Implementation of an automation method for the syntheses of 6-[18F] fluoro-meta-tyrosine and [18F] fluorodopa

Purpose: For quantitative assessment and visualization of basal ganglia integrity, we at Tzu Chi Hospital routinely prepare 6-[18F]fluoro-meta-tyrosine (FMT) and [18F]fluorodopa (FDOPA) for imaging using PET. In our radiopharmacy, we have equipped a synthesis module (GE Medical System 'TRACERlab FX-FE') for production of both tracers, and we wish to share our experience and discuss some issues relating to the use of this module for their automatic syntheses. As it appears, the GE Medical System markets aggressively this synthesis module, therefore, this brief discussion provides a good reference for many other institutes which have just installed or planed to install the exact same model. Methods: For both FDOPA and FMT syntheses, we followed the previously reported fluorodestannylation method (Appl. Rad. Isot. 51:389, 1999). Briefly, the [18F]F2 gas was led through CFC13 containing precursor at -20 degree C in the reaction vessel. Upon the completion of the fluorodestannylation reaction, the solvent was evaporated and the hydrolysis of the protective groups was done in the same vessel by HBr at 130 degree C. After the hydrolysis, the reaction mixture was adjusted for pH by NH4OH and concentrated phosphate buffer before the HPLC purification. Results: In regard to the aspects of chemical reaction and purification, this method successfully eliminates the need for the purification of the reaction intermediate and yields products with good purity. Nevertheless, to make this method successful in the routine automatic production using this specific module, the control programs for the syntheses have to be carefully crafted. Prior to our smooth routine production, several problems were encountered and they are now discussed below. (1) Dryness of [18F]F2 delivery line Because the delivery line submerging into -20 degree C solution, we have found that slight moisture (as often occurred in a humid environment like ours) was enough to cause the blockage of the delivery due to ice formation in the line. Therefore, extra effort was taken to ensure the dryness of the line prior to the synthesis. (2) Addition of the reagents. We found that the speed of the helium-driven solution addition often varied due to several reasons, such as the unexpected partial obstruction of the reaction vessel vent. At first, our control program was written as such that the duration of the addition was controlled by timer, and we suffered many times from the incomplete addition of either CFC13 solution or HBr solution. The incompletion of CFC13 addition resulted in the lowered [18F]F2 trapping efficiency, and that of HBr resulted in the incompletion of hydrolysis. The failure of total hydrolysis also caused the formation of precipitation upon neutralization of the mixture and thus disturbed the following HPLC injection. We changed the program so that the operator determined when to stop the addition by visually inspecting the reagent vessels via the video monitor that revealed the inside of the hot cell, (3) Evaporation of CFC13. The complete evaporation of the solvent is essential in preventing the salt precipitation in the following aqueous reaction. One needs to be sure that the duration of vacuum evaporation set in the program is long enough. We have experienced that five seconds too short resulted in about 1 mL of solvent left and subsequently causing problem during HPLC injection. (4) Temperature sensor. Many steps in the synthesis program were only advanced to the next depending on the correct temperature sensor activation. We have experienced occasionally the failure of the correct sensor activation resulting in the stalled synthesis sequence. When this happened, the operator intervention measure was taken to advance the synthesis. (5) Fluid detector. Fluid detector was designed in this module to activate the injection of the reaction mixture into HPLC column at the end of the mixture flow. We have encountered two types of fluid detector faults. One was the fail of detection at the end of the flow and resulting in no activation of injection. Consequently, the entire reaction mixture flowed through the injection loop and into the over-flow bottle. The other was the false detection of no flow (or the detection of air gap) and resulting in the injection of only 40% to 70% of the entire reaction mixture. Despite careful adjustment for the sensitivity of the detector, due to high salt content and color of the mixture, we had limited success in using the fluid detector. We changed the activation of the injection to timer-controlled. Because the set volume of the mixture and the needle-valve-controlled helium flow for pushing solution through the loop (unlike helium for the addition of reagents), the time for the total transfer of the reaction to the loop did not vary much, and thus, possible for using set duration for transfer. Conclusion: Using TRACERlab FX-FE module, both FDOPA and FMT were now routinely produced in our radiopharmacy. With extra pampering to the module at pre-and post-synthesis routines and the use of carefully crafted control program, this machine showed us rather well performance in terms of the production yield (44±5%, decay corrected; specific activity of 34.0±17.2 GBq/mmo). (authors)