IntroductionZirconium-89 (89Zr) is a positron emitter with several advantages over other shorter-lived positron emission tomography (PET) compatible radiometals such as gallium-68 or copper-64. These include practically unlimited availability, extremely low cost, greatly facilitated distribution logistics, positron energy fit for medical PET imaging, and sufficiently long physical half-life to enable PET imaging at later time points for patient-specific dosimetry estimations. Despite these apparent benefits, the reception of 89Zr in the nuclear medicine community has been tepid. The driving factor for the absence of broader adaptation is mostly routed in its final formulation — [89Zr]zirconium oxalate. While serving as a suitable precursor solution for the gold standard chelator deferoxamine (DFO), [89Zr]Zr-oxalate is inaccessible for the most commonly used chelators, such as the macrocyclic DOTA, due to its pre-chelated state.Consequently, pioneering work has been conducted by multiple research groups to create oxalate-free forms of [89Zr]Zr4+, either via chemical conversion of oxalate into other counterion forms or via direct radiochemical isolation of [89Zr]ZrCl4, showing that [89Zr]Zr-DOTA complexes are possible and stable. However, this success was accompanied by challenges, including complex and labor-intensive radiochemical processing and radiolabeling procedures as well as the relatively minuscule conversion rates. Here, we report on the direct production of [89Zr]ZrCl4 avoiding oxalate and metal contaminants to enable efficient radiolabeling of DOTA constructs. MethodsWe based our direct production of [89Zr]ZrCl4 on previously reported methods and further optimized its quality by including an additional iron-removing step using the TK400 Resin. Here, we avoided using oxalic acid and effectively minimized the content of trace metal contaminants. Our two-step purification procedure was automated, and we confirmed excellent radionuclide purity, minimal trace metals content, great reactivity over time, and high specific molar activity. In addition, DOTA-based PSMA-617 and DOTAGA-based PSMA-I&T were radiolabeled to demonstrate the feasibility of direct radiolabeling and to estimate the maximum apparent specific activities. Lastly, the biodistribution of [89Zr]Zr-PSMA-617 was assessed in mice bearing PC3-PIP xenografts, and the results were compared to the previously published data. ResultsA total of 18 batches, ranging from 6.9 to 20 GBq (186 to 541 mCi), were produced. The specific molar activity for [89Zr]ZrCl4 exceeded 0.96 GBq (26 mCi) per nanomole of zirconium. The radionuclidic purity was >99 %, and the trace metals content was in the <1 ppm range. The [89Zr]ZrCl4 remained in its reactive chemical form for at least five days when stored in cyclic olefin polymer (COP) vials. Batches of 11.1 GBq (300 mCi) of [89Zr]Zr-PSMA-617 and 14.4 GBq (390 mCi) of [89Zr]Zr-PSMA-I&T, corresponding to specific activities of 11.1 MBq/μg (0.3 mCi/μg), and 14.4 MBq/μg (0.39 mCi/μg), respectively, were produced. [89Zr]Zr-PSMA-617 animal PET imaging results were in agreement with the previously published data. ConclusionIn this work, we report on a suitable application of TK400 Resin to remove iron during [89Zr]ZrCl4 radiochemical isolation. The breakthrough allows for direct radiolabeling of DOTA-based constructs with [89Zr]ZrCl4, leading to high apparent molar activities and excellent conversion rates.