Otto: a 4.04 GBq (109 mCi) 68Ge/68Ga generator, first of its kind - extended quality control and performance evaluation in the clinical production of [68Ga]Ga-PSMA-11

Background Here we report on the comprehensive quality control of a 4.04 GBq (109 mCi) generator supplied by itG (Munich, Germany), and used for routine production of [68Ga]Ga-PSMA-11 for clinical imaging. The performance of the 4.04 GBq itG 68Ge/68Ga generator was studied for a year and parameters including elution yield, elution profile, radioactive and stable contaminants were collected. The production yields of a series of 175 [68Ga]Ga-PSMA-11 clinical batches are also reported herein. Results This first-of-its-kind GMP grade 68Ge/68Ga generator from itG with a nominal activity of 4.04 GBq (109 mCi) showed a stable 68Ga elution profile with elution efficiency averaging 58.3 ± 3.7%. 68Ge contaminant in the eluent slightly increased over time but remained 100x lower than those reported for comparable 1.85 GBq (50 mCi) itG generators. Metal impurities were found in concentrations lower than 100 ng/ml (ppb) throughout the study. [68Ga]Ga-PSMA-11 was obtained in 89 ± 4% radiochemical yields and > 99% radiochemical and chemical purities. Conclusion 4.04 GBq (109 mCi) itG 68Ge/68Ga generator is suitable for routinely produced 68Ga tracers used in the clinic. Up to 30% higher amount of final drug product was obtained as compared to the 1.85 GBq (50 mCi) itG generator, and as a result larger number of studies could be performed, while reducing the synthetic burden.

IMPLICATIONS FOR PATIENT CARE: This 68 Ge/ 68 Ga generator doubles the initial activity of existing generators accommodating higher patient volumes and resulting a longer shelf life while still performing according to specifications.

Introduction
The value of PSMA-targeted diagnosis and therapy monitoring of prostate cancer by means of PET/CT imaging is undeniable (Hana et al. 2018). While several groups are working on an 18 F-labeled substitute for PSMA imaging Giesel et al. 2017;Szabo et al. 2015), [ 68 Ga]Ga-PSMA-11 (a.k.a. [ 68 Ga]Ga-PSMA-HBED-CC or [ 68 Ga]Ga-DKFZ-PSMA-11) is the current gold standard (Hana et al. 2018). However, PET/CT imaging with [ 68 Ga]Ga-PSMA-11 is becoming a victim of its own success, and the increasing patient volume is calling for either the increase in generator production or the availability of generators containing higher initial activity, or both (Smith et al. 2013). Despite efforts to directly produce Gallium-68 ( 68 Ga) in cyclotrons and because of many technical and financial complications (Pandey et al. 2014), currently 68 Ga can only be reliably produced using a 68 Ge/ 68 Ga generator (Amor-Coarasa et al. 2016McElvany et al. 1984;Amor-Coarasa et al. 2018). To date, the commercially available 68 Ge/ 68 Ga generators do not exceed the capacity of 1.85 GBq (50 mCi) (Amor-Coarasa et al. 2016McElvany et al. 1984;Roesch 2013;Greene and Tucker 1961). Here we report a comprehensive quality control of a 4.04 GBq (109 mCi) 68 Ge/ 68 Ga generator produced by Isotopen Technologies Garching GmbH (itG GmbH, Munich, Germany); herein lovingly and appropriately referred to as "Otto" (Fig. 1). We also evaluate its use in the routine clinical production of [ 68 Ga]Ga-PSMA-11 in combination with an iQS Fluidic Labeling Module.  For labeling, Otto was eluted with 4 ml 0.05 M HCl, making sure an elution had been performed at least 24 h in advance. Generator elutions for quality control purposes were performed on a weekly basis -preferably on Mondays after weekend inactivityusing 6 ml 0.05 M HCl and collecting 6 × 1 ml fractions. Collected fractions were assayed for 68 Ga activity content in a CRC-15 PET Capintec dose calibrator and left to decay for at least 24 h. All decayed fractions were counted to determine 68 Ge breakthrough (reported as nominal activity, activity concentration, or as % of the total 68 Ge activity in the generator at the time of elution) using a Wallace Wizard 3″ 1480 well-counter, and a 4.118 kBq (111.3 nCi; calibrated on 8/7/2017) 68 Ge NIST traceable source was used for quantification. Fractions from elutions performed on days 41, 77, 111, 200 and 322 postcalibration were randomly selected (a representative sample spread over the year of study) and their 68 Ga and 68 Ge elution profiles are presented in the Results section. The same decayed fractions were analyzed by ICP-MS to determine the amounts of stable Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, Ge, and Al contaminants per elution and per fraction.
As stated before, the generator was eluted at least 24 h in advance of any patient study to eliminate excess 68 Zn from 68 Ga decay and radiolysis products. To further test generator's performance, [ 68 Ga]Ga-PSMA-11 was labelled using the itG's iQS 68 Ga Fluidic Labeling Module and itG's 68 Ga Peptide Radiolabeling kit at 95°C for 5 min as described previously (Amor-Coarasa et al. 2016). Briefly, 5μg of PSMA-11 were added to 1 ml NaOAc buffer solution included in the kit package. [ 68 Ga]Ga-PSMA-11 was purified using a reverse phase C18 Sep-Pak Light (Waters, USA) and filtered for sterilization through a Millipore Cathivex-GV 0.22 μm membrane before undergoing quality control testing. All QC testing was also performed as previously described (Amor-Coarasa et al. 2016), and included bubble point test, pH, sterility, decay, MCA, HPLC and pyrogen testing (Additional file 1: Table S4).

Results
The 68 Ge/ 68 Ga generator studied herein contained 4040 MBq (109.2 mCi) of 68 Ge at calibration. This generator was used extensively in our department for almost a year, having undergone 230 elutions for clinical [ 68 Ga]Ga-PSMA-11 production and generator quality control as well as > 100 additional elutions for preclinical research (the latter data not included in this study). The average 68 Ga elution efficiency for this generator was 58.3 ± 3.7% (all reported values are decay corrected). Over the studied period, the elution efficiency remained remarkably consistent, as shown in Fig. 2 (slope ≈0). The maximum elution yield was 65.2% registered at day 103 post-calibration, while the minimum 43.0% was obtained at day 274 (Fig. 2). In contrast to the stable and reproducible 68 Ga elution yield shown by Otto, the amount of 68 Ge in the eluting solution increased over time, ranging from 4.8 × 10 − 6 % on day 82 to 7.9 × 10 − 5 % on day 350 postcalibration (and average of 6× increase within the studied period, expressed as % of 68 Ge present in the generator at the time of elution) (Fig. 2). Despite this increase of 68 Ge content with time, the amounts always remained under 0.001%, with an average value of (3.4 ± 1.8)·10 − 5 % (Fig. 2).
During the first 100 days of use, 69.5±5.6% of the eluted 68 Ga activity was found in fractions 3 and 4. The elution profile started changing gradually after day 100 with the bulk of the 68 Ga activity eluted moving towards the elution front; 83.4±3.7% of the activity was found in fractions 2 and 3 (with a reduction to 34.4±13.6% in fractions 3 and 4) (Fig. 3a). The 68 Ge elution profile also changed in a similar manner, accompanied by an overall increase in the eluted activity (Fig. 3b). Raw data collected is shown in tables in the Additional file 1: Table S1.
The concentrations of metal impurities, such as Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, Ge, and Al, present in elutions 41, 77, 111, 200, and 322 were extremely low, always under 100 ng/ml (ppb) as shown in Fig. 4. The main impurity present was Zinc, mainly due to 68 Ga decay. A comprehensive table containing the raw values presented in Fig. 4 is included in the Additional file 1: Table S2. The average radiochemical yield for 175 [ 68 Ga]Ga-PSMA-11 clinical preparations using this generator was 89 ± 4%. The individual radiochemical yields over time (vs. generator age in days post calibration) are presented in Fig. 5. During clinical preparations, the radioactivity found in the waste vial accounted for only 3.4 ± 1.2% of total eluted activity -presumed to be free ionic 68 Gaand was not further tested. The C-18 sep-pak lite (used for final drug purification and reformulation) contained 5.5 ± 3.2% of the eluted activity while less than 2% of the activity (presumed [ 68 Ga]Ga-PSMA-11, but not extracted for testing) was retained in the 0.22 μm Cathivex filter. The 68 Ge radionuclidic impurity was not detected in the final drug product (< 50 Bq/ml or 1.5 nCi/ml: detection limit for 68 Ge in our well-counter) and was found to be at similar levels in the waste vial during synthesis than that in the quality control elutions (Figs. 2 and 3). The radiochemical and chemical purity of the drug product was > 99% for all preparations of [ 68 Ga]Ga-PSMA-11, as determined by radio-HPLC.
A table containing the list with the Batch Release Acceptance Criteria for [ 68 Ga]Ga-PSMA-11 along with the average results obtained in 175 production syntheses is included in the Additional file 1: Table S4. A table containing the values plotted in Fig. 5 is also included in the Additional file 1: Table S3.

Discussion
Otto is yet another example of the outstanding performance achieved by modern 68 Ge/ 68 Ga generators. The 100% increase in 68 Ge activity at calibration when compared to any other reported 68 Ge/ 68 Ga generator, did not led to any measurable increase of   -Coarasa et al. 2016-Coarasa et al. , 2018McElvany et al. 1984;Roesch 2013;Greene and Tucker 1961). The 100% increase in 68 Ge activity at calibration, resulted in only a 20-30% increase in eluted 68 Ga activity, due to a decreased elution yield when compared to previously published reports of similar generators from the same manufacturer (Amor-Coarasa et al. 2016. These elution yields did not appear to change significantly over time, eluting more than 1 TBq (≈27 mCi) of 68 Ga a year after calibration (Fig. 2). In our clinical setting, this increase in overall eluted activity allowed us to prepare multiple doses of [ 68 Ga]Ga-PSMA-11 out of a single clinical production run, thus reducing the overall cost of drug production as well as the labor involved. The decrease observed in both the 68 Ge content and the elution yield cannot be explained since we had no part in the production of this generator, but we could speculate that it maybe is the result of incorporating an enlarged column to accommodate the higher initial activity. Interestingly, and contrary to previous reports, the amount of 68 Ge breakthrough increased along with the generator age, as shown in Fig. 2. Nevertheless, the maximum 68 Ge breakthrough observed (1.32 kBq or 0.036 μCi), accounts for only 8 × 10 − 5 % of the total 68 Ge activity present in the generator at the time of elution, which is almost 100-fold lower than the one observed with previous generators at their purest 68 Ge elution levels.
The elution profiles for both 68 Ga and the 68 Ge impurity changed over time. The highest activity concentration was initially found in fraction 3, and later moved to fraction 2. This change is again contrary to what was reported before for smaller generators from the same manufacturer, for which the elution profile was extended with time. While the elution profiles were determined in the 6 ml quality control elution, the elution yields were determined with all elutions (performed with both 6 and 4 ml). Hence, this initially extended profile could have reduced the overall yield measured when eluting with 4 ml 0.05 M HCl for labeling (Fig. 3a). This change in profile can also be partially responsible for the "stable" elution yield observed over time, as well as the minor elution yield variabilities here reported (Fig. 2).
The extended ICP-MS metal contamination study performed here revealed: i) the amounts of Iron contaminant found (main interference in the labeling of [ 68 Ga]Ga-PSMA-11) were 10 times lower than the ones reported for previous 68 Ge/ 68 Ga generators from this manufacturer (Amor-Coarasa et al. 2016), ii) the Zinc contaminant was found in similar quantities to previously reported data for previous 68 Ge/ 68 Ga generators -most likely the direct result of accumulation due to 68 Ga decay and iii) Of all other metals studied, Aluminum concentrations were always found to be the most prominent, however never exceeding 30 ng/ml (ppb). The amounts of metal contaminants did not change significantly during the studied period (p > 0.05) and did not showed a marked elution profile (p > 0.5, between fractions for all metals), which indicates that fractioning should perhaps be avoided as a purification method for this generator, given that there will not be a reduction in 68 Ge amounts either (Fig. 3), and valuable 68 Ga activity will be lost. Another important consideration is that the determination of metal contaminants presented in this report was based exclusively to quality control elution samples collected without the 24 h pre-elution that routinely precedes the clinical production runs of [ 68 Ga]Ga-PSMA-11. Therefore, the concentrations reported herein for metal contaminants represent the "worst case scenario" and are estimated to be significantly lower in production elutions.
[ 68 Ga]Ga-PSMA-11 syntheses were reproducibly performed with activity eluted from the 4.04 GBq (109 mCi) 68 Ge/ 68 Ga Generator and with an average radiochemical yield of 89 ± 4%. A few lower yield outliers could most likely be linked to operator manipulation errors. As stated before, the 68 Ge breakthrough in the final drug product was found < 50 Bq/ml (< 5·10 − 6 % of 68 Ge activity in the generator) at all instances, which is > 200 times below the acceptance criteria of 0.001% for [ 68 Ga]Ga-PSMA-11. The waste vial from [ 68 Ga]Ga-PSMA-11 production was found to contain the bulk of the 68 Ge breakthrough from the elution. No radio or UV impurities were noticed in any of the 68 GaPSMA chromatograms, and all batches showed > 99% radiochemical and chemical purity. The pure and reliable 68 Ga produced by Otto resulted in a year of reproducible drug production for clinical use. Although typically the manufacturer specified shelf life of 68 Ge/ 68 Ga generators is set to 1 year due to the decrease of 68 Ga elution yield and the parallel increase in 68 Ge breakthrough , this type of 68 Ge/ 68 Ga Generators (Containing approximately 3.7 GBq or 100 mCi, Otto-like) could easily surpass it while still performing according to specifications.

Conclusion
Otto, the first-of-its-kind GMP grade itG 68 Ge/ 68 Ga Generator with a nominal activity of 4.04 GBq (109 mCi) at calibration, was studied over a year. Otto's performance showed unparallel reproducibility over the studied period and afforded 68 Ga activity at an almost stable 58.3 ± 3.7% elution efficiency. Although amounts of 68 Ge in the elution slightly increased over time, they always remained approximately 100-fold lower than previously reported for generators with lower 68 Ge load (Amor-Coarasa et al. 2016McElvany et al. 1984;Roesch 2013;Greene and Tucker 1961). Also, the amounts of other metal impurities were lower than the ones measured in previous reports (Amor-Coarasa et al. 2016McElvany et al. 1984;Roesch 2013;Greene and Tucker 1961). A total of 175 clinical productions of [ 68 Ga]Ga-PSMA-11 were performed with an 89 ± 4% average radiochemical yield and > 99% radiochemical and chemical purity. Up to 30% more drug product activity was obtained when compared to a typical 1.85 GBq (50 mCi) generator, accommodating higher patient volumes.