[en] Complete text of publication follows. Dear Sir, We read with interest the [¹⁸F]FDG review, recently published in the Journal of Radioanalytical Nuclear Chemistry, relating the most important aspects of [¹⁸F]FDG production over the last 40 years [1]. However, we would like to add some comments and to mention some references which did not appear in the bibliography of this publication though our lab's work is well implemented in several synthesizers from the market used for the routine synthesis of this radiopharmaceutical. 1. As stated by the authors, Hamacher's nucleophilic approach is undoubtedly the [¹⁸F]FDG method on which most of the current automated synthesis modules are based [2]. This original [¹⁸F]FDG process was reported in 1986 by the Julich group and some modification were implemented later to improve the radiochemical yield (RCY). For example, a basic hydrolysis in liquid phase rather than with HCl has been proposed by Fuchtner et al. [3]. In 1997, our laboratory presented at the XIIth International Symposium on Radiopharmaceutical, a synthetic approach that combines the two procedures and which overcome the disadvantages of the acid and basic aqueous methods [4]. In the CRC process, the first labelling step is performed classically according to the Hamacher's method. After labelling, the crude 2-[¹⁸F]fluoro-1,3,4,6-tetra-O-acetyl-D-glucose ([¹⁸F]FTAG) is trapped on a ^tC18 solid phase extraction (SPE) cartridge. In a second step, the removal of the acetyl protecting groups of the [¹⁸F]FTAG is performed by basic hydrolysis on this solid phase support (^tC18) [4,5,6]. The main advantages of the CRC approach is that the [¹⁸F]FTAG is not eluted from the SPE but kept on it and the hydrolysis reaction conducted on this ^tC18 cartridge. This basic hydrolysis on the solid phase is faster (90s) than in solution and proceeds at room temperature, without any [¹⁸F]FDG epimerization [7, 8] what greatly facilitates the automation of the process (no second evaporation is required for solvent removal and no additional heating is required). After acidification, the very near to neutral solution eluted from the ^tC18 support allows the use of a buffer solution rather than a cation exchange or a retardation resin to afford the final isotonic solution. Therefore, the organic resins of the other methods for which microbial controls were not very easy to realize were discarded. As a result, in addition to the ^tC18 hydrolysis SPE, the other cartridges of the process are a Sep-Pak Accell Plus QMA™ (silica-based anion exchanger from Waters) for the [¹⁸F]fluoride recovery and a ^tC18/alumina for the final purification (see details in the paper [5]). All these chemistry improvements have greatly simplified the hardware and the software, decreased the duration of the synthesis and opened the way to the first GMP synthesis of [¹⁸F]FDG. This patented hydrolysis method [9] was firstly fully automated with a disposable kit assembly made of standard commercially available single-use components and the FDG synthesizer from Coincidence Technologies (in use for routine production, since September 1998, in our Cyclotron Research Centre) [5, 6, 9], the TRACERlab MX™ from GE Medical Systems and the FASTlab™ module of GE Healthcare. Up until now, this synthesis is still largely used in the world for the [¹⁸F]FDG preparation. Moreover this approach is now also implemented on other synthesizers from the market [i.e., AIO (Trasis)]. 2. We would like also to draw your attention on the following point: even if the author's claim that the presence of the [¹⁸F]FDM (2-deoxy-2-[¹⁸F]fluoro-D-mannose) has to be evaluated after basic hydrolysis, the review reports only papers where the stability of [¹⁸F]FDG towards epimerization is investigated in liquid phase. In 2001, we demonstrated that no epimerization of [¹⁸F]FDG to [¹⁸F]FDM occurs on the solid phase support under standard routine conditions [i.e. 2 min, room temperature and NaOH (2 N)] and even with NaOH (12 M) and hydrolysis duration extended up to one hour [7, 8]. 3. The table 1 of the review lists the commercial modules, grouped by type, with the synthesis duration and radiochemical yields at the end of synthesis (EOS). For information, the first commercial module with cassette and reagent kit was the Coincidence module from JL Morelle (1997) developed in collaboration with the CRC/ULg. After the acquisition of Coincidence by GE in October 2001, the FDG Coincidence module was renamed TRACERlab MX™ in mid-2002. We think that the work carried out in our laboratory, briefly summarized in this letter, and more in details in the following papers [5, 8] was at the origin of the state-of-the-art automation for the GMP synthesis of [¹⁸F]FDG. With best regards, Yours sincerely, C. Lemaire, A Luxen. (author)