- Email: [email protected]
A “dose on demand” Biomarker Generator for automated production of [18F]F− and [18F]FDG
Applied Radiation and Isotopes 89 (2014) 167–175
Contents lists available at ScienceDirect
Applied Radiation and Isotopes journal homepage: www.elsevier.com/locate/apradiso
A “dose on demand” Biomarker Generator for automated production of [18F]F and [18F]FDG V. Awasthi a,n, J. Watson b, H. Gali a, G. Matlock a, A. McFarland b, J. Bailey b, A. Anzellotti b a b
Department of Pharmaceutical Sciences, College of Pharmacy, The University of Oklahoma Health Sciences Center, Oklahoma City, OK 73117, USA ABT Molecular Imaging, 3024 Topside Business Park Drive, Louisville, TN 37777, USA
H I G H L I G H T S
Biomarker Generator (7.5 MeV cyclotron) for “dose on demand” production of PET biomarkers. Fully automated FDG production and quality control, requiring minimum operator input. Ion source comprising of a detachable anode, enabling simple swap and limiting maintenance issues. Simple and easy to replace internal target.
art ic l e i nf o
a b s t r a c t
Article history: Received 1 October 2013 Received in revised form 30 January 2014 Accepted 16 February 2014 Available online 6 March 2014
The University of Oklahoma—College of Pharmacy has installed the first Biomarker Generator (BG75) comprising a self-shielded 7.5-MeV proton beam positive ion cyclotron and an aseptic automated chemistry production and quality control module for production of [18F]F and clinical [18F]FDG. Performance, reliability, and safety of the system for the production of “dose on demand” were tested over several months. No-carrier-added [18F]F was obtained through the 18O(p,n)18F nuclear reaction by irradiation (20–40 min) of a 4 95% enriched [18O]H2O target (280 μl) with a 7.5-MeV proton beam (3.5– 5.0 μA). Automated quality control tests were performed on each dose. The HPLC-based analytical methods were validated against USP methods of quality control. [18F]FDG produced by BG75 was tested in a mouse tumor model implanted with H441 human lung adenocarcinoma cells. After initial installment and optimization, the [18F]F production has been consistent since March 2011 with a maximum production of 400 to 450 mCi in a day. The average yield is 0.61 mCi/min and 0.92 mCi/min at 3.8 mA and 5 mA, respectively. The current target window has held up for over 25 weeks against 4400 bombardment cycles. [18F]FDG production has been consistent since June 2012 with an average of six doses/day in an automated synthesis mode (RCY E50%). The release criteria included USP-specified limits for pH, residual solvents (acetonitrile/ethanol), kryptofix, radiochemical purity/identity, and filter integrity test. The entire automated operation generated minimal radiation exposure hazard to the operator and environment. As expected, [18F]FDG produced by BG75 was found to delineate tumor volume in a mouse model of xenograft tumor. In summary, production and quality control of “[18F]FDG dose on demand” have been accomplished in an automated and safe manner by the first Biomarker Generator. The implementation of a cGMP quality system is under way towards the ANDA submission and first clinical use of [18F]FDG produced by BG75. & 2014 Elsevier Ltd. All rights reserved.
Keywords: Cyclotron FDG PET
1. Introduction The use of positron emission tomography (PET) and 2-deoxy2-[18F]fluoro-D-glucose ([18F]FDG) has gradually transformed the
n
Corresponding author. E-mail address: [email protected] (V. Awasthi).
http://dx.doi.org/10.1016/j.apradiso.2014.02.015 0969-8043 & 2014 Elsevier Ltd. All rights reserved.
clinical diagnostics landscape since coming into existence more than 30 years ago. Starting primarily as a research tool, PET imaging with [18F]FDG has now become an accepted, needed, and reimbursable clinical procedure for oncology and other medical indications (Bar-Shalom et al., 2003; Hoh, 2007; Bohnen et al., 2012; Fletcher et al., 2008; U.S. Pharmacopeial Convention, 2011). Currently hybrid technologies such as PET/CT or PET/MRI are poised to keep expanding the potential of PET (Townsend, 2008; Ratib and Beyer,
168
V. Awasthi et al. / Applied Radiation and Isotopes 89 (2014) 167–175
2011). PET imaging centers are capital intensive and are dependent on the regular and predictable supply of imaging radiopharmaceuticals from independent sources. Since [18F]FDG production itself requires a cost-prohibitive infrastructure consisting of medical cyclotrons and accompanying radiosynthesis, quality control, and dispensing units, much of the production is delegated to specialized radiopharmaceutical vendors, such as IBA, PETNet, Cardinal Health, etc. It may not be out of place to cite a lack of access to radiopharmaceuticals as one of the reasons for impeded growth of PET centers in remote regions, and patient migration into the busy citybased imaging centers is routine. The current paradigm for [18F]FDG production and supply is largely based on strategically located commercial sites, producing, and distributing clinical quality product. However given the relatively short half-life of the fluorine-18 (110 min), consistent access to [18F]FDG in hospitals or imaging centers out of the distribution grid is very challenging. Such problems are magnified and are of immense importance for health care delivery in developing and underdeveloped countries. Considering the limitations posed by currently accepted network of commercial manufacturers-to-PET center distribution channel, a need exists of enabling PET centers to fulfill their own requirement of [18F]FDG and other [18F]-labeled radiopharmaceuticals in an independent, cost-effective, safe, and quality conscious manner. In this regard, Biomarker Generator (BG75), developed by ABT Molecular Imaging, is capable of providing [18F]FDG doses ondemand. The dose on demand concept in relation to radiochemistry on chip has been recently reported by Pascali's group (Arima et al., 2013; Pascali et al., 2013). However, integrating the radiochemistry with a relatively low energy cyclotron (7.5 MeV) for clinical production of imaging biomarkers is a novelty described in this paper and has been in development for over 5 years. The College of Pharmacy in the University of Oklahoma Health Sciences Center (OUHSC) recently installed a BG75 to supplement their SPECT Nuclear Pharmacy business and to service pre-clinical demands for PET imaging in its Research Imaging Facility. As the first site to install and commission the BG75, in this report we share the installation requirements, performance, reliability, and safety of BG75 for production of [18F]F and [18F]FDG. The data reported were collected over a period of several months after the successful commissioning of the cyclotron.
2. Materials and methods 2.1. General All chemicals were obtained commercially and were used without further purification. [18O]H2O was obtained from Rotem Industries Ltd. (Israel). All chemicals used for [18F]FDG production were obtained from ABX GmbH (Radeberg, Germany). The [18F] FDG Dose Synthesis Cards (DSCs) and kits were obtained from ABT Molecular Imaging (Knoxville, TN). The synthesis kit consisted of vials containing acetonitrile, phase transfer catalyst (PTC, Kryptofix 2.2.2 in acetonitrile), 2 M HCl, and water for injection. Each kit was meant to last for approximately eight [18F]FDG synthesis runs. 2.2. Biomarker Generator (BG75) BG75 (ABT Molecular Imaging, Knoxville, TN) comprises a selfshielded 7.5-MeV proton beam, positive ion cyclotron and an aseptic single-use card-based automated chemistry production and quality control module (Fig. 1). An internal beam current of r5 mA can be achieved for [18F]F production, using three nonsimultaneous internal target ports. It has a pole diameter of 74.8 mm, an extraction radius of 35 cm, and four dees operating
at a voltage of 16 kV max and frequency of 72 MHz. A magnet of mass 3.5 t generates the average magnetic field of 1.2 T (1.8 Tmax). The entire cyclotron of height 0.37 m and diameter 1.25 m is enclosed in a high efficiency shielding. The two-part (upper and lower) self-shielding of BG75 consists of a 0.64 cm steel exterior enclosing high density boronated concrete and polyethylene. The physical dimensions of the 21 t shielding are: 2.39 m diameter 1.63 m height. The enclosed cyclotron is accessed for servicing by mechanically parting the upper shield from the lower stationary shield using three permanently installed 1 HP screw drive motors. The cyclotron room exhaust was passed through an active dedicated charcoal filter (life span of 3–5 years). A continuous stack monitoring at the outlet of the College of Pharmacy building exhaust system is performed using a Ludlum Model 375-20 radiation monitoring system. The system consists of a Ludlum Model 375 digital controller coupled to two shielded 5.1 cm diameter by 5.1 cm thick NaI(Tl) scintillation detectors enclosed in weather-proof enclosures. The controller is mounted on a wall inside the cyclotron room, about 15 m away from the overhead exhaust. In addition, the institutional radiation safety office monitors the level of gamma and neutron radiations emanating from the cyclotron by monitoring radiation exposure on four walls of the cyclotron room on monthly basis. The gamma dose levels and the secondary neutrons around the cyclotron shield in the cyclotron room were also monitored with a calibrated Ludlum gamma/neutron survey meter Model 2363 with the Model 42-41L PRESCILA neutron detector (Ludlum Measurements, Inc., Sweetwater, TX). 2.3. [18F]F and [18F]FDG production No-carrier-added [18F]F was produced through the 18O(p,n)18F nuclear reaction by irradiation (20–40 min) of a 495% enriched [18O]H2O target (280 μl) with a 7.5-MeV proton beam (3.5–5.0 μA). The bombardment time (in μA min) was a selectable parameter. The process was initiated by selecting the end-product ([18F]F and [18F] FDG) and pushing a start button on an all-in-one computer touch screen. If producing [18F]F , the process was aborted after bombardment and the product was dispensed in the user-selectable container. For [18F]FDG production, the built-in program guided the operator to take further steps to prepare the DSC and chemistry production and quality control module while the target bombardment was in progress. Each run of [18F]FDG production used new sterile-packed DSC. The DSC was aseptically placed in the Card Chemistry System (CCS) of the chemistry production and quality control module (Fig. 2) and made ready by following the menudriven instructions. The preparation of CCS included the installation of a synthesis kit. Once the bombardment was done, the [18F]F containing [18O] water was transferred to a reaction vial contained in the DSC (Fig. 2b). [18F]FDG was synthesized using nucleophilic substitution of 1,3,4,6-tetra-O-acetyl-2-O-trifluoro-methanesulfonyl-β-D-mannopyranose (mannose triflate precursor) followed by acid hydrolysis as reported by Hamacher et al. (1986). However, the reaction conditions were optimized for our system. The [18F] labeling was achieved in acetonitrile at 80 1C for 3 min. The hydrolysis was achieved by using 2 M hydrochloric acid at 110 1C for 5 min. The product was purified by solid phase extraction using a cartridge containing layers of strong cation/anion exchange, neutral alumina, and C18 resins. The product was then passed through a 0.2 mm filter and delivered into a sterile syringe or a vial. 2.4. [18F]FDG quality control As part of the automated process, the quality control of each batch of [18F]FDG was performed without any user intervention. An aliquot of the final product was sampled for quality control in
V. Awasthi et al. / Applied Radiation and Isotopes 89 (2014) 167–175
169
Fig. 1. Biomarker Generator and its constituents. (a) BG75 installed at the OUHSC—College of Pharmacy campus. (b) The 7.5 MeV cyclotron inside the berite shield. (c) The target showing the replaceable insert (arrow). (d) A detachable anode of the ion source. A ruler (cm) is shown against the anode for comparison.
Fig. 2. The components of Chemistry Cart. (a) The pre-packed synthesis kit is supplied with various venting and delivery needles, and vials with acetonitrile, water for injection, 2 M HCl, phase transfer catalyst, and precursor mannose triflate. An additional vial with orange-colored ion exchange resin is also provided for connecting the vent tubing of the reaction vial. Eight sterile-packed Dose Synthesis Cards (DSC) come packaged with the kit. Overall, the entire kit is designed to enable eight production runs of [18F]FDG. (b) A DSC showing a reaction vial and the collection syringe connected via a two-way stop cock. (c) Reagent Metering System (RMS) carrying various vials is shown. (d) The Card Chemistry System (CCS) showing the bank where DSC is placed and secured by a lever and gas pressure. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)
170
V. Awasthi et al. / Applied Radiation and Isotopes 89 (2014) 167–175
the quality control module (QCM), which consisted of a pH microelectrode, refractive index, UV–vis, and radioactivity detectors. Ion exclusion high pressure liquid chromatography (HPLC) analyses were performed on a system equipped with a Knauer Smartline Pump 100, Knauer Smartline Degasser, Rheodyne TitanHT™ High Pressure Injector, Phenomenex ThermaSphere™ TS-130 HPLC Column Heater, Knauer Smartline UV Detector 2500, and a Knauer Smartline RI Detector 2300. A Rezex™ ROA-Organic Acid Hþ (8%) LC column (Phenomenex, Torrance, CA, 8 mm, 4.6 mm 250 mm) was used with HPLC grade water as mobile phase at a flow rate of 0.3 ml/min. The column temperature was set at 76 1C and the UV detector was set at 210 nm. The following release tests were performed in the QCM in an automated fashion: pH, residual solvents (acetonitrile and ethanol), radiochemical identity/purity, Kryptofix 2.2.2 (EP), and filter integrity. A diluted sample was separately provided for endotoxin test. The remaining release tests, i.e. visual inspection, Kryptofix (USP), 2-chlorodeoxyglucose, radionuclidic identity/purity, sterility, and endotoxin were performed according to the USP/FDA guidelines (U.S. Pharmacopeial Convention, 2011). For validation purposes, additional tests were performed for residual solvents using a Model 8610C Capillary FID GC system (SRI Instruments, Torrance, CA), pH determination using pH strips, and Kryptofix 2.2.2. using a spot test (Celltech Srl, Italy). The current production schedule enables all quality control tests to be performed in an automated fashion except endotoxin, sterility, and particle visualization. Since sterility takes 14 days of incubation, we release the product for pre-clinical work after endotoxin (Endosafe-PTS, Charles Liver Lab, Horsham, PA) and visual inspection for particulate material.
2.5. Animal studies The animal study was conducted in accordance with the protocols approved by the OUHSC institutional animal care and use committee. Human lung adenocarcinoma H441 cells were obtained from American Type Culture Collection (Manassas, VA). The cells were cultured in ATCC-formulated RPMI-1640 Medium supplemented with 10% fetal calf serum (Invitrogen Corporation, Carlsbad, CA) and 50 μg/ml gentamicin. Male athymic nu/nu mice (4 weeks old) were obtained from the breeding colony of Harlan Laboratories (Indianapolis, IN). Before initiating the experiment, we let the mice acclimatize for at least 5 days. For xenograft implantation, H441 cells were harvested, washed in serum-free medium and re-suspended in 0.1 M phosphate-buffered saline (PBS, pH 7.2). A single cell suspension with 490% viability was used for subcutaneous injections in mice (2 106 per implant). Two to three weeks after inoculation of the tumor cells, when the tumors reached 0.5–1 cm in diameter, the mice were recruited for PET/CT imaging studies. The small animal PET/CT was conducted in the OUHSC College of Pharmacy Research Imaging Facility using a Flex™ X-O™ X-PET™ Pre-Clinical Platform (Gamma Medica, Northridge, CA). A H441 tumor-bearing mouse was anesthetized initially using 2% isoflurane in oxygen at 2 L/min, in a polypropylene induction chamber. When fully anesthetized, the mouse was placed on the scanner bed and anesthesia was maintained through a nose cone with 1% isoflurane in oxygen at 2 L/min. Body temperature was maintained at 37 71 1C by using a water-circulated pad under the animal. A dose of 70 mCi of [18F]FDG in 100 ml normal saline was injected into the tail vein. Static small animal PET imaging data was acquired for a period of 10 min after 2 h post-injection. An X-ray computed tomography (CT) was also acquired for 1 min to enable anatomic localization of radioactivity. The PET data was reconstructed using a 2D filtered back-projection algorithm. The
PET and CT images were fused together and visualized using Amira 3.1 software (Visage Imaging Inc., San Diego, California, USA).
3. Results 3.1. Installation The installation of BG75 was preceded by a site planning guide and radiation safety guidance provided by the vendor. Space specifications for the BG75 were 4.6 m 4.6 m which resulted in a good fit with the available space in the basement of the OUHSC-College of Pharmacy. No additional construction was needed for the installation, except for floor reinforcement to specifications and minor renovations. The installation in an existing basement was a bit challenging because the heavy equipment and cyclotron parts could not be transported in the elevators. The only walkout from the basement was through an approximately 100 m underpass with a couple of 901 turns and three doors. The cyclotron and shields were taken through the underpass from the loading area to the cyclotron room. The main connections of electricity and water for the cyclotron, CCS and QCM were established to power the system and start the installation and operation qualification of the BG75. The entire rigging operation and physical installation of the system took approximately one week. The operation of the cyclotron was performed under a permit from Oklahoma Department of Environmental Quality issued to the OUHSC —College of Pharmacy. The vendor supplied berite concrete as the shielding material. As shown in Fig. 3, the self-shielding system of the BG75 reduced the gamma and neutron fields generated during operation of the cyclotron well within the permissible exposure rate. The gamma dose levels and the secondary neutrons around the cyclotron shield in the cyclotron room were monitored with a calibrated Ludlum gamma/neutron survey meter Model 2363 with the Model 42-41L PRESCILA neutron detector (Ludlum Measurements, Inc., Sweetwater, TX). The continuous stack monitoring of exhaust provided no evidence of radiation exceeding background levels. 3.2. [18F]Fluoride production The performance of the cyclotron was evaluated for [18F]F production and compared to the indicated specification of 1 mCi/ min. In order to accommodate initial optimization processes, we have
Fig. 3. The radiation profile for BG75 in operation. Exposure rate of less than 1 mR/h at 1 m from the shield surface enables general occupancy outside of the Biomarker Generator room and limits occupational exposure to the operator within the permissible levels. The measurements are shown in mR/h.
V. Awasthi et al. / Applied Radiation and Isotopes 89 (2014) 167–175
171
The ion source power supply usually runs at 49.1 70.8 mA and 1382.34 78.8 V. 3.4. [18F]FDG synthesis
Fig. 4. The [18F]F output with respect to the bombardment time in the BG75. The data for all BG75 runs at all levels of integrated-ion current levels (mA) as a function of bombardment time, regardless of downtime, daily run number, “crowbar frequency” (loss of cyclotron synchronization), variations in different runs, and target unloading times. These data are from a long period of time using a number of different ion sources and targets since installation. The average yield at 3.8 mA was 0.61 mCi/min (n ¼186), whereas approximately 0.92 mCi/min was obtained at 5.0 mA (n ¼202).
presented the data at two bombardment currents of 3.8 mA and 5.0 mA (Fig. 4). We found that bombardment at 5.0 mA provided approximately 0.92 mCi/min. Rinsing of target with additional 100 ml of enriched water could increase the yield by approximately 10%, but this practice was not followed on routine basis.
3.3. Target rebuilding and ion source replacement The single most dreaded task for a cyclotron operator is the target rebuilding. The process of target rebuilding in BG75 is much easier than that in the conventional cyclotrons. The internal target consisting of a grid-less terminal window of 20 mm Haver foil (Fig. 1c) is simple to rebuild and replace. The target is essentially a disposable rectangular clamp with a thin silicone gasket held by four screws, carrying maximum 300 ml of liquid under approximately 320 psi argon. The target body material is a non-magnetic undisclosed alloy cooled by circulating water. The placement of target inside the cyclotron is optimized for maximum beam extraction by rotational and vertical displacements. In our experience, the target lasts more than 775 mA h, which is comparable to that reported for targets in the conventional medical cyclotrons (Pant and Senthamizhchelvan, 2007). The simplicity of target design is extended to the construction of ion source as well. The ion source supports the ion beam by providing a stable supply of ions to the accelerator. Typically, the ion sources in medical cyclotrons are characterized by beam intensity, stability (both short and long terms), reproducibility, simplicity of operation, convenience of maintenance, reliability, and safety. Harnessing optimal performance from an ion source at a reasonable cost is a continuous challenge. BG75 has an innovative ion source design (Fig. 1d) that allows simple swapping of the detachable anode. It requires no specialized expertise and significantly limits the maintenance demands. The main body of the ion source carries H2 at a typical vacuum reading of 3 10 4 Torr.
The CCS unit present in the production module is in-charge of producing clinical grade [18F]FDG. The reagents are placed on top of the CCS in a unit called the Reagent Metering System (RMS) which includes a special color-coded vial support and three metering pumps paired with corresponding distribution heads (Fig. 2c). The vial support houses the vials in receptacles with a unique color/size combination which minimizes the possibility of vial misplacement by the user, as was noted during its routine use. The three metering pumps were tailored for deliveries of: (i) [18O] H2O to the cyclotron target, (ii) organic reagents to the DSC, and (iii) aqueous reagents to the DSC. The pumps were connected to the vials via sterile needles and tube lines. The DSC is the sterile and disposable unit (Fig. 2b) in which [18F]FDG manufacturing takes place. It comes double-bagged and is installed in the CCS following aseptic techniques. It is a closed system and contains the reaction vial and purification column required for [18F]FDG manufacturing. Unlike other commonly used synthesis systems, it does not employ trap-and-release of [18F]F . The water is evaporated by producing an azeotrope with acetonitrile, and the labeling is performed via a nucleophilic substitution on the mannose triflate precursor at 110 1C. This labeling step is followed by acid hydrolysis of the acetyl groups using hydrochloric acid producing the unpurified [18F]FDG. The solution is then passed through a purification column and a 0.2 mm syringe filter, to finish in a sterile syringe shielded in a tungsten shield. Overall, the average dose-to-dose workflow was 40 min, producing 12–15 mCi [18F]FDG/dose in 3 ml volume (Fig. 5a). However, the first dose of day takes approximately 1 h. The sterility and bioburden evaluations have been performed for this workflow using a third party laboratory. It passed sterility requirements according to USP general chapter /71S (U.S. Pharmacopeial Convention, 2013); less than 4 CFU/sample were recovered in total bioburden panel tests. 3.5. [18F]FDG quality control The QCM automatically performs pH determination, residual solvents (acetonitrile and ethanol), radiochemical identity/purity, Kryptofix 2.2.2. determination, and filter integrity test. The system is HPLC-based and is prepared for analysis concurrent to the ongoing target bombardment and [18F]FDG synthesis. First, the QCM primes the delivery lines and monitors the pressure in the HPLC pump. After the priming, the system executes a System Suitability Test (SST) where the pH detector is verified and the performance of the HPLC system is assessed in terms of three criteria: 1) tailing factor of the acetonitrile peak, 2) resolution factor between the acetonitrile and ethanol peaks, and 3) the value reported for the concentration of a kryptofix 2.2.2./acetonitrile/ ethanol standard. After these initial checks, the QCM is ready for the analysis of the [18F]FDG batch. The system automatically withdraws a sample (180 ml) from the DSC and performs pH determination via a microelectrode with a tolerance within 0.3 pH units. The HPLCbased system performs Kryptofix 2.2.2. and residual solvent determination plus radiochemical purity/identity. The analytical methods on the QCM have been previously validated according to specifications set by the International Conference on Harmonization in terms of accuracy, specificity, linearity, repeatability, range, robustness and detection/quantification limits (Anzellotti et al., 2013; Food and Drug Administration, 2013). A typical QC report generated for [18F]FDG is shown in Fig. 6.
172
V. Awasthi et al. / Applied Radiation and Isotopes 89 (2014) 167–175
Fig. 5. (a) A representative workflow to produce an [18F]FDG dose. (b) A representative QC chromatogram of a [18F]FDG dose manufactured with the BG75. A typical run time is 15 min.
3.6. Problems and troubleshooting BG75 being the first cyclotron of its kind installed for medical [18F]F production, the documentation and analysis of its operating data are a continuous process necessary to predict yields, identify inefficiencies, track progresses from frequent upgrades, and identify opportunities for further improvements. Over the last one year of BG75 operation, there were several instances when the intended production could not be accomplished. The failures, their frequencies, and the possible causes are enlisted in Table 1. The most frequent occurrence of target failure was an issue early during the optimization process, when the beam current was being stabilized. Since then, in the last 6–9 months of operation, we have not observed unanticipated target failure. Empirically, we have observed that the [18F]F production yield starts to gradually decrease as the target window approaches its end of life. A significant increase in pressure reading (i.e. vacuum leakage) was the typical sign of target failure. We believe that stable beam, courtesy to the recent RF coil and ion source upgrades, and the optimized beam current of 3.8 mA have a lot to do with the current success. The second critical problem was that of ion source failure, mostly attributed to the shorting of anode and cathode or the erosion of cathode. The novel design upgrade of screw-on, smallsized ion source has simplified the replacement of ion source parts. As far as the synthesis of [18F]FDG is concerned, the foremost concern was the malfunction of DSCs. The intake ports in a DSC are abutted against the solvent delivery ports in CSC by a set of mini orings to prevent leakage. Occasionally, the loss of o-rings, or their skewed placement, results in leakage and loss of reagents. The leakage may also occur if the silicone delivery lines inside the DSC become loose or come off under pressure. Whereas the placement of o-rings could be ensured before placing the DSC card in the intended groove, the failure of silicone delivery lines is impossible to determine without compromising the sealed nature of the DSCs. To address this concern, the current version of production program
pre-tests the integrity of DSC card under argon pressure. Since the CCS is completely shielded with 2.5 cm thick lead, any accidental leaks do not expose the outside environment or the operator. 3.7. Animal study In order to biologically validate the diagnostic efficacy of [18F] FDG produced by the novel automated process, we tested [18F]FDG in a mouse model of xenografted H441 lung adenocarcinoma. [18F] FDG is a biomarker of the metabolic turnover in tumor tissue. Fig. 7 shows the coronal, sagittal and transaxial views of a representative mouse carrying tumor in the right hind limb. Clearly, the tumor accumulated large amounts of [18F]FDG and the distribution in the other parts of the body was also similar to the one reported for [18F] FDG produced by other devices.
4. Discussion Installation of compact medical cyclotrons to produce [18F]F and other positron-emitting radionuclides is rapidly increasing worldwide due to the increased utilization of PET imaging in clinical practice. In this regard, a BG75 was installed at OUHSC to support both clinical and preclinical requirements of [18F]-radiopharmaceuticals. BG75 consists of a 7.5 MeV positive ion cyclotron with an internal target configuration. The positive ion beam, coupled with an internal target, enables BG75 to exist as an ultra-compact cyclotron. It is a self-shielded machine that could be operated without any concerns of undue radiation exposure to the operator. As described above, the BG75 work-flow, from production to quality control is completely automated, and requires minimum manual input. After initial placement of reagents, DSC, and collection vessel, the entire operation (consisting of bombardment and followed by FDG synthesis) starts with a push of a button and ends with an appropriately shielded dose of [18F]FDG. The process could be
V. Awasthi et al. / Applied Radiation and Isotopes 89 (2014) 167–175
173
Fig. 6. A quality control sheet is automatically printed for each [18F]FDG dose to enable the operator to objectively release or restrain to [18F]FDG dose. A representative QC sheet is shown.
Table 1 Problems encountered during installation and optimization of BG75 in the period between Jan. 2012 and Aug. 2013. Problem
Probable cause
Target failure
Bad Havar window, normal wear and tear, no cooling water flow, deliberate tests at high ion source current, low target load volume Erratic ion source operation, bad upper or lower cathode, defective high voltage ion source cable
Ion source failure
DSC-related failures (low or FDG yield below 20% for no apparent reason) RF generator malfunction
Mass (argon) flow controller breakdown Power failure
Frequency in 610 operations (2012–8/2013) 6
System error message, high vacuum indication, un-load volume zero; requires target rebuild
7
System error message, low voltage indication with high current, unstable voltage, large spike in target current; requires replacement of detachable anode DSC does not pass validation test, low pressure or mass flow controller reading, erratic reaction signal patterns, visible leakage, bad filter RF failed to resonate message, only plate volts indicated on RF controller
57 Defective ports, defective junctions, DSC pressure plate failure, defective argon line fitting, reagent concentration problems, syringe failure RF component (triode), RF amplifier, RF controller, circuit 5 breaker, cable, “D tips” touching ‘Hill sides’, bad high voltage capacitor 3 Defective unit, power surge, intermittent flow controller malfunctions, bad pressure controller, partially blocked or defective fittings Severe thunderstorms, unscheduled building 3 maintenance
repeated depending on the patient schedule and the need of subsequent doses, potentially enhancing the cost-effectiveness of imaging services. In addition, compared to the regular smallest 10 MeV
Signs/troubleshooting
Low mass flow readout and erratic graph signals; poor FDG yields Low vacuum reading, hardware/software errors
medical cyclotron, BG75 provides greater flexibility in terms of footprint, resources, and personnel needed for the entire manufacturing operation. Furthermore, BG75 would be highly suitable for
174
V. Awasthi et al. / Applied Radiation and Isotopes 89 (2014) 167–175
Fig. 7. A representative set of PET/CT images of a mouse with lung adenocarcinoma H441 cell xenograft. The mouse was injected with BG75-generated [18F]FDG and imaged at 4 h after injection. The PET/CT images were acquired using the Flex-CT/ PET system (Gamma Medica Ideas, CA) and rendered with Amria 3.1 software. Arrows show uptake of FDG in implanted tumor.
the production of low-volume [18F]-labeled radiopharmaceuticals (e.g. [18F]FMISO) that are not routinely prescribed and, therefore, are costly to procure. Due to the ultra-compact nature of the BG75, we were able to install it in an existing room of 5.5 m 5.5 m with minimal site modifications. In addition, the cyclotron, chemistry production, and quality control module, and other ancillary instrumentation were accommodated in the same room. Since installation and optimization of the operating conditions, we have been able to obtain [18F]F production yield of 0.61 mCi/min at 3.8 mA current. However, the [18F]F production has been consistent since March 2011 with a maximum production of 400 to 450 mCi in one single day (E1 mCi/3.5 mA min). Apparently, the frequency of use affects the yield, and our observations suggest that the yield increases with more frequent bombardment. During [18F]F production runs, the parameters such as [18O]H2O enrichment (495% 18O), target volume (280 ml), and beam energy (7 MeV) were kept constant while beam current and bombardment time were changed. As shown in Fig. 4, the [18F]F production yields increased almost linearly with increased bombardment. Significantly, the current target window has held up for over 25 weeks against 4400 bombardment cycles. The radiation safety requirements of a cyclotron facility are more stringent than a traditional nuclear pharmacy due to highly penetrating nature of 511 keV gamma photons, higher specific gamma ray dose constant of positron emitters, and secondary neutrons emitted from the target during a production run. In this regard, the BG75 presents several advantages over other compact medical cyclotrons mainly due to the small scale production of [18F]F and a single dose of a radiopharmaceutical per production run. In other words, much lower amount of 18F activity is produced compared to other cyclotrons ( 50 mCi versus 1–4 Ci). The
radiation shield of this cyclotron is composed of dense concrete and polyethylene (30.5 cm thick) which provides acceptable dose levels during the operation. The dose rates due to secondary neutrons always remained well within the permissible limits around the shield during the operation for the maximum beam current in use (5 mA) and for the longest duration of production run (1 h), indicating that the shield adequately minimizes the exposure from neutrons around the cyclotron. The intensity of the gamma rays and secondary neutron field produced by a cyclotron primarily depend on the beam strength and since we used o5 mA beam current in all production runs those intensities were minimal. The radiation dose levels were minimal (o 1 mR/h) around the chemistry production module due to 2.5 cm thick lead shielding. In addition to the personnel safety, it is essential to monitor the exhaust system for environmental safety, which was achieved by a continuous stack monitoring system installed at the outlet of the building exhaust system. The entire operation was deemed satisfactory from the radiation safety viewpoint. The [18F]FDG production was achieved by using the chemistry production module (Fig. 1a), which is located next to the cyclotron. This module is self-shielded, installed in a sterile environment, and thus, requires no separate hot cell. The module is equipped with a laminar flow hood to provide a low particulate environment during the production process. As the module uses prepackaged and certified DSC and reagent kit (Fig. 2), the [18F]FDG production process can be achieved under full cGMP compliance. [18F]FDG production has been consistent since June. 2012 with the maximum capacity of six doses/day in automated synthesis mode (RCYE 50%). Except for the first dose of the day, which takes approximately 60 min, a typical [18F]FDG production run takes about 40 min (Fig. 5). The synthesis conditions were optimized to obtain a desired [18F]FDG dose of 10–13 mCi/run in a final volume of 3 ml. The release criteria included USP-specific limits for pH, residual solvents, Kryptofix, radiochemical purity/identity and filter integrity test.
5. Conclusions Installation of the first ABT Biomarker Generator system, BG75, at OUHSC was successfully achieved with minimum site modifications. The cyclotron operation parameters were optimized to achieve a [18F]F production yield of 0.61 mCi/min at 3.8 mA. Since this was the first installation of BG75 in the world, some teething problems were expected. However, once those problems are overcome, BG75 is expected to reliably produce “[18F]FDG dose on demand”. The implementation of a cGMP quality system is under way towards the first ANDA submission and clinical use of BG75-produced [18F]FDG in the USA.
Acknowledgments We gratefully acknowledge the continuous support of the University of Oklahoma, College of Pharmacy and its administrative staff. References Anzellotti, A.I., McFarland, A.R., Ferguson, D., Olson, K.F., 2013. Towards the full automation of QC release tests for [18F]fluoride-based radiotracers. Curr. Org. Chem. 17, 2153–2158. Arima, V., Pascali, G., Lade, O., Kretschmer, H.R., Bernsdorf, I., Hammond, V., Watts, P., De Leonardis, F., Tarn, M.D., Pamme, N., Cvetkovic, B.Z., Dittrich, P.S., Vasovic, N., Duane, R., Jaksic, A., Zacheo, A., Zizzari, A., Marra, L., Perrone, E., Salvadori, P. A., Rinaldi, R., 2013. Radiochemistry on chip: towards dose-on-demand synthesis of PET radiopharmaceuticals. Lab Chip 13, 2328–2336.
V. Awasthi et al. / Applied Radiation and Isotopes 89 (2014) 167–175
Bar-Shalom, R., Yefremov, N., Guralnik, L., Gaitini, D., Frenkel, A., Kuten, A., Altman, H., Keidar, Z., Israel, O., 2003. Clinical performance of PET/CT in evaluation of cancer: additional value for diagnostic imaging and patient management. J. Nucl. Med. 44, 1200–1209. Bohnen, N.I., Djang, D.S., Herholz, K., Anzai, Y., Minoshima, S., 2012. Effectiveness and safety of 18F-FDG PET in the evaluation of dementia: a review of the recent literature. J. Nucl. Med. 53, 59–71. Fletcher, J.W., Djulbegovic, B., Soares, H.P., Siegel, B.A., Lowe, V.J., Lyman, G.H., Coleman, R.E., Wahl, R., Paschold, J.C., Avrill, N., Einhorn, L.H., Suh, W.W., Samson’O, D., Delbekell, D., Gorman, M., Shields, A.F., 2008. Recommendations on the use of F-18-FDG PET in oncology. J. Nucl. Med. 49, 480–508. Food and Drug Administration, Guidance for Industry: Process Validation: General Principles and Practices, 〈http://www.fda.gov/downloads/drugs/guidancecom plianceregulatoryinformation/guidances/ucm070336.pdf〉, Rockville, MD; 2011 (Accessed Dec. 2013). Hamacher, K., Coenen, H.H., Stocklin, G., 1986. Efficient stereospecific synthesis of no-carrier-added 2-[18F]-fluoro-2-deoxy-D-glucose using aminopolyether supported nucleophilic substitution. J. Nucl. Med. 27, 235–238.
175
Hoh, C.K., 2007. Clinical use of FDG PET. Nucl. Med. Biol. 34, 737–742. Pant, G.S., Senthamizhchelvan, S., 2007. Initial experience with an 11 MeV selfshielded medical cyclotron on operation and radiation safety. J. Med. Phys. 32, 118–123. Pascali, G., Watts, P., Salvadori, P.A., 2013. Microfluidics in radiopharmaceutical chemistry. Nucl. Med. Biol. 40, 776–787. Ratib, O., Beyer, T., 2011. Whole-body hybrid PET/MRI: ready for clinical use? Eur. J. Nucl. Med. Mol. Imaging 38, 992–995. Townsend, D.W., 2008. Positron emission tomography/computed tomography. Semin. Nucl. Med. 38, 152–166. U.S. Pharmacopeial Convention, U.S. Pharmacopeia 34th ed., The National Formulary 29th ed., 2011. U.S. Pharmacopeial Convention, U.S. Pharmacopeia 36th ed., The National Formulary 31st ed., General Chapter o 71 4Sterility Tests, 2013.
Contents lists available at ScienceDirect
Applied Radiation and Isotopes journal homepage: www.elsevier.com/locate/apradiso
A “dose on demand” Biomarker Generator for automated production of [18F]F and [18F]FDG V. Awasthi a,n, J. Watson b, H. Gali a, G. Matlock a, A. McFarland b, J. Bailey b, A. Anzellotti b a b
Department of Pharmaceutical Sciences, College of Pharmacy, The University of Oklahoma Health Sciences Center, Oklahoma City, OK 73117, USA ABT Molecular Imaging, 3024 Topside Business Park Drive, Louisville, TN 37777, USA
H I G H L I G H T S
Biomarker Generator (7.5 MeV cyclotron) for “dose on demand” production of PET biomarkers. Fully automated FDG production and quality control, requiring minimum operator input. Ion source comprising of a detachable anode, enabling simple swap and limiting maintenance issues. Simple and easy to replace internal target.
art ic l e i nf o
a b s t r a c t
Article history: Received 1 October 2013 Received in revised form 30 January 2014 Accepted 16 February 2014 Available online 6 March 2014
The University of Oklahoma—College of Pharmacy has installed the first Biomarker Generator (BG75) comprising a self-shielded 7.5-MeV proton beam positive ion cyclotron and an aseptic automated chemistry production and quality control module for production of [18F]F and clinical [18F]FDG. Performance, reliability, and safety of the system for the production of “dose on demand” were tested over several months. No-carrier-added [18F]F was obtained through the 18O(p,n)18F nuclear reaction by irradiation (20–40 min) of a 4 95% enriched [18O]H2O target (280 μl) with a 7.5-MeV proton beam (3.5– 5.0 μA). Automated quality control tests were performed on each dose. The HPLC-based analytical methods were validated against USP methods of quality control. [18F]FDG produced by BG75 was tested in a mouse tumor model implanted with H441 human lung adenocarcinoma cells. After initial installment and optimization, the [18F]F production has been consistent since March 2011 with a maximum production of 400 to 450 mCi in a day. The average yield is 0.61 mCi/min and 0.92 mCi/min at 3.8 mA and 5 mA, respectively. The current target window has held up for over 25 weeks against 4400 bombardment cycles. [18F]FDG production has been consistent since June 2012 with an average of six doses/day in an automated synthesis mode (RCY E50%). The release criteria included USP-specified limits for pH, residual solvents (acetonitrile/ethanol), kryptofix, radiochemical purity/identity, and filter integrity test. The entire automated operation generated minimal radiation exposure hazard to the operator and environment. As expected, [18F]FDG produced by BG75 was found to delineate tumor volume in a mouse model of xenograft tumor. In summary, production and quality control of “[18F]FDG dose on demand” have been accomplished in an automated and safe manner by the first Biomarker Generator. The implementation of a cGMP quality system is under way towards the ANDA submission and first clinical use of [18F]FDG produced by BG75. & 2014 Elsevier Ltd. All rights reserved.
Keywords: Cyclotron FDG PET
1. Introduction The use of positron emission tomography (PET) and 2-deoxy2-[18F]fluoro-D-glucose ([18F]FDG) has gradually transformed the
n
Corresponding author. E-mail address: [email protected] (V. Awasthi).
http://dx.doi.org/10.1016/j.apradiso.2014.02.015 0969-8043 & 2014 Elsevier Ltd. All rights reserved.
clinical diagnostics landscape since coming into existence more than 30 years ago. Starting primarily as a research tool, PET imaging with [18F]FDG has now become an accepted, needed, and reimbursable clinical procedure for oncology and other medical indications (Bar-Shalom et al., 2003; Hoh, 2007; Bohnen et al., 2012; Fletcher et al., 2008; U.S. Pharmacopeial Convention, 2011). Currently hybrid technologies such as PET/CT or PET/MRI are poised to keep expanding the potential of PET (Townsend, 2008; Ratib and Beyer,
168
V. Awasthi et al. / Applied Radiation and Isotopes 89 (2014) 167–175
2011). PET imaging centers are capital intensive and are dependent on the regular and predictable supply of imaging radiopharmaceuticals from independent sources. Since [18F]FDG production itself requires a cost-prohibitive infrastructure consisting of medical cyclotrons and accompanying radiosynthesis, quality control, and dispensing units, much of the production is delegated to specialized radiopharmaceutical vendors, such as IBA, PETNet, Cardinal Health, etc. It may not be out of place to cite a lack of access to radiopharmaceuticals as one of the reasons for impeded growth of PET centers in remote regions, and patient migration into the busy citybased imaging centers is routine. The current paradigm for [18F]FDG production and supply is largely based on strategically located commercial sites, producing, and distributing clinical quality product. However given the relatively short half-life of the fluorine-18 (110 min), consistent access to [18F]FDG in hospitals or imaging centers out of the distribution grid is very challenging. Such problems are magnified and are of immense importance for health care delivery in developing and underdeveloped countries. Considering the limitations posed by currently accepted network of commercial manufacturers-to-PET center distribution channel, a need exists of enabling PET centers to fulfill their own requirement of [18F]FDG and other [18F]-labeled radiopharmaceuticals in an independent, cost-effective, safe, and quality conscious manner. In this regard, Biomarker Generator (BG75), developed by ABT Molecular Imaging, is capable of providing [18F]FDG doses ondemand. The dose on demand concept in relation to radiochemistry on chip has been recently reported by Pascali's group (Arima et al., 2013; Pascali et al., 2013). However, integrating the radiochemistry with a relatively low energy cyclotron (7.5 MeV) for clinical production of imaging biomarkers is a novelty described in this paper and has been in development for over 5 years. The College of Pharmacy in the University of Oklahoma Health Sciences Center (OUHSC) recently installed a BG75 to supplement their SPECT Nuclear Pharmacy business and to service pre-clinical demands for PET imaging in its Research Imaging Facility. As the first site to install and commission the BG75, in this report we share the installation requirements, performance, reliability, and safety of BG75 for production of [18F]F and [18F]FDG. The data reported were collected over a period of several months after the successful commissioning of the cyclotron.
2. Materials and methods 2.1. General All chemicals were obtained commercially and were used without further purification. [18O]H2O was obtained from Rotem Industries Ltd. (Israel). All chemicals used for [18F]FDG production were obtained from ABX GmbH (Radeberg, Germany). The [18F] FDG Dose Synthesis Cards (DSCs) and kits were obtained from ABT Molecular Imaging (Knoxville, TN). The synthesis kit consisted of vials containing acetonitrile, phase transfer catalyst (PTC, Kryptofix 2.2.2 in acetonitrile), 2 M HCl, and water for injection. Each kit was meant to last for approximately eight [18F]FDG synthesis runs. 2.2. Biomarker Generator (BG75) BG75 (ABT Molecular Imaging, Knoxville, TN) comprises a selfshielded 7.5-MeV proton beam, positive ion cyclotron and an aseptic single-use card-based automated chemistry production and quality control module (Fig. 1). An internal beam current of r5 mA can be achieved for [18F]F production, using three nonsimultaneous internal target ports. It has a pole diameter of 74.8 mm, an extraction radius of 35 cm, and four dees operating
at a voltage of 16 kV max and frequency of 72 MHz. A magnet of mass 3.5 t generates the average magnetic field of 1.2 T (1.8 Tmax). The entire cyclotron of height 0.37 m and diameter 1.25 m is enclosed in a high efficiency shielding. The two-part (upper and lower) self-shielding of BG75 consists of a 0.64 cm steel exterior enclosing high density boronated concrete and polyethylene. The physical dimensions of the 21 t shielding are: 2.39 m diameter 1.63 m height. The enclosed cyclotron is accessed for servicing by mechanically parting the upper shield from the lower stationary shield using three permanently installed 1 HP screw drive motors. The cyclotron room exhaust was passed through an active dedicated charcoal filter (life span of 3–5 years). A continuous stack monitoring at the outlet of the College of Pharmacy building exhaust system is performed using a Ludlum Model 375-20 radiation monitoring system. The system consists of a Ludlum Model 375 digital controller coupled to two shielded 5.1 cm diameter by 5.1 cm thick NaI(Tl) scintillation detectors enclosed in weather-proof enclosures. The controller is mounted on a wall inside the cyclotron room, about 15 m away from the overhead exhaust. In addition, the institutional radiation safety office monitors the level of gamma and neutron radiations emanating from the cyclotron by monitoring radiation exposure on four walls of the cyclotron room on monthly basis. The gamma dose levels and the secondary neutrons around the cyclotron shield in the cyclotron room were also monitored with a calibrated Ludlum gamma/neutron survey meter Model 2363 with the Model 42-41L PRESCILA neutron detector (Ludlum Measurements, Inc., Sweetwater, TX). 2.3. [18F]F and [18F]FDG production No-carrier-added [18F]F was produced through the 18O(p,n)18F nuclear reaction by irradiation (20–40 min) of a 495% enriched [18O]H2O target (280 μl) with a 7.5-MeV proton beam (3.5–5.0 μA). The bombardment time (in μA min) was a selectable parameter. The process was initiated by selecting the end-product ([18F]F and [18F] FDG) and pushing a start button on an all-in-one computer touch screen. If producing [18F]F , the process was aborted after bombardment and the product was dispensed in the user-selectable container. For [18F]FDG production, the built-in program guided the operator to take further steps to prepare the DSC and chemistry production and quality control module while the target bombardment was in progress. Each run of [18F]FDG production used new sterile-packed DSC. The DSC was aseptically placed in the Card Chemistry System (CCS) of the chemistry production and quality control module (Fig. 2) and made ready by following the menudriven instructions. The preparation of CCS included the installation of a synthesis kit. Once the bombardment was done, the [18F]F containing [18O] water was transferred to a reaction vial contained in the DSC (Fig. 2b). [18F]FDG was synthesized using nucleophilic substitution of 1,3,4,6-tetra-O-acetyl-2-O-trifluoro-methanesulfonyl-β-D-mannopyranose (mannose triflate precursor) followed by acid hydrolysis as reported by Hamacher et al. (1986). However, the reaction conditions were optimized for our system. The [18F] labeling was achieved in acetonitrile at 80 1C for 3 min. The hydrolysis was achieved by using 2 M hydrochloric acid at 110 1C for 5 min. The product was purified by solid phase extraction using a cartridge containing layers of strong cation/anion exchange, neutral alumina, and C18 resins. The product was then passed through a 0.2 mm filter and delivered into a sterile syringe or a vial. 2.4. [18F]FDG quality control As part of the automated process, the quality control of each batch of [18F]FDG was performed without any user intervention. An aliquot of the final product was sampled for quality control in
V. Awasthi et al. / Applied Radiation and Isotopes 89 (2014) 167–175
169
Fig. 1. Biomarker Generator and its constituents. (a) BG75 installed at the OUHSC—College of Pharmacy campus. (b) The 7.5 MeV cyclotron inside the berite shield. (c) The target showing the replaceable insert (arrow). (d) A detachable anode of the ion source. A ruler (cm) is shown against the anode for comparison.
Fig. 2. The components of Chemistry Cart. (a) The pre-packed synthesis kit is supplied with various venting and delivery needles, and vials with acetonitrile, water for injection, 2 M HCl, phase transfer catalyst, and precursor mannose triflate. An additional vial with orange-colored ion exchange resin is also provided for connecting the vent tubing of the reaction vial. Eight sterile-packed Dose Synthesis Cards (DSC) come packaged with the kit. Overall, the entire kit is designed to enable eight production runs of [18F]FDG. (b) A DSC showing a reaction vial and the collection syringe connected via a two-way stop cock. (c) Reagent Metering System (RMS) carrying various vials is shown. (d) The Card Chemistry System (CCS) showing the bank where DSC is placed and secured by a lever and gas pressure. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)
170
V. Awasthi et al. / Applied Radiation and Isotopes 89 (2014) 167–175
the quality control module (QCM), which consisted of a pH microelectrode, refractive index, UV–vis, and radioactivity detectors. Ion exclusion high pressure liquid chromatography (HPLC) analyses were performed on a system equipped with a Knauer Smartline Pump 100, Knauer Smartline Degasser, Rheodyne TitanHT™ High Pressure Injector, Phenomenex ThermaSphere™ TS-130 HPLC Column Heater, Knauer Smartline UV Detector 2500, and a Knauer Smartline RI Detector 2300. A Rezex™ ROA-Organic Acid Hþ (8%) LC column (Phenomenex, Torrance, CA, 8 mm, 4.6 mm 250 mm) was used with HPLC grade water as mobile phase at a flow rate of 0.3 ml/min. The column temperature was set at 76 1C and the UV detector was set at 210 nm. The following release tests were performed in the QCM in an automated fashion: pH, residual solvents (acetonitrile and ethanol), radiochemical identity/purity, Kryptofix 2.2.2 (EP), and filter integrity. A diluted sample was separately provided for endotoxin test. The remaining release tests, i.e. visual inspection, Kryptofix (USP), 2-chlorodeoxyglucose, radionuclidic identity/purity, sterility, and endotoxin were performed according to the USP/FDA guidelines (U.S. Pharmacopeial Convention, 2011). For validation purposes, additional tests were performed for residual solvents using a Model 8610C Capillary FID GC system (SRI Instruments, Torrance, CA), pH determination using pH strips, and Kryptofix 2.2.2. using a spot test (Celltech Srl, Italy). The current production schedule enables all quality control tests to be performed in an automated fashion except endotoxin, sterility, and particle visualization. Since sterility takes 14 days of incubation, we release the product for pre-clinical work after endotoxin (Endosafe-PTS, Charles Liver Lab, Horsham, PA) and visual inspection for particulate material.
2.5. Animal studies The animal study was conducted in accordance with the protocols approved by the OUHSC institutional animal care and use committee. Human lung adenocarcinoma H441 cells were obtained from American Type Culture Collection (Manassas, VA). The cells were cultured in ATCC-formulated RPMI-1640 Medium supplemented with 10% fetal calf serum (Invitrogen Corporation, Carlsbad, CA) and 50 μg/ml gentamicin. Male athymic nu/nu mice (4 weeks old) were obtained from the breeding colony of Harlan Laboratories (Indianapolis, IN). Before initiating the experiment, we let the mice acclimatize for at least 5 days. For xenograft implantation, H441 cells were harvested, washed in serum-free medium and re-suspended in 0.1 M phosphate-buffered saline (PBS, pH 7.2). A single cell suspension with 490% viability was used for subcutaneous injections in mice (2 106 per implant). Two to three weeks after inoculation of the tumor cells, when the tumors reached 0.5–1 cm in diameter, the mice were recruited for PET/CT imaging studies. The small animal PET/CT was conducted in the OUHSC College of Pharmacy Research Imaging Facility using a Flex™ X-O™ X-PET™ Pre-Clinical Platform (Gamma Medica, Northridge, CA). A H441 tumor-bearing mouse was anesthetized initially using 2% isoflurane in oxygen at 2 L/min, in a polypropylene induction chamber. When fully anesthetized, the mouse was placed on the scanner bed and anesthesia was maintained through a nose cone with 1% isoflurane in oxygen at 2 L/min. Body temperature was maintained at 37 71 1C by using a water-circulated pad under the animal. A dose of 70 mCi of [18F]FDG in 100 ml normal saline was injected into the tail vein. Static small animal PET imaging data was acquired for a period of 10 min after 2 h post-injection. An X-ray computed tomography (CT) was also acquired for 1 min to enable anatomic localization of radioactivity. The PET data was reconstructed using a 2D filtered back-projection algorithm. The
PET and CT images were fused together and visualized using Amira 3.1 software (Visage Imaging Inc., San Diego, California, USA).
3. Results 3.1. Installation The installation of BG75 was preceded by a site planning guide and radiation safety guidance provided by the vendor. Space specifications for the BG75 were 4.6 m 4.6 m which resulted in a good fit with the available space in the basement of the OUHSC-College of Pharmacy. No additional construction was needed for the installation, except for floor reinforcement to specifications and minor renovations. The installation in an existing basement was a bit challenging because the heavy equipment and cyclotron parts could not be transported in the elevators. The only walkout from the basement was through an approximately 100 m underpass with a couple of 901 turns and three doors. The cyclotron and shields were taken through the underpass from the loading area to the cyclotron room. The main connections of electricity and water for the cyclotron, CCS and QCM were established to power the system and start the installation and operation qualification of the BG75. The entire rigging operation and physical installation of the system took approximately one week. The operation of the cyclotron was performed under a permit from Oklahoma Department of Environmental Quality issued to the OUHSC —College of Pharmacy. The vendor supplied berite concrete as the shielding material. As shown in Fig. 3, the self-shielding system of the BG75 reduced the gamma and neutron fields generated during operation of the cyclotron well within the permissible exposure rate. The gamma dose levels and the secondary neutrons around the cyclotron shield in the cyclotron room were monitored with a calibrated Ludlum gamma/neutron survey meter Model 2363 with the Model 42-41L PRESCILA neutron detector (Ludlum Measurements, Inc., Sweetwater, TX). The continuous stack monitoring of exhaust provided no evidence of radiation exceeding background levels. 3.2. [18F]Fluoride production The performance of the cyclotron was evaluated for [18F]F production and compared to the indicated specification of 1 mCi/ min. In order to accommodate initial optimization processes, we have
Fig. 3. The radiation profile for BG75 in operation. Exposure rate of less than 1 mR/h at 1 m from the shield surface enables general occupancy outside of the Biomarker Generator room and limits occupational exposure to the operator within the permissible levels. The measurements are shown in mR/h.
V. Awasthi et al. / Applied Radiation and Isotopes 89 (2014) 167–175
171
The ion source power supply usually runs at 49.1 70.8 mA and 1382.34 78.8 V. 3.4. [18F]FDG synthesis
Fig. 4. The [18F]F output with respect to the bombardment time in the BG75. The data for all BG75 runs at all levels of integrated-ion current levels (mA) as a function of bombardment time, regardless of downtime, daily run number, “crowbar frequency” (loss of cyclotron synchronization), variations in different runs, and target unloading times. These data are from a long period of time using a number of different ion sources and targets since installation. The average yield at 3.8 mA was 0.61 mCi/min (n ¼186), whereas approximately 0.92 mCi/min was obtained at 5.0 mA (n ¼202).
presented the data at two bombardment currents of 3.8 mA and 5.0 mA (Fig. 4). We found that bombardment at 5.0 mA provided approximately 0.92 mCi/min. Rinsing of target with additional 100 ml of enriched water could increase the yield by approximately 10%, but this practice was not followed on routine basis.
3.3. Target rebuilding and ion source replacement The single most dreaded task for a cyclotron operator is the target rebuilding. The process of target rebuilding in BG75 is much easier than that in the conventional cyclotrons. The internal target consisting of a grid-less terminal window of 20 mm Haver foil (Fig. 1c) is simple to rebuild and replace. The target is essentially a disposable rectangular clamp with a thin silicone gasket held by four screws, carrying maximum 300 ml of liquid under approximately 320 psi argon. The target body material is a non-magnetic undisclosed alloy cooled by circulating water. The placement of target inside the cyclotron is optimized for maximum beam extraction by rotational and vertical displacements. In our experience, the target lasts more than 775 mA h, which is comparable to that reported for targets in the conventional medical cyclotrons (Pant and Senthamizhchelvan, 2007). The simplicity of target design is extended to the construction of ion source as well. The ion source supports the ion beam by providing a stable supply of ions to the accelerator. Typically, the ion sources in medical cyclotrons are characterized by beam intensity, stability (both short and long terms), reproducibility, simplicity of operation, convenience of maintenance, reliability, and safety. Harnessing optimal performance from an ion source at a reasonable cost is a continuous challenge. BG75 has an innovative ion source design (Fig. 1d) that allows simple swapping of the detachable anode. It requires no specialized expertise and significantly limits the maintenance demands. The main body of the ion source carries H2 at a typical vacuum reading of 3 10 4 Torr.
The CCS unit present in the production module is in-charge of producing clinical grade [18F]FDG. The reagents are placed on top of the CCS in a unit called the Reagent Metering System (RMS) which includes a special color-coded vial support and three metering pumps paired with corresponding distribution heads (Fig. 2c). The vial support houses the vials in receptacles with a unique color/size combination which minimizes the possibility of vial misplacement by the user, as was noted during its routine use. The three metering pumps were tailored for deliveries of: (i) [18O] H2O to the cyclotron target, (ii) organic reagents to the DSC, and (iii) aqueous reagents to the DSC. The pumps were connected to the vials via sterile needles and tube lines. The DSC is the sterile and disposable unit (Fig. 2b) in which [18F]FDG manufacturing takes place. It comes double-bagged and is installed in the CCS following aseptic techniques. It is a closed system and contains the reaction vial and purification column required for [18F]FDG manufacturing. Unlike other commonly used synthesis systems, it does not employ trap-and-release of [18F]F . The water is evaporated by producing an azeotrope with acetonitrile, and the labeling is performed via a nucleophilic substitution on the mannose triflate precursor at 110 1C. This labeling step is followed by acid hydrolysis of the acetyl groups using hydrochloric acid producing the unpurified [18F]FDG. The solution is then passed through a purification column and a 0.2 mm syringe filter, to finish in a sterile syringe shielded in a tungsten shield. Overall, the average dose-to-dose workflow was 40 min, producing 12–15 mCi [18F]FDG/dose in 3 ml volume (Fig. 5a). However, the first dose of day takes approximately 1 h. The sterility and bioburden evaluations have been performed for this workflow using a third party laboratory. It passed sterility requirements according to USP general chapter /71S (U.S. Pharmacopeial Convention, 2013); less than 4 CFU/sample were recovered in total bioburden panel tests. 3.5. [18F]FDG quality control The QCM automatically performs pH determination, residual solvents (acetonitrile and ethanol), radiochemical identity/purity, Kryptofix 2.2.2. determination, and filter integrity test. The system is HPLC-based and is prepared for analysis concurrent to the ongoing target bombardment and [18F]FDG synthesis. First, the QCM primes the delivery lines and monitors the pressure in the HPLC pump. After the priming, the system executes a System Suitability Test (SST) where the pH detector is verified and the performance of the HPLC system is assessed in terms of three criteria: 1) tailing factor of the acetonitrile peak, 2) resolution factor between the acetonitrile and ethanol peaks, and 3) the value reported for the concentration of a kryptofix 2.2.2./acetonitrile/ ethanol standard. After these initial checks, the QCM is ready for the analysis of the [18F]FDG batch. The system automatically withdraws a sample (180 ml) from the DSC and performs pH determination via a microelectrode with a tolerance within 0.3 pH units. The HPLCbased system performs Kryptofix 2.2.2. and residual solvent determination plus radiochemical purity/identity. The analytical methods on the QCM have been previously validated according to specifications set by the International Conference on Harmonization in terms of accuracy, specificity, linearity, repeatability, range, robustness and detection/quantification limits (Anzellotti et al., 2013; Food and Drug Administration, 2013). A typical QC report generated for [18F]FDG is shown in Fig. 6.
172
V. Awasthi et al. / Applied Radiation and Isotopes 89 (2014) 167–175
Fig. 5. (a) A representative workflow to produce an [18F]FDG dose. (b) A representative QC chromatogram of a [18F]FDG dose manufactured with the BG75. A typical run time is 15 min.
3.6. Problems and troubleshooting BG75 being the first cyclotron of its kind installed for medical [18F]F production, the documentation and analysis of its operating data are a continuous process necessary to predict yields, identify inefficiencies, track progresses from frequent upgrades, and identify opportunities for further improvements. Over the last one year of BG75 operation, there were several instances when the intended production could not be accomplished. The failures, their frequencies, and the possible causes are enlisted in Table 1. The most frequent occurrence of target failure was an issue early during the optimization process, when the beam current was being stabilized. Since then, in the last 6–9 months of operation, we have not observed unanticipated target failure. Empirically, we have observed that the [18F]F production yield starts to gradually decrease as the target window approaches its end of life. A significant increase in pressure reading (i.e. vacuum leakage) was the typical sign of target failure. We believe that stable beam, courtesy to the recent RF coil and ion source upgrades, and the optimized beam current of 3.8 mA have a lot to do with the current success. The second critical problem was that of ion source failure, mostly attributed to the shorting of anode and cathode or the erosion of cathode. The novel design upgrade of screw-on, smallsized ion source has simplified the replacement of ion source parts. As far as the synthesis of [18F]FDG is concerned, the foremost concern was the malfunction of DSCs. The intake ports in a DSC are abutted against the solvent delivery ports in CSC by a set of mini orings to prevent leakage. Occasionally, the loss of o-rings, or their skewed placement, results in leakage and loss of reagents. The leakage may also occur if the silicone delivery lines inside the DSC become loose or come off under pressure. Whereas the placement of o-rings could be ensured before placing the DSC card in the intended groove, the failure of silicone delivery lines is impossible to determine without compromising the sealed nature of the DSCs. To address this concern, the current version of production program
pre-tests the integrity of DSC card under argon pressure. Since the CCS is completely shielded with 2.5 cm thick lead, any accidental leaks do not expose the outside environment or the operator. 3.7. Animal study In order to biologically validate the diagnostic efficacy of [18F] FDG produced by the novel automated process, we tested [18F]FDG in a mouse model of xenografted H441 lung adenocarcinoma. [18F] FDG is a biomarker of the metabolic turnover in tumor tissue. Fig. 7 shows the coronal, sagittal and transaxial views of a representative mouse carrying tumor in the right hind limb. Clearly, the tumor accumulated large amounts of [18F]FDG and the distribution in the other parts of the body was also similar to the one reported for [18F] FDG produced by other devices.
4. Discussion Installation of compact medical cyclotrons to produce [18F]F and other positron-emitting radionuclides is rapidly increasing worldwide due to the increased utilization of PET imaging in clinical practice. In this regard, a BG75 was installed at OUHSC to support both clinical and preclinical requirements of [18F]-radiopharmaceuticals. BG75 consists of a 7.5 MeV positive ion cyclotron with an internal target configuration. The positive ion beam, coupled with an internal target, enables BG75 to exist as an ultra-compact cyclotron. It is a self-shielded machine that could be operated without any concerns of undue radiation exposure to the operator. As described above, the BG75 work-flow, from production to quality control is completely automated, and requires minimum manual input. After initial placement of reagents, DSC, and collection vessel, the entire operation (consisting of bombardment and followed by FDG synthesis) starts with a push of a button and ends with an appropriately shielded dose of [18F]FDG. The process could be
V. Awasthi et al. / Applied Radiation and Isotopes 89 (2014) 167–175
173
Fig. 6. A quality control sheet is automatically printed for each [18F]FDG dose to enable the operator to objectively release or restrain to [18F]FDG dose. A representative QC sheet is shown.
Table 1 Problems encountered during installation and optimization of BG75 in the period between Jan. 2012 and Aug. 2013. Problem
Probable cause
Target failure
Bad Havar window, normal wear and tear, no cooling water flow, deliberate tests at high ion source current, low target load volume Erratic ion source operation, bad upper or lower cathode, defective high voltage ion source cable
Ion source failure
DSC-related failures (low or FDG yield below 20% for no apparent reason) RF generator malfunction
Mass (argon) flow controller breakdown Power failure
Frequency in 610 operations (2012–8/2013) 6
System error message, high vacuum indication, un-load volume zero; requires target rebuild
7
System error message, low voltage indication with high current, unstable voltage, large spike in target current; requires replacement of detachable anode DSC does not pass validation test, low pressure or mass flow controller reading, erratic reaction signal patterns, visible leakage, bad filter RF failed to resonate message, only plate volts indicated on RF controller
57 Defective ports, defective junctions, DSC pressure plate failure, defective argon line fitting, reagent concentration problems, syringe failure RF component (triode), RF amplifier, RF controller, circuit 5 breaker, cable, “D tips” touching ‘Hill sides’, bad high voltage capacitor 3 Defective unit, power surge, intermittent flow controller malfunctions, bad pressure controller, partially blocked or defective fittings Severe thunderstorms, unscheduled building 3 maintenance
repeated depending on the patient schedule and the need of subsequent doses, potentially enhancing the cost-effectiveness of imaging services. In addition, compared to the regular smallest 10 MeV
Signs/troubleshooting
Low mass flow readout and erratic graph signals; poor FDG yields Low vacuum reading, hardware/software errors
medical cyclotron, BG75 provides greater flexibility in terms of footprint, resources, and personnel needed for the entire manufacturing operation. Furthermore, BG75 would be highly suitable for
174
V. Awasthi et al. / Applied Radiation and Isotopes 89 (2014) 167–175
Fig. 7. A representative set of PET/CT images of a mouse with lung adenocarcinoma H441 cell xenograft. The mouse was injected with BG75-generated [18F]FDG and imaged at 4 h after injection. The PET/CT images were acquired using the Flex-CT/ PET system (Gamma Medica Ideas, CA) and rendered with Amria 3.1 software. Arrows show uptake of FDG in implanted tumor.
the production of low-volume [18F]-labeled radiopharmaceuticals (e.g. [18F]FMISO) that are not routinely prescribed and, therefore, are costly to procure. Due to the ultra-compact nature of the BG75, we were able to install it in an existing room of 5.5 m 5.5 m with minimal site modifications. In addition, the cyclotron, chemistry production, and quality control module, and other ancillary instrumentation were accommodated in the same room. Since installation and optimization of the operating conditions, we have been able to obtain [18F]F production yield of 0.61 mCi/min at 3.8 mA current. However, the [18F]F production has been consistent since March 2011 with a maximum production of 400 to 450 mCi in one single day (E1 mCi/3.5 mA min). Apparently, the frequency of use affects the yield, and our observations suggest that the yield increases with more frequent bombardment. During [18F]F production runs, the parameters such as [18O]H2O enrichment (495% 18O), target volume (280 ml), and beam energy (7 MeV) were kept constant while beam current and bombardment time were changed. As shown in Fig. 4, the [18F]F production yields increased almost linearly with increased bombardment. Significantly, the current target window has held up for over 25 weeks against 4400 bombardment cycles. The radiation safety requirements of a cyclotron facility are more stringent than a traditional nuclear pharmacy due to highly penetrating nature of 511 keV gamma photons, higher specific gamma ray dose constant of positron emitters, and secondary neutrons emitted from the target during a production run. In this regard, the BG75 presents several advantages over other compact medical cyclotrons mainly due to the small scale production of [18F]F and a single dose of a radiopharmaceutical per production run. In other words, much lower amount of 18F activity is produced compared to other cyclotrons ( 50 mCi versus 1–4 Ci). The
radiation shield of this cyclotron is composed of dense concrete and polyethylene (30.5 cm thick) which provides acceptable dose levels during the operation. The dose rates due to secondary neutrons always remained well within the permissible limits around the shield during the operation for the maximum beam current in use (5 mA) and for the longest duration of production run (1 h), indicating that the shield adequately minimizes the exposure from neutrons around the cyclotron. The intensity of the gamma rays and secondary neutron field produced by a cyclotron primarily depend on the beam strength and since we used o5 mA beam current in all production runs those intensities were minimal. The radiation dose levels were minimal (o 1 mR/h) around the chemistry production module due to 2.5 cm thick lead shielding. In addition to the personnel safety, it is essential to monitor the exhaust system for environmental safety, which was achieved by a continuous stack monitoring system installed at the outlet of the building exhaust system. The entire operation was deemed satisfactory from the radiation safety viewpoint. The [18F]FDG production was achieved by using the chemistry production module (Fig. 1a), which is located next to the cyclotron. This module is self-shielded, installed in a sterile environment, and thus, requires no separate hot cell. The module is equipped with a laminar flow hood to provide a low particulate environment during the production process. As the module uses prepackaged and certified DSC and reagent kit (Fig. 2), the [18F]FDG production process can be achieved under full cGMP compliance. [18F]FDG production has been consistent since June. 2012 with the maximum capacity of six doses/day in automated synthesis mode (RCYE 50%). Except for the first dose of the day, which takes approximately 60 min, a typical [18F]FDG production run takes about 40 min (Fig. 5). The synthesis conditions were optimized to obtain a desired [18F]FDG dose of 10–13 mCi/run in a final volume of 3 ml. The release criteria included USP-specific limits for pH, residual solvents, Kryptofix, radiochemical purity/identity and filter integrity test.
5. Conclusions Installation of the first ABT Biomarker Generator system, BG75, at OUHSC was successfully achieved with minimum site modifications. The cyclotron operation parameters were optimized to achieve a [18F]F production yield of 0.61 mCi/min at 3.8 mA. Since this was the first installation of BG75 in the world, some teething problems were expected. However, once those problems are overcome, BG75 is expected to reliably produce “[18F]FDG dose on demand”. The implementation of a cGMP quality system is under way towards the first ANDA submission and clinical use of BG75-produced [18F]FDG in the USA.
Acknowledgments We gratefully acknowledge the continuous support of the University of Oklahoma, College of Pharmacy and its administrative staff. References Anzellotti, A.I., McFarland, A.R., Ferguson, D., Olson, K.F., 2013. Towards the full automation of QC release tests for [18F]fluoride-based radiotracers. Curr. Org. Chem. 17, 2153–2158. Arima, V., Pascali, G., Lade, O., Kretschmer, H.R., Bernsdorf, I., Hammond, V., Watts, P., De Leonardis, F., Tarn, M.D., Pamme, N., Cvetkovic, B.Z., Dittrich, P.S., Vasovic, N., Duane, R., Jaksic, A., Zacheo, A., Zizzari, A., Marra, L., Perrone, E., Salvadori, P. A., Rinaldi, R., 2013. Radiochemistry on chip: towards dose-on-demand synthesis of PET radiopharmaceuticals. Lab Chip 13, 2328–2336.
V. Awasthi et al. / Applied Radiation and Isotopes 89 (2014) 167–175
Bar-Shalom, R., Yefremov, N., Guralnik, L., Gaitini, D., Frenkel, A., Kuten, A., Altman, H., Keidar, Z., Israel, O., 2003. Clinical performance of PET/CT in evaluation of cancer: additional value for diagnostic imaging and patient management. J. Nucl. Med. 44, 1200–1209. Bohnen, N.I., Djang, D.S., Herholz, K., Anzai, Y., Minoshima, S., 2012. Effectiveness and safety of 18F-FDG PET in the evaluation of dementia: a review of the recent literature. J. Nucl. Med. 53, 59–71. Fletcher, J.W., Djulbegovic, B., Soares, H.P., Siegel, B.A., Lowe, V.J., Lyman, G.H., Coleman, R.E., Wahl, R., Paschold, J.C., Avrill, N., Einhorn, L.H., Suh, W.W., Samson’O, D., Delbekell, D., Gorman, M., Shields, A.F., 2008. Recommendations on the use of F-18-FDG PET in oncology. J. Nucl. Med. 49, 480–508. Food and Drug Administration, Guidance for Industry: Process Validation: General Principles and Practices, 〈http://www.fda.gov/downloads/drugs/guidancecom plianceregulatoryinformation/guidances/ucm070336.pdf〉, Rockville, MD; 2011 (Accessed Dec. 2013). Hamacher, K., Coenen, H.H., Stocklin, G., 1986. Efficient stereospecific synthesis of no-carrier-added 2-[18F]-fluoro-2-deoxy-D-glucose using aminopolyether supported nucleophilic substitution. J. Nucl. Med. 27, 235–238.
175
Hoh, C.K., 2007. Clinical use of FDG PET. Nucl. Med. Biol. 34, 737–742. Pant, G.S., Senthamizhchelvan, S., 2007. Initial experience with an 11 MeV selfshielded medical cyclotron on operation and radiation safety. J. Med. Phys. 32, 118–123. Pascali, G., Watts, P., Salvadori, P.A., 2013. Microfluidics in radiopharmaceutical chemistry. Nucl. Med. Biol. 40, 776–787. Ratib, O., Beyer, T., 2011. Whole-body hybrid PET/MRI: ready for clinical use? Eur. J. Nucl. Med. Mol. Imaging 38, 992–995. Townsend, D.W., 2008. Positron emission tomography/computed tomography. Semin. Nucl. Med. 38, 152–166. U.S. Pharmacopeial Convention, U.S. Pharmacopeia 34th ed., The National Formulary 29th ed., 2011. U.S. Pharmacopeial Convention, U.S. Pharmacopeia 36th ed., The National Formulary 31st ed., General Chapter o 71 4Sterility Tests, 2013.