An improved preparation of [18F]AV-45 by simplified solid-phase extraction purification

Article abstract of DOI:10.1002/jlcr.3813

Amyvid (florbetapir f18, [¹⁸F]AV-45, [¹⁸F]5) was the first FDA-approved positron emission tomography imaging agent targeting β-amyloid (Aβ) plaques for assisting the diagnosis of Alzheimer disease. This work aimed to improve the [

Full text of DOI:10.1002/jlcr.3813

Kung Hank F (Orcid ID: 0000-0002-8811-4402)
An improved preparation of [18F]AV-45 by simplified solid-phase
extraction purification
Lifang Zhang1#, Aili Zhang1#, Xinyue Yao1,3, Yan Zhang1, Futao Liu1, Haiyan Hong1, Zhihao
Zha2,3, Yajing Liu2, Zehui Wu2, Jinping Qiao1, Lin Zhu1,2*, Hank F. Kung2,3*
1Key Laboratory of Radiopharmaceuticals, Ministry of Education, College of Chemistry,
Beijing Normal University, Beijing, 100875, P. R. China
2Beijing Institute of Brain Disorders, Capital Medical University, Beijing 100069, P. R.
China
3Department of Radiology, University of Pennsylvania, Philadelphia, PA 19104, USA
#
Lifang Zhang and Aili Zhang contributed equally to this work and were considered co-first
authors.
*Corresponding author contact information:
Hank F. Kung, PhD, Department of Radiology, University of Pennsylvania, 3700 Market
Street, Room 305, Philadelphia, PA 19104, USA
Tel.: +1 215 662 3989; fax: +1 215 349 5035; email: kunghf@gmail.com
Lin Zhu, PhD, Key Laboratory of Radiopharmaceuticals (Beijing Normal University),
Ministry of Education, No. 19, XinJieKouWai Street, Haidian District, Beijing 100875, P. R.
China
This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process which may
lead to differences between this version and the Version of Record. Please cite this article as
doi: 10.1002/jlcr.3813
This article is protected by copyright. All rights reserved.

Tel.: +86 10 62207789; email: zhulin@bnu.edu.cn
Total pages: 28, Word count: 6685, Schemes: 2, Figures: 3, Tables: 3.
Abstract:
Amyvid (florbetapir f18, [18F]AV-45, [18F]5) was the first FDA approved positron
emission tomography (PET) imaging agent targeting β-amyloid (Aβ) plaques for assisting the
diagnosis of Alzheimer’s disease. This work aimed to improve the [18F]AV-45 ([18F]5)
preparation by using solid-phase extraction (SPE) purification. [18F]AV-45 ([18F]5) was
synthesized by direct nucleophilic radiofluorination of O-tosylated precursor (1 mg) at
120 °C in anhydrous dimethyl sulfoxide (DMSO), followed by acid hydrolysis of the N-Boc
protecting group. Purification was accomplished by loading the crude reaction mixture to a
cartridge (Oasis HLB 3 cc) and eluting with different combinations of solvents. This method
removed the chemical impurity while leaving [18F]AV-45 ([18F]5) on the cartridge. The final
dose was eluted by ethanol. [18F]AV-45 ([18F]5) was produced within 51 min (radiochemical
yield 42.7 ± 5.9%, decay corrected, n = 3), and the radiochemical purity was >95%. Total
chemical impurity per batch (24.1 ± 2.7 μg per batch) was below the limit described in the
package insert of Amyvid, florbetapir f18 (chemical mass: less than 50 μg/dose). In
summary, [18F]AV-45 ([18F]5) was produced efficiently and reproducibly using a
cartridge-based SPE purification. This method brings the process closer for routine
preparation, similar to the commercially used [18F]FDG.
Key Words: Fluorine-18, solid-phase extraction purification, florbetapir f18,
automated radiosynthesis
This article is protected by copyright. All rights reserved.

1. Introduction
Advances in nuclear medicine and biomedical imaging depend on the development of
diverse radiopharmaceuticals. Production of radiopharmaceuticals requires a strategy that
optimizes a combination of factors, including the radionuclide half-life, structural
characteristics of the final molecule or its precursor(s), radiolabeling procedures, and their
clinical in vivo application.1 As a logical evolution of improving the preparation of
18F-radiopharmaceuticals in routine production for clinical application, several obstacles need
to be resolved. First, the initial radiochemical yield via direct fluorination reaction catalyzed by
crown ethers needs to be optimized. Second, based on the reaction kinetics, the reaction time
and temperature need to be adjusted to maximize the production of the desired product and to
reduce the amount of side products. Third, the final product needs to be simply and readily
separated. The separation needs to reduce or eliminate chemical and radiochemical impurities,
which is usually accomplished via high-performance liquid chromatography (HPLC)
purification. HPLC offers a generally applicable purification method for producing PET
radiotracers. However, this method is relatively time consuming, and the dedicated HPLC
equipment requires additional space inside the hot cells and regular maintenance. For routine
preparation, clinical doses would be best accomplished by a simple, fast, and readily available
solid-phase extraction (SPE) purification. Disposable cartridges or solid-phase columns are
used only once and conveniently disposed after completion of the purification procedure. This
progression is necessary for routine preparation of many radiopharmaceuticals.2-4 One only
needs to examine the history of the development of a method for [18F]FDG preparation,
because it is the most widely produced radiopharmaceutical for PET imaging. Previously, in
This article is protected by copyright. All rights reserved.

the 80s and 90s, the production was based on HPLC purification, and the process was time
consuming and produced low yield.5 Improvements reported in the late 80s demonstrated the
use of a SN2 fluorination reaction with triflate precursor. The new fluorination of precursor,
1,3,4,6-tetra-O-acetyl-2-O- trifluoromethanesulfonyl--D-mannopyranose, enhanced the
radiochemical yield and led to significant scaling up of the production from 100 s MBq to 100
s of GBq. After hydrolysis of the acetyl protecting by either acid or base, the final product,
[18F]FDG, was “purified” by SPE through different combinations of alumina, C-18, cation
exchange, and/or ion retardation. This optimized preparation is now the main production
method that is regularly used in commercial or hospital radiopharmacy.5
Amyvid ([18F]AV-45, [18F]5) was the first amyloid-β PET imaging agent approved by
the FDA in 2012. It is now available for routine clinical use in assisting diagnosis of patients
suspected of Alzheimer’s disease (AD) and also for selecting patients in drug trials.6-10
Florbetapir f18 PET imaging study is now being used to determine the presence of amyloid
load in the brain as a significant risk factor for developing AD. Amyloid-β plaques in the
brain are associated with mild cognitive impairment or dementia of uncertain etiology.11,12
The original method for preparing [18F]AV-45 ([18F]5) was based on the final HPLC
purification.13,14 Similar to the [18F]FDG production described above, we previously reported
an effort to streamline the [18F]AV-45 ([18F]5) production.15-17 We wish to develop a routine
production method, which would be amendable for radiopharmacy operation. The SPE method
would likely help make the routine clinical doses of [18F]AV-45 ([18F]5) readily available. This
situation is a necessary progression for maturing radiochemistry technology to convert a
research tracer from laboratory to routine clinical application.
This article is protected by copyright. All rights reserved.

The recommended dose for Amyvid is 370 MBq (10 mCi), containing a maximum of
50 μg chemical mass in one dose, administered as a single intravenous bolus with a total
volume of 10 mL or less. Limited Aβ binding sites exist in the brain. Therefore, the molecular
activity (chemical amount of carrier or pseudo-carrier is a key determining factor) would be an
important factor for consideration. Other chemical impurities may compete for the same Aβ
binding sites in the brain.15 In our previous work, we suggested the use of the SPE method
instead of HPLC purification to prepare [18F]AV-45 ([18F]5) operated with a semi-automatic
synthesizer (module BNU F-A2).18 The separation procedure included loading of [18F]AV-45
([18F]5) crude reaction mixture on an Oasis HLB 3 сс cartridge and fractional elution of the
cartridge with aqueous solutions containing increasing percentage of ethanol. This procedure
produced [18F]AV-45 ([18F]5) with excellent radiochemical purity (RCP) >95%, in which the
“cold” pseudo carrier of AV-136 (4, 200–500 μg/batch) remained in the final dose (see Scheme
1).18 As reported, AV-136, 4 also displayed good binding affinity to Aβ plaques (Ki = 6.37
nM) and might inhibit the binding of [18F]AV-45 ([18F]5) (Ki = 2.87 nM) in the human brain
to a certain degree.15-17 Removal or minimizing of this chemical impurity from the final dose
formulation is an important goal on meeting the [18F]AV-45 ([18F]5) specification.18
Based on the previous SPE purification results, additional efforts were made to further
reduce the total chemical mass (total amount of chemical impurities) in the final [18F]AV-45
([18F]5) doses. Ultimately, the goal was to maximize the radiochemical yield (RCY) and
minimize the chemical impurity to below 50 μg/batch as defined by the package insert of
Amyvid. One of the major nonradioactive side products, AV-136, 4, a pseudo-carrier, was
less lipophilic than [18F]AV-45 ([18F]5). Thus, it could be better eluted from the reversed-phase
This article is protected by copyright. All rights reserved.

cartridge by manipulating the polarity of the solvent used for eluting the cartridge. Taking
advantage of these differences, we have developed a new SPE procedure. In the current study,
we reported an optimization of automated radiosynthesis of [18F]AV-45 ([18F]5) via an
efficient and reliable SPE purification, by which the majority of chemical impurities could be
removed from the final doses.
2. Experimental
2.1 General
2.1.1 Reagents and solvents
All reagents and solvents were of analytical reagent grade and purchased from
commercial sources (Aldrich, Acros, or Alfa Inc.). They were used without further
purification unless otherwise specified.
2.1.2 Instrument and consumable materials
Thin layer chromatography (TLC) was run on pre-coated plates of a silica gel 60 F254
(Merck, Darmstadt, Germany). 1H NMR spectra were obtained on a 600 MHz
JMTC-600/54/JJ (Japan Superconductor Technology, Inc., Japan). Chemical shifts are
reported as δ values (parts per million) relative to residual protons of deuterated solvent.
Coupling constants are reported in Hertz. The multiplicity is defined as singlet (s), doublet
(d), triplet (t), broad (br), or multiplet (m). Liquid chromatography–mass spectrometry (LC–
MS) analysis was performed on a Triple TOF 5600 (AB SCIEX, Foster City, CA, USA) mass
spectrometry with an electrospray ionization (ESI) source, and the LC inlet was Thermo
This article is protected by copyright. All rights reserved.

Scientific UltiMate 3000 chromatographic system. Analytical HPLC was conducted on the
Shimadzu LC-20A HPLC system equipped with a multiwavelength UV detector and
radiation detector. SPE cartridges (Oasis HLB 3 cc cartridge, Sep-Pak light QMA cartridge)
were obtained from Waters (Milford, MA, USA). The [18F]fluoride ion was provided by the
Peking University Cancer Hospital.
2.1.3 Liquid chromatography
LC was performed on Thermo Scientific UltiMate 3000 chromatographic system with
an autosampler and a column oven. A Phenomenex Luna C18 column (4.6 mm × 150 mm, 5
μm) was employed. Chromatographic separation was achieved with gradient elution by using
a mobile phase composed of water containing 10 mM ammonium formate (A) and
acetonitrile (ACN) (B). A flowrate of 1 mL/min with the following gradients: from 0 min to 2
min, A 95% and B 5%; from 2 min to 5 min, A 95%–30% and B 5%–70%; from 5 min to 10
min, A 30%–1% and B 70%–99%; from 10 min to 15 min, A 1%–95% and B 99%–5%; and
from 15 min to 20 min, A 95% and B 5%. The total HPLC run time was 20 min. The
injection volume was 10 μL. Additionally, the compound was determined by UV at 350 nm.
The product RCP was confirmed by analytical HPLC (Shimadzu Prominence LC-20AT and
SPD-20A, Shimadzu Corp., Japan) connected to a radiation detector (Gabi Star, Raytest
Corp., Japan). The analytical HPLC conditions were the same as described above. The RCP
of the final product was also measured by Radio-TLC, with a mobile phase ethyl acetate.
This article is protected by copyright. All rights reserved.

2.1.4 Mass spectrometry
The description of LC-MS instrument was included in Section 2.1.2. The ion spray
voltage floating was 5500 V, and the source temperature was 600 °C. The ion source gas 1
and gas 2 were both 40 psi, and the curtain gas was 25 psi. The compound declustering
potential was 100 V, and the collision energy was 10 eV. The dwell time was set at 0.2750 s
per transition. All data collected were acquired and processed using Analyst TF1.7 Software
with QuanLynx program. A LCT Premier XE mass spectrometer was used to identify and
confirm the chemical impurity of the product. The exact mass measurement was carried out
in positive V mode. In the full scan mode, data were acquired in the mass range m/z 100–
2000, with a scan time of 1.0 s. A lock spray was used for external calibration with a solution
of the leucine-enkefaline as reference. Accurate masses were calculated and used to
determine the elementary composition of the components with fidelity of 5 ppm.
2.2 Synthesis
The precursor AV-105, 1, nonradioactive standard AV-45, 5, and AV-136, 4 were
synthesized according to previous reports (Scheme 1).19,20
(E)-4-(2-(6-(2-(2-(2-Chloroethoxy)ethoxy)ethoxy)pyridin-3-yl)vinyl)-N-methylaniline
(Cl-derivative) 2
(E)-2-(2-(2-((5-(4-(Methylamino)styryl)pyridin-2-yl)oxy)ethoxy)ethoxy)ethyl
4-methylbenzenesulfonate (OTs-derivative) 3
A mixture of
(E)-2-(2-(2-((6-(4-((tert-butoxycarbonyl)(methyl)amino)styryl)-pyridin-3-yl)oxy)ethoxy)etho
xy)ethyl 4-methylbenzenesulfonate (AV-105, 1) (100 mg, 0.16 mmol) was stirred in
This article is protected by copyright. All rights reserved.

dichloromethane (20 mL) at room temperature, 1 M HCl (0.5 mL) was added, and the
reaction mixture was stirred for 3 days. The solvent was evaporated, and the saturated
solution of sodium carbonate (20 mL) was added and extracted with ethyl acetate (20 mL ×
3) thrice. The organic layers were combined and dried over anhydrous sodium sulfate and
filtered. Then, the filtrate was evaporated and purified by a combiflash (silica gel 25 g) (ethyl
acetate:petroleum ether = 0% to 50%) to obtain the Cl-derivative 2 (15 mg, yield 25%) as a
yellow solid and OTs-derivative 3 (25 mg, yield 29%) as a yellow solid.
Cl-derivative 2: 1H NMR (600 MHz, CDCl3) δ: 8.13 (s, 1H), 7.75 (d, J = 8.7 Hz, 1H),
7.34 (d, J = 7.3 Hz, 2H), 6.89 (d, J = 16.3 Hz, 1H), 6.80 (d, J = 16.4 Hz, 1H), 6.75 (d, J = 8.6
Hz, 1H), 6.61 (d, J = 7.4 Hz, 2H), 4.50–4.45 (m, 2H), 3.86–3.85 (m, 2H), 3.76–3.74 (m, J =
6.0, 2H), 3.73–3.66 (m, 4H), 3.63–3.60 (m, 2H), 2.86 (s, 3H). HRMS (ESI) calculated for
C20H26ClN2O3 [M+H]+, 377.1626; found, 377.1611.
OTs-derivative 3 (OTs-derivative): 1H NMR (600 MHz, CDCl3) δ: 8.11 (d, J = 2.1
Hz, 1H), 7.78 (d, J = 8.1 Hz, 2H), 7.71 (dd, J = 8.7, 2.1 Hz, 1H), 7.34 (d, J = 8.3 Hz, 2H),
7.31 (d, J = 8.2 Hz, 2H), 6.92 (d, J = 7.7 Hz, 2H), 6.84 (d, J = 5.2 Hz, 2H), 6.73 (d, J = 8.6
Hz, 1H), 4.46–4.40 (m, 2H), 4.15–4.14 (m, 2H), 3.81–3.79 (m, 2H), 3.69–3.67 (m, 3H),
3.63–3.59 (m, 4H), 2.91 (s, 3H), 2.41 (s, 3H). HRMS (ESI) calculated for C27H33N2O6S
[M+H]+, 513.2053; found, 513.2018.
This article is protected by copyright. All rights reserved.

2.3 Radiosynthesis of [18F]AV-45 ([18F]5) and the optimization of SPE purification
Aqueous solution of the [18F]fluoride ion produced by the cyclotron was passed
through a Sep-Pak light QMA cartridge. The cartridge was previously activated with 10 mL 1
18
M NaHCO3 and 10 mL water and dried with N2. Then, the F activity was eluted with 1.1
mL of Kryptofix 2.2.2 (K2.2.2)/K2CO3 solution (11 mg K2.2.2 and 2 mg K2CO3 in 0.93 mL
ACN and 0.17 mL H2O). The eluent was then evaporated at 110 °C under an N2 stream.
Additionally, the residue was azeotropically dried twice with 1 mL anhydrous ACN.
2.3.1 SPE purification of [18F]AV-45 ([18F]5) by using an Oasis HLB cartridge
To the dried [18F]KF/K2.2.2, 1 mg precursor (AV-105, 1) dissolved in 1 mL
anhydrous DMSO was added. The mixture was heated at 110 °C for 10 min, followed by
hydrolysis with 3 M HCl (1 mL) at 100 °C for 5 min. After cooling down to room
temperature, 0.3 M NaOH (10 mL) was added for neutralization. The crude products were
evenly divided into two parts and passed through an Oasis cartridge (pretreated with 10 mL
EtOH, followed by 10 mL sterile water). The Oasis cartridges were washed with 2 × 10 mL
water. One Oasis cartridge was washed with 2 mL different concentrations of
acetonitrile/water (ACN/H2O, 10%–50%). The other was washed with 2 mL different
concentrations of ethanol/water (EtOH/H2O, 10%–70%). The eluent activity and the retained
activity on the Oasis cartridge were recorded. The eluent was also analyzed by HPLC.
Further gradient elution was then explored, including the washing conditions of 8 mL
35% ACN/H2O, 4 mL 40% ACN/H2O, 2 mL 45% ACN/H2O, 4 mL 55% EtOH/H2O, and 4
mL 60% EtOH/H2O. The pure product was eluted by 1 mL EtOH. The eluent activity and the
This article is protected by copyright. All rights reserved.

retained activity on the Oasis cartridge were recorded. The eluent and product were analyzed
by HPLC.
Finally, we investigated the optimum volume of 35% ACN/H2O as the elution. The
Oasis cartridge was washed with increasing volume (4–8 mL) of 35% ACN/H2O, followed
by elution with 1 mL EtOH. The eluent and retained activities on the Oasis cartridge were
recorded. The eluent and product were analyzed by HPLC.
2.3.2 Optimization of fluorination temperature and deprotection conditions
To the dried [18F]KF/K2.2.2 (185–370 MBq), 1 mg precursor (AV-105, 1) dissolved
in 1 mL anhydrous DMSO was added. The mixture was heated at 110 °C, 120 °C, 130 °C or
140 °C for 10 min, followed by hydrolysis with 3 M HCl (1 mL) at 100 °C for 5 min. After
cooling down to room temperature, 0.3 M NaOH (10 mL) was added to neutralize. The crude
product was passed through an Oasis cartridge. The Oasis cartridge was washed with 10 mL
water, 6 mL 35% ACN/H2O, and 20 mL water in turn. The pure product was eluted by 2 mL
EtOH.
To the dried [18F]KF/K2.2.2 (185–370 MBq), 1 mg precursor (AV-105, 1) dissolved
in 1 mL anhydrous DMSO was added. The mixture was heated at 120 °C for 10 min,
followed by hydrolysis with 2 M H2SO4 (1 mL) at 100 °C for 5 min. After cooling down to
room temperature, 0.4 M NaOH (10 mL) was added for neutralization. The purification
method was the same as that described above.
This article is protected by copyright. All rights reserved.

2.4 Automated radiosynthesis of [18F]AV-45 ([18F]5) with SPE purification
The fully automated synthesis of [18F]AV-45 ([18F]5) was performed using a BIBD-F
synthesizer (Figure 1). The device was built with flexibility in mind and can implement
various synthetic procedures. The sequence program of the whole system included the main
synthesis unit, purification unit, and formulation unit. Before starting the synthesis, the
reagents for the production of [18F]AV-45 ([18F]5) were stored in the reagent vials, as
follows: (P01) 11 mg K2.2.2 and 2 mg K2CO3 in 1.1 mL (0.93 mL ACN and 0.17 mL H2O)
solution; (P02, P03) 1 mL anhydrous ACN; (P04) 1 mg (AV-105, 1) dissolved in 1 mL
anhydrous DMSO; (P05) 1 mL 2 M H2SO4 solution; (P06) 10 mL 0.4 M NaOH solution;
(P07) 10 mL H2O; (P08) 6 mL 35% ACN/H2O; (P09) 20 mL H2O; and (P10) 2 mL EtOH.
The automated synthesis module (Figure 1) was operated in the following sequence. The
[18F]fluoride ion (1.85–3.70 GBq) was trapped on a Sep-Pak Light QMA cartridge. Then, the
18F activity was eluted with 1.1 mL eluent (P01) into the reaction vessel. The mixture
solution in the reaction vessel was evaporated under a vacuum and N2 gas flow at 110 °C.
The residue was azeotropically dried twice with 1 mL anhydrous ACN (P02, P03) under the
same condition. A solution of precursor (1 mg AV-105, 1 in 1 mL anhydrous DMSO, P04)
was added to the evaporation residue in the reaction vessel and heated to 120 °C for 10 min.
After [18F] fluorination, the reaction mixture was cooled to 90 °C, and 1 mL of 2 M H2SO4
acid solution (P05) was added to remove the N-Boc group at 100 °C for 5 min. The 10 mL
0.4 M NaOH solution (P06) was added for neutralization, and the mixture was subsequently
passed through an Oasis HLB 3 cc cartridge. Subsequently, the cartridge was rinsed with 10
mL deionized water (P07), 6 mL 35% ACN/H2O (P08), and 20 mL deionized water (P09) in
This article is protected by copyright. All rights reserved.

turn. After drying the cartridge under a N2 flow, the pure product was eluted with 2 mL EtOH
(P10) into the transfer vial containing 7 mL of 0.9% saline solution. The final product was
moved to the final product vial containing 11 mL of 0.9% saline solution. The solution was
passed through a 0.22 μm sterile filter (Medical Millex-GV; Millipore, Billerica, MA, USA)
into a sterile vial to generate the dose for injection.
2.5 Quality control
The chemical impurities and specific activity of [18F]AV-45 ([18F]5) were measured
and quantified with calibration curves of F-19 standards by HPLC (Figure S1). RCP was
measured by radio-HPLC and radio-TLC, and the conditions were as described above
(Section 2.1.3). The Rf value of AV-45, 5, on the TLC plate was 0.7. Residual K2.2.2 was run
on pre-coated plates of silica gel 60 F254 with a mobile phase EtOH/30% ammonia (v/v =
9/1). In addition, the spots were stained with iodine vapor to make them visible. The pH of
the final product was measured by applying a small volume (20 μL) to a pH-indicator strip
(pH range: 0.0–14.0) and compared with the standard color chart provided by the test
manufacturer. The residual organic solvents (ACN and DMSO) were quantified using a gas
chromatograph system (Agilent 6820 GC, Agilent Technologies, USA) equipped with an FID
detector and a DB-FFAP capillary column (30 m × 0.32 mm × 0.5 μm, Agilent Technologies,
USA).
This article is protected by copyright. All rights reserved.

3. Results and Discussion
3.1 Chemical synthesis
The synthesis of Cl-derivative 2 and OTs-derivative 3 is illustrated in Scheme 1.
Precursor, AV-105, 1, was hydrolyzed by 1 M HCl to give OTs-derivative 3 in 29% yield. In
the same reaction, the Cl-derivative 2 was produced in 25% yield. The “cold” AV-45, 5, was
1
prepared as described previously.20 All compounds were characterized by H NMR and
HRMS. The results were consistent with the proposed structures. The authentically prepared
AV-45, 5, and related derivatives, 2, 3, and 4, were employed as standards for establishing
the analytical HPLC methods to measure the residual chemical quantity. The calibration
curves are included in the Supporting information.
3.2 Radiosynthesis of [18F]AV-45 ([18F]5) and the optimization of SPE purification
3.2.1 SPE purification of [18F]AV-45 ([18F]5) by using an Oasis HLB cartridge
The [18F]AV-45 ([18F]5) radiosynthesis was performed according to the reaction
shown in Scheme 2. The radiotracer was produced from the O-tosylated precursor (AV-105,
1) via a standard nucleophilic substitution fluorination reaction. After an acid hydrolysis of
the N-Boc protection group, the crude reaction mixture was subjected to SPE purification.
Many recent reports describe the SPE purification method for different PET
radiotracers. This technique has been successfully applied to prepare a number of clinically
relevant radioactive tracers, including [18F]FDG,4,21 [18F]FMISO,22 [11C]PiB,23 [18F]FMZ2,
and [18F]FES.24 A SPE purification method for preparing [18F]AV-45 ([18F]5) had been
previously reported without HPLC purification.18 However, the reported SPE purification
This article is protected by copyright. All rights reserved.

method did not fully remove the hydroxyl derivative, AV-136, 4, which could potentially
interfere with the detection of Aβ plaques in the brain. Due to its relatively high binding
affinity (Ki = 6.37 ± 3.75 nM),15 AV-136, 4, might compete for the same Aβ binding sites in
the brain.
Since the hydroxyl derivative, AV-136, 4 is more hydrophilic than [18F]AV-45
([18F]5), we employed the following separation steps: 1) the crude reaction mixture was
loaded on an Oasis HLB 3 cc cartridge; and 2) the Oasis HLB 3 cc cartridge was fractionally
eluted by aqueous solutions with increasing ethanol or acetonitrile content. In addition, LC–
MS technique and calibration curves of standards with HPLC were used to identify and
quantify the residual chemical impurities in the SPE-purified [18F]AV-45 ([18F]5) product.
Using the optimized synthesis, the [18F]AV-45 ([18F]5) synthesis was finally implemented on
an automated synthesis module (BIBD-F).
Commercially available Oasis HLB 3 cc cartridge is a reversed-phase SPE cartridge
packed with polymer-based sorbent, which was selected for this study. Compared with the
relatively hydrophilic AV-136, 4, the desired [18F]AV-45 ([18F]5) was not washed out by the
selected eluent but rather retained on the cartridge. To facilitate purification, we used
different combinations of EtOH/H2O or ACN/H2O to elute the SPE cartridge and remove the
side product, AV-136, 4. This study aimed to maximize the removal of the impurity, AV-136,
4, while retaining the desired product, [18F]AV-45 ([18F]5), on the cartridge. Afterward, the
desired product, [18F]AV-45 ([18F]5), was left on this Oasis cartridge and finally eluted by
ethanol.
This article is protected by copyright. All rights reserved.

To test the efficiency of AV-136, 4, removal from the cartridge, various combinations
of eluent were tested. As shown in Figure 2A, when the Oasis cartridge was eluted with 2 mL
of increasing ACN/H2O concentrations of 10%, 20%, 30%, 40%, and 50%, the removal of
impurity AV-136, 4, and product loss of [18F]AV-45 ([18F]5) were measured and calculated.
When ACN/H2O at 10% to 30% was selected as the eluent for rinsing the cartridge, the
removal of impurity AV-136, 4, was <10%. Meanwhile, the desired product, [18F]AV-45
([18F]5), still remained on the cartridge (the loss was <5%). When the ACN content was
increased to 40% in the eluent, more than 80% of AV-136, 4 was removed, and the loss of
product, [18F]AV-45 ([18F]5), was <10%. Moreover, as expected, the use of 50% ACN/water
as the eluent resulted in a loss of 70% of the product. Using 30%–40% ACN/H2O as the
eluent might remove most of the undesired impurity, AV-136, 4.
When the Oasis cartridge was eluted with 2 mL of either 10%, 20%, 30%, 40%, 50%,
60%, or 70% EtOH/H2O, the removal of impurity AV-136, 4 and loss of product were
calculated (Figure 2B). When the cartridge was eluted with either 10% to 50% EtOH/H2O,
the removal of impurity AV-136, 4 and the loss of product, [18F]AV-45 ([18F]5) were very
low. When the EtOH fraction was increased to 60%, the AV-136, 4 removal was
approximately 50%, and less than 10% of the product [18F]AV-45 ([18F]5) was lost. When the
EtOH fraction was increased to 70%, 21% of the product [18F]AV-45 ([18F]5) was lost.
To further optimize the SPE purification, we explored the combinations of a narrow
range of EtOH/H2O and ACN/H2O as the eluting solvent. When 8 mL of 35% ACN/H2O was
used to wash the Oasis cartridge, close to 100% of AV-136, 4, was removed, and close to
89% of the desired product was retained on the cartridge (Figure 2C). When the volume of
This article is protected by copyright. All rights reserved.

35% ACN/H2O was adjusted, 6 mL eluent afforded the best results (Figure 2D), i.e., more
than 90% of impurity was removed, and >90% of the product remained on the cartridge. With
this elution condition for SPE purification, eluting the Oasis HLB 3 cc cartridge with 6 mL of
35% ACN/H2O resulted in the removal of approximately 95% of AV-136, 4, and more than
90% of the desired product, [18F]AV-45 ([18F]5), was retained on the cartridge. The final
doses of [18F]AV-45 ([18F]5) on the Oasis cartridge were completely eluted by ethanol,
thereby resulting in a 100% pure product.
3.2.2 Optimization of fluorination temperature and deprotection conditions
Two other main chemical impurities, compounds 2 and 3, were found in the
SPE-purified product. The fluorination reaction most likely produced the Cl-derivative 2 and
the OTs-derivative, 3, (Scheme 1), which were confirmed by testing the crude reaction
mixture by a high-resolution LCT Premier XE mass spectrometer. The protonated molecular
mass of impurity A (theoretical 377.1626 Da, measured 377.1631 Da) showed a 1.2 ppm
error. In addition, the molecular formula was C20H26ClN2O4, which was consistent with the
Cl-derivative 2. The protonated molecular mass of impurity B (theoretical 513.2053 Da,
measured 513.2047 Da) showed a 1.5 ppm error, and the molecular formula was
C27H33N2O6S. This result was consistent with the OTs-derivative 3. The corresponding Cl-
and OTs-derivatives, 2 and 3, were authentically synthesized, and the concentrations of 2 and
3 in the final formulated solution were calculated by the calibration curves of standards with
HPLC. The conditions are described in Section 2.1.3. The calibration curves of Cl-derivative
2, OTs-derivative 3, AV-136, 4, and AV-45, 5, by using authentically synthesized samples
are shown in the Supporting Information (Figure S1).
This article is protected by copyright. All rights reserved.

To minimize the residual amount of OTs-derivative 3, fluorination reaction was
carried out for 10 min at 110 °C, 120 °C, 130 °C, or 140 °C. Various reaction temperatures
were tested to determine their influence on the residual amount of OTs-derivative 3, synthesis
time, the total mass of impurities, and the RCYs (decay corrected). The results are
summarized in Table 1. When the fluorination reaction temperature was increased to 110 °C–
140 °C, comparable decay corrected RCYs of [18F]AV-45 ([18F]5), remained mass of
AV-136, 4 (13–16 μg per batch) and Cl-derivative 2 (14–18 μg per batch) in the final
formulated solution were observed, while the total mass of OTs-derivative, 3, was
significantly reduced. When the fluorination reaction was carried out at ≥130 °C, no
OTs-derivative, 3, was found in the reaction mixture. Considering that the synthesis module
was not made to sustain the temperature above 120 °C, we selected 120 °C as the fluorination
reaction temperature, at which 1–2 μg of OTs-derivative 3, remained in the final dose.
As speculated, the impurity Cl-derivative 2, might be produced from the 3 M HCl
acid solution. To eliminate the production of Cl-derivative 2, which was possibly caused by
using HCl, 2 M H2SO4 was used instead for the acid hydrolysis step to remove the N-Boc
protection group. The amount of Cl-derivative 2 was significantly reduced to 1–2 μg per
batch by using 2 M H2SO4 (Table 2). No significant difference was observed in the total mass
of AV-136, 4 (13–17 μg per batch) and the OTs-derivative 3 (1–3 μg per batch) in the final
formulated product. It is worth mentioning that, in theory, when H2SO4 was used for acid
hydrolysis, without introducing chloride ion, the final formulated solution should not contain
any Cl-derivative 2. However, the experimental results showed otherwise. Traces of
Cl-derivative 2 (1–2 μg) were still observed in the final product (Table 2). It is likely that the
This article is protected by copyright. All rights reserved.

Cl sources could have been derived from either during the synthesis of O-tosylated precursor,
1, or from the [18F]fluoride ion solution derived from cyclotron. The amount of Cl-derivative
2 in the final formulation of [18F]AV-45 ([18F]5) was very low (less than 5 μg per batch). At
this low level, it might not play an important role in inhibiting the binding of [18F]AV-45
([18F]5) in the brain.
The optimized procedure for preparing [18F]AV-45 ([18F]5) was established as
follows. [18F]fluoride (1.85–3.70 GBq) was eluted from a QMA cartridge with the
K2.2.2/K2CO3 solution. The solution was azeotropically dried twice with 1 mL anhydrous
ACN under N2 gas. The O-tosylated precursor, 1, (1 mg in 1 mL of anhydrous DMSO) was
added to this reaction vessel containing the activated [18F]fluoride. The reaction mixture was
heated to 120 °C for 10 min. After the fluorination reaction, the reaction mixture was cooled
to 90 °C, and 1 mL of 2 M H2SO4 acid solution was added to remove the N-Boc group at
100 °C for 5 min. NaOH at 10 mL of 0.4 M solution was added for neutralization, and the
crude reaction mixture was subsequently loaded on an Oasis HLB 3 cc cartridge. The
cartridge was sequentially rinsed with 10 mL of deionized water first and then 6 mL 35%
ACN/H2O and 20 mL deionized water. After drying the cartridge under N2, the pure product
was eluted by 2 mL of EtOH into a vial containing 7 mL of 0.9% saline solution. The
solution was transferred to the final product vial containing 11 mL of 0.9% saline solution (a
total of 20 mL volume). The final [18F]AV-45 ([18F]5) product solution was tested for quality
control and then released (Table 3).
This article is protected by copyright. All rights reserved.

3.3 Radiosynthesis of [18F]AV-45 ([18F]5) with SPE purification on an automated synthesizer
The radiosynthesis for [18F]AV-45 ([18F]5) was tested further by using an automated
synthesizer (BIBD-F, Figure 1). The radioactivity for the preparation was increased (18F−
activity ranging from 1.85 GBq to 3.70 GBq). The RCY of [18F]AV-45 ([18F]5) was 42.7 ±
5.9% (decay corrected, n = 3), and the total synthesis time was 51 ± 2 min. Radio-HPLC and
radio-TLC analysis showed that the RCP of [18F]AV-45 ([18F]5) was >95%. The specific
activity at the end of synthesis was 254 ± 99 GBq/mg. The radiochemical identity was
confirmed by analytical HPLC with the nonradioactive standard (spike), AV-45, 5, which
showed identical retention time under the HPLC conditions described in the Methods section
(Figure 3). All of the necessary quality control steps for release criteria, including physical
characteristics, pH value, residual K2.2.2, residual acetonitrile, and DMSO of the solution
met the drug release requirements (Table 3). The total mass of chemical impurities in the
product was calculated (24.1 ± 2.7 μg per batch). In addition, the impurity concentration in
the final formulation was lower than 3 μg/mL in a total of 20 mL volume. Chemical mass
dose per injection met the purity requirement, i.e., the total amount of chemical was lower
than 50 µg as defined in the package insert of Amyvid ([18F]AV-45, [18F]5).
4. Conclusions
An improved process for preparing Amyvid ([18F]AV-45, [18F]5) by using a SPE
purification method was developed. The desired product was prepared with a high
radiochemical yield (42.7 ± 5.9%, decay corrected) and high radiochemical purity (>95%) in
51 min. Specifically, the total chemical mass in the final dose met the release criteria (total
This article is protected by copyright. All rights reserved.

chemical lower than 50 µg/dose). This modified radiosynthesis might be suitable for routine
preparation for clinical application.
Acknowledgments
This work was supported in part by grants from the National Key Research and
Development Program of China (2016YFC1306304; 2018YFC1312302), Guangdong
Province Key R&D Program (2018B030337001).
References
1.
2.
3.
4.
Molavipordanjani S, Tolmachev V, Hosseinimehr SJ. Basic and practical concepts of
radiopharmaceutical purification methods. Drug Discov Today. 2019;24(1):315-324.
Vaulina D, Nasirzadeh M, Gomzina N. Automated radiosynthesis and purification of [18F]flumazenil
with solid phase extraction. Appl Radiat Isot. 2018;135:110-114.
Mueller D, Klette I, Kalb F, Baum RP. Synthesis of O-(2-[18F]fluoroethyl)-L-tyrosine based on a
cartridge purification method. Nucl Med Biol. 2011;38(5):653-658.
Lemaire C, Plenevaux A, Aerts J, et al. Solid phase extraction - An alternative to the use of rotary
evaporators for solvent removal in the rapid formulation of PET radiopharmaceuticals. J Labelled
Compd Radiopharm. 1999;42(1):63-75.
5.
6.
Fowler JS, Ido T. Initial and subsequent approach for the synthesis of 18FDG. Semin Nucl Med.
2002;32(1):6-12.
Salloway S, Honigberg LA, Cho W, et al. Amyloid positron emission tomography and cerebrospinal
fluid results from a crenezumab anti-amyloid-beta antibody double-blind, placebo-controlled,
randomized phase II study in mild-to-moderate Alzheimer's disease (BLAZE). Alzheimers Res Ther.
2018;10(1):96.
7.
8.
Devous MD, Sr., Joshi AD, Navitsky M, et al. Test-Retest Reproducibility for the Tau PET Imaging
Agent Flortaucipir F 18. J Nucl Med. 2018;59(6):937-943.
Clark CM, Schneider JA, Bedell BJ, et al. Use of florbetapir-PET for imaging β-amyloid pathology.
JAMA. 2011;305(3):275-283.
9.
Mallik A, Drzezga A, Minoshima S. Clinical amyloid imaging. Semin Nucl Med. 2017;47(1):31-43.
Pontecorvo MJ, Devous MD, Kennedy I, et al. A multicentre longitudinal study of flortaucipir (18F) in
normal ageing, mild cognitive impairment and Alzheimer's disease dementia. Brain.
2019;142(6):1723-1735.
10.
11.
Whittington A, Gunn RN. Amyloid Load: A more sensitive biomarker for amyloid imaging. J Nucl
This article is protected by copyright. All rights reserved.

Med. 2019;60(4):536-540.
12.
Rabinovici GD, Gatsonis C, Apgar C, et al. Association of amyloid positron emission tomography with
subsequent change in clinical management among medicare beneficiaries with mild cognitive
impairment or dementia. JAMA. 2019;321(13):1286-1294.
13.
14.
Yao CH, Lin KJ, Weng CC, et al. GMP-compliant automated synthesis of [18F]AV-45 (Florbetapir F 18)
for imaging β-amyloid plaques in human brain. Appl Radiat Isot. 2010;68(12):2293-2297.
Hayashi K, Tachibana A, Tazawa S, et al. Preparation and stability of ethanol-free solution of
[18F]florbetapir ([18F]AV-45) for positron emission tomography amyloid imaging. J Labelled Comp
Radiopharm. 2013;56(5):295-300.
15.
16.
Choi SR, Golding G, Zhuang Z, et al. Preclinical properties of 18F-AV-45: a PET agent for Aβ plaques
in the brain. J Nucl Med. 2009;50(11):1887-1894.
Choi SR, Schneider JA, Bennett DA, et al. Correlation of amyloid PET ligand florbetapir F 18
(18F-AV-45) binding with β-amyloid aggregation and neuritic plaque deposition in postmortem brain
tissue. Alzheimer Dis Assoc Disord. 2012;26(1):8-16.
17.
18.
19.
20.
21.
22.
Wong DF, Rosenberg PB, Zhou Y, et al. In vivo imaging of amyloid deposition in Alzheimer Disease
using the radioligand 18F-AV-45 (Flobetapir F 18). J Nucl Med. 2010;51(6):913-920.
Liu Y, Zhu L, Plössl K, et al. Optimization of automated radiosynthesis of [18F]AV-45: a new PET
imaging agent for Alzheimer's disease. Nucl Med Biol. 2010;37(8):917-925.
Zhang W, Oya S, Kung M, Hou C, Maier D, Kung HF. F-18 Polyethyleneglycol stilbenes as PET
imaging agents targeting Aβ aggregates in the brain. Nucl Med Biol. 2005;32(8):799-809.
Zhang W, Kung M, Oya S, Hou C, Kung HF. 18F-labeled styrylpyridines as PET agents for amyloid
plaque imaging. Nucl Med Biol. 2007;34(1):89-97.
Kuge Y, Nishijima K, Nagatsu K, et al. Chemical impurities in [18F]FDG preparations produced by
solid-phase 18F-fluorination. Nucl Med Biol. 2002;29(2):275-279.
Lee SJ, Hyun JS, Oh SJ, Yu KH, Ryu JS. Development of a new precursor-minimizing base control
method and its application for the automated synthesis and SPE purification of [18F]fluoromisonidazole
([18F]FMISO). J Labelled Comp Radiopharm. 2013;56(14):731-735.
23.
24.
Boudjemeline M, Hopewell R, Rochon PL, et al. Highly efficient solid phase supported radiosynthesis
of [11C]PiB using tC18 cartridge as a "3-in-1" production entity. J Labelled Comp Radiopharm.
2017;60(14):632-638.
Fedorova O, Nikolaeva V, Krasikova R. Automated SPE-based synthesis of 16α-[18F]fluoroestradiol
without HPLC purification step. Appl Radiat Isot. 2018;141:57-63.
This article is protected by copyright. All rights reserved.

Table 1 Optimization of fluorination reaction temperature (n = 3)
Fluorination
RCY* of
[18F]AV-45 (%)
34.2 ± 2.0
Total time
(min)
Mass of
AV -136, 4 (μg)**
13.2 ± 1.25
Mass of Cl-derivative
2 (μg)**
Mass of OTs-derivative
temperature (°C)
3 (μg)**
110
45 ± 3
49 ± 6
54 ± 2
55 ± 4
17.7 ± 6.34
21.7 ± 10.6
120
30.2 ± 2.0
14.3 ± 1.78
17.2 ± 6.61
1.32 ± 2.28
130
27.5 ± 1.8
16.5 ± 0.75
14.1 ± 3.15
0
0
140
30.5 ± 3.7
14.5 ± 2.48
14.1 ± 1.25
*Decay corrected;
**Total mass per batch, amounts of chemical impurities were calculated based on calibration curves (supporting Figure S1)
Table 2 Optimization of acid hydrolysis (n = 3)
Acid for
Hydrolysis
2 M H2SO4
3 M HCl
RCY* of
[18F]AV-45 (%)
34.2 ± 4.8
Total time
(min)
Mass of
AV -136, 4 (μg)**
17.2 ± 3.71
Mass of OTs-derivative
3 (μg)**
Mass of Cl-derivative 2
(μg)**
50 ± 3
3.44 ± 1.98
2.46 ± 1.35
30.2 ± 2.0
49 ± 6
14.3 ± 1.78
1.32 ± 2.28
17.2 ± 6.61
*Decay corrected;
**Total mass per batch, amounts of chemical impurities were calculated based on calibration curves (supporting Figure S1)
Table 3 QC tests, specifications and results from three validation batches of [18F]AV-45 ([18F]5) (n = 3)
Parameter
Value
Parameter
Value
Synthesis time
51 ± 2 min
Radiochemical purity
Radionuclide identity
96.1 ± 0.5%
Radiochemical yield (EOS) *
42.7 ± 5.9%
18F (t1/2 = 111.9 ± 1.7 min)
Clear, colorless liquid;
no suspended particles
6–7
Physical characteristics
Residual DMSO
<0.5%
pH value
Residual acetonitrile
Content of K2.2.2
Chemical purity**
<0.041%
Activity concentration (EOS)
Specific activity (EOS)**
*Decay corrected
34.5 ± 3.6 MBq/mL
254 ± 99 MBq/μg
<50 μg/mL
1.21 ± 0.13 μg/mL
**Amounts of chemical impurities and the specific activity were calculated based on calibration curves (supporting Figure
S1)
This article is protected by copyright. All rights reserved.

Scheme 1 Synthesis of Cl-derivative 2 and OTs-derivative 3.
Reagents and conditions: (a) 1 M HCl, DCM, rt, 3 d
Scheme 2 Radiosynthesis of [18F]AV45 ([18F]5).
Reagents and conditions: (a) [18F]KF/K2.2.2, K2CO3, DMSO, 110–140 °C, 10 min; (b) 3 M HCl (or 2 M
H2SO4), 100 °C, 5 min.
This article is protected by copyright. All rights reserved.

FIGURE 1 Module diagram (BIBD-F) of automated radiosynthesis of [18F]AV-45 ([18F]5).
This article is protected by copyright. All rights reserved.

FIGURE 2 After loading the crude reaction mixture on to a solid-phase cartridge (Oasis HLB 3 cc), it was
eluted with different solvents. The results of a series of eluents with different solvent combinations for
solid-phase extraction (SPE): (A) Eluates: 2 mL of 10%–50% ACN/H2O; (B) Eluates: 2 mL of 10%–70%
EtOH/H2O; (C) Different ratio of ACN/H2O or EtOH/H2O; (D) different volume of 35% ACN/H2O. The
optimal combinations of eluent, maximizing removal of chemical impurity (AV136, 4) in black and
retaining the desired product, [18F]AV45 ([18F]5) in stripe, were indicated by red dotted frame.
This article is protected by copyright. All rights reserved.

FIGURE 3 Analytical HPLC profile of SPE-purified [18F]AV-45 ([18F]5) sample (undiluted into a
formulation): trace A was the radio trace of [18F]AV-45 ([18F]5, tR = 12.2 min); trace B was the UV trace
of sample (@350 nm), the black arrow represent AV-136 (4, tR = 10.5 min), the green arrow represent
AV-45 (5, tR = 11.7 min), the red arrow represent Cl-derivative (2, tR = 12.3 min), and the blue arrow
represent OTs-derivative (3, tR = 12.8 min). Analytical HPLC profile of “cold” standard AV-45, 5: trace C
was the UV trace of AV-45 (5, tR = 11.6 min, @350 nm). HPLC was acquired using a Phenomenex Luna
C18 column (4.6 mm × 150 mm, 5 μm) and UV spectroscopic detection at 350 nm with a flow rate of 1
mL/min and the following gradient: from 0 min to 2 min, A 95% and B 5%; from 2 min to 5 min, A 95%–
30% and B 5%–70%; from 5 min to 10 min, A 30%–1% and B 70%–99%; from 10 min to 15 min, A 1%–
95% and B 99%–5%; and from 15 min to 20 min, A 95% and B 5%. Mobile phase A: water with 10 mM
ammonium formate buffer and mobile phase B: ACN. The total HPLC run time was 20 min. The injection
volume was 10 μL.
This article is protected by copyright. All rights reserved.