Automated and manual synthesis of [68Ga]Ga-DOTA-4AH29 intended for PET-imaging of FAP-expression Authors: Marion Berdal, Surasa Nagachinta, Ximena Langsberg, Vaiva Gaspariunaite, Sam Massa, Ana Rita Pombo Antunes, Dimitrios Mantzilas, Matthias D’Huyvetter, Laurent Navarro Published: 2026 ABSTRACT AbstractBackgroundFibroblast activation protein is a well-established target for tumour imaging, and single-domain antibodies (sdAb) offer favourable pharmacokinetic properties for PET applications. However, the translation of sdAb-based radiotracers into robust and clinically applicable manufacturing processes remains limited. Previously, preclinical protocol for the production of [68Ga]Ga-DOTA-4AH29 has been reported. In this study, this process was adapted for clinical use. Both a manual, kit-based approach and automated synthesis on two different platforms were developed. Process robustness was assessed through a stress study evaluating key parameters, and compatibility with two 68Ga generator systems was investigated.ResultsEfficient radiolabelling was achieved at 50 °C for 10 min, with radiochemical conversion exceeding 90% and no need for purification. Both automated and manual processes showed consistent performance, with radiochemical purities of 99.5 ± 0.5% (n = 9) and 96.3 ± 1.5% (n = 8) as determined by iTLC and SE-HPLC respectively, with associated non-decay-corrected radiochemical yields of 78.3 ± 3.3% (n = 9) and 76.9 ± 2.1% (n = 8). The main source of activity loss was the radioactive decay of 68Ga.The process remained stable across a range of activities and buffer conditions. Comparable results were obtained with both 68Ga generator systems, with no observable impact on product quality. The final product was stable for at least 3 h at RT, with preserved binding capacity.ConclusionA robust and reproducible manufacturing process for [68Ga]Ga-DOTA-4AH29 was established, supporting both automated and manual clinical production. This approach will facilitate broader access to sdAb-based PET imaging across clinical centres.Graphical abstract Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s41181-026-00465-9.Keywords: 68Ga, sdAb, FAP, Automation, Radiosynthesis Introduction Fibroblast activation protein (FAP) is widely expressed by cancer-associated fibroblasts (CAFs) across a broad range of malignancies, including breast, ovarian, and colorectal cancers (Fitzgerald and Weiner 2020). FAP is generally low or absent in most healthy adult tissues but can be upregulated in pathological tissue-remodelling processes, including cancer-associated stroma and selected inflammatory or fibrotic conditions. In certain tumour types, such as sarcoma and melanoma, FAP expression has also been reported directly on tumour cells (Puré and Blomberg 2018). Beyond its expression pattern, FAP contributes to tumour progression, metastatic dissemination, and modulation of the tumour immune microenvironment (Shi et al. 2021). These characteristics have led to increasing interest in FAP as a target for molecular imaging and therapeutic applications. In particular, FAP-targeted positron emission tomography (PET) enables non-invasive assessment of the tumour microenvironment and has demonstrated clinical utility in both oncological and non-oncological indications (Lugat et al. 2022; Windisch et al. 2020, 2021). Several classes of targeting vectors have been investigated for FAP imaging, including small molecule inhibitors, peptides, monoclonal antibodies, antibody fragments, and single-domain antibodies (sdAbs). Among these, sdAbs have emerged as promising tools for imaging applications. Their small size (~ 15 kDa) allows rapid blood clearance and efficient tumour penetration, resulting in high target-to-background ratios while maintaining strong affinity and specificity. In addition, sdAbs are generally stable, soluble, and amenable to engineering (De Pauw et al. 2023; Liu et al. 2021a). These features make them particularly suitable for labelling with short-lived radionuclides such as gallium-68 and fluorine-18 (Liu et al. 2020, 2021b). Gallium-68 is widely used in PET imaging, primarily because it is available from 68Ge/68Ga generator systems, eliminating the need for a cyclotron. Several 68Ga-labelled radiopharmaceuticals, including [68Ga]Ga-DOTA-TOC and [68Ga]Ga-PSMA-11, are now routinely used in clinical practice, highlighting the growing importance of this radionuclide in nuclear imaging (Tirosh and Kebebew 2018; Hennrich and Benešová 2020; Hennrich and Eder 2021). The FAP-targeting sdAb 4AH29 has previously demonstrated promising preclinical characteristics, including rapid tumour uptake, fast blood clearance via the kidneys, and low uptake in normal tissues (Dekempeneer et al. 2023). While these results support its potential as a PET imaging agent, establishing a robust, clinically applicable manufacturing process remains an important step toward clinical translation. In this study, we describe the development of a clinical production process for [68Ga]Ga-DOTA-4AH29. The approach consisted of translating preclinical labelling conditions to automated synthesis platforms to ensure robustness and standardization, followed by the development of a manual, kit-based method to increase accessibility for centres without automated manufacturing option. A stress study was conducted to identify critical process parameters, and process performance runs were performed to confirm reproducibility under clinically relevant conditions. In addition, compatibility with different 68Ge/68Ga generator systems was evaluated. Together, these efforts aim to support the reliable, broadly implementable production of [68Ga]Ga-DOTA-4AH29. Results Automated clinical manufacturing process for [68Ga]Ga-DOTA-4AH29 - ML Eazy & PharmTracer modules Verification of reaction conditions A clinical manufacturing process for [68Ga]Ga-DOTA-4AH29 was developed based on previously established preclinical conditions (Dekempeneer et al. 2023), consisting of a 10 min reaction at 50 °C in NH4OAc solution. Before transferring the process to automated modules, a series of manual experiments was conducted to confirm its suitability for clinical production. The precursor amount was increased to 3 mg to enable the production of (limited) multi-patient batches. Also, increasing the administered protein dose to 1 mg has been shown to improve biodistribution, notably by decreasing liver uptake (Keyaerts et al. 2016), while maintaining appropriate molar activity and radiochemical stability. Labelling was performed using a GalliaPharma 68Ge/68Ga generator system, with elution in 5 mL of 0.1 N HCl. First tests were performed including only the labelling step to assess the efficiency of these conditions. Since no GMP-compliant ammonium acetate (NH4OAc, 2 M, pH 6) preparation kit was available, a commercially available sodium acetate (NaOAc, 2 M, pH 6) solution was tested as an alternative. This substitution had no measurable effect on labelling efficiency (radiochemical conversion (RCC) ≥ 97%; Table 1) and was thus used for further experiments. Table 1Results of the manufacturing of [68Ga]Ga-DOTA-4AH29 after the introduction of additional process modifications#TestStart. Act. (MBq)RCC/RCP (%)iTLCRCC/RCP (%)SE-HPLC% HMMS1Referencea, e928–106299.4 ± 0.2c94.7 ± 0.7c0.79 ± 0.382Replacing NH4OAc by NaOAc (extra buffer)b95899.6c97.0c0.053Sterile filtrationb98199.6d97.8dNDReaction conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, NH4OAc 0.2 N pH 6), 68Ga (HCl 0.1 N, 5.0 mL), 50 °C, Extra solution (2 M, pH 6, 1.05–1.25 mL) with aextra solution = NH4OAc and bextra solution = NaOAc, cRCC, dRCP, n = 1 unless otherwise specified, en = 4 Results of the manufacturing of [68Ga]Ga-DOTA-4AH29 after the introduction of additional process modifications Reaction conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, NH4OAc 0.2 N pH 6), 68Ga (HCl 0.1 N, 5.0 mL), 50 °C, Extra solution (2 M, pH 6, 1.05–1.25 mL) with aextra solution = NH4OAc and bextra solution = NaOAc, cRCC, dRCP, n = 1 unless otherwise specified, en = 4 For clinical use, sterile filtration was performed at the end of the process and did not affect the final radiochemical purity (RCP) (> 97%; Table 1). Taken together, these results confirm that the initial labelling conditions (50 °C, 10 min) are directly applicable to clinical production. The substitution of NH4OAc with NaOAc was maintained. Since high RCPs were achieved, no purification step was required. Process development After confirming the reaction conditions, the process was translated into the Modular-Lab Eazy synthesizer (see Fig. 1A). To ensure smooth fluid flow and avoid air blockage during transfer, a vented filter was added upstream of the sterile filter. No statistical calculation was applied for runs comparison. A Modular-Lab Eazy Radiosynthesizer, B Cassette used for producing [68Ga]Ga-DOTA-4AH29 with (a) connection to 68Ga generator, (b) 0.9% saline or ascorbic acid solution in 0.9% saline, (c) connection between 68Ga generator and precursor + extra NaOAc inlet, (d) connection to the final vial through filters, (e) overflow vial, (f) precursor + extra NaOAc inlet, and (g) reaction vial. C Connection to the final vial via filters with (h) vented filter, (i) sterile filter, and (j) final vial. D ML PharmTracer Radiosynthesizer. E Cassette used to produce [68Ga]Ga-DOTA-4AH29 with ML PharmTracer Radiosynthesizer, including (a) reaction vial, (b) ascorbic acid solution, (c) connection to 68Ga-generator, (d) final vial, (e) sterile filter, (f) vented filter, and (g) vent and connection to pressure test. F HMI scheme of the programmed process for the ML PharmTracer The cassette configuration was designed as shown in Fig. 1B–C. The precursor solution and NaOAc solution were preloaded into the reaction vial prior to synthesis (Fig. 1Bg). The system was directly connected to the 68Ge/68Ga generator system (Fig. 1Ba), allowing direct elution into the reaction vial. An overflow vial was included as a safety feature. The reaction vial and final vial (Fig. 1Cj) were connected (Fig. 1Bd) through a vented filter (Fig. 1Ch) and a sterile filter (Fig. 1Ci). The process started with eluting 68Ga from the generator. Initial automated runs showed efficient radiolabelling (RCP > 97%) but revealed activity losses during filtration (7–14.4%), resulting in a moderate overall recovery (RCY < 80%; Table 2, runs 1–3). To address this, a rinsing step was introduced by adding a dedicated vial containing 0.9% NaCl to the cassette. This solution was used to flush the reaction vial and filters after product transfer. Table 2Results of [68Ga]Ga-DOTA-4AH29 productions on ML Eazy Radiosynthesizer and ML PharmTracer radiosynthesizerML EazyML PharmTracerRun123412Start. Act. (MBq)887809853817709.4764.5End. Act. (MBq)698585650695553.5602.4RCP – iTLC-SG (%)99.799.399.7> 99.999.099.3RCP – SE-HPLC (%)99.399.299.399.296.798.2RCY – iTLC-SG (%)a78.671.876.085.077.278.2RCY – SE-HPLC (%)a78.271.775.784.375.477.4Reaction conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, NH4OAc 0.2 N pH 6), NaOAc (2 M, pH 6, 1.0 mL), 68Ga (HCl 0.1 N, 5.0 mL), 50 °C, 10 min, n = 1. aRCY is expressed non-decay corrected Results of [68Ga]Ga-DOTA-4AH29 productions on ML Eazy Radiosynthesizer and ML PharmTracer radiosynthesizer Reaction conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, NH4OAc 0.2 N pH 6), NaOAc (2 M, pH 6, 1.0 mL), 68Ga (HCl 0.1 N, 5.0 mL), 50 °C, 10 min, n = 1. aRCY is expressed non-decay corrected The addition of this rinsing step reduced residual activity within the system (losses < 2%) and improved recovery (RCY > 80%; Table 2, run 4), while maintaining high RCP (> 97%). Residual activity in the reaction vial remained low (< 4%), confirming efficient transfer. The optimized process was subsequently implemented on a second automated platform, the Modular-Lab PharmTracer system (Fig. 1D), using the same sequence of steps, including the rinsing procedure. At this stage, the final formulation was further refined by replacing saline with an ascorbic acid solution in saline to limit radiolysis. This resulted in a final product volume of 10 mL with a defined ascorbic acid concentration (10 mg/mL). In a previous clinical trial with radiolabelled sdAb, the final formulation contained 5 mg/mL ascorbic acid (D’Huyvetter et al. 2021; Gondry et al. 2024). In the present study, the concentration was increased to 10 mg/mL to reduce the risk of radiolysis when using higher starting activities (3–4 GBq) in future applications while not presenting any risk for the patient. Prior to each run, cassette integrity was verified through a pressure test, with pressure monitored over 1 min. Production on the Modular-Lab PharmTracer yielded results comparable to those obtained with the Modular-Lab Eazy, with consistently high RCP (> 95%) and reproducible yields (RCY > 75%; Table 2). The final product exhibited a pH of 4.8–4.9, which is lower than in the previous reported studies linked to higher ascorbic acid concentration (reported range 5.2–6.2). According to these studies, the concentration could be lowered if needed. Visual inspection performed by eye using a black and a white screen revealed a clear, colourless solution, practically free of visible particles. Stress study Prior to process performance campaign, the robustness of the manufacturing procedure was assessed through a stress study performed on the Modular-Lab Eazy platform. The aim was to evaluate the impact of variations in key process parameters, including reaction temperature, reaction time, buffer conditions (volume, molarity, pH), and starting activity. Results were compared to the reference process, with quality control performed at end of synthesis (EOS). It as to be noted that no statistical analysis was performed. However, clear trends could be assessed. Under standard conditions, high RCP and recovery were consistently obtained (RCP > 95%, RCY > 75%). The influence of reaction temperature was investigated between 25 °C and 70 °C (Fig. 2). At lower temperatures (≤ 35 °C), labelling efficiency decreased, leading to reduced RCP (< 92%). At higher temperatures (≥ 60 °C), an incomplete transfer of the reaction mixture was observed due to filter blockage associated with air bubble formation and protein precipitation. This resulted in increased activity retention and reduced recovery (RCY < 70%). Optimal performance was consistently achieved at 50 °C. Influence of reaction temperature on the RCP and RCY of [68Ga]Ga-DOTA-4AH29. Standard conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, CAR 2.23, 0.2 M NH4OAc pH 6), NaOAc (2 M, pH 6, 1.0 mL), 68Ga (684–1080 MBq, 0.1 N HCl, 5.0 mL), 10 min, 25–70 °C, n = 1 Reaction time was evaluated between 5 and 30 min (Fig. 3). Short reaction times (5 min) resulted in incomplete labelling (RCP < 80%), while longer reaction times did not improve product quality and led to decreased recovery due to radioactive decay (RCY < 70%). A reaction time of 10 min, therefore, provided the best balance between efficiency and yield. Influence of reaction time on the recovery of [68Ga]Ga-DOTA-4AH29. Standard conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, CAR 2.23, 0.2 M NH4OAc pH 6), NaOAc (2 M, pH 6, 1.0 mL), 68Ga (891–1054 MBq, 0.1 N HCl, 5.0 mL), 5–30 min, 50 °C, n = 1 The role of the NaOAc solution conditions was assessed by varying volume, molarity, and pH (Table 3). In the absence of buffer, the reaction medium remained strongly acidic (pH ≈ 1.3), resulting in negligible labelling (RCP < 10%, RCY < 5%). In contrast, variations in the solution molarity, pH (5.9–9.0), or volume had little impact on performance, with RCP and recovery remaining high (RCP > 95%, RCY > 70%). These findings indicate that while this solution is essential to ensure appropriate reaction conditions, the process is tolerant to variations in its composition. Table 3Results of stressed [68Ga]Ga-DOTA-4AH29 manufacturing runs varying NaOAcTestiTLC-SGSE-HPLCRCP (%)RCY (%)aRCP (%)RCY (%)a1Reference> 99.979.595.876.222 mL NaOAc 2 M pH 696.175.796.375.93No extra solution5.84.9< 0.1< 0.141 mL NaOAc 2 M pH 997.872.697.772.551 mL NaOAc 5 M pH 5.998.380.797.580.0Standard conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, CAR 2.23, NH4OAc 0.2 M pH 6), NaOAc (2–5 M, pH 5.9–9, 0–2.0 mL), 68Ga (806–1054 MBq, HCl 0.1 N, 5.0 mL), 10 min, 50 °C, n = 1. aRCY presented are non-decayed corrected Results of stressed [68Ga]Ga-DOTA-4AH29 manufacturing runs varying NaOAc Standard conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, CAR 2.23, NH4OAc 0.2 M pH 6), NaOAc (2–5 M, pH 5.9–9, 0–2.0 mL), 68Ga (806–1054 MBq, HCl 0.1 N, 5.0 mL), 10 min, 50 °C, n = 1. aRCY presented are non-decayed corrected The effect of starting activity was evaluated between 400 MBq and 2 GBq (Table 4). The process showed consistent performance across this range, maintaining high RCP (> 95%) and recovery (RCY > 70%). A slight decrease in RCP was observed at the lowest activity, likely reflecting reduced reaction kinetics at lower radionuclide concentration. Table 4Results of stressed [68Ga]Ga-DOTA-4AH29 manufacturing runs varying the starting 68Ga activityTestStart. Act. (MBq)iTLC-SGSE-HPLCRCP (%)RCY (%)aRCP (%)RCY (%)a142193.074.096.777.021054 (Reference)> 99.979.595.876.23146298.676.4> 99.977.54191598.877.8> 99.978.7Standard conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, CAR 2.23, 0.2 M NH4OAc pH 6), NaOAc (2 M, pH 6, 1.0 mL), 68Ga (0.1 N HCl, 5.0 mL), 10 min, 50 °C, n = 1. a RCY presented are non-decayed corrected Results of stressed [68Ga]Ga-DOTA-4AH29 manufacturing runs varying the starting 68Ga activity Standard conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, CAR 2.23, 0.2 M NH4OAc pH 6), NaOAc (2 M, pH 6, 1.0 mL), 68Ga (0.1 N HCl, 5.0 mL), 10 min, 50 °C, n = 1. a RCY presented are non-decayed corrected To conclude, this stress study identified reaction temperature, reaction time, and buffer presence as key process parameters for producing [68Ga]Ga-DOTA-4AH29. The process showed robustness to variations in buffer composition and starting activity, supporting its suitability for clinical use. Process performance campaign Process performance runs were performed under non-GMP conditions on both automated synthesizers, each starting with approximately 1 GBq of 68Ga. The production of [68Ga]Ga-DOTA-4AH29 was completed within 17–18 min, depending on the system, and met all predefined acceptance criteria (Table 5). Table 5PP runs resultsAcceptance criteriaModular-Lab EazyModular-Lab PharmTracerPP 1PP 2PP 3PP 1PP 2PP 3Start. Act. (MBq)–106910668931145.91121.81080.6End. Act. (MBq)–876864706902.1875.9859.5iTLCRCP (%)> 92%98.999.199.6> 99.998.8> 99.9RCY (n.d.c.) (%)> 70%81.080.478.878.777.279.5RCY (d.c.) (%)≥ 80%96.593.695.693.992.094.8SE-HPLCRCP (%)> 92%96.497.496.393.395.298.2RCY (n.d.c.) (%)> 70%79.079.076.673.474.478.1RCY (d.c.) (%)≥ 80%94.191.992.987.688.693.1ELISARBP (%)> 60%133.2109.0110.6106.997.6100.2Standard conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, 0.2 M NH4OAc pH 6), NaOAc (2 M, pH 6, 1.0 mL), 68Ga (0.1 N HCl, 5.0 mL), 50 °C, 10 min, n = 1 PP runs results Standard conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, 0.2 M NH4OAc pH 6), NaOAc (2 M, pH 6, 1.0 mL), 68Ga (0.1 N HCl, 5.0 mL), 50 °C, 10 min, n = 1 RCP consistently exceeded specifications (RCP > 98% by iTLC-SG and > 95% by SE-HPLC, representative chromatograms available in Supplementary, Figs. S1 and S2), except for PP N°1 on the Modular-Lab PharmTracer, where a slightly lower value was observed by SE-HPLC (RCP > 92%). This variation is due to reagents and buffers being prepared a day before synthesis and stored at room temperature (RT) for this specific run, while fresh preparations were used for all other runs. NaOAc solution may be the reagent affecting purity, as it is the only component of the labelling mixture that needs to be prepared. Furthermore, previous internal run showed a better labelling efficiency when using a fresh solution compared to an oldest one. This observation underscores the importance of reagent freshness, although all acceptance criteria were still met. RCY were consistent across all runs (> 70% non-decay corrected and > 85% decay corrected; Table 5). Limited activity was lost in the reaction vial (< 5%) and filters (< 5%). The final product consistently exhibited a pH between 4.5 and 5.0. and was obtained as a clear, colourless solution, practically free of visible particles. Filter integrity was confirmed by bubble point testing, and microbiological analysis demonstrated sterility of the final product. Endotoxin levels, measured after decay, were below the detection limit (< 0.5 EU/mL). The relative binding potency, assessed by ELISA after complete decay, confirmed preservation of sdAb 4AH29 functionality (RBP > 95%) (Supplementary Figs. S3 and S4). Stability studies showed that RCP remained within specification over 3 h at RT (≥ 92%), while the proportion of high-molecular-mass species remained stable (Fig. 4). Stability of [68Ga]Ga-DOTA-4AH29 obtained from the PP runs performed with ML Eazy and ML PharmTracer over 3 h at RT (n = 3) with A RCP of [68Ga]Ga-DOTA-4AH29 obtained via iTLC-SG and SE-HPLC (radio-detection) and B % HMMS evolution by SE-HPLC UV – 280 nm channel detection These results demonstrate the reproducibility and robustness of the automated manufacturing process under clinically relevant conditions. Manual clinical manufacturing process for [68Ga]Ga-DOTA-4AH29 Process translation To facilitate implementation in centres without automated synthesis platforms, a manual, kit-based process was developed. The procedure involved direct elution of 68Ga into a reaction vial containing DOTA-4AH29 and NaOAc buffer, followed by heating at 50 °C for 10 min. After labelling, the reaction mixture was diluted with an ascorbic acid solution in saline and then sterile-filtered to produce the final drug product. Initial manual runs showed efficient radiolabelling, with high RCP and consistent recovery (RCP > 98%, RCY > 70%; Table 6). Activity losses were mainly associated with the reaction vial, tubing, and filtration step, but remained limited overall, confirming efficient product transfer. Table 6Test manual runs using a kit for manual production of [68Ga]Ga-DOTA-4AH29Acceptance criteriaRun 1Run 2Start. Act. (MBq)–844.3768.9End. Act. (MBq)–649.0574.3iTLCRCP (%)> 92%99.499.2RCY (n.d.c.) (%)> 70%76.474.1RCY (d.c.) (%)≥ 80%86.7N/AStandard conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, 0.2 M NH4OAc pH 6), NaOAc (2 M, pH 6, 1.0 mL), 68Ga (0.1 N HCl, 5.0 mL), 50 °C, 10 min, n = 1 Test manual runs using a kit for manual production of [68Ga]Ga-DOTA-4AH29 Standard conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, 0.2 M NH4OAc pH 6), NaOAc (2 M, pH 6, 1.0 mL), 68Ga (0.1 N HCl, 5.0 mL), 50 °C, 10 min, n = 1 The final product exhibited a pH between 4.5 and 5.0 and met all visual inspection criteria, appearing as a clear, colourless solution without visible particles. Filter integrity was confirmed, supporting the suitability of the process for clinical use. Process performance campaign Three process performance runs were performed under non-GMP conditions (Table 7). All runs met predefined acceptance criteria, with consistently high RCP (> 98% by iTLC-SG and > 95% by SE-HPLC) and yields within specification (RCY > 70% non-decay-corrected and > 80% decay-corrected), with limited loss in the reaction vial (< 5%). Table 7PP results of the manual production of [68Ga]Ga-DOTA-4AH29Acceptance criteriaPP 1PP 2PP 3Start. Act. (MBq)–107512371182End. Act. (MBq)–757.3977.1953.1iTLCRCP (%)> 92%99.699.499.9RCY (n.d.c.) (%)> 70%70.178.580.5RCY (d.c.) (%)≥ 80%80.190.091.5SE-HPLCRCP (%)> 92%*97.596.7RCY (n.d.c.) (%)> 70%*77.077.9RCY (d.c.) (%)≥ 80%*88.288.6ELISARBP (%)> 60%85.899.295.4Standard conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, 0.2 M NH4OAc pH 6), NaOAc (2 M, pH 6, 1.0 mL), 68Ga (0.1 N HCl, 5.0 mL), 50 °C, 10 min, n = 1, *HPLC issue PP results of the manual production of [68Ga]Ga-DOTA-4AH29 Standard conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, 0.2 M NH4OAc pH 6), NaOAc (2 M, pH 6, 1.0 mL), 68Ga (0.1 N HCl, 5.0 mL), 50 °C, 10 min, n = 1, *HPLC issue A lower recovery observed during PP N°1 was attributed to the premature interruption of the filtration step, which led to increased activity retention in the filter. Due to a technical issue, SE-HPLC data at EOS were not available for this run; however, measurements at 1 h confirmed that RCP remained within specification, consistent with the other process performance runs. For the PP N°2 and N°3, less than 5% of the activity was lost in the filter. The final product met all quality requirements, including sterility and endotoxin levels below the detection limit. Binding functionality was preserved as showed by ELISA (RBP > 80%) (Supplementary Fig. S5), and stability studies confirmed that RCP was maintained for at least 3 h at RT (RCP ≥ 92%), with minimal formation of high-molecular-mass species (Fig. 5). Stability of [68Ga]Ga-DOTA-4AH29 produced over the PP runs of manual runs over 3 h (n = 3) at RT with A RCP of [68Ga]Ga-DOTA-4AH29 by iTLC-SG and SE-HPLC (radio detection) and B % HMMS evolution by SE-HPLC UV–280 nm channel detection These results show that the manual kit-based process provides a reliable and consistent alternative to the automated production of [68Ga]Ga-DOTA-4AH29. The performance observed during manual process performance runs was similar to that of the automated systems, with comparable RCP, yields, and stability profiles. This consistency across different production methods further supports the reliability of the manufacturing process and its suitability for use in various clinical settings. Generator compatibility The clinical manufacturing process was modified in-house to accommodate two commonly used 68Ge/68Ga generator systems, the GalliaPharma generator (Eckert & Ziegler) and the Galli Eo generator (IRE). This was performed to assess the robustness and transferability of the process. For this part of the study, all experiments used the manual production setup. These systems mainly differ in elution volume. GalliaPharma delivers 68Ga in 5 mL of 0.1 N HCl, while Galli Eo delivers it in 1 mL of 0.1 N HCl. To ensure comparable reaction conditions and final formulation, we adjusted the volumes of NaOAc buffer and the final formulation components, including the concentration of ascorbic acid (Table 8). Other than these changes, the manufacturing process stayed the same. Table 8Comparison of production parameters between E&Z and IRE 68Ge/68Ga generator systemsStepMaterialE&Z conditionIRE conditionLabelling68Ga in 0.1 N HCl (mL)5.01.02 M NaOAc pH 6 (mL)1.00.2Protein (mL)0.60.6Volume total (mL)6.61.8Final formulationAA concentration (mg/mL)29.412.2Adding volume (mL)3.48.2Total volume (mL)10.010.0 Comparison of production parameters between E&Z and IRE 68Ge/68Ga generator systems Production runs performed with both generators showed similar performance (Table 9). RCP and RCY consistently met acceptance criteria (RCP > 92%; RCY > 70% non-decay-corrected and > 85% decay-corrected). No significant activity losses were observed during the process. The final product had a pH between 4.6 and 4.7. Table 9manufacturing and QC results for [68Ga]Ga-DOTA-4AH29 using both 68Ge/68Ga generator systemsAcceptance criteriaE&Z conditionsIRE conditionsStart. Act. (MBq)–903889End. Act. (MBq)–681664iTLCRCP (%)> 92%99.798.2RCY (n.d.c.) (%)> 70%75.273.4RCY (d.c.) (%)≥ 80%92.389.1SE-HPLCRCP (%)> 92%97.696.1RCY (n.d.c.) (%)> 70%73.671.8RCY (d.c.) (%)≥ 80%90.487.2ELISARBP (%)≥ 60%117.62135.23Standard conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, 0.2 M NH4OAc pH 6), NaOAc (2 M, pH 6, 1.0 mL), 68Ga (0.1 N HCl, 5.0 mL), 50 °C, 10 min, n = 1 manufacturing and QC results for [68Ga]Ga-DOTA-4AH29 using both 68Ge/68Ga generator systems Standard conditions: DOTA-4AH29 (5.0 mg/mL, 3 mg, 0.6 mL, 0.2 M NH4OAc pH 6), NaOAc (2 M, pH 6, 1.0 mL), 68Ga (0.1 N HCl, 5.0 mL), 50 °C, 10 min, n = 1 Product quality was further evaluated using several analytical methods, including ELISA and a bead-based assay (BBA). All results confirmed the preservation of binding capacity (RBP > 90%; Table 10), indicating that the process changes did not affect the biological functionality of the sdAb. Table 10Potency of [68Ga]Ga-DOTA-4AH29 obtained by BBA at t0 and t = 3 h(%)ProcessConditionst0t = 3 hIRE process60 min-4 °C89.591.8E&Z process60 min-4 °C88.091.0n = 1 Potency of [68Ga]Ga-DOTA-4AH29 obtained by BBA at t0 and t = 3 h(%) n = 1 Stability studies showed that RCP stayed within specification for 3 h at RT (RCP ≥ 92%). There was minimal formation of high-molecular-mass species (< 3%, Fig. 6). Stability of [68Ga]Ga-DOTA-4AH29 obtained by IRE and E&Z processes over 3 h at RT, n = 1. A %RCP on iTLC-SG and SE-HPLC (radio channel) and B %HMMS (UV–280 nm channel) These results confirm that the manufacturing process is robust and transferable across different 68Ge/68Ga generator systems, enabling flexible implementation in diverse clinical environments. Discussion This study establishes a practical and robust manufacturing process for [68Ga]Ga-DOTA-4AH29, enabling both automated and manual production without requiring purification. The consistency of the results across platforms, production methods, and generator systems supports the practical implementation of this tracer in routine clinical settings. FAP remains a highly relevant target within the tumour microenvironment, and a wide range of FAP-directed tracers are currently being investigated. Among them, various FAP inhibitors (FAPis), such as FAPI-04 (Lindner et al. 2018), FAPI-46 (Loktev et al. 2019), FAPI-2286 (Zboralski et al. 2022), and FAPI-74 (Giesel et al. 2021; Naka et al. 2021; Watabe et al. 2023), have been developed for imaging. However, their limited tumour retention could affect their therapeutic effectiveness in patients compared to sdAbs (Dumauthioz et al. 2026). Furthermore, labelling of such molecules requires purification through a cartridge, which can complicate the implementation of the manufacturing into the clinic. In this context, sdAbs represent a particularly attractive class of targeting vectors, combining high binding specificity with favourable pharmacokinetics. The 4AH29 construct builds on these advantages, showing rapid tumour targeting and low background uptake, which are key features for high-contrast PET imaging. From a manufacturing perspective, one of the main contributions of this work is demonstrating that sdAb-based tracers can be produced using workflows comparable in simplicity to those used for small-molecule radiopharmaceuticals. Unlike more complex methods, such as prosthetic group-based labelling strategies (e.g., [18F]FB-4AH29), which involve multiple synthesis and purification steps with limited overall yields (~ 5%) (Dierick et al. 2024), the present approach uses a direct, single-step radiolabelling process. This leads to high radiochemical conversion (RCC > 95%) and removes the need for purification, present is some already reported 68Ga-radiolabelled sdAbs used in clinic (Gondry et al. 2024; Keyaerts et al. 2016; Zhang et al. 2025), providing a significant advantage for clinical translation with easier manufacturing and higher RCYs. Only minimal modifications were needed to adapt the preclinical process for clinical use. Increasing the precursor amount to 3 mg ensured compatibility with clinical dosing (Keyaerts et al. 2016) while keeping efficient labelling. Replacing NH4OAc with a GMP-compatible NaOAc buffer simplified implementation without affecting process efficacy. Importantly, adding sterile filtration, which is required for clinical use, did not impact product quality. The stress study clearly defined the operational range of the process. Reaction temperature and time were identified as key factors influencing labelling efficiency, with optimal results at 50 °C and 10 min. Conditions outside this range caused either incomplete 68Ga complexation or operational problems such as filter blockage and decreased recovery. Conversely, the process demonstrated high tolerance to variations in buffer composition and starting activity, an important feature for routine clinical production. The transition to automated synthesis required only minor process adjustments. Adding a rinsing step effectively addressed initial activity losses, enhancing overall recovery without affecting RCP. The use of ascorbic acid in the final formulation further improved product stability by reducing radiolysis. These changes enabled consistent, reproducible production on both automated platforms. PP runs confirmed the robustness of the manufacturing procedure, with all runs meeting predefined acceptance criteria. Importantly, similar performance was observed between automated and manual production methods. This equivalence is an important finding because it supports flexible implementation across sites with varying infrastructure levels. The presence of a manual, kit-based approach greatly reduces the barrier to adoption in centres without automated systems. The successful use of the process with two different 68Ga generator systems shows its versatility. Despite differences in elution profiles and formulation requirements, there was no effect on radiochemical performance or biological function. This compatibility is especially important for multi-center studies and wider clinical use. From an analytical perspective, the combination of iTLC-SG and SE-HPLC offered complementary information. This allowed for quick assessments and detailed characterization of the product. This dual approach ensured consistent quality control throughout the study. Some limitations should be noted. This work focuses on a single sdAb construct, so more studies will be needed to confirm if this method works for other sdAb-based tracers. Also, while the radiochemical and biological properties of [68Ga]Ga-DOTA-4AH29 seem promising, clinical studies are still necessary to fully determine its diagnostic value. In addition, future work could explore the use of alternative chelators such as NOTA to enable efficient RT complexation with gallium-68, which may further improve radiolabelling robustness and facilitate translation to more temperature-sensitive constructs. Overall, this work demonstrates that sdAb-based radiopharmaceuticals can be produced using practical, scalable processes. Beyond the specific case of 4AH29, it provides a framework that could help use protein-based tracers in routine clinical settings. Conclusion In this study, we developed a robust and reproducible manufacturing process for [68Ga]Ga-DOTA-4AH29, which was successfully validated for clinical use. The process enables both automated and manual production without purification and is compatible with multiple 68Ga generator systems. Consistent product quality was achieved, with high RCP and yield, and preserved binding capacity (RCP ≥ 92–95%, RCY ≥ 70% non-decay-corrected). The identification of critical process parameters, including reaction temperature, time, and buffer conditions, further supports the robustness and reliability of the manufacturing process within a well-defined operational space. Beyond the specific case of 4AH29, this research demonstrates the potential to develop protein-based radiopharmaceuticals using manufacturing methods nearly as simple and reliable as those for traditional small-molecule tracers. This work demonstrates that sdAb-based radiopharmaceuticals can be manufactured using clinically scalable workflows that approach the simplicity and robustness of small-molecule tracers. Although additional constraints remain due to the protein nature of sdAbs, such as increased sensitivity to temperature and handling conditions, as well as more complex quality control requirements (e.g., SEC-based analysis)—the process can nevertheless be implemented using straightforward and reliable procedures compatible with routine clinical production. Materials and methods A detailed description of materials and methods can be found in supplementary information. Materials All reagents and solvents were obtained from commercial suppliers at the highest available purity and used as received unless stated otherwise. The 68Ga generators Galli Eo and GalliaPharma were supplied by IRE (Fleurus, Belgium) and Eckert & Ziegler (Berlin, Germany), respectively. The Modular-Lab Eazy module, the Modular-Lab PharmTracer radiosynthesizer, and the corresponding cassette components were obtained from Eckert & Ziegler (Berlin, Germany). Solvents used for quality control, including 0.1 M sodium citrate, pH 4.5 for iTLC-SG and PBS 1X, pH 7.4 for SE-HPLC, were stored at 4 °C. Precursor molecule DOTA-4AH29 in 0.2 M NH4OAc, pH 6, with a chelator/sdAb ratio of 2.23 and protein concentration of 5.02 mg/mL was stored at − 80 °C. All other reagents were stored according to the manufacturer’s recommendations. The pH was measured using pH indicator strips in combination with a QUANTOFIX Relax reader (Macherey-Nagel, Feucht, Germany). Endotoxin measurements were performed using the EndoSafe PTS system (Charles River, Sulzfeld, Germany) with licensed endotoxin cartridges. FAP pre-coated beads were purchased from CD Bioparticles (Shirley, NY, USA), and magnetic separation was performed using a DynaMag-2 magnet (Thermo Fisher Scientific, Waltham, MA, USA). ELISA measurements were acquired on a SpectraMax ID3 plate reader (Molecular Devices, San Jose, CA, USA). Radioactivity measurements were performed using a gamma counter (PerkinElmer, Shelton, CT, USA) and a dose calibrator (COMECER, Castel Bolognese, Italy). Methods Radiochemistry Automated production Automated synthesis was performed using Modular-Lab software (version ML6.2) on both ML Eazy and ML PharmTracer synthesizers. ML Eazy module transfer liquids by gaz pressure. ML PharmTracer module transfer liquids by syringe module. Filter and cassette integrity tests were carried out prior to each production run. 0.6 mL of DOTA-4AH29 and 1.0 mL of 2 M NaOAc pH 6 were manually added in the reaction vial.68Ga was eluted from the GalliaPharma generator with 5 mL of 0.1 N HCl directly into the reaction vial. The reaction mixture was heated at 50 °C for 10 min. After radiolabelling, the mixture was transferred through a Cathivex®-GV filter unit 0.22 μm vented filter and a Millex®-GV filter unit 0.22 μm sterile filter into the final product vial. The reaction vial was rinsed with 3.4 mL of ascorbic acid solution in saline (29.4 mg/mL), and the rinse was also transferred through the filters into the final vial. Activities in the final vial, reaction vial, sterile filter and vented filter were measured separately at the end of synthesis. Starting activity was estimated by decay-correction of the addition of all these activities. Manual production Manual productions were performed using either the Galli Eo or GalliaPharma 68Ga generator. For the Galli Eo-based process, 1.0 mL of 68Ga in 0.1 N HCl was eluted into a 10 mL glass vial containing 0.6 mL of DOTA-4AH29 and 0.2 mL of 2 M NaOAc pH 6. The reaction was carried out at 50 °C for 10 min. After labelling, the reaction mixture was diluted with 8.2 mL of ascorbic acid in saline (12.2 mg/mL) and sterile-filtered using a Millex®-GV filter unit 0.22 μm sterile filter into the final product vial. Activities in the final vial, reaction vial and sterile filter were measured separately at the end of synthesis. Starting activity was estimated by decay-correction of the addition of all these activities. For the GalliaPharma-based manual process, 5.0 mL of 68Ga in 0.1 N HCl was eluted into a 10 mL glass vial containing 0.6 mL of DOTA-4AH29 and 1.0 mL of 2 M NaOAc pH 6. The mixture was heated at 50 °C for 10 min, diluted with 3.4 mL of ascorbic acid in saline (29.4 mg/mL), and sterile-filtered using a Millex®-GV filter unit 0.22 μm sterile filter into the final product vial. Activities in the final vial, reaction vial, sterile filter and vented filter were measured separately at the end of synthesis. Starting activity was estimated by decay-correction of the addition of all these activities. Study design The study used a step-by-step approach. First, preclinical labelling conditions were adapted into a process suitable for clinical use, including testing buffer substitution, sterile filtration, and transfer to automated systems. A stress test was then carried out to evaluate the effects of key process variables. Process performance runs were later performed for both automated and manual methods. Finally, compatibility with two different 68Ga generator systems was tested. Quality control iTLC-SGRCP was assessed by iTLC-SG using 0.1 M sodium citrate pH 4.5 as the mobile phase. A 5 µL aliquot of the drug product was applied to the strip and developed under standard conditions. Radioactive species were detected using either a MiniGita Dual radio-TLC scanner (Raytest-Elysia) or, where applicable, an Eckert & Ziegler radio-TLC system. RCP was calculated from relative peak areas using total-area normalization. This method was used to distinguish [68Ga]Ga-DOTA-4AH29 from free 68Ga species. iTLC-SG RCP was assessed by iTLC-SG using 0.1 M sodium citrate pH 4.5 as the mobile phase. A 5 µL aliquot of the drug product was applied to the strip and developed under standard conditions. Radioactive species were detected using either a MiniGita Dual radio-TLC scanner (Raytest-Elysia) or, where applicable, an Eckert & Ziegler radio-TLC system. RCP was calculated from relative peak areas using total-area normalization. This method was used to distinguish [68Ga]Ga-DOTA-4AH29 from free 68Ga species. SE-HPLCSE-HPLC analyses were performed to assess radiochemical identity, RCP, chemical purity, and the presence of high-molecular-mass species. Radio-SE-HPLC was carried out using either an Agilent or Shimadzu HPLC system equipped with UV (280 nm) and radioactivity detectors. Depending on the analytical setup, a Superdex 75 Increase 10/300 GL or Superdex 30 Increase 10/300 GL column (Cytiva) was used with PBS as the mobile phase. The flow rate was 0.7 mL/min for the Eckert & Ziegler setup and 1.0 mL/min for the internal setup.Radiochemical identity was established by comparison of the retention time of the main radioactive peak in the drug product with that of the DOTA-4AH29 reference peak detected by UV at 280 nm. RCP was determined from the radiochromatogram by total area normalization, while chemical purity was assessed from the UV chromatogram. Detailed preparation of quality control samples is described in the Supplementary Information. SE-HPLC SE-HPLC analyses were performed to assess radiochemical identity, RCP, chemical purity, and the presence of high-molecular-mass species. Radio-SE-HPLC was carried out using either an Agilent or Shimadzu HPLC system equipped with UV (280 nm) and radioactivity detectors. Depending on the analytical setup, a Superdex 75 Increase 10/300 GL or Superdex 30 Increase 10/300 GL column (Cytiva) was used with PBS as the mobile phase. The flow rate was 0.7 mL/min for the Eckert & Ziegler setup and 1.0 mL/min for the internal setup. Radiochemical identity was established by comparison of the retention time of the main radioactive peak in the drug product with that of the DOTA-4AH29 reference peak detected by UV at 280 nm. RCP was determined from the radiochromatogram by total area normalization, while chemical purity was assessed from the UV chromatogram. Detailed preparation of quality control samples is described in the Supplementary Information. pH and visual inspectionpH was determined using indicator strips over the range 2.0–9.0. Visual inspection was performed after complete decay and included assessment of container-closure integrity, visible particles, colour, and the overall appearance of the final drug product. pH and visual inspection pH was determined using indicator strips over the range 2.0–9.0. Visual inspection was performed after complete decay and included assessment of container-closure integrity, visible particles, colour, and the overall appearance of the final drug product. Filter integrityFor automated productions, filter integrity was verified by bubble point testing using the filter test module provided by Eckert & Ziegler. For manual productions, bubble point testing was performed using a Millex integrity tester (Merck Millipore). The test was considered successful if the bubble point was > 3.6 bar. Filter integrity For automated productions, filter integrity was verified by bubble point testing using the filter test module provided by Eckert & Ziegler. For manual productions, bubble point testing was performed using a Millex integrity tester (Merck Millipore). The test was considered successful if the bubble point was > 3.6 bar. Potency assayBinding functionality of [68Ga]Ga-DOTA-4AH29 was assessed by ELISA and, where applicable, by bead-based assay. Potency assay Binding functionality of [68Ga]Ga-DOTA-4AH29 was assessed by ELISA and, where applicable, by bead-based assay. ELISA100 µL of recombinant antigen at 1 µg/mL in 100 mM NaHCO3 was immobilized overnight at 4 °C into separate wells of a Maxisorb 96-well microtiter plate (NUNC, Rochester, NY). Non-immobilized antigen was removed and wells were washed 3 times with PBS + 0.1% (v/v) Tween-20 before being overcoated with 250 µL of protein-free T20 Blocking buffer for 2 h at RT. Wells were washed as previously and incubated with serial dilutions of reference DOTA-4AH29 and final product (100 µL in PBS + 0.1% Tween-20). The plate was incubated 1 h at RT. Wells were washed as previously. Wells were then incubated first with 100 µL of rabbit polyclonal antiserum against 4AH29 and then 100 µL of goat anti-Rabbit IgG-heavy and light chain antibody HRP conjugated, each 1 h at RT with washing in between. After new washing cycle, wells were incubated with 100 µL of TMB and the plate was left 10 min in the dark. Finally, 50 µL of H2SO4 were added in each well. The relative binding potency of [68Ga]Ga-DOTA-4AH29 towards their reference was calculated. Therefore, the obtained results had to be constrained after which the resulting EC50 were used in the calculation ((EC50 DOTA-4AH29 / EC50 [68Ga]Ga-DOTA-4AH29) × 100) (White et al. 2019). ELISA 100 µL of recombinant antigen at 1 µg/mL in 100 mM NaHCO3 was immobilized overnight at 4 °C into separate wells of a Maxisorb 96-well microtiter plate (NUNC, Rochester, NY). Non-immobilized antigen was removed and wells were washed 3 times with PBS + 0.1% (v/v) Tween-20 before being overcoated with 250 µL of protein-free T20 Blocking buffer for 2 h at RT. Wells were washed as previously and incubated with serial dilutions of reference DOTA-4AH29 and final product (100 µL in PBS + 0.1% Tween-20). The plate was incubated 1 h at RT. Wells were washed as previously. Wells were then incubated first with 100 µL of rabbit polyclonal antiserum against 4AH29 and then 100 µL of goat anti-Rabbit IgG-heavy and light chain antibody HRP conjugated, each 1 h at RT with washing in between. After new washing cycle, wells were incubated with 100 µL of TMB and the plate was left 10 min in the dark. Finally, 50 µL of H2SO4 were added in each well. The relative binding potency of [68Ga]Ga-DOTA-4AH29 towards their reference was calculated. Therefore, the obtained results had to be constrained after which the resulting EC50 were used in the calculation ((EC50 DOTA-4AH29 / EC50 [68Ga]Ga-DOTA-4AH29) × 100) (White et al. 2019). Bead-based assayThe immunoreactive fraction (IRF) of the final product was determined using a bead-based assay with antigen-coated magnetic particles and a serial dilution. Briefly, magnetic beads conjugated with the target antigen (hFAP) were reconstituted and diluted to produce a range of antigen concentrations. Simultaneously, the final product was diluted to a specific working concentration in DPBS containing 1% BSA. Aliquots of the drug product were then incubated with the different bead dilutions in triplicate, next to control samples for non-specific binding, for 1 h at 4 °C. After incubation, bead-bound and unbound fractions were separated using a magnetic rack, and the supernatant was collected. Both fractions were measured for radioactivity using a gamma counter. The IRF was calculated from the ratio of bound to total activity across the antigen dilution series. Bead-based assay The immunoreactive fraction (IRF) of the final product was determined using a bead-based assay with antigen-coated magnetic particles and a serial dilution. Briefly, magnetic beads conjugated with the target antigen (hFAP) were reconstituted and diluted to produce a range of antigen concentrations. Simultaneously, the final product was diluted to a specific working concentration in DPBS containing 1% BSA. Aliquots of the drug product were then incubated with the different bead dilutions in triplicate, next to control samples for non-specific binding, for 1 h at 4 °C. After incubation, bead-bound and unbound fractions were separated using a magnetic rack, and the supernatant was collected. Both fractions were measured for radioactivity using a gamma counter. The IRF was calculated from the ratio of bound to total activity across the antigen dilution series. EndotoxinsEndotoxin testing was performed as part of release and validation testing, in accordance with the analytical workflow described in the study. In short, a sample of the final product was diluted 5 times in LAL reagent and buffer BGB. 25 µL of the diluted sample was added in each of the 4 wells of the cartridge. Endotoxins Endotoxin testing was performed as part of release and validation testing, in accordance with the analytical workflow described in the study. In short, a sample of the final product was diluted 5 times in LAL reagent and buffer BGB. 25 µL of the diluted sample was added in each of the 4 wells of the cartridge. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements We thank Eckert & Ziegler team for experimentations. Author contributions MB, SN, XL performed experiments, analysed the results, prepared the first draft of the manuscript, contributed to the experimental design and data interpretation. VG, ARPA contributed to experiments and analysis. SM contributed to the development and production of the sdAb protein used in this study. DM contributed by participating in data interpretation. MD provided supervision, contributed to the experimental design and data interpretation, and reviewed the manuscript. LN provided supervision, contributed to the experimental design, data interpretation and reviewed the manuscript the experimental design and data interpretation. Funding Not applicable. Data availability The datasets used during the current studies are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests MB, SN, XL, VG, SM, ARPA, MD and LN are employees, and DM is a consultant of Precirix NV; and MD hold ownership interest (including patents) in sdAb radiotherapeutics. No other potential conflict of interest relevant to this article was reported. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Marion Berdal and Surasa Nagachinta have contributed equally to this work. Matthias D’Huyvetter and Laurent Navarro contributed equally to this work. References Associated Data Data Availability Statement The datasets used during the current studies are available from the corresponding author on reasonable request.