Optimization of Automated Synthesis of 2-[18F]Fluoro-2-deoxy-D-glucose Involving Base Hydrolysis
ABSTRACT Synthesis of 2-[18F]fluoro-2-deoxy-D-glucose ([18F]FDG) involving base hydrolysis was optimized. Fluorine-18 was isolated from irradiated water to more than 90% by sorption of [1
8F]fluoride on QMA anion-exchange resin, which was followed by elution with a 96 : 4 (by volume) acetonitrile-water mixture containing Kryptofix 2.2.2 and potassium carbonate (molar ratio 2 : 1). This composition is the best for preparing the complex [K/K2.2.2]+
18F- used in nucleophilic fluorinations. No additional azeotropic drying is required. Base hydrolysis under optimized conditions (40-45C), followed by neutralization with HCl, removal of traces of the solvent, and purification of the final product on a combined SCX/Alumina N column, yielded [1
8F]FDG of high radiochemical (>99%) and chemical purity with minimal product loss. With an RB-86 robotic system (Anatech, Sweden), the synthesis time was 38 min. The procedure is used in the Institute of Human Brain, Russian Academy of Sciences for routine synthesis of FDG; the radiochemical yield of the product by the end of synthesis (EOS) is reproducibly high: 633% (n = 40).
- [Show abstract] [Hide abstract]
ABSTRACT: Introduction: The diagnosis of infection and the ability to distinguish bacterial infection from nonbacterial inflammation by positron emission tomography (PET) have gained interest in recent years, but still few specific radiopharmaceuticals are available for use. In this study, we developed a new radiosynthesis method of 2-deoxy-2-[18F]fluoroacetamido-D-glucopyranose ([18F]FAG) by applying microwave irradiation and demonstrated that [18F]FAG could be a potential radiopharmaceutical to distinguish bacterial infection from nonbacterial inflammation. Methods: 1,3,4,6-Tetra-O-acetyl-2-deoxy-2-bromoacetamido-D-glucopyranose was used as precursor, and labeling was performed under microwave irradiation conditions followed by alkaline hydrolysis and high-performance liquid chromatography (HPLC) purification. In vitro uptake of [18F]FAG by Escherichia coli was performed. Tissue biodistribution of [18F]FAG was performed in mice. Moreover, PET imaging acquisition of E. coli infection and nonbacterial inflammation models was performed in rats. Tissue radiotracer-accumulated sites were analyzed by hematoxylin and eosin staining and anti–E.coli immunostaining. Results: The radiosynthesis of [18F]FAG was achieved with microwave irradiation, and the radiochemical yield was 9.7%±2.8% end of bombardment (EOB); the radiochemical purity was more than 98%, and the total synthesis time was 62 min. Compared with control group, in vitro uptake of [18F]FAG by E. coli was significantly decrease in inhibition group (Pb.05). Biodistribution studies in mice showed rapid clearance of [18F]FAG from the animal body. [18F]FAG clearly visualized the infection areas but not nonbacterial inflammation areas in PET studies. Quantitative analysis revealed that the uptake of [18F]FAG into infection areas was significantly higher than that of [18F]FAG into inflammation areas (Pb.05). Histological analysis demonstrated the presence of bacterial cells at the sites of accumulation of [18F]FAG. Conclusions: Using 1,3,4,6-tetra-O-acetyl-2-deoxy-2-bromoacetamido-D-glucopyranose as a precursor, the new radiosynthesis method of [18F]FAG was achieved in fewer steps and with a shorter synthesis time than previously reported. Furthermore, [18F]FAG was able to distinguish bacterial infection from nonbacterial inflammation.Nuclear Medicine and Biology 07/2011; · 2.41 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: A novel fully automated radiosynthesis procedure for the fluorine-18 analog of 1-alpha-D-(5'-deoxy-5'-fluoro-(1S,2R,3S,4S) arabinofuranosyl)-2-nitroimidazole ([(18)F]FAZA) using a commercially available combination column - Chromabond Set V (FDG-base-hydrolysis) - for purification was developed. [(18)F]FAZA was prepared via a one-pot, two-step synthesis using a nuclear interface nucleophilic synthesis module. Nucleophilic fluorination of the precursor molecule, 1-(2,3-di-O-acetyl-5-O-tosyl-alpha-D-arabinofuranosyl)-2-nitroimidazole, with no-carrier added [(18)F]fluoride followed by hydrolysis of the protecting groups with 0.3M NaOH was performed. Purification was carried out using the Chromabond Set V column without any modifications. The overall radiochemical yield obtained was 21.98+/-1.40% (n=5, without decay correction) within a total synthesis period of 51+/-1 min. The radiochemical purity was greater than 95% and devoid of any other chemical impurities. The reported method can easily be adapted in any commercial FDG synthesis module.Applied radiation and isotopes: including data, instrumentation and methods for use in agriculture, industry and medicine 10/2010; 68(10):1944-9. · 1.09 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: The aim of this study was to develop a direct fluorination method for the preparation of [18F]-(E)-N-(3-iodoprop-2-enyl)-2β-carbofluoroethoxy-3β-(4′-methyl-phenyl)nortropane ([18F]FE-PE2I (VI). The synthesis procedure relies on the conventional Kryptofix-mediated nucleophilic 18F-substitution of the tosylate group in the precursor, TsOE-PE2I (V). Out of reaction conditions tested, the highest fluorination efficiency was obtained in dimethyl sulfoxide at 140°C. The reaction mixture was purified by semi-preparative HPLC, followed-up by a standard Sep-Pak SPE procedure. On average, 1.0 GBq of [18F]FE-PE2I was produced from 5-min irradiation at 35 μA (dimethyl sulfoxide, 5 min/140°C). Decay-uncorrected yield of the product after HPLC purification and formulation was in the order of 20%. Specific radioactivity of [18F]FE-PE2I at 15 min after EOS was 3.3–5.1 Ci/µmol (n = 3); radiochemical purity was >98% (n = 4). This direct nucleophilic fluorination strategy is well suited for the automation of the entire synthesis of [18F]FE-PE2I in a modern PET synthesizer for human PET application. In addition, the 18F-incorporation rate into TsOE-PE2I was evaluated using radio-thin layer chromatography (TLC) and radio-HPLC. The suggested HPLC method (ACE 5 C18-HL column and acetonitrile/0.1 M NH4CO2 (80:20)) was found to be suitable for evaluation of ‘free’ 18F-fluoride in the reaction mixture; in addition, this method allowed the detection of three radiolabelled by-products that were not discernable with the TLC approach. Therefore, we conclude that the HPLC approach may serve as a good alternative to traditional radio-TLC technique as it provides more detailed information about the fluorination process in the reaction kinetics or optimization studies.Journal of Labelled Compounds and Radiopharmaceuticals 05/2012; 55:206-210. · 1.19 Impact Factor
1066-3622/02/4404-0403$27.00?2002 MAIK ?Nauka/Interperiodica?
Radiochemistry, Vol. 44, No. 4, 2002, pp. 403?409. Translated from Radiokhimiya, Vol. 44, No. 4, 2002, pp. 366?372.
Original Russian Text Copyright ? 2002 by Gomzina, Vasil’ev, Krasikova.
Optimization of Automated Synthesis
of 2-[18F]Fluoro-2-deoxy-D-glucose Involving Base Hydrolysis
N. A. Gomzina, D. A. Vasil’ev, and R. N. Krasikova
Institute of Human Brain, Russian Academy of Sciences, St. Petersburg, Russia
Received February 12, 2002
Abstract?Synthesis of 2-[18F]fluoro-2-deoxy-D-glucose ([18F]FDG) involving base hydrolysis was opti-
mized. Fluorine-18 was isolated from irradiated water to more than 90% by sorption of [18F]fluoride on QMA
anion-exchange resin, which was followed by elution with a 96 : 4 (by volume) acetonitrile?water mixture
containing Kryptofix 2.2.2 and potassium carbonate (molar ratio 2 : 1). This composition is the best for pre-
paring the complex [K/K2.2.2]+18F?used in nucleophilic fluorinations. No additional azeotropic drying is
required. Base hydrolysis under optimized conditions (40?45?C), followed by neutralization with HCl, re-
moval of traces of the solvent, and purification of the final product on a combined SCX/Alumina N column,
yielded [18F]FDG of high radiochemical (>99%) and chemical purity with minimal product loss. With an
RB-86 robotic system (Anatech, Sweden), the synthesis time was 38 min. The procedure is used in the Insti-
tute of Human Brain, Russian Academy of Sciences for routine synthesis of FDG; the radiochemical yield of
the product by the end of synthesis (EOS) is reproducibly high: 63?3% (n = 40).
2-Fluoro-2-deoxy-D-glucose ([18F]FDG), a flu-
orinated analog of D-glucose labeled with positron-
emitting18F (T1/2= 110 min), is a radiopharmaceuti-
cal (RP) used in positron emission tomography (PET).
[18F]FDG is phosphorylated similarly to glucose, but,
in contrast to glucose, it does not take part in sub-
sequent metabolic stages and is accumulated in cells.
Thanks to this feature, in combination with favorable
18F half-life, PET investigations with [18F]FDG fur-
nish unique information on the regional rate of glu-
cose consumption in normal and pathological tissues.
This radiotracer, proposed in 1977 by the Brookhaven
National Laboratory, is still the most popular in PET
investigations. [18F]FDG is used in various fields of
PET diagnostics: in cardiology and oncology, in diag-
nostics of psychiatric diseases, motor system distur-
bances, and other pathologies [1, 2], and also as refer-
ence in development of new radiotracers . Today,
[18F]FDG is virtually the only PET RP delivered from
radiochemical centers to clinics that have no cyclotron
and no radiochemical laboratory. The enormous sig-
nificance of [18F]FDG for PET provides powerful
impetus to development of new synthetic approaches
and improvement of the existing synthesis procedures.
Several procedures, based on nucleophilic substi-
tution and electrophilic addition, have been proposed
for preparing [18F]FDG; for their detailed review, see
[4?6]. Nucleophilic substitution by [18F]F?of such
leaving groups as nitro, triflato, halo, tosylato, and
some other groups is the most common way for18F
labeling of organic substrates to prepare PET RPs.
Radiofluorine is produced by the nuclear reaction
18O(p,n)18F in a cyclotron target filled with18O-
enriched water. Transfer of [18F]F?from the aqueous
to organic phase is promoted by adding phase-transfer
catalysts: tetrabutylammonium salts, crown ethers,
or cryptands [7, 8]. In Hamacher’s procedure ,
which is the most widely used route to [18F]FDG,
the phase-transfer catalyst is Kryptofix 2.2.2 (K2.2.2),
and the counterion for [18F]fluoride is potassium
introduced in the form of carbonate. The resulting
reactive complex [K/K2.2.2]+18F?is a fluorinating
agent substituting the triflate group in the molecule of
D-mannopyranose (mannose triflate) (see scheme).
From intermediate 2-[18F]fluoro-1,3,4,6-tetra-O-ace-
groups are removed by acid hydrolysis at 120?130?C,
yielding [18F]FDG. Hamacher’s procedure allows
preparation of carrier-free [18F]FDG and is used in
automated modules (?black boxes?) for [18F]FDG
synthesis; in the process, the yield of the product, dis-
regarding the18F decay (by the end of synthesis,
EOS), is as high as 50%. In some PET centers, in-
cluding the Institute of Human Brain, the synthesis is
performed with an RB-86 robotic system. The yield
of [18F]FDG in the Institute of Human Brain reached
30?35% (EOS), at the synthesis time of 70?75 min
RADIOCHEMISTRYVol. 44No. 4 2002
404GOMZINA et al.
Scheme of [18F]FDG synthesis.
An important stage in improving the [18F]FDG
synthesis process was replacement of acid hydrolysis
of [18F]TAG by base hydrolysis, which can be per-
formed under milder conditions . This allowed us
to considerably simplify the synthesis process and to
reduce the loss by performing the nucleophilic fluori-
nation and hydrolysis in the same vessel, without
intermediate purification. As shown in [13, 14], the
use of this procedure in robotic systems shortened the
synthesis time and increased the [18F]FDG yield. The
base hydrolysis appeared to be especially efficient in
automated modules; the product activity was as high as
3.5 Ci (EOS yield 70%, synthesis time 25 min) .
However, in contrast to acid hydrolysis, base hy-
drolysis is very sensitive to reaction conditions (tem-
perature, time, reactant concentrations), and, if the
optimal conditions are not strictly maintained, the
product yield becomes poorly reproducible, and its
chemical and radiochemical purity is drastically im-
In this study, we optimized the automated synthesis
of [18F]FDG by base hydrolysis of [18F]TAG and
achieved a high and reproducible yield of the product,
with its quality meeting the requirements of the Rus-
sian, European, and US Pharmacopoeias. The results
were preliminarily reported at conferences [16, 17].
Equipment. MC-17 cyclotron (Scanditronix, Swe-
den); RB-86 laboratory robot (Anatech, Sweden) with
Scanditronix workstations (Sweden); CalRad 2 iso-
tope calibrator (Victoreen, the United States); hot cell
(Van Gahlen, the Netherlands); Tru Motion Mini Map
manipulators (CGA, the United States); Varian 3400
gas chromatograph (the United States); Varian 4400
integrator (the United States); and BDBSZ-1eM well
scintillation ?-counter and PSO-5 scaler (both Russia).
Chemicals and materials. Water 95% enriched in
18O (Global Scientific Technologies, Sosnovy Bor,
Leningrad oblast, Russia); QMA Accell Plus quaterna-
ry methylammonium anion-exchange resin (Waters);
LC-SCX cation-exchange resin and LC-Alumina-N
neutral alumina (Supelclean Supelco); 1,3,4,6-tetra-O-
nose (St. Petersburg State Technological Institute
[8.8.8]hexacosane (Kryptofix 2.2.2) and anhydrous
potassium carbonate (Aldrich); acetonitrile (Krio-
khrom, St. Petersburg, Russia), which was distilled
from barium hydroxide in an inert gas stream and
placed in dry vessels hermetically stoppered with
Teflon-coated lids; Millipore Millex GS sterilizing
filter with the pore diameter of 0.22 ?m.
Analysis conditions. The radiochemical purity
(RCP) of the product was determined by thin-layer
chromatography (TLC) on Sorbfil silica gel plates on
plastic support (Krasnodar, Russia), with a 95 : 5 (by
volume) acetonitrile?water mixture as eluent. The ra-
dioactivity distribution was measured with a ?-counter.
The retention factors (Rf) of [18F]F?, [18F]FDG, and
[18F]TAG were 0.05, 0.45, and 0.85, respectively.
The content of Kryptofix 2.2.2 was determined by
TLC with a 9 : 1 (by volume) mixture of methanol
and concentrated aqueous ammonia as eluent. The
TLC plate was developed with iodine vapor, and the
spot coloration intensity was compared to that ob-
tained with a reference 0.2 mg ml?1solution of Kryp-
tofix 2.2.2. Under these conditions, Rfof Kryptofix
2.2.2 was 0.30.
The acetonitrile content in the sample was deter-
mined by GLC under the following conditions: 1000?
2.5-mm column packed with Porapak Q (Serva, Swe-
den), particle size 80?100 mesh; flame ionization
detector; carrier gas helium, flow rate 40 ml min?1;
column temperature 110?C; and vaporizer temperature
200?C. The retention time of acetonitrile under these
conditions was 3.0?0.4 min.
Preparation of a column with QMA anion-
exchange resin. QMA anion-exchange resin was con-
verted to the hydrocarbonate form by treatment with
a tenfold volume of a saturated NaHCO3solutin with
stirring for 10 min. The modified resin was washed
with deionized water (Milli Q) to negative reaction for
chloride ions (test with silver nitrate) and stored under
70?80% ethanol in a refrigerator. About 0.5 ml of the
suspension (15?20 mg of dry resin) was placed in a
RADIOCHEMISTRYVol. 44 No. 42002
OPTIMIZATION OF AUTOMATED SYNTHESIS OF 2-[18F]FLUORO-2-DEOXY-D-GLUCOSE405
Table 1. Results of [18F]FDG synthesis involving acid and base hydrolysis of [18F]TAG
Series no. ? [18F]TAG ? Sorption of ? Azeotropic drying ?
(number of re-? hydrolysis ? [18F]F?on ? of the complex ?
plicate runs) ? procedure ? QMA resin ? [K/K2.2.2]+18F??
1 (n = 25) ?Acid
2 (n = 5) ??
3 (n = 10) ?Base
4 (n = 6) ??
5 (n = 15) ??
6 (n = 40) ?Base*
* Amount of mannose triflate taken 0.02 mmol (10?12 mg).
?Synthesis ? Yield of
? % (EOS)
Teflon tube 10?12 cm long, 0.33 cm i.d. The thus
prepared column was arranged, using appropriate fit-
tings, in a unit which, in turn, was arranged in the line
for delivery of irradiated water from the target. The
unit, including two three-way electric valves , was
controlled with a computer according to the [18F]FDG
synthesis program. Prior to synthesis, the column was
purged with nitrogen (150 ml min?1) for 5 min to
remove traces of solvents.
Production of18F. Fluorine-18 was produced by
the reaction18O(p,n)18F. A target filled with18O-
enriched water was irradiated with 17-MeV protons.
Target volume about 1 ml, beam current ?15 ?A,
irradiation time 50?10 min.
Preparation of the fluorinating agent, complex
After irradiation, the target con-
tents was passed through a column packed with QMA
resin, and the enriched water was collected in a sep-
arate vessel for the subsequent regeneration. The col-
umn was purged with helium for 5 min. The [18F]flu-
oride adsorbed on the resin was eluted into a reaction
vial with 2 ml of a 96 : 4 (by volume) acetonitrile?
water mixture containing 9.5?0.4 mg (0.026 mmol)
of K2.2.2 and 1.7?0.2 mg (0.013 mmol) of K2CO3
; the solution was then evaporated to dryness in a
nitrogen stream at 130?C for 4 min. The eluent was
prepared in the following proportions (amounts for ten
syntheses): 99?1 mg (0.25 mmol) of Kryptofix 2.2.2,
20?1 mg (0.12 mmol) of K2CO3, 0.90?0.05 ml of
Milli Q deionized water, and 20 ml of acetonitrile.
Nucleophilic fluorination. A solution of 10 mg
(0.02 mmol) of mannose triflate in 0.8 ml of acetoni-
trile was added to the dry residue containing the com-
plex [K/K2.2.2]+18F?. The reaction was performed in
a closed vessel at 80?C for 5 min. The mixture was
evaporated to a volume of 0.1?0.2 ml by passing
nitrogen for 2 min and then cooled to 40?45?C. Some
experiments were performed with 20 mg (0.04 mmol)
of mannose triflate and 1.2 ml of acetonitrile.
Hydrolysis of [18F]TAG. To the reaction vessel
was added 0.9 ml of 0.3 M NaOH. The mixture was
stirred for 10 s on a Wortex station (a part of the
RB-86 system) and hydrolyzed for 100 s in a nitrogen
stream (20 ml min?1). Then the mixture was neutra-
lized with 0.8 ml of 0.5 M HCl. The temperature was
raised to 120?C, and traces of acetonitrile were re-
moved by passing nitrogen (200 ml min?1) for 4 min.
Purification of [18F]FDG by solid-phase extrac-
tion. The reaction mixture was diluted with 3 ml of
water and passed through a column packed with 0.5 g
of LC-SCX cation-exchange resin and 0.5 g of LC-
Alumina N neutral alumina. At the column outlet, we
arranged a sterilizing filter and a vial for the product.
The column was washed with an additional 5 ml of
water; the total volume of the ready preparation was
Synthesis of FDG involving acid hydrolysis .
For comparison, we prepared [18F]FDG using acid
hydrolysis (series 1, Table 1) . The procedure
involved distillation of irradiated [18O]water from a
mixture containing 4.6 mg of K2CO3and 26 mg of
K2.2.2, followed by azeotropic distillation of residual
water with 3 ml of acetonitrile (0.3 ml ? 10) at 130?C;
reaction with 20 mg of mannose triflate at 80?C for
5 min, followed by evaporation of acetonitrile to 0.3?
0.5 ml; passing of the reaction mixture diluted with
1 ml of ether through a column packed with 0.5 g of
Supelco silica gel to remove unchanged [18F]F?; elu-
tion of [18F]TAG retained on the silica gel with 2 ml
of ether, followed by solvent evaporation; acid hydrol-
ysis of the resulting [18F]TAG by heating with 1.5 ml
of 2 M HCl at 120?C for 10 min, followed by acid
RADIOCHEMISTRYVol. 44No. 42002
406GOMZINA et al.
Table 2. Isolation of [18F]fluoride from irradiated [18O]H2O on various anion-exchange resins
? [18F]F?? [18F]F?de- ?
Dowex 1 ? 10 (CO3
AB-17 (OH?), 100 mg
Dowex AG1 ? 8 (CO3
AG 1 ? 8 (CO3
* QMA Sep-Pak Light cartridge, Waters, 130 mg.
? K2.2.2/K2CO3, mmol ?MeCN/H2O, vol % ?
eluent, % ?
resin, % ?
2?), 12 mg
0.026/0.018 (1.4) ?
? This work
2?), 10 mg?
2?), 36 mg?
?), 15?20 mg
evaporation; and purification of the product on a com-
bined column packed with 3 g of BioRad resin and
0.5 g of C18. In some experiments (series 2, Table 1),
the fluorinating complex [K/K2.2.2]+18F?was pre-
pared as described above using QMA resin, and all
the other stages were performed according to .
RESULTS AND DISCUSSION
When optimizing the automated synthesis, we
introduced the most significant changes into the fol-
lowing stages, which will be considered in more
detail: preparation of the fluorinating agent, reactive
complex [K/K2.2.2]+18F?; fluorination of mannose
triflate; hydrolysis of acetylated glucose derivative;
and purification of the resulting product.
Preparation of the complex [K/K2.2.2]+18F?and
nucleophilic fluorination. The complex [K/K2.2.2]+?
18F?is extensively used for radiofluorination of a
wide range of aliphatic and aromatic compounds;
therefore, optimization of procedures for its prepara-
tion is of general interest for development of new
routes to PET RPs. When choosing the synthetic pro-
cedure, it should be taken into account that nucleo-
philic fluorination is very sensitive to the presence of
traces of moisture and metal ions produced by irradia-
tion of the target . Furthermore, the procedure
should ensure regeneration with a high efficiency of
expensive [18O]H2O with a minimal loss of18F. In
early studies on synthesis of [K/K2.2.2]+18F?, the
irradiated water was collected in a reaction vessel
containing K2CO3and Kryptofix, after which water
was distilled off in a vacuum or in a nitrogen stream,
and the remaining traces of water were removed by
azeotropic distillation with acetonitrile, which was
added in portions to compensate for its evaporation.
Distillation of enriched water from the reaction mix-
ture allows no more than 80% recovery of water, and
azeotropic drying prolongs the synthesis and is one of
the sources of the18F loss.
Today, [18F]F?is isolated from irradiated [18O]-
water by sorption on anion-exchange resins followed
by elution with an appropriate eluent. In this case, the
loss of the enriched water is as low as 2?5%, with
virtually complete isolation of [18F]fluoride. The most
widely used is QMA resin, which is placed in a Tef-
lon tube (column) or is used as ready Sep-Pak cart-
ridges. It was shown that the use of freshly prepared
columns ensures the stable yield of the eluted fluoride
ion, and 15?20 mg of the resin is quite sufficient for
adsorption of several curies of [18F]F?. The use
of QMA Sep-Pak cartridges is more convenient for
routine syntheses, but in this case uncontrollable loss
of the activity is sometimes observed, which may be
due to shortcomings of the cartridge design . We
preferred to prepare QMA microcolumns (15?20 mg)
just before the synthesis, since in this case we were
able to elute 99.2?0.5% of the sorbed [18F]F?with
high reproducibility (n = 200).
Adsorbed [18F]fluoride is usually eluted from the
anion-exchange resin with an aqueous solution of
K2CO3or with an acetonitrile?water mixture contain-
ing potassium carbonate and Kryptofix in various
molar ratios (Table 2). The solution containing
[18F]fluoride desorbed from the column is evaporated.
To remove traces of moisture and enhance the reactiv-
ity of the complex, acetonitrile is added, and azeotrop-
ic distillation is performed. This stage of azeotropic
drying is present in virtually all the published proce-
dures of nucleophilic fluorination with [K/K2.2.2]+?
We have performed experiments on optimization
of this procedure with the aim to shorten the time for
RADIOCHEMISTRYVol. 44No. 42002
OPTIMIZATION OF AUTOMATED SYNTHESIS OF 2-[18F]FLUORO-2-DEOXY-D-GLUCOSE 407
preparation of the reactive complex and attain high
and reproducible yields in fluorination. In contrast to
other studies (Table 2), we chose the eluent containing
the minimal amount of water (4?5 vol % relative to
acetonitrile) necessary to dissolve the calculated
amount of K2CO3. This eluent was used to desorb
[18F]fluoride from both the QMA microcolumn and
QMA cartridge (Table 2); the degree of desorption
was high at the eluent volume as small as 2 ml. It is
of principal importance that evaporation of the eluent
containing acetonitrile and water in a volume ratio of
96 : 4 ensured complete removal of water without
additional azeotropic drying, which was confirmed by
high efficiency of the subsequent nucleophilic fluori-
An important parameter of the eluent composition
is the molar ratio of Kryptofix to K2CO3, which was
2 : 1 in our case, as recommended by Coenen et al.
 who added these agents to a solution containing
hydrated18F ions. Our experiments showed that, for
the subsequent nucleophilic fluorination to be effi-
cient, this ratio should also be kept when the complex
[K/K2.2.2]+18F?is prepared from [18F]fluoride sorbed
on the anion-exchange resin. The figure shows that, in
a narrow range of molar ratios of Kryptofix to carbo-
nate (2?2.5), the yield of [18F]FDG was as high as
50?72% (EOS). With the Kryptofix-to-carbonate ratio
increased to 4, the yield decreased to 30?35% and
then remained at approximately the same level. The
influence exerted by the eluent composition on the
yield of the final product is also clearly seen from
comparison of series 4 and 5 in Table 1.
The reaction of mannose triflate with the fluorinat-
ing complex [K/K2.2.2]+18F?was performed in aceto-
nitrile at 80?C for 5 min. The results of synthesizing
[18F]FDG using the optimized procedure for preparing
the fluorinating complex (series 2, 4?6) and the tradi-
tional procedure with additional azeotropic drying
(series 1, 3) are compared in Table 1. Along with
increase in the product yield (6 and 15% in the case of
acid and base hydrolysis, respectively), the optimized
procedure for preparing the fluorinating complex
made the synthesis time 10 min shorter (series 1 and
2, 3 and 5). Our experiments also showed that the
yield of the final product is largely affected by the
ratio of Kryptofix and mannose triflate. At their equi-
molar ratio (0.02 mmol of mannose triflate, series 6),
the yield of [18F]FDG was, on the average, 8?10%
higher than at their 1 : 2 ratio (0.04 mmol of mannose
triflate, series 5).
Hydrolysis of [18F]TAG. The protective acetyl
groups are removed from the [18F]TAG molecule by
Radiochemical yield of [18F]FDG as a function of
treating the reaction mixture with 0.3 M NaOH as
recommended in . As shown in that study, with
an alkali of such a concentration the hydrolysis at
room temperature is virtually complete in 1 min.
Particular attention should be given to evaporation of
the reaction mixture, as affecting the acetonitrile con-
tent in the final product. Our experience shows that
evaporation of the reaction mixture to dryness prior to
hydrolysis results in a drastically decreased yield of
the final product; a similar trend was noted in .
The highest and well reproducible yield was attained
when a volume of 0.1?0.2 ml was left. However, in
this case the content of acetonitrile in the final product
(5?8 mg ml?1) considerably exceeded the maximum
permissible concentration in this product (0.3 mg ml?1
[26, 27]). For azeotropic distillation of traces of aceto-
nitrile, the reaction mixture should be heated to 100?
120?C, but, under conditions of base hydrolysis, this
causes significant loss of the product [28, 29]. For
example, hydrolysis of [18F]TAG with 0.33 M NaOH
at 80?C for 5 min is accompanied by 30% epimeriza-
tion of [18F]FDG into 2-deoxy-2-[18F]fluoro-D-man-
nose ([18F]FDM) and partial decomposition with the
release of 16% of [18F]F?. To avoid the product
loss and deterioration of its radiochemical purity upon
heating, we performed the hydrolysis under mild con-
ditions (0.3 M NaOH, 100 s, 40?45?C) and then
neutralized the mixture with HCl. After neutralization,
the temperature could be safely raised to 120?C to
remove residual acetonitrile by azeotropic distillation;
the acetonitrile content in the resulting preparation
was as low as 0.05?0.10 mg ml?1. Also, optimization
of the hydrolysis and neutralization (proper choice of
the concentrations and volumes of reagents, tempera-
ture, time, and stirring mode) increased the degree of
conversion of [18F]TAG into [18F]FDG to >99%
without further transformations of the final product.
Purification of [18F]FDG. Replacement of acid
hydrolysis by base hydrolysis substantially simplified
the [18F]FDG synthesis process, because intermediate
RADIOCHEMISTRY Vol. 44No. 42002
408 GOMZINA et al.
purification of [18F]TAG on silica gel  became
unnecessary. In automated modules, removal of im-
purities is usually provided by several series-con-
nected cartridges packed with sorbents such as alumi-
na, cation-exchange resins, or octadecyl silica gel
(C18), whereas in the RB-86 robotic system it is more
convenient to use a single column packed with several
sorbents. In this study, we used a combined extraction
column packed with LC-SCX cation-exchange resin
for retention of K2.2.2 and neutralization of the solu-
tion and with neutral alumina (Alumina N) to remove
traces of unchanged [18F]fluoride. With this purifica-
tion procedure, the radiochemical purity of [18F]FDG
exceeded 99%, and the content of Kryptofix 2.2.2 was
less than 0.20 mg ml?1, meeting the Pharmacopoeia
requirements [27, 30]. Special experiments showed
that additional packing of the column with C18 sor-
bent used in [13, 14] did not further improve the
Evaluation of the radioactivity balance recalculated
on the end of bombardment (EOB) gave the following
results (for a typical experiment of series 6): [18F]-
FDG, 84%; purification column, 9%; reaction vessel,
1%; and unaccounted loss, 6%. Considerably smaller
loss of the radioactivity in the procedure involving
base hydrolysis is due to advantages of the procedure
as a whole and to optimization of each stage. For
comparison, the losses of radioactivity in the synthe-
sis of [18F]FDG involving acid hydrolysis (series 1)
are as follows: two extraction columns, ?20%; reac-
tion vessel after fluorination, up to 20%; reaction
vessel after hydrolysis, ?1?3%; unaccounted loss,
10?15%; and total loss, more than 50% (EOB).
Thus, we propose a procedure involving the use of
an optimal eluent containing Kryptofix 2.2.2 and
potassium carbonate (molar ratio 2 : 1) in a 96 : 4 (by
volume) acetonitrile?water mixture; preservation of
the equimolar ratio of Kryptofix 2.2.2 and the sub-
strate in the fluorination stage; and thorough observ-
ance of the base hydrolysis conditions. This procedure
is used today in the Institute of Human Brain to pre-
pare [18F]FDG for routine PET diagnostic investiga-
tions of brain . It allows the desired product to be
obtained within 38 min in 63?3% yield (EOS, n =
40) and with >99% radiochemical purity.
Our results were obtained at the initial [18F]fluo-
ride activity of 0.2?0.4 Ci, which was restricted by
the target capacity. However, our procedure was also
successfully tested with higher (0.4?2.0 Ci) initial18F
activities (PETtrace cyclotron) on an FDG Microlab
automated module (GEMS, Uppsala, Sweden) .
The procedure proposed for preparing the reactive
complex using the optimized eluent composition is
used today not only in synthesis of [18F]FDG, but
also in other nucleophilic fluorinations with substrates
containing various leaving groups [17, 31]. The proce-
dure affords a high and well reproducible yield of the
fluorinated product at shorter synthesis time and sim-
pler process (due to elimination of the traditionally
used stage, azeotropic drying of the complex).
1. Meyer, G.-J., Waters, S.L., Coenen, H.H., et al., Eur.
J. Nucl. Med., 1995, vol. 22, no. 12, pp. 1420?1432.
2. Pauwels, E.K.J., Ribeiro, M.J., Stoot, J.H.M.B., et al.,
Nucl. Med. Biol., 1996, vol. 25, pp. 317?322.
3. Amano, S., Inoue, T., Tomiyoshi, K., et al., J. Nucl.
Med., 1998, vol. 39, pp. 1424?1427.
4. Coenen, H.H., Pike, V.W., Stocklin, G., and Wag-
ner, R., J. Radiat. Appl. Instrum., Part A, 1987,
vol. 38, no. 8, pp. 605?610.
5. Helus, F., Maier-Borst, W., and Oberdorfer, F., Devel.
Nucl. Med., 1992, vol. 27, pp. 235?238.
6. Beuthien-Baumann, B., Hamacher, K., Oberdorfer, F.,
and Steinbach, J., Carbohydr. Res., 2000, vol. 327,
nos. 1?2, pp. 107?118.
7. Block, D., Klatte, B., and Knochel, A., J. Label.
Comp. Radiopharm., 1986, vol. 23, no. 5, pp. 467?
8. Kilbourn, M.R., Fluorine-18 Labelling of Radiophar-
National Academy, 1990.
9. Hamacher, K., Coenen, H.H., and Stocklin, G.,
J. Nucl. Med., 1986, vol. 27, pp. 235?238.
10. Kiselev, M.Yu., Solov’ev, D.V., and Korsakov, M.V.,
Radiokhimiya, 1992, vol. 34, no. 2, pp. 129?135.
11. Krasikova, R.N., Radiokhimiya, 1998, vol. 40, no. 4,
12. Fuchtner, F., Steinbach, J., Mading, P., and Johan-
sen, B., Appl. Radiat. Isot., 1996, vol. 47, no. 1,
13. Bogni, A., Pascali, C., Iwata, R., et al., J. Label.
Comp. Radiopharm., 1997, vol. 40, pp. 219?221.
14. Zijlsta, S., Medema, J., Elsinga, P.H., et al., J. Label.
Comp. Radiopharm., 1997, vol. 40, suppl. 1, pp. 229?
15. Mosdzianovski, C., Lemaire, C., Lauricella, B., et al.,
J. Label. Comp. Radiopharm., 1999, vol. 42, suppl. 1,
16. Gomzina, N.A., Krasikova, R.N., Korsakov, M.V.,
and Dahlstrom, K., Eur. J. Nucl. Med., 2000, vol. 27,
no. 8, p. 1220.
RADIOCHEMISTRYVol. 44No. 42002
OPTIMIZATION OF AUTOMATED SYNTHESIS OF 2-[18F]FLUORO-2-DEOXY-D-GLUCOSE 409
17. Gomzina, N.A., Zaitsev, V.V., and Krasikova, R.N.,
J. Label. Comp. Radiopharm., 2001, vol. 44, suppl. 1,
18. Studentsov, E.P., Korsakov, M.V., Shchukin, E.V.,
et al., RF Patent Appl. 2000116601/14.
19. Korsakov, M.V., Solov’ev, D.V., and Udalov, V.B.,
Radiokhimiya, 1994, vol. 36, no. 5, pp. 434?439.
20. Schlyer, D.J., Firouzbakht, M.L., and Wolf, A.P.,
Appl. Radiat. Isot., 1993, vol. 44, no. 12, pp. 1459?
21. Proc. Seventh Workshop on the Targetry and Target
Chemistry, Heidelberg, 1997, pp. 167?172.
22. Jewett, D.M., Toorongian, S.A., Bachelor, M.A., and
Kilbourn, M.R., Appl. Radiat. Isot., 1990, vol. 41,
no. 6, pp. 583?586.
23. Schlyer, D.J., Bastos, M.A.V., Alexoff, D., and
Wolf, A.P., Int. J. Appl. Radiat. Isot., 1990, vol. 41,
no. 6, pp. 531?533.
24. Paskali, C., Bogni, A., Decise, D., et al., Anticancer
Res., 1998, vol. 18, pp. 1563?1568.
25. Znang, Z.Y., Zigler, S.S., and Kabalka, G.W., J. La-
bel. Comp. Radiopharm., 1997, vol. 40, suppl. 1,
26. Channing, M.A., Huang, B.X., and Eckelman, W.C.,
Nucl. Med. Biol., 2001, vol. 28, pp. 469?471.
27. Pharmeuropa, 1996, vol. 8, no. 4, pp. 514?517.
28. Meyer, G.-J., Martzke, K.H., Hamacher, K., et al.,
Appl. Radiat. Isot., 1999, vol. 51, pp. 37?41.
29. Varelis, P. and Barnes, R.K., Appl. Radiat. Isot.,
1996, vol. 47, no. 8, pp. 731?733.
30. FSP (Pharmacopoeia Item of Enterprise) 42-0257-
1062-01: Fluorodeoxyglucose,18F: Solution for Injec-
tion, Inst. of Human Brain, Russian Acad. Sci., 2001.
31. Zaitsev, V.V., Fedorova, O.S., Mosevich, I.K., et al.,
Radiokhimiya, 2002, vol. 44, no. 4, pp. 358?365.