ionization-chamber arrays, plastic scintillation detectors, and TLD/OSLD systems—that have been benchmarked against TPS-calculated dose maps for ¹⁹²Ir. Roussakis and Anagnostopoulos (168) wrote a mini-review on the aspects of the Iridium-Knife, detailing the key physical properties of the ¹⁹²Ir HDR source and illustrating how these underlie its characteristic steep dose gradients.

Nohara et al. (169) reported that 166 localized prostate cancer patients treated with a 44 Gy EBRT and 3 × 6 Gy ¹⁹²Ir HDR boost achieved a 5-year biochemical recurrence-free survival of 93.0%. Shigehara et al. (170) observed a 4-year overall survival of 87.2% and PSA progression-free survival of 82.6% in 84 prostate patients receiving 18 Gy ¹⁹²Ir HDR and 44 Gy EBRT. Chin et al. (171) treated 65 prostate cancer patients with EBRT plus two 8.5 Gy ¹⁹²Ir HDR fractions, reporting a 3-year biochemical disease-free rate of 90.8%. Potter et al. (172) used CT-planned ¹⁹²Ir HDR and 48.6–50 Gy EBRT in 189 cervical cancer patients, achieving 3-year pelvic control of 77.6% and disease-specific survival of 68.6%. Ott et al. (173) demonstrated that interstitial ¹⁹²Ir accelerated partial breast irradiation (APBI) in 69 early-stage breast cancer patients which yielded 100% 2-year local control, minimal acute and late toxicity, in 90% of cases.

Abtahi et al. (174) conducted a systematic review (1984–2020) between ¹⁹²Ir and ⁶⁰Co in GYN cancers. They reported that the 5-year overall survival (OS), local control, disease-free survival (DFS) and high-grade GI/GU toxicity were statistically equivalent between the two (174). Wen et al. (175) compared miniaturized HDR sources for cervical brachytherapy and found nearly identical dose distributions within 25 mm of the source, with equivalent clinical outcomes and toxicity rates. Strohmaier and Zwierzchowski (176) reviewed the physical and logistical aspects of ⁶⁰Co vs. ¹⁹²Ir, concluding that the two radionuclides matched in radial dose function, while delivering equivalent clinical efficacy. Tantivantana and Rongsriyam (177) performed a retrospective analysis of 480 stage IB2-IIIB cervical cancer patients treated between 2004 and 2014, comparing outcomes following HDR brachytherapy with ¹⁹²Ir (274 patients; 57.1%) or ⁶⁰Co sources (206 patients; 42.9%). The study found no statistically significant differences in OS, recurrence rate, or DFS between the ¹⁹²Ir and ⁶⁰Co cohorts (177).

The production for ^103mRh is shown in Table 5. Epperson et al. (179) introduced a rapid, high-yield generator for ^103mRh by solvent-solvent extraction of RuO₄ into CCl₄ achieving 94 ± 0.6% ^103mRh yield with 3.8 ± 0.7% ¹⁰³Ru contamination in a single, 15-min extraction. This method contrasts with earlier ion-exchange and distillation approaches referenced by the authors, offering a practical foundation for routine on-demand ^103mRh availability (179). Bartos et al. (178) similarly used reactor-produced ¹⁰³Ru (from natural ruthenium irradiation of 36 h, yielding 466 MBq) and separated ^103mRh from RuO₄ extraction. This work laid the foundation for supplying short-lived ^103mRh in sufficient quantities for further studies (178). Thery et al. (180) reported the recent progress in ruthenium chemistry for the ¹⁰³Ru/^103mRh generator for Auger therapy, describing the main limiting factor being an effective separation between the two radionuclides due to the unpredictable, misunderstood chemistry. Their work overcame prior barriers in earlier solvent-extraction and speciation studies, establishing optimal conditions for examining the experimental feasibility of the generator in the future (180).

More recently, Jensen et al. (118) demonstrated a solid-phase ¹⁰³Pd/^103mRh generator using neutron-activated ¹⁰²Pd targets. They chelated carrier-added ¹⁰³Pd with a lipophilic macrocycle, 16aneS₄, and loaded it on a C18 cartridge (118). The optimal elution performance for ^103mRh was achieved with 1.0 M HCl, yielding a radiochemical purity of 99%, an apparent molar activity of 26.6 MBq/nmol, and an elution yield of 5.81% (118). Despite the potential, the low elution yield indicates that further optimization is necessary to utilize the generator for extended use, particularly in clinical applications (118). Ohya et al. (181) improved on this by testing various anion-exchange resins—inspired by Berk (182) and Mamadaliev et al. (183)—following a separation method described in Ohya et al. (101). Four commercially available gel-type anion-exchange resins with comparable functions groups and matrixes were investigated: IRA410 and SA20A (dimethylethanol ammonium), and IRA904 and SA11AL (trimethyl ammonium) (181). Of these, SA11AL delivered the best performance, with a raw yield of 39% and lowest ¹⁰³Pd breakthrough of 0.29% over 32 milking cycles spanning eight weeks (181). More recently, Zagryadsky et al. (184) performed measurements of the ¹⁰²Pd(n,γ)¹⁰³Pd and ¹⁰²Ru(n,γ)¹⁰³Ru reactions in the IR-8 Reactor for the purpose of ¹⁰³Ru/^103mRh and ¹⁰³Pd/^103mRh generators. They indicated the experimental channel of the IR-8 reactor will be capable of achieving sufficiently ¹⁰³Ru and ¹⁰³Pd for the utilization of ^103mRh in radiopharmaceuticals (184).

Bernhardt et al. (185) performed Monte Carlo simulations to model the metastatic growth of tumor sizes for radionuclide therapy, comparing between high-energy electron emitter ⁹⁰Y (t_1/2 = 64.05 h), medium-energy electron emitter ¹⁷⁷Lu (t_1/2 = 6.65 d), and the low-energy electron emitter ^103mRh. They observed for low tumor-to-normal (TNC) tissue activity concentrations, ^103mRh performed slightly better compared to ¹⁷⁷Lu; however, for high TNC values, ^103mRh was the best choice for tumor treatment (185). However, as the authors noted, the short half-life (t_1/2 = 56.1 min) may be a limitation in the adaptation as an optimal radiotherapeutic (185).

Described in Table 5, Jia et al. (191) developed a scalable route to n.c.a. ¹⁰⁵Rh by irradiating enriched ¹⁰⁴Ru in the MURR reactor (ϕ = 8 × 10¹³ n cm⁻² s⁻¹, 72 h), achieving average yields of ∼5 mCi per mg Ru and >85% total recovery of ¹⁰⁵Rh. Their MgO adsorption method eliminated the need for chlorine gas and the formation of RuO₄—required in an earlier approach (189)—while delivering a ruthenium decontamination factor of 16,600, supporting the reliable availability of ¹⁰⁵Rh in large quantities (191). Subsequently, Unni et al. (188) developed a methodology for the production and purification of carrier-free ¹⁰⁵Rh by irradiating natural Ru (99.9%) at a thermal flux of 3 × 10¹³ n cm⁻² s⁻¹ for 5–7 days, followed by a 24 h decay of ¹⁰⁵Ru to ¹⁰⁵Rh, achieving within 5% of ≈24 mCi predicted by Bateman's equation. The authors oxidized the Ru matrix (⁹⁷Ru, ¹⁰³Ru, and trace ¹⁹²Ir) to volatile RuO₄ (KIO₄/KOH at 70°C, 20 min), performed successive solvent extractions with CCl₄ (retaining 97.8 ± 0.78% of ¹⁰⁵Rh in aqueous phase), and then applied 100% TBP extraction to obtain 95.35 ± 0.78% of ¹⁰⁵Rh (aqueous phase) and 96.6 ± 0.8% of ¹⁹²Ir (organic phase) (188). A co-precipitation of ¹⁰⁵Rh with Fe(III) as hydroxide using KOH recovered 89.4 ± 2.2% of ¹⁰⁵Rh, and a three-stage Fe removal—using cationic exchange chromatography—delivered a final overall recovery of ∼80% (15–20 mCi) of carrier-free ¹⁰⁵Rh (188). Okoye et al. (192) demonstrated a comprehensive strategy to reclaim, purify, and reuse enriched ¹⁰⁴Ru targets—originally captured as RuO₄ in 3 M HCl from decades of ¹⁰⁵Rh production—for economical, high-yield ¹⁰⁵Rh manufacture. The recycled metal retained 98.84% ¹⁰⁴Ru enrichment—a slight decrease from their original—and enabled up to 97.3% ¹⁰⁵Rh recovery (19.10 mCi) (192). The isolated ¹⁰⁵Rh was subsequently used in radiolabeling experiments with two previously developed chelators (193, 194), yielding radiochemical efficiencies of 91.0% ± 1.5 for 222-S₄-diAcOH (Figure 7A) and 80.9% ± 0.4 for 16S₄-diol (Figure 7B) (192).

Inagaki et al. (195) investigated the production of ¹⁰⁵Rh via two distinct routes: neutron irradiation of ^natRuO₄ powder (ϕ = 4.5 × 10¹² n cm⁻² s⁻¹, 10 min) through the ¹⁰⁴Ru(n,γ)¹⁰⁵Ru reaction; and bremsstrahlung photon irradiation of natural Pd foils (5 × 5 mm²) at 20–40 MeV using an electron linear accelerator (linac), inducing the ¹⁰⁶Pd(γ,p)¹⁰⁵Ru reaction. To enable comparison, the authors normalized yield data to equivalent target masses, beam currents, and irradiation times, reporting 77 ± 2 kBq of ¹⁰⁵Rh via the reactor method (10 mg) and 88 ± 5 kBq at 40 MeV from the linac method (50 mg, 100 μA, 10 min) (195). Furthermore, extrapolation to clinical-scale conditions using the linac method—10 g Pd target, 1 mA current, and 24 h irradiation—predicted a ¹⁰⁵Rh yield of approximately 20.1 GBq, far exceeding the 0.148 GBq typically required for diagnostic or therapeutic applications, as described in Sciuto et al. (195, 196). Kazakov et al. (197) investigated a method for producing carrier-free ¹⁰⁵Rh using a 55 MeV electron accelerator, analyzing the isotopic composition of irradiated PdCl₂ and optimizing separation methods. Irradiation of 270 mg PdCl₂ in 5 ml solution at 100 nA for 1 h yielded 73.7 kBq/μAh of total rhodium activity, with ¹⁰⁵Rh containing 82% (60 kBq/μAh, 2.1 kBq (197). When compared to Inagaki et al. (195), who reported 88 kBq from 50 mg ^natPd foil for at 40 MeV (10 min, 100 nA), both demonstrated feasible accelerator-based alternatives to reactor or cyclotron production for medical applications (197). Nonetheless, the irradiated PdCl₂ was dissolved in 2 M HCl and passed through extraction chromatography columns using either DGA-Normal or TEVA resins (197). Column and distribution coefficient studies showed DGA-Normal offered superior performance, eluting ≥98% of ¹⁰⁵Rh in 2 M HCl and enabling complete Pd stripping with 11 M HCl (Pd/Rh separation factor >10⁵), while TEVA failed to achieve sufficient Pd/Rh separation (197).

Khandaker et al. (198) reported the first experimental measurement of ^natPd(p,x)¹⁰⁵Rh excitation function from 4 to 40 MeV using stacked-foil activation, observing significant discrepancies between measured cross-sections and nuclear model predictions from TALYS (49) and ALICE-IPPE (159). From the experimental data, thick target yield calculations suggest that low-energy cyclotrons (E < 20 MeV) can effectively produce ¹⁰⁵Rh, primarily via the ¹⁰⁸Pd(p,α)¹⁰⁵Rh reaction (198).

Jurisson et al. (187) investigated ¹⁰⁵Rh radiopharmaceutical development by exploring a suite of cis- and trans-[RhCl₂l]⁺ complexes using tetradentate thioether ligands. Brooks et al. (190) reported the synthesis and purification of novel ¹⁰⁵Rh-bleomycin (¹⁰⁵Rh-BLM) complex, demonstrating >80% complexation yield, high in vitro stability, and rapid biphasic in vivo clearance with minimal non-specific retention. Although ¹⁰⁵Rh-BLM achieved tumor uptake approximately four-fold greater than contralateral muscle, its potential for targeted radiotherapy is limited by significant levels and prolonged retention in the kidneys relative to tumor (190). The study by Ando et al. (199) evaluated ¹⁰⁵Rh as a candidate for radiotherapeutic applications targeting bone metastases by leveraging its favorable decay properties and investigating its biological behavior when chelated to EDTMP shown in Figure 7C. Radiolabeling with EDTMP achieved >99% labeling efficiency, with no dissociation observed for up to 5 days at room temperature (199). Compared to a study using ^99mTc-MDP by Sanada et al. (200), ¹⁰⁵Rh-EDTMP demonstrated comparable bone uptake, but exhibited faster clearance from circulation and significantly higher bone-to-tissue ratios (199). Mentioned in Okoye et al. (192), a variety of chelates have been evaluated (186, 193, 194, 201–213), along with preclinical biological distribution studies have been highlighted in Li et al. (209) and Goswami et al. (193) for ¹⁰⁵Rh clinical utility towards advancing therapeutic radiopharmaceuticals.

The production routes for ⁹⁷Ru are listed in Table 6. Zaitseva et al. (229) measured excitation functions for ⁹⁷Ru production via the ⁹⁹Tc(p,3n)⁹⁷Ru reaction, using a stacked-foil technique (50–100 nA) from 20 to 99 MeV. They measured a 438 ± 66 mb peak at 32 MeV—corresponding to a thin-target yield of ∼934 μCi/μAh—and a cumulative yield of ∼10.49 mCi/μAh when degrading protons from 99 MeV to the threshold (E_th = 18.3 MeV) (229). An optimal 19–50 MeV window maximized ⁹⁷Ru production (∼7 mCi/μAh) while higher-energy beams (>50 MeV) could push yields beyond 10.5 mCi/μAh for Ci-scale production at higher currents (229). Building on this, Zaitseva et al. (230) optimized a radiochemical separation for metallic Tc targets irradiated at 50 MeV (∼8 μA, 1 h), isolating 40–50 mCi of ⁹⁷Ru. A four-step process—dissolution, acid conversion, oxidation-distillation, and absorption—reduced Ru(VIII) to Ru(III) and recovered 95%–98% of Ru with >10⁴ purity after 6–7 h (230). An estimated delivery of ≥150 mCi of ⁹⁷Ru is needed (50 MeV, 6–8 μA, 8 h) for clinical purposes to be feasible ∼70 h after EOB (230).

Ditroi et al. (231) and Tarkanyi et al. (232) explored α-induced routes on natural molybdenum, measuring ⁹⁷Ru excitation functions up to 40 MeV. Both found peaks near 39 MeV (182.4 ± 20.5 mb and 232 ± 26 mb, respectively), along with good agreement from previous results by Levkovskij (233) and Graf and Munzel (234) across all energy ranges, and with Rapp et al. (235) at low energies (231, 232). Model comparisons (TENDL-2011/TENDL-2015 (49), ALICE-IPPE (159), and EMPIRE-3.1 (96)) generally agreed in trend, with (232) calculating thick-target yields reaching 2 GBq/C (0.19 mCi/μAh), and potential to increase the yield by a factor of three through isotopic enrichment favoring the ⁹⁴Mo(α,n), ⁹⁵Mo(α,2n), and ⁹⁶Mo(α,3n) reactions. Thick target yields were described by Abe et al. (236), thereby obtaining a yield of 126 μCi/μAh via the ⁹⁴Mo(α,n) reaction using 30 MeV α-particles (232). Sitarz et al. (237) extended α-induced production of ⁹⁷Ru to 67 MeV, confirming a 237 ± 20 mb at 41.8 MeV, agreeing with (232) and (231) below 40 MeV. Most recently, Happl et al. (238) demonstrated ⁹⁷Ru production for α-induced irradiation of ^natMo for 10 h to yield >300 MBq end of irradiation (EOI). Post-irradiation, the target foil was dissolved and bulk Mo was removed using two sequential ion exchange columns; obtaining trace impurities of Mo (0.9–2.0 μg) and minor radionuclidic contaminants including ⁹⁵Ru (t_1/2 = 1.6 h), ^95mTc (t_1/2 = 61.0 h), and ⁹⁵Tc (t_1/2 = 20.0 h) in the ⁹⁷Ru eluate (238). The reported radiochemical yield of ⁹⁷Ru was 40%–56%, resulting in deliverable activities of 87–123 MBq (74–106 MBq/ml) (238).

Furthermore, Maiti and Lahiri (239) introduced a novel ¹²C + ⁸⁹Y production route for n.c.a ⁹⁷Ru, while avoiding co-production of longer-lived radionuclides to achieve tracer-level yields after cooling. Furthermore, the authors developed a two-separation scheme—a solid-liquid extraction in 1 M HCl and sequential 0.1 M/6 M HCl column chromatography—yielding 88% n.c.a. ⁹⁷Ru and resulting distinct Ru(IV)/Ru(III) speciation under certain conditions (239).

Oster et al. (240) evaluated ⁹⁷Ru-DTPA as a potential imaging agent for cerebrospinal fluid by injecting 0.4 mCi of the compound into the cisterna magna of dogs, while comparing the performance with ¹¹¹In-DTPA. From their study, they established ⁹⁷Ru-DTPA to be superior to ¹¹¹In-DTPA as it delivered approximately half the absorbed dose to the tissues, along with better imaging capabilities (240). Som et al. (241) labeled transferrin with ⁹⁷Ru (⁹⁷Ru-TF) and compared its biodistribution to ⁶⁷Ga-citrate, ¹²³I-transferrin, ^99mTc-plasmin, ¹²⁵I-fibrinogen, and ¹³¹I-albumin in tumor and abscess bearing animals. Notably, the difference between ⁹⁷Ru-TF and ⁶⁷Ga-citrate were of particular focus, as tumor concentrations of ⁹⁷Ru-TF increased substantially with time, whereas the ⁶⁷Ga concentration did not (241). The authors noted although there were no significant advantages using ⁹⁷Ru over ⁶⁷Ga, the nuclear characteristics of ⁹⁷Ru may improve imaging quality (241).

More recently, as a potential radiopharmaceutical, Borisova et al. (242) reported the first ⁹⁷Ru complex with pyridine-2,6-dicarboxamide conjugate shown in Figure 8A (243). Happl et al. (238) further explored the same method from Happl et al. (243) for a three-step synthesis for radiolabeling BOLD-100 (Figure 8B) with c.a. [⁹⁷Ru]RuCl₃ (0.2–0.5 MBq/μmol). The radiochemical purity of all three intermediates was >99% (238). The final product exhibited an overall radiochemical yield of 8% and an overall chemical yield of 13%, based on the mass of isolated intermediates and products (238). Additionally, the specific activity at the end of synthesis was 0.1 MBq/mg, with a molar activity of 0.05 MBq/μmol (238). Although radiolabeling BOLD-100 with ⁹⁷Ru was successful, the radiochemical yield and specific activity must be improved to enable SPECT imaging using c.a. [⁹⁷Ru]BOLD-100 (238).

The production routes for obtaining ¹⁰³Ru are highlighted in Table 6. The measured excitation function in Ditroi et al. (231) demonstrated a maximum cross-section of 10.6 ± 1.2 mb at 13.8 ± 0.6 MeV, then gradually declined and plateaued between 18 and 40 MeV, with cross sections ranging from 0.5 to 5 mb (231). The experimental results aligned closely with earlier measurements by Graf and Munzel (234) and Esterlund and Pate (244), though discrepancies in peak values were observed across the studies (231). TENDL-2011 (49) underestimated the experimental cross-sections and exhibited a shift towards the lower energies for the maximum, while EMPIRE-3.1 (96) better replicated the shape of the curve but slightly overestimated the maximum value (231). Integral yield data indicated that α-induced production of ¹⁰³Ru is inefficient—due to its low cross-sections—compared to ⁹⁷Ru (mentioned in 8.1.1), with practical yields falling well below the MBq/μAh range (231). Tarkanyi et al.'s (232) experiment demonstrated a rise in cross section from threshold to a peak of 15.6 ± 1.7 mb at 13.79 ± 0.6 MeV, following a gradual decline and plateau between 18 and 40 MeV, with values ranging from approximately 6.2 to 1.2 mb (232). The authors reported good agreement with the corrected data of Ditroi et al. (231) and earlier measurements by Graf and Munzel (234) and Esterlund and Pate (244), except for a discrepancy by a factor of two near the absolute maximum (232). The TENDL-2011 and TENDL-2015 (49) libraries were found to underpredict the experimental cross sections and shifted the peak position toward lower energies, whereas EMPIRE-3.1 (Rivoli) (96) best reproduced both the shape and magnitude of the experimental excitation curve (232).

Mastren et al. (245) developed a two-step chromatographic purification scheme for obtaining ¹⁰³Ru from proton irradiation on a thorium target. Elution with 30 ml of 10 M HNO₃ (fractions 8–15) recovered 85 ± 5% of ¹⁰³Ru with a radiochemical purity of 82%, where they reported main impurities of ^117mSn and ^125,126Sb with trace amounts of ^230,233Pa, ⁹⁵Nb, and ⁹⁵Zr in this fraction (245). To remove those impurities, the DGA resin was incorporated, yielding a final ¹⁰³Ru recovery of 83 ± 5% with a radiochemical purity of >99.9% (245).

Blicharska et al. (246) developed a streamlined separation process to obtain ¹⁰³Ru as a surrogate for fission produced ¹⁰⁶Ru for the utilization in brachytherapy sources. The authors explored the Ru extraction efficiency in various oxidizing solutions, reporting H₅IO₆ to demonstrate the highest conversion of Ru(III/IV) to RuO₄ with 86.1% extraction (246). The method proved to be sufficiently scalable to produce hundreds GBq of ¹⁰⁶Ru per liter of PUREX raffinate (246). More recently, Happl et al. (243) obtained [¹⁰³Ru]RuCl₃• xH₂O by neutron activation with the Production Neutron Activation (PNA) installation at the spallation neutron source SINQ at Paul Scherrer Institute. The irradiation occurred over a three-week period at a neutron flux of 4 × 10¹³ n cm⁻² s⁻¹ with five ampoules containing 40–50 mg ^natRuCl₃• xH₂O that were then dissolved in concentrated hydrochloric acid, thereby resulting in activities up to 185 MBq (3.7–4.7 MBq/mg) (243). Happl et al. (238) improved their methods by obtaining ¹⁰³Ru via thermal neutron irradiation (5–10 × 10¹⁴ n cm⁻² s⁻¹) for 6–8 d using ^natRu metal foils enclosed in quart ampoules, yielding 1,049 MBq at end of irradiation (EOI). From this, c.a.¹⁰³Ru was recovered with radiochemical yields of 81%–82% (up to 648 MBq), molar activities up to 19.4 MBq μmol⁻¹ (249 MBq ml⁻¹), and a radionuclide purity of >99.9% (238).

Tanabe (247) reported ¹⁰³Ru scintigraphy in 37 patients with various types of malignant tumors. In the cohort of four lung cancer patients, ¹⁰³Ru failed to reliably differentiate carcinoma from inflammatory lesions under the study conditions. Wenzel et al. (248) synthesized a metallocene-based analog of iodo-hippuran—ruthenocenoyl-glycine (ruppuran) shown in Figure 9A—and labeled it with ¹⁰³Ru to directly compare its renal clearance kinetics with ¹²⁵I-labeled hippuran. The authors reported similar renal and plasma clearance pattern between the two compounds (248). Moreover, they did report absorbed doses to kidney and bladder with using ⁹⁷Ru-ruppuran as well, achieving slightly lower than that of ¹²³I-hippuran, with the results of clearance studies and dose estimates encouraging further kidney scintigraphy and secretory renal function measurements regarding the ⁹⁷Ru-labeled compound (248). Weiss et al. (249) demonstrated radiolabeled [¹⁰³Ru]RAPTA-C (Figure 9B) to be a promising compound for translation to clinical evaluation as it rapidly cleared from the organs and the excreted by the kidneys.

Happl et al. (243) modified a three-step synthesis—published and patented for non-radioactive BOLD-100 in 2018 (250)—of [¹⁰³Ru]BOLD-100 using 1.8–4.2 MBq/mg c.a. [¹⁰³Ru]RuCl₃, obtaining a >93% radiochemical purity of all three compounds and >38% overall radiochemical yield in the final product. Cytotoxicity of BOLD-100 and [¹⁰³Ru]BOLD-100 were compared in human colon carcinoma (HCT116) and murine colon carcinoma (CT26) cell lines using the colorimetric MTT assay with an exposure time of 96 h (243). The authors reported no effects to the biological activity in vitro even at low specific activities of 0.5–1.4 MBq/mg for [¹⁰³Ru]BOLD-100 (243). Furthermore, biodistributions studies with both BOLD-100 and [¹⁰³Ru]BOLD-100 were conducted in Balb/c mice bearing CT26 allografts over a period of 72 h (243). The authors reported from their tissue distribution studies that sub-equimolar amounts of c.a. [¹⁰³Ru]BOLD-100 achieved a higher and prolonged tumor uptake over 72 h, establishing a potential theragnostic approach with ¹⁰³Ru and ⁹⁷Ru once diagnostic SPECT imaging studies with c.a. [⁹⁷Ru]BOLD-100 are performed (238).