Synthesis of [F-18]-ZD1839 as a PET Imaging Agent for Epidermal Growth Factor Receptors

Full text of DOI:10.1002/jlcr.639

JOURNAL OF LABELLED COMPOUNDS AND RADIOPHARMACEUTICALS
J Label Compd Radiopharm 2003; 46: 1–27.
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jlcr.639
Review Article
The production of a broader palette of PET
tracers
R. J. Nickles*
Medical Physics Department, University of Wisconsin, Madison WI, USA
Table of contents:
1. Introduction
1.1. Historical benchmarks
1.2 PET returns to biochemistry
2. Technical considerations in the production of proton-rich tracers
2.1. Theoretical predictions
2.2 Practical consequences affecting yields
2.3 124I as an example
2.4 Watching the pot
2.4.1 Physical measurements
2.4.1.1 Current
2.4.1.2. Target temperature
2.4.1.3 Characteristic radiation during bombardment
2.4.1.4 Characteristic radiation following bombardment
2.4.2 Chemical separation maneuvers
2.4.2.1 Dry distillation
2.4.2.2 Sparging
2.4.2.3 Wet chemistry
3. A survey of results from proton irradiation at 11 MeV
4. Conclusions
5. Full circle
6. References
*Correspondence to: R. J. Nickles, Medical Physics Department, University of Wisconsin, Room
1530 Medical Sciences Madison, WI 53706, USA.
Copyright # 2002 John Wiley & Sons, Ltd.
Received 8 May 2001
Revised 17 July 2002
Accepted 24 July 2002

2
R. J. NICKLES
1. Introduction
The application of cyclotrons to medical inquiry, accompanied by the
rise of positron emission tomography, has engaged a growing cadre of
researchers, chemists, physicists, and clinicians. They join in a quest to
bring the tracer technique to bear on the basic understanding of the
disease process. The recent clinical embrace is welcomed, albeit with it’s
faster pace driven by the scent of revenue. There is a sharpening focus
on tracers that fit into the template epitomized by [F-18]-FDG,
analogues of critical bio-compounds, trapped at a rate-limiting step to
reveal a key process. The elemental constituents of native biological
compounds leads to PET labeling with a short list of radionuclides, 11C,
13N, 15O and 18F. This focus poses the threat that radionuclides off the
main sequence (CNOF) may be orphaned, neglected due to unfami-
liarity of production techniques, or less-than-perfect decay character-
istics. This review would address that concern, casting some brief
attention to the technical aspects of making a wider array of starting
materials from which to fashion radiotracers capable of illuminating
physiological function.
1.1. Historical benchmarks
The discovery of radioactivity has passed the century mark; the
application of radioactive tracers is soon to follow. The naturally
occurring radionuclides are mostly daughters along the alpha decay
chains, throttled at their headwaters by a rate-limiting uranium or
thorium with a half-life of cosmological scale, 1010 years. The arduous
separation of the shorter-lived, and thus more active, decay products
required remarkable perseverance,1 in a day before chromatography
and solvent extraction techniques were known. Little time passed
between the isolation of radium and the application of lead-2102 in a
carefully thought-out, dynamic measurement of lead transport in
plants. Soon after, the energetic alphas from polonium were used for
transmutation experiments,3 such as
27Alða; nÞ30P
with the elemental identity of the 2-min, positron-emitting phosphorous
confirmed by the its chemical conversion to phosphine. It is truly
humbling that the pioneers of radiochemistry made such giant strides in
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

BROADER PALETTE OF PET TRACERS
3
technique and imagination, in a time prior to the discovery of the
vacuum triode which would usher in the age of electronics.
The progression of nuclear physics, and the availability of radio-
isotopic tracers, closely follows the development of accelerators. The
cyclotron4 and the electrostatic generator5,6 provided charged particle
beams with exquisite spatial and energy control, permitting the
systematic production of a wide variety of products. It is little realized
that the biological sciences made extensive use of 20-min carbon-11 in
the pre-war years, with a rich literaure7 covering topics that were not re-
discovered until 40 years later.
Post-war access to nuclear reactors gave rise to longer-lived carbon-
14, tritium, P-32 and S-35, which allowed the bench chemist to conduct
his tracer studies at a more relaxed pace. Reactors consume kilograms
of fissile fuel, making moles of neutrons with fluxes of the order of
1013–15 n/s-cm2. Thermal neutron capture cross sections are of the order
of barns, rather than mbarn for typical charged particle reactions.
Finally, target thicknesses can be g/cm2, a thousand-fold greater than
charged particle ranges. This typically results in kCi activities from
reactors, rather than mCi from most accelerators. This 6-decade scale-
up in feedstock opened the doors for the large-scale application of
radiotracers in medicine, biology, industry8 and agricultural science.9–11
Clearly, the same advantages that radionuclide imaging brings to
medicine, namely non-invasiveness, tracer authenticity and mass-less
concentrations, can bear fruit in other disciplines as well.
The last four decades of the medical application of radiotracers have
been shaped by the symbiotic development12 of imaging instruments
(e.g. the Anger camera) and the concurrent rise of agents labeled with
ideal imaging characteristics, most notably 99mTc.13 It is a credit to
technetium chemists that, starting with an element that is not native to
our solar system, such a wide array of useful radiopharmaceuticals
could be fashioned.
1.2. PET returns to biochemistry
The logical sequence that underlies PET’s appeal to the biological
scientist is grounded in three overwhelming virtues. First, its tracers are
authentic, with the pivotal elements of carbon, nitrogen, oxygen and
hydrogen respectively represented by the radioactive 11C, 13N, 15O and
the stand-in 18F. Second, the measurements are non-invasive, with
quantitative, in vivo imaging in transverse section isolation with spatial
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

4
R. J. NICKLES
resolution on the millimeter scale. And finally, the studies do not
perturb the microsystem, being possible at true tracer concentrations
below any pharmacological threshold. The high specific activities are a
direct consequence of charged particle transmutation ðDZ ¼ ꢀ1Þ and
the short physical half-lives, offsetting their logistical disadvantages.
In the conventional wisdom, the four PET tracers, 11C, 13N, 15O and
18F, form the basic inventory from which to synthesize imaging agents
to trace the body’s metabolic pathways. However, this review would
invert the cornucopia, deferring any judgment until all of the candidates
are shaken out. From this broader perspective, we then select the pearls
from those precursor radioisotopes that we can actually make.
Magnesium-28, iron-52 and copper 67 exert a siren’s song, with a
perfect fit to important measurements of physiological function. But
their inaccessibility to all but the largest accelerators places them
effectively out of reach. We should focus our attentions on those
products that could lie within our grasp, limited by fluxes in real
reactors and beams from real accelerators. Even the smallest cyclotron
dedicated to PET proves to be amazingly versatile, capable of providing
mCi levels of dozens of radioisotopes up to the rare earths. With these
tracers, we can prime the pump for radiopharmaceutical development,
provide tracers for ‘working applications’ in industry, agriculture and
nutritional applications, as well as providing labeled agents for
translational basic research.
Examples of these translational investigations include the autoradio-
graphic imaging of Rhizobial 13N2-fixation14,15 and mitochondrial
15O-oxydative phosphorylation16 where the demands for spatial and
temporal resolution surpass even the most challenging PET studies of
today.
Finally, nuclear astrophysicists study the reactions that occur in the
violence of supernovae, where a ten billion year-old star is playing out
its last hundred seconds, releasing a major fraction of its stellar mass as
radiation and alpha condensates. The ‘hot CNO cycle’ precedes this
finale, where rapid proton accretion on 12C leads to 14O. This 71-s
oxygen, bathed in a flux of energetic alphas, can enter into a breakout
reaction,
14Oða; pÞ17F
vaulting over the 16O turning point that had limited nucleo-genesis
during eons of hydrogen-burning on the main sequence. To study the
details of such reactions under controlled laboratory conditions,
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

BROADER PALETTE OF PET TRACERS
5
physicists need prodigious activities of feedstock to make radioactive
ion beams (RIB’s) in dual-accelerator experiments. It is not a
coincidence that rapid proton burning in stellar interiors pass through
the very same radionuclides that PET requires. Both fields, PET and
nuclear astrophysics, have benefited from the high yield product ion of
11C17, 13N18 and 18F19.
These examples remind us that radioactively labeled tracers have a
rich history of application in a far-wider reach of disciplines than
medicine. A clear lesson should be to keep an open mind, and welcome
these parallel fields for the leavening of spirit that they provide. The
basic ‘nuts and bolts’ discussion below runs the risk of being
sophomoric, yet needs re-visiting as the recent explosion of clinical
PET brings in those unfamiliar with the successes and mistakes of the
past.
2. Technical considerations in the production of proton rich tracers
The chart of the nuclides20 distills the essential details of stable and
radioactive isotopes, ordered in a two-dimensional plane with neutron
number N along the abscissa, and atomic number Z along the ordinate.
If we expand the chart into the third dimension to reflect the (log) half-
life for physical decay, Figure 1 provides this overview of decay
systematics, in the limited range of 05Z520. This departure from a
binary, hot-or-not, picture of radioactivity allows for a sense of nuance,
or gray scale, inviting a search for the best match between decay rate
and biological kinetics at 1015t1/25105 s.
2.1. Theoretical predictions
The binding energy systematics that determine the topography of the
valley of beta stability are embodied in the basic form of the Myers–
Swiatecki mass formula,21 balancing the attractive and disruptive
contributions to the binding energy of a charged liquid drop arising
from volume, surface, Coulomb and pairing effects. The quadratic
dependence on the atomic number Z of Equation (2)–(5) of Reference21
imparts the parabolic shape on the canyon walls.22 The physicist plays
the role of an Alpine climber, judiciously choosing his path to the
desired destination, a small ledge perched ꢁ 8 MeV above the valley
floor for each neutron dislodged.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

6
R. J. NICKLES
Figure 1 A graphic rendition of the valley of beta stability, from 0<Z<20,
showing the decay half lives in depth perspective in log scale
A useful theoretical prediction of the reaction yield is given by the
code ALICE,23 which assumes that the reaction proceeds purely
through the compound nucleus
X þ aꢂ > Cꢂ > Y þ b
in a two-step mechanism, formation of the compound nucleus, followed
by the evaporation into product particles. The first process depends on
incoming energy through its penetration of Coulomb and centrifugal
barriers. The second process is characterized by the partial widths and
level density of final states in the residual nucleus. Approximating these
variables from the Fermi gas model leads to equations in closed form,
albeit with major considerations neglected. In particular, contributions
from direct reactions are not included, nor resonance reactions that
account for the fine structure in the reaction cross section sðEÞ, defined
as the probability of the reaction occurring per target nucleus per unit
beam flux. These shortcomings limit the credibility of ALICE
predictions of absolute yields, but they still provide valuable insight
into the competition between various allowable exit channels, which
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

BROADER PALETTE OF PET TRACERS
7
dictate the radionuclide purity of the product. The smooth energy
dependence of the actual yield curve is well predicted, and can guide the
experimentalist when such landmark compendia as Landolt-Bornstein24
are silent.
2.2. Practical consequences affecting yields
Inspection of the yields of such reaction processes as
Xðp; nÞY or Xðp; aÞZ
ð1Þ
show that the cross section sðEÞ climbs steeply as the incident energy
rises above threshold, then peaks and begins a descent at ꢁ 10 MeV
above threshold, where additional exit channels (e.g. ((p,2n), (p,an))
open up to compete for the available flux in the entrance channel. After
the reaction Xða; bÞY is chosen, the best bombarding energy incident on
the target can be deduced.
This exercise in variational calculus underlines the search for the
accelerator best suited to make any given set of radionuclides. The result
determines whether multiparticle capability is essential, or what incident
beam energy E0 is needed so that the desired activity A0 can be achieved
in an acceptable irradiation time T with a realistic beam current I.
Clearly, one size does not fit all. The search must start with a realistic
institutional self perception, recognizing that there is no need to be able
to produce every radionuclide on-site. Globalization in isotope supply
means that the ratio between activity A needed at the user site and the
activity A0 produced at end of irradiation at some distant production
site is given by
A=A0 ¼ expðꢂl ꢃ range=velocityÞ
ð2Þ
For A=A051, this relation sets the range which can separate a producer
from the consumer, and explicitly involves the match between the
cyclotron’s capability and the user’s real needs. On the other hand,
A=A041 implies a negative range, a clear signal of unrealistic
expectations, such as the widespread use of 67Cu for therapy.
A figure of merit approach maximizes a rational function, represent-
ing the best ‘‘bang/buck.’’ The numerator would favor high yields of the
desired set of radionuclides. The denominator would disfavor high
startup (offset) costs and elevated operating budgets. The yields are
subject to the immutable laws of physics. The economics contain the
vagaries of local politics. The yield is given by the integral (over energy)
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

8
R. J. NICKLES
of the cross section s times the target thickness n dx, divided by the
particle stopping power,
Z
This expression clearly favors:
E0
A0 ꢁ I
ðs=dE=dxÞn dx:
ð3Þ
0
*
refractory targets that can tolerate high beam currents without
destruction,
*
reactions with a large cross section s, principally involving neutron
evaporation,
*
*
high natural isotopic abundance, favoring odd-Z targets, and
low stopping power of the incident beam, with dE=dx going as the
particle z2.
The economic denominator of the figure of merit favors naturally
abundant targets, since isotopic enrichment by electromagnetic separa-
tion generally results in material costs of the order of ten $ US per
milligram. If the beam strike covers a cm2 and target thicknesses are of
the order of 100 mg/cm2, then a thousand dollar target inventory is at
stake. Post-irradiation recycling becomes mandatory, imposing tight
constraints on the initial chemical steps. 10B and 13C are notable
exceptions, where enrichment costs are driven down ( ꢁ $100/g)
by gaseous diffusion on hydrides, satisfying the large demand that
exists.
The scaling of accelerator cost with energy is covered by Schaarf.25
Simple power laws govern the cost of cyclotron magnet steel and copper
conductor, going as mass ꢁ E3=2. A technical breakthrough, such
as the strong vertical focussing inherent in deep-valley cyclotrons,
permits a threefold reduction in the vertical dimension of the magnet
gap, with a dramatic reduction in steel, copper, power and cooling
requirements.
2.3. 124I as an example
These cost/benefit factors can be seen in an extreme case. There is an
increasing demand for 124I as a positron emitting iodine, suitable for a
wide variety of biological tracers. Two reaction pathways present
themselves:
124Teðp; nÞ124I and natSbða; nÞ124I
Since the 4.2 day half life of 124I readily permits air transport to
anywhere in the world, the production of this isotope should naturally
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

BROADER PALETTE OF PET TRACERS
9
fall to a few laboratories willing to provide the service for a fair return.
The first reaction can well employ a modern PET cyclotron26 with a
proton beam of 10–20 MeV, irradiating an enriched 124Te oxide target.
Roughly one hundred such cyclotrons populate PET centers today, but
the recent explosion of clinical studies has most sites making 18FDG for
oncology studies virtually non-stop, with precious little beam time
available for long-term irradiation of low yield reactions. Furthermore,
only a few27 are ready to invest a beam port and target inventory
( ꢁ $20/mg 124Te) that act as a ten k$ startup cost, even with rigorous
recycling.
On the other hand, many older cyclotrons, particularly in the eastern
bloc, have very serviceable alpha beams. Irradiation of natural
antimony was the original means28 of making 123I, albeit contaminated
with 124I. This base of legacy cyclotrons could embrace a new mission,
fueling the world’s needs for 124I in exchange for much-needed foreign
currency to revitalize their basic research sector. The criterion of
radionuclidic purity determines the balance point. If a user insists on
very low levels of the contaminant 126I anticipating human applications,
then thin targets or enriched 121antimony would be needed. On the other
hand, for animal studies or basic radiopharmaceutical chemistry, the
natural antimony route should suffice. Table 1 lists saturated yields on
elemental targets of 100% enrichment, as well as some of the side
constraints that influence the choice of production route. The
disadvantageous thermal properties of tellurium (melting point,
conductivity, vapor pressure) and its oxides that limit the beam power
dissipation during irradiation are evident. The advantages of a vertically
directed beam cannot be overstated in the irradiation of small samples
of solid target material with a low melting point. Even a shallow
downward deflection can suffice to keep a molten substrate from sliding
Table 1. Target properties leading to 124I
124Te(p,n) 124I
4%
121Sb(a,n) 124I
60%
Natural isotopic abundance
Cost of target material (natural) ꢁ $20/mg (95%)
ꢁ $3/mg (95%), ꢁ 0
Thick target yield @ E0
Target compositron
Melting point
Thermal conductivity
Temperature at P ¼ 1 mm Hg
Need to recycle
36 mCi/mA @ 11 MeV 17 mCi/mA(p) @ 26 MeV
TeO2 melt
4498C
3
Plated Sb on Al substrate
6308C
24
520
Essential
886
Unnecessary if natural Sb
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

10
R. J. NICKLES
down a wedge-shaped support. Innovative use29 of sloping beams and
Al2O3 stiffening additives to the 124TeO2 glassy melt have allowed
beams to reach tens of mA, but volatile losses of both the target
substrate and the iodine product are still the rate-limiting bottleneck in
the cyclotron irradiation.
The choice of production route obviously must be made on the
basis of what accelerator is available, and at what cost, with the 4 day
half-life inviting a global perspective. This type of choice reoccurs for
most of the Group VII halogens, where the production of other versatile
PET tracers (34mCl, 76Br) by proton irradiation confront very real
challenges in dealing with molten, toxic, volatile Group VI target
materials.
2.4. Watching the pot
2.4.1. Physical measurements. The effective use of the major laboratory
resource, the cyclotron beam I in the yield relation (Eq. (3)), calls for a
number of sensors to monitor the irradiation and its effects, from
immediate to more remote. This information warrants data logging, to
provide the cyclotron operator with a running commentary on the
progress of the beam, the target conditions, the resulting radiation and
the activity expected at end of bombardment.
2.4.1.1. Beam current. First and most obvious is the beam
current itself, from an accelerator that is a very close approximation
to a perfect current source. The beam on target must be measured
by a high quality electrometer with an internal impedance that
is not compromised by any leakage current from the target to
ground through spurious paths such as conducting O-rings or cooling
water traces. A more subtle source of error in target current
measurement is an imbalance of secondary electrons originating from
slits immediately upstream, and those back-ejected from the target. This
correction is assisted by re-entrant target design, fortuitous cyclotron
fringe fields or more heroic efforts, such as several kV of electrostatic
electron suppression. The charge integration of this beam current-on-
target
Z
Q ¼ i dt ðCoulombsÞ
ð4Þ
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

BROADER PALETTE OF PET TRACERS
11
simultaneously with its measurement with a high impedance electro-
meter is not trivial electronically, but straightforward with current-to-
frequency conversion and data logging.
More relevant is the ‘leaky integral’
Z
Q0 ¼ i expðꢂltÞ dt
ð5Þ
where the decay constant l corresponds to the physical decay of the
isotope being produced. Parameter Q0 is the solution to the differential
equation describing the growth of the product activity, decremented by
decay that has occurred since the beginning of bombardment. This
activity expectation can be formed analogically30 with an RC ¼ 1=l
network across the integrating capacitor, with an up-down counter31 or
by utilizing more discriminating signals than beam current, as described
below. The utility of performing this operational integration is to
maintain an accurate prediction of the activity in the target, in spite of
fluctuations in beam current iðtÞ.
The shape of the beam strike on target is readily apparent with a
beam profile monitor directly upstream, by TV telemetry on thin foils or
by the much-maligned ‘tape-burn’. The variation of the beam profile
with time, and as the beam current is raised, can be monitored by
irradiation of a segmented mosaic32 of isolated carbon rods. Finally, the
irradiation of almost any metallic foil (Cu, Havar, stainless steel) results
in induced activity that can be read out over a wide dynamic range with
a brief exposure to a BaFBr (Ce) phosphor plate, a very useful resource
in the accelerator lab.
2.4.1.2. Target temperature. The target temperature in the beam strike
can be remotely monitored by an infra-red thermocouple.33 The
advantage over a conventional TC is that being out of the beam strike,
it does not perturb the thermal ballistics of the target material. It does
require the design of a germanium IR window into the target body, and
some effort to keep the window clear of deposits of sublimed material.
The measurement of target temperature is critical to the understanding
of diffusion phenomena that underlie the transport of volatiles, of
particular importance in the release of radionuclides of very short
10
lifetime, such as 10CO2 from molten B2IO3.34
A more refined analysis of the optical spectra from gas targets has
been used by the Turku group35 to identify the electronic and
vibrational states of the target species in the beam path, useful in
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

12
R. J. NICKLES
quantitating trace impurities and excited intermediates that dictate the
hot atom fate of the recoiling product. This same group used laser
measurement36 of the index of refraction of the target gas to reveal the
temperature profile during bombardment. These measurements explain
the near-cyclone convection conditions that operate when tens of
mA ꢃ 10 MeV ꢁ hundreds of watts of beam power are dissipated in
conventional cyclotron gas targets. The measured target gas pressure p
averages over the local complexity, following the ideal gas law,
p ¼ nRT=V ¼ rRT, thus concealing the spatial dependence of the
temperature T, and density r, of the stopping material. Visually, the
luminescent Bragg peak is clearly seen at the end of range of the
‘banana-shaped’ beam path, with a transient ‘bounce’ at the instant of
beam introduction that reflects the elastic properties of the confined gas.
2.4.1.3. Characteristic radiations during bombardment. Neutrons: Neu-
trons emerge from the target chamber, signalling the reaction in
progress. In the proton irradiation of 18O, as H218O or 18O2, the neutron
flux leading to the low-lying states in 18F overwhelms any other source,
such as beam lost on slits or surfaces of the target chamber. Since one
18F is created for every neutron, the observed neutron counting rate is
the ideal input into the leaky integrator equation (Eq. (5)). Tracking the
beam current i is useful, but does not alert the operator to problems
inside the target, such as vapor voids, water loss or low enrichment,
whereas watching the neutron counting rate assures us that the desired
precursor 18F is being formed.
Two neutron detectors are convenient for this purpose, the 235U-
foiled fission proportional counter37 with a sensitivity selectable in the
10ꢂ4–10ꢂ8 cps/n/cm2 s range or the classical38 BF3-filled ‘long counter’,
with a sensitivity several decades higher. Both have essentially
quantitative neutron/gamma discrimination. The small footprint of
the fission counter is ideal for mounting within the shielding of an
enclosed cyclotron. Various efforts have been made to absolutely
calibrate the detection efficiency against the known neutron fluence
from a 241Am–Be source, so as to achieve an absolute prediction of
expected activity from Eq. (5), but errors of the order of a factor of two
persist from the differing neutron energy spectra.
The logging of the neutron counting rate is useful even if the desired
exit channel does not involve neutrons. Proton irradiation of a natural
nitrogen gas target is the source of 11carbon for a wide variety of
organic syntheses, proceeding through the 14N(p,a)11C reaction.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

BROADER PALETTE OF PET TRACERS
13
Simultaneously, the 14N(p,n)14O reaction is taking place in the target,
with a neutron signature that is perfectly suitable for monitoring the
concurrent production of 11C.
These measured variables of beam power distribution, target
temperature and neutron counting rate all enter into a detailed
description34 of the thermal ballistics of a molten 10B2O3 target,
employed in the production of 19 s 10CO2. The probability of release
eðTÞ of 10CO2 is determined by the diffusion out of the melt over a
characteristic distance x0, and is given by the Arrhenius relation.
ln ½ln ð1=eÞꢄ ¼ ln ðlx20Þ=D0 þ Ea=kT
ð6Þ
arising from the one dimensional solution to the diffusion equation.
Release efficiency reaches ꢁ 16% at 15 mA of 11 MeV protons on a
10B2O3 glassy melt at ꢁ 14008C. The log–log relationship (Eq. (6)) is
confirmed experimentally, with an enthalpy of activation fitted to
DH ¼ 590R ¼ 4:9 kJ=mol, indicating convective effects in the melt,
similar to earlier reports on iodine migration out of heated TeO2
pellets.39
Gammas: The measurement of the energy spectrum with the gamma
signature of a beam+target pair can be more problematic during
production runs, where tens of mA on target result in radiation
exposures of ꢁ 103 R/h. The intense neutron flux would quickly destroy
a germanium spectrometer, and activate a NaI detector. Both BGO and
YAP appear to be robust against the brutal conditions inside the
cyclotron vault. However, several meters of distance and extensive
shielding are needed to collimate the view of the target chamber against
the omnipresent ꢁ 8 MeV gamma flux from the (n,g) capture events
occuring in the walls of the accelerator vault. A laser-aligned BGO
detector, housed in a 250 kg lead shield, subtends ꢁ 10 msr at 3 m from
the target, with adequate energy resolution to identify the major gamma
rays emerging from the irradiation of the thick targets listed in Table 2
below. The prompt gammas Eg ðMevÞ arise predominately from inelastic
proton scattering (p,p0) and alpha reactions to low-lying excited states.
The delayed gammas Eg ðkeVÞ and the ubquitous annihilation radiation
observed in the relative calm following beam-off, come from the desired
products, or unavoidable side channels.
The prompt gamma signal is heavily contaminated by background
from the neutron capture events occurring in the laboratory walls, yet
the gamma/neutron counting rate ratio is a characteristic of the each
target/beam combination. Calibration of this ratio with modest beam
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

14
R. J. NICKLES
Table 2. Signature gamma rays during and following proton irradiation of light
targets
Target
Reaction
Eg (prompt)
Eg (delayed)
10B2O3
14N2
16O(p,p0)
7.2 MeV
2.31
4.43
7.2
5.25
719 keV (10C), 478 (7Be)
511, 2314 (14O)
511 (15O)
14N(p,p0)
15N2
15N(p,a0)12C
16O(p,p0)
H126O
H128O
511 (13N)
18O(p,a0)15N
18O(p,p0)
511 (15F)
1.98
1.63
20Ne
20Ne(p,p0)
511 (17F)
can provide a useful indicator when problems arise at higher beam
currents. In particular, proton irradiation of a solid target such as a thin
slice of elemental germanium provides useful activities of 72As, needed
for phantom work to characterize different PET scanners. Pressing the
wafer against a water-cooled aluminum cold finger, cushioned with an
indium conduction pad, permits beam currents of roughly five mA at
11 MeV before the Ge melts. The destruction of the target is signaled by
the abrupt departure of the neutron/gamma ratio from its usual value,
indicating that the beam is now no longer incident on the desired
material. This effect is heightened if the beam stop is made of gold or
tantalum, high Z materials with low neutron production.
Data logging of the neutron and gamma counting rates makes use of
myriad personal computers, equipped with counter and A/D boards. A
unified approach to this task is essential, since the same problem arises
at every turn in the PET cyclotron lab, from logging dose calibrators,
area monitors, radiation detectors in HPLC’s, GC’s and metabolite
analyzers. The natural tendency to consolidate these data logging tasks
into a single central computer may be ill-advised, since the various
activities are separated in time and scattered in different distant labs,
with a need for immediate visual feedback. A cost-effective solution in
an academic setting is to utilize retired PC’s, populated with a generic
card set40 running a common data acquisition program. A dozen such
stations are spread out over the cyclotron, chemistry labs
and PET imaging suite, with a unit cost of less than $500. Along
with extensive inter-lab cabling and intercom linkups, all members
of the group are kept aware of the progress of all activities
in a laboratory situation that could otherwise quickly degenerate into
chaos.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

BROADER PALETTE OF PET TRACERS
15
2.4.1.4. Characteristic radiation following bombardment. A wide variety
of specialized detectors have been developed to measure the presence of
radioactive products, with their sensitivities and operating features
listed in Table 3. All of the radiation detectors are research instruments
employing high quality electronics for their readout. An example of this
is the ionization chamber filled with 10 atmospheres of xenon, read out
by an electrometer44 with 10ꢂ16 ampere sensitivity. When surrounded
with a 200 kg lead shield, the advantage of this combination becomes
apparent with a low background that permits positron source activities
of 10 nCi to be quantitatively measured, with liter-sized sources in a 4p-
well geometry. The massive shielding is essential to exploit this wide
dynamic range in a high level cyclotron chemistry lab environment, with
Ci of activity nearby.
Another re-entrant detector is made from a 20 cm diameter ꢃ 30 cm
high cylinder of BC-400 plastic scintillator, with a 12 cm diameter
ꢃ 25 cm deep well, suitable for the whole body assay of a rabbit. Three
photomultipliers tubes are fanned into an electrometer for the 10 nCi-
100 mCi measurement range. A PIN diode, calibrated against a built-in
LED, extends this activity limit into the multi-Ci range.
Table 3. Process detectors
Application
Detector
Readout
Dynamic range Ref
Process control
CaWO4/PM
NAI/PM
Electrometer
NIM electronics 10 nCi–10 mCi/ml
mCi–10 mCi/ml
Quantitative assay Ar ion chamber
Xe ion chamber
Plastic well/PM
Electrometer
Electrometer
Electrometer
mCi–Ci
10 nCi–Ci
10 nCi–100 mCi 41
NaI well/PM
High purity Ge
BGO/fiber optics/PM Electrometer
NIM electronics nCi–mCi
NIM electronics nCi–mCi
mCi–Ci
Hot spot loc’n
42
CsI(T1)/PIN
Si beta det
NaI/NaI coinc
Hand held DMM mCi–Ci
NIM electronics nCi–10 mCi
NIM electronics nCi–mCi
TLC
GC
Thermal conductivity A/D
mmol/ml
Flame ionization
Electron capture
NaI/NaI coinc
UV absorption
Conductivity
A/D
Counter
NIM electronics nCi–mCi
A/D
A/D
nmol/ml
sub nmoles/ml
HPLC
nmol/ml
mmol/ml
nmol/ml
mmol/ml
Pulsed amperometric A/D
Evap light scattering A/D
Metabolite analysis Beta–gamma–gamma NIM electronics 100 pCi/ml
43
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

16
R. J. NICKLES
At the other extreme, a ‘point geometry’ detector suitable for tracking
multi-mCi bands migrating down a prep HPLC column is made by
mounting a small disc of CsI(Tl) onto a silicon PIN diode,45 light
shielded and read through its optogalvanic response with a hand-held
multimeter.46 The simplicity, low ($400) parts cost, audio signal and
bar-graph display make this detector indispensable in the PET
chemistry lab.
Flow-through detectors are constructed as needed from a crib of
NIM electronics, and detectors. Examples range from simple, shielded
NaI spectrometers distinguishing 511 keV positron annihilation radia-
tion from the signature gammas of Table 2 in passing eluates to the
elaborate triple coincidence b ꢂ g ꢂ g metabolite analyzer with its sub-
nanoCi sensitivity and vanishing (10ꢂ2 cps) background.
The anchor in radionuclide quantitative assay is the high purity,
intrinsic germanium detector, with its energy resolution (1.65 keV
FWHM@ 1332 keV) capable of resolving complex mixtures of products
and contaminants, over a wide dynamic range on a 3-m, shielded
counting stage. The efficiency, nominally 15% at 1332 keV, is carefully
calibrated at various standard geometries against NIST-traceable
standard sources, from 59 keV5Eg51886 keV. Other sources (24Na
and 152Eu) with known relative intensities are splined to extend the
efficiency calibration to 2750 keV. At lower energies, a Be-windowed
Si(Li) X-ray spectrometer covers the 5–100 keV X-ray range, although
absolute activity quantitation is much more difficult.
2.4.2. Chemical separation maneuvers
2.4.2.1. Dry distillation. Following the irradiation, the radioactive
products must be separated from the target substrate, on a time scale
that is set by the decay constant and the pace of the application. The
techniques are dictated by the scale of the activity, and the need to
conserve isotopically enriched target material. The dry distillation of
10CO2 into a helium stream from a molten 10B2O3 melt during
irradiation, the thermochromatographic separation of 94mTc from
123,
94MoO3,47,48 or
124I from TeO2,49 are ready examples of fast,
selective separation with minimal loss of the target feedstock. To its
disadvantage, the additional stopping power of the oxygen lowers the
yield, and the poor thermal properties of the oxides limit the beam
current that the target tolerates without losing the desired activity to
volatilization during the irradiation.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

BROADER PALETTE OF PET TRACERS
17
2.4.2.2. Sparging. A particularly constrained situation occurs when the
simultaneous production of a matched pair of products is desired, as for
example the oxygen pair 19O2/15O520 or the pair of short-lived kryptons,
81mK/79mK, with 13 and 55 s half-lives, respectively. Steady state
equilibrium with a constant infusion of such a pair results in a relative
concentration at any point in the body that is a measure of the tracer
‘age’ T. The limited solubility and absence of any chemical binding of
krypton make lung imaging particularly easy to interpret, with a gamma
camera well suited to form simultaneous planar images of the 190 and
127 keV gammas. The two radiotracers are sparged into a helium
stream, bubbling through a liquid bromoform target under proton
bombardment. Optimal gas transport can deliver mCi activities to a
ballast tank upstream of the subject, included in the camera’s field of
view. Steady breathing from this gas supply results in an ‘age image’.
Tðx; yÞ ¼ T ¼ ðl1 ꢂ l2Þꢂ1ln ððA10=A20Þ=ðA1=A2ÞÞ
ð7Þ
where bold type implies an ðx; yÞ image, 1 refers to 81mKr, 2 refers to
79mKr, the null superscript refers to the ballast tank average, and the
reduced decay constant ðl1 ꢂ l2Þ ¼ ln 2=17 s. The age T ranges from 23
to 15 s as a normal subject switches from normal respiration to
hyperventilation, with a remarkable uniformity from base to apex.
2.4.2.3. Wet chemistry. If wet chemistry is needed to separate the no-
carrier-added product from the tens of milligrams of isotopically
enriched target material, considerable care is needed to minimize
inventory losses in closing the entire loop from one irradiation to the
next. If the initial irradiation starts out with a metal foil, and the end
stage of the solution-extraction-precipitation process results in the
precious target stock as a small pile of oxide, then the micro-scale
lanthanum thermite reduction in vacuo51 can close the loop. Similarly,
hydrogen reduction in an induction furnace can often return metal
oxides (e.g. Fe3O4) back to a metallic pellet. Reforming a foil target for
the next irradiation can employ a rolling mill, electroplating, evapora-
tion or argon-sputtering, with their attendant losses. Two techniques
that offer near-quantitative deposition of adherent planar films of
molecular compounds, particularly multi-layer composites (e.g. Ti3
15N2/graphite), are jet vapor deposition52 and high-voltage molecular
plating from organic solvents.53
The production of the 23-min PET tracer 94mTcO4ꢂ, for kit-
preparation of any number of clinical radiopharmaceuticals, provides
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

18
R. J. NICKLES
an informative example. If the study involves laboratory animals, a
2-fold additional absorbed radiation dose from radionuclidic impurities
is not a compelling concern. In this case, 94mTc can be made from the
proton irradiation of a natural molybdenum foil.54 Electrochemical
dissolution and solvent extraction from base into methyl ethyl ketone
(MEK) takes ꢁ 10 min, with ꢁ 70% activity yield after drydown as
94mTcOꢂ4 . This is re-suspended in saline, ready for incorporation into
the appropriate kit, much as if the pertechnetate had been eluted from a
99Mo/99mTc generator.
If, on the other hand, the Tc-labeled pharmaceutical is destined for
human use,55 then the additional radiation dose is a concern. A layer of
95% enriched 94MoO3 is deposited in the bottom of a glassy carbon
crucible which forms the interior liner of the target chamber. After
irradiation, the 94mTcO4ꢂ can be separated by dry distillation at 9008C,
or by careful dissolution into a small volume of concentrated
NH4OH+H2O2, followed by solvent extraction into MEK. In this
case, the aqueous fraction is reserved, dried and fired to return the
enriched 94MoO3 for the next irradiation. In reality, recovery of greater
than 80% of the starting 94MO mass is quite difficult in the turmoil of
making a patient dose. The truth of these wet chemistry endeavors to
achieve loss-free recyling is belied by legions of small bottles, stored and
labeled, ‘94MoO3595% probably OK’.
3. A survey of yield results from proton irradiation at 11 MeV
A broad-brush approach to the measurement of over a hundred proton
induced reactions was carried out, in order to ascertain the versatility of
an 11 MeV proton cyclotron for general purpose activation. In general,
target materials were of natural abundance, and elemental in nature, if
possible. Beam current was kept low, of the order of one microamp, in
cases where either the target material or the products were volatile.
Charge integration employed a charge pump digitizer as well as
digital logging. In a low-level laboratory 50 m distant from the
cyclotron, the efficiency-calibrated high purity intrinsic germanium
spectrometer separated the signature gammas, identifying the many
products formed by (p,n), (p,a), (p,pn) and (p,an) reactions on multi-
isotopic targets. This task was made simpler with time resolved multi-
channel analysis. Figure 2 shows the results of a brief irradiation of
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

BROADER PALETTE OF PET TRACERS
19
Figure 2 A sequence of high resolution gamma spectra showing the time evolution
of the various radioisotopes of iodine produced by a short irradiation of natural
tellurium with 11 MeV protons
natural Te, with the six iodine activities evolving over two weeks
following irradiation.
In order to form a consistent basis for comparison, all measured
yields are extrapolated to end-of-saturated bombardment (EOSB), and
further extrapolated to the yields that would be expected from the
irradiation of pure elemental targets of 100% isotopic enrichment. This
facilitates the identification of those reactions that are intrinsically
advantageous, beyond the fortuitous natural abundance. Finally,
charge integration, secondary electron suppression and target transfer
contribute to the ꢀ 20% uncertainties in the stated activity A, in
Table 4 below, building on an earlier compendium.56
4. Conclusions
Several conclusions can be drawn from a re-ordering of the empirical
thick target yields of Table 4. First, at least 70 of the 100 (p,n) reactions
provide saturation yields above ꢁ 20 mCi/mA ꢁ 10ꢂ5 reaction/pro-
ton=10 ppm. This somewhat arbitrary watershed signals the effective
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

20
R. J. NICKLES
Table 4. Experimental thick target yields at Ep=11 MeV at end-of-saturated-
bombardment for elemental targets of 100% enrichment
Product
t1=2
Reaction
Abund (%)
Q(Mev)
Eg
A (mCi/mA)
7Be
10C
11C
53 d
19 s
20 m
10B(p,a)
20
20
80
99
1
99
99
0.3
100
90
0.2
3
+1.9
ꢂ4.4
ꢂ2.8
ꢂ2.9
ꢂ3.0
ꢂ5.2
ꢂ5.9
ꢂ3.5
ꢂ7.5
ꢂ4.1
ꢂ2.4
ꢂ5.0
ꢂ6.4
ꢂ6.7
ꢂ3.0
ꢂ4.7
ꢂ4.4
ꢂ0.5
ꢂ2.8
ꢂ3.8
ꢂ4.8
ꢂ1.5
ꢂ3.1
ꢂ5.9
ꢂ5.5
ꢂ2.2
ꢂ9.2
ꢂ1.3
ꢂ5.3
ꢂ1.6
ꢂ0.3
ꢂ3.1
ꢂ6.9
ꢂ3.0
+0.8
ꢂ4.7
ꢂ2.5
ꢂ4.1
ꢂ2.1
ꢂ7.9
ꢂ6.0
ꢂ1.8
ꢂ3.7
ꢂ3.0
ꢂ7.0
ꢂ5.1
ꢂ1.1
ꢂ3.3
ꢂ1.7
ꢂ0.1
477
717
511
511
511
511
2314
511
511
511
511
511
2127
2167
373
271
1157
984
511
511
1312
320
511
1434
744
835
511
931
847
122
122
811
1332
283
283
511
511
669
115
992
1039
93
170
8
93
80
120
7
2
70
1.1
14
120
116
12
5
11
1.7
2.9
40
10B(p,n)
11B(p,n)
14N(p,a)
13C(p,n)
16O(p,a)
14B(p,n)
13N
10 m
14O
15O
71 s
2 m
15N(p,n)
19F(p,an)
20Ne(p,a)
18O(p,n)
30Si(p,n)
34S(p,n)
17F
66 s
2 h
18F
30P
2.5 m
32 m
7.7 m
3.9 h
2.4 d
3.9
34mCl
38K
4
38Ar(p,n)
43Ca(p,n)
44Ca(p,n)
44Ca(p,n)
48Ca(p,n)
45Sc(p,n)
47Ti(p,n)
48Ti(p,n)
51V(p,n)
54Fe(p,a)
52Cr(p,n)
52Cr(p,n)
54Cr(p,n)
54Fe(p,n)
58Ni(p,a)
56Fe(p,n)
57Fe(p,n)
60Ni(p,a)
58Fe(p,n)
60Ni(p,n)
61Ni(p,n)
64Zn(p,a)
62Ni(p,n)
64Ni(p,n)
63Cu(p,n)
65Cu(p,n)
64Zn(p,n)
66Zn(p,n)
67Zn(p,n)
68Zn(p,n)
69Ga(p,n)
70Ge(p,n)
72Ge(p,n)
73Ge(p,n)
74Ge(p,n)
76Ge(p,n)
78Kr(p,a)
0.06
0.1
2
43Sc
44mSc
44Sc
2
48Sc
43 h
3 h
0.2
100
8
74
99
6
84
84
2.4
2.4
68
92
2
26
0.3
26
1
49
3.6
0.9
69
31
49
28
4
19
60
20
27
8
45Ti
47
47V
33 m
16 d
28 d
46 m
21 m
5.7 d
312 d
1.4 m
17 h
79 d
271 d
100
108
140
4.6
67
48V
51Cr
51Mn
52mMn
52Mn
54Mn
54mCo
55Co
56Co
57Co
15
131
2.9
1.3
77
129
5.8
150
40
58Co
60Cu
61Cu
71 d
23 m
3.4 h
76
5.2
130
73
116
228
10
105
51
177
103
45
62Cu
64Cu
63Zn
65Zn
64Ga
66Ga
67Ga
68Ga
69Ge
70As
72As
73As
74As
76As
75Br
9.8 m
13 h
38 m
244 d
2.6 m
9.4 h
78 h
68 m
38 h
53 m
26 h
80 d
1077
574
1040
834
53
596
560
285
102
60
176
41
18 d
26 h
95 m
36
8
0.3
0.664
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

BROADER PALETTE OF PET TRACERS
21
Table 4. (continued )
76Br
16 h
57 h
4.4 h
35 h
6.3 h
83 h
20 m
33 d
19 d
38 m
15 h
13 h
80 h
108 d
78 h
15 h
10 d
23 h
7 h
76Se(p,n)
9
8
50
9
11
12
57
57
17
0.6
10
7
ꢂ5.4
ꢂ2.1
ꢂ2.6
ꢂ0.9
ꢂ5.2
ꢂ1.6
ꢂ3.9
ꢂ3.5
ꢂ1.3
ꢂ7.1
ꢂ6.1
ꢂ2.9
ꢂ2.5
ꢂ4.4
ꢂ3.6
ꢂ6.9
ꢂ2.9
ꢂ0.6
ꢂ3.6
ꢂ8.8
ꢂ8.9
ꢂ5.0
ꢂ2.5
ꢂ2.4
ꢂ3.7
ꢂ7.7
ꢂ7.2
ꢂ5.0
ꢂ2.9
ꢂ2.9
ꢂ4.4
ꢂ1.5
ꢂ3.2
ꢂ1.3
ꢂ2.2
ꢂ1.0
ꢂ4.7
ꢂ1.9
ꢂ3.4
ꢂ0.4
ꢂ0.9
ꢂ2.4
ꢂ0.3
ꢂ2.0
ꢂ5.8
ꢂ5.3
ꢂ2.6
ꢂ4.7
ꢂ3.5
ꢂ2.4
ꢂ1.4
ꢂ2.4
ꢂ2.1
ꢂ1.1
560
239
616
776
776
520
248
881
1076
793
1076
381
485
898
909
2319
934
778
685
1510
871
871
204
766
778
140
833
652
1261
528
540
307
478
20
37
30
17
44
15
23
8
38
65
31
70
83
157
96
100
44
88
66
1.7
17
50
10
21
50
95
12
18
56
40
18
63
50
66
94
66
48
179
54
139
0.23
143
5.5
0.18
4.3
51
4.3
39
0.048
2.2
20
77Br
77Se(p,n)
80mBr
82Br
80Se(p,n)
82Se(p,n)
82mRb
83Rb
84mRb
84Rb
86Rb
84Y
82Kr(p,n)
83Kr(p,n)
84Kr(p,n)
84Kr(p,n)
86Kr(p,n)
84Sr(p,n)
86Y
86Sr(p,n)
87mY
87Y
87Sr(p,n)
87Sr(p,n)
7
88Y
88Sr(p,n)
83
100
51
17
3
100
15
15
9
16
16
17
10
5
89Zr
83Y(p,n)
90Nb
92mNb
96Nb
93mMo
92Tc
90Zr(p,n)
92Zr(p,n)
96Zr(p,n)
93Nb(p,n)
92Mo(p,n)
94Mo(p,n)
94Mo(p,n)
95Mo(p,n)
95Mo(p,n)
96Mo(p,n)
100Mo(p,2n)
96Ru(p,n)
98Ru(p,n)
99Ru(p,n)
99Ru(p,n)
100Ru(p,n)
101Ru(p,n)
102Ru(p,n)
103Rh(p,n)
107Ag(p,n)
109Ag(p,n)
110Cd(p,n)
111Cd(p,n)
112Cd(p,n)
113In(p,p0)
113Cd(p,n)
114Cd(p,n)
115In(p,p0)
113In(p,n)
116Sn(p,n)
116Sn(p,n)
117Sn(p,n)
118Sn(p,n)
120Sn(p,n)
122Sn(p,n)
124Sn(p,n)
121Sb(p,n)
121Sb(p,n)
123Sb(p,n)
4 m
53 m
5 h
61 d
20 h
4 d
94mTc
94Tc
95mTc
95Tc
96Tc
99mTc
96Rh
98Rh
99mRh
99Rh
100Rh
101mRh
102mRh
103Pd
107Cd
109Cd
110In
111In
112mIn
113mIn
6 h
10 m
9 m
5 h
15 d
20 h
4 d
206 d
17 d
6 h
453 d
69 m
2.8 d
21 m
1.6 h
2
13
13
13
17
32
100
52
48
12
13
24
4
12
29
95
4
15
15
8
824
88
658
245
155
392
392
725
336
79
1294
1294
159
253
1172
564
603
212
573
159
114mIn
115mIn
113mSn
116mSb
116Sb
49 d
4 h
21 m
60 m
15 m
3 h
5 h
6 d
3 d
60 d
154 d
16 d
120 d
117Sb
118mSb
120Sb
24
32
5
122Sb
124Sb
5
46
16
21
17
121mTe
121Te
57
57
43
123mTe
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

22
R. J. NICKLES
Table 4. (continued )
120I
1.3 h
13 h
4.2 d
13 d
25 m
12 h
69 s
120Te(p,n)
123Te(p,n)
124Te(p,n)
126Te(p,n)
128Te(p,n)
130Te(p,n)
127I(p,n)
0.09
0.9
4
19
32
ꢂ3.2
ꢂ2.1
ꢂ4.0
ꢂ2.9
ꢂ2.0
ꢂ1.2
ꢂ1.6
ꢂ1.3
ꢂ1.0
560
159
603
389
443
536
125
202
166
17
32
36
43
13
55
7
123I
124I
126I
128I
130I
35
127mXe
127Xe
139Ce
100
100
100
36 d
137 d
127I(p,n)
8
8
139La(p,n)
use of accelerator time. Sifting out the 59 cases of Xðp; nÞY reactions on
even-Z, even-N targets shows a tenuous Gaussian dependence on the
reaction Q-value (in MeV) peaked as ꢁ 165 mCi=mA expðꢂðQ ꢂ 3:5Þ=
2:1Þ2, if one removes the outliers at major shell closure. No such obvious
energy systematics underlie the (p,n) reactions on odd-A targets,
indicating the complexity of the many-particle states characterizing the
compound nucleus. With the exception of the very important 14N/(p,a)
doorway to 11C, most of the other eight (p,a) reactions of Table 6 have a
saturation yield of about 5 mCi/mA. Nonetheless, the naturally occuring
target abundances make these reactions attractive for the production of
7Be, 17F, 55,57Co and 61Cu.
Second, if one focusses the search for radionulcides with a real
potential in biological research, a considerable number emerge in useful
Table 5. Partial list based on biological potential
Radionuclide
Application
References
10C
Regional cerebral blood flow
rCBF, oxygen metabolism
rCBF
Electrolyte balance
Plant metabolism
Myocardial perfusion
Stent activation
Cell death
Flow and hypoxia agents
Nutrition
Monoclonal antibodies
General labeling
PET m-spheres
Monoclonal antibodies
Bridging PET and SPECT
Specialty brachytherapy sources
General iodination for PET
57
58
59
14O
17F
34mCl
45Ti
60
61
62
63
52mMn
52Mn
55Co
60,61,62Cu
63Zn
66Ga
72As
86Y
64,65
66
67
68
54,69
89Zr
94mTc
103Pd
124I
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

BROADER PALETTE OF PET TRACERS
23
activities, accessible with a small accelerator. The entries in Table 5
reflect this, listing their applications and some representative references.
5. Full circle
A collegial atmosphere pervaded the early years of PET, with cyclotron
chemists and imaging scientists free to think ‘outside the box’. This was
epitomized by such giants as Paul Harper and Katherine Lathrop at
Chicago. Basic research took precedence over ‘pretty pictures’, and a
tradition of quantitative measurement of metabolism was born.
PET has now come of age. Re-imbursement for clinical studies, and
the turf struggles that follow, threaten to cloud the scientific vision that
marked the technique’s origins. Today’s young researchers often find
that cyclotron and scanner time are dominated by commercial interests.
FDG factories serving cancer clinics are the logical fruit of PET’s
success, with a singleness of purpose that is in striking contrast to the
1970’s and 1980’s.
And yet, as commercial cyclotrons are taking over the high volume
production of a few PET tracers for routine studies, those surviving
academic cyclotrons may well experience a second Renaissance. Freed
from the daily responsibilities of making a single agent, the resources
and skilled chemists are now liberated to return to basic science, and
with it, a fresh look at the chart of the nuclides. Clearly, great strides are
being made in the understanding of the subtle kinetics of selective
neuroligands, labeled with the ‘conventional’ 11C and 18F. But it is also
clear that similar advances await the PET chemist willing to broaden his
palette of tracers.
References
1. Curie P, Sklodowska Curie M. Sur une Substance nouvelle Radio-active,
contenue dans la Pechblende. Compt Rend 1898; 127: 175.
2. Hevesy G. Radioactive indicators: Application to Biochemistry, Animal
Physiology and Pathology. Interscience: New York, 1948.
3. Joliot F, Curie I. Artificial production of a new kind of radio-element.
Nature 1934; 133: 201.
4. Lawrence EO, Livingston MS. Phys Rev 1932; 40: 19.
5. Van de Graaff RJ. Phys Rev 1931; 38: 1919.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

24
R. J. NICKLES
6. Herb RG, Parkinson DB, Kerst DW. Phys Rev 1937; 51: 75.
7. Calvin M, Heidelberger C, Reid JC, Tolbert BM, Yankwich PR. Isotopic
Carbon. Wiley: New York, 1949.
8. IAEA Technical Report Series #136, 1990.
9. Tiller KG. Applications of isotopes to micronutrient studies. IAEA/FAO
Symposium on the Use of Isotopes and Radiation in Research on Soil–Plant
relationships, Colombo Sri Lanka, Dec 1978.
10. Nye PH, Tinker PB. Local movements of solutes in soils. Solute
Movements in the Soil–Root System. Blackwell: Oxford, 1977.
11. Hanson JE. Influence and application rate on atrazine fate in coarse and
fine-textured Wisconsin soils, PhD Thesis, Univ of Wisconsin, 1997.
12. Wagner HN. Principles of Nuclear Medicine. WB Saunders and Co:
Philadelphia, 1968; 15.
13. Harper PV, Lathrop KA, Gottschalk A. In Radioactive Pharmaceuticals,
Andrews GA, Kniseley JW, Wagner HN (eds). US Atomic Energy
Commission, 1966.
14. Root JW, Krohn KA. Short-lived radionuclides in chemistry and biology.
Advances in Chemistry Series, Vol. 197, Amer Chem Soc: Washington DC,
1981.
15. Thomas J, Meeks JC, Wolk CP, Shaffer CP, Austin SM, Chien W-S.
J Bacteriol 1977; 129: 1545.
16. Nickles RJ, Kiuru AJ, Schuster SM, Holden JE. A catalytic generator for
the production of H214O and H125O. Prog Nucl Med 1978; 4: 72.
17. Powell J, Guo FQ, Haustein PE, et al. In Duggan JL and Morgan IL (eds).
BEARS: a radioactive ion beam initiative at LBNL. AIP Conference
Proceedings, Vol. 475, Applications of Accelerators in Research and
Industry, Am Inst Phys: CP475, 1999; 318.
18. Decrock P, Delbar Th, Duhamel P, Determination of the 13N(p,g)14O
reaction cross section using a 13N radioactive ion beam. Phys Rev Lett
1991; 67: 808.
19. Rehm KE, Paul M, Roberts A, et al. The astrophysical reaction rate for
the 18F(p,a)15O reaction. Phys Rev C 1996; 53: 438.
20. General Electric Co, Chart of Nuclides (12th edn). Schenectady: New
York, 1977.
21. W.D. Myers and W.J. Swiatecki, Nuclear masses and deformation. Nucl
Phys 1966; 81: 1.
22. Friedlander G, Kennedy JW, Macias ES, Miller JM. Nuclear and
Radiochemistry. (3rd edn). Wiley: New York, 1981; 41.
23. Blann M. Corrections to Overlaid ALICE. University of Rochester,
Nuclear Structure Research Lab, UR-NSRL-181, 1978.
24. Landolt-Bornstein, Numerical Data and Functional Relationships in
Science and Technology, New Series, Vol. 5. Hellewege K-H (ed),
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

BROADER PALETTE OF PET TRACERS
25
Q-values and Excitation Functions of Nuclear Reactions, Part b, Keller
KA, Lange J, Munzel H, Pfennig G (eds), Springer: Berlin, 1973.
25. Schaarf W. Particle Accelerators. Applications in Technology and Research.
Research Studies Press, Ltd., Wiley: New York, 1989.
26. Scholten B, Kovacs Z, Tarkanyi F, Qaim SM. Excitation functions of
124Te(p,xn) 124,123I reactions from 6–31 MeV with special reference
to the production of 124I at a small cyclotron. Appl Rad Isotopes 1995;
46: 225.
27. Sheh Y, Koziorowski J, Balatoni J, Lom C, Dahl JR, Finn RD. Low
energy cyclotron production and chemical separation of ‘no-carrier-added’
iodine-124 from a Reusable enriched tellurium-124 dioxide/aluminum
Solid Solution Target. Radiochim Acta 2000; 88: 169.
28. Watson IA, Waters SL, Silvester DL. Excitation functions for the reactions
producing 121I, 123I and 124I from Irradiation of Natural Antimony with
4
3He and He Particles with Energies up to 30 MeV. J Inorg Nucl Chem
1973; 35: 3047.
29. Reischl G, Roesch F, Machulla HJ. Production of iodine-124 with a new
target system at a low energy cyclotron. J Nucl Med 2000; 41: 252P.
30. Snowdon SC. Excitation curve for protons in the reaction 19F(d,p) 20F.
Phys Rev 1950; 78: 299.
31. Nickles RJ and Hichwa RD. A digital leaky integrator. Nucl Instr Meth
1979; 158: 609.
32. Hichwa Rd, Watkins GL, Ponto LL, Cyclotron produced 18F fluoride: a
target’s view of the beam. J Nucl Med 2001; 42: 18P.
33. Excergen Corp., Watertown, MA.
34. Schueller MJ. An investigation of 10CO2 and H214O as cerebral blood flow
tracers. PhD Thesis, University of Wisconsin, 2001.
35. Aho K, Light emission during cyclotron bombardment of gases used in
radionuclide production. PhD Thesis, Abo Akademi, Abo, 1998.
36. Solin O, Interaction of high energy charged particles with gases, PhD
Thesis, Abo Akademi: Abo, 1988.
Heselius S-J, Studies of density reduction in gas targets, PhD Thesis, Abo
Akademi: Abo, 1986.
37. LND Inc., 3230 Lawson Blvd. Oceanside: New York, 11572.
38. Knoll GF. Radiation Detection and Measurement (2nd edn). Wiley: New
York, 1979; 487.
39 Beyer G-J, Pimental G. Physiocochemical and radiochemical aspects of
separation of radioiodine from TeO2-targets. Radiochim Acta 2000; 88: 175.
40. Keithley Instruments Inc., 28775 Aurora Rd., Cleveland OH 44139. Model
CTM-05/A and DAS-801.
41. Nickles RJ, Votaw JR. An extended range well counter. IEEE Trans Nucl
Sci 1985; 32(1): 60.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

26
R. J. NICKLES
42. Nickles RJ, Solin OH. A wide-range flexible gamma probe. Appl. Radiat
Isotopes 1992; 43: 1169.
43. Nickles RJ, Solin OH, Harparaanta M, DeJesus OT. A flow-through
detector for nano-curie activities encountered in HPLC analysis of PET
tracer metabolites. IEEE Trans Nucl Sci 1992; 39: 2316–2321.
44. Keithley Instruments, Inc., 28775 Aurora Rd, Cleveland OH 44139, Model
6517A.
45. Hamamatsu Photonics Corp, 360 Foothill Road, PO Box 6910, Bridge-
water, NJ 08807, Model S2281-01.
46. Fluke Corp., PO Box 909-, 6920 Seaway Blvd, Everett WA 98203, Model
79.
47. Nickles RJ, Christian BT, Nunn AD, Stone CK. Cyclotron production of
high purity Tc-94m by in situ sublimation. J Label Compd Radiopharm
1993; 32: 447–448.
48. Roesch F, Novgorodov AF, Qaim SM. Thermochromatographic separa-
tion of 94mTc from enriched molybdenum targets and its large scale
production for nuclear medicine application. Radiochim Acta 1994; 64:
113.
49. Beyer GJ, Pimental G, Solin O, et al. Production of 123I for medical use
with small accelerators. Isotopenpraxis 1988; 24: 297.
50. Nickles RJ and Au YF. The oxygen clock, a dual tracer physiological
timer. Phys Med Biol 1975; 20: 54.
51. Cotton FA, Wilkinson G. Advanced Inorganic Chemistry (4th edn). Wiley:
New York, 1980; 989.
52. Zhang J-Z, Boorse SR, Halpern BL, Schmitt JJ, Gai M. Novel targets for
radioisotope generation made by the jet vapor deposition process. In Link
JM, Ruth TJ (eds). Proceedings of the Sixth Workshop on Targetry and
Target Chemistry, Vancouver, 1995; 147.
53. Friedlander G, Kennedy JW, Macias ES, Miller JM. Nuclear and
Radiochemistry (3rd edn). Wiley: New York, 1981; 290.
54. Nickles RJ, Nunn AD, Stone CK, Christian BT. Technetium-94m-
Teboroxime: synthesis, dosimetry and initial PET imaging studies. J Nucl
Med 1993; 34: 1058.
55. Stone CK, Christian BT, Nickles RJ, Perlman SB. Technetium-94m
methoxy isobutyl isonitrile: dosimetry and resting cardiac imaging with
positron emission tomography. J Nucl Cardiol 1994; 1: 425.
56. Nickles RJ. Production of a broad range of radionuclides with an 11 MeV
proton cyclotron. J Label Compd Radipharm 1991; 30: 120.
57. Law I, Jensen M, Holm S, Nickles RJ, Paulson OB, Using 10CO2 for single
subject characterization of the stimulus frequency dependence in visual
cortex – a nover PET tracer for human brain mapping. J Cereb B1 Flow
Metab 2001; 21: 1003–1012.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27

BROADER PALETTE OF PET TRACERS
27
58. Nickles RJ, Jensen M, Paans AMJ, Holm S, Schueller MJ, Vaalburg W.
Doppler broadening revealing the chemical environment following oxygen-
14 decay. Radiochimica Acta 2000; 88(3–4), 189–193.
59. Roberts AD, Dabbs KA, Schueller MJ, Dick DW, Nickles RJ. Production
of the short-lived flow tracer fluorine-17 fluoromethane from proton
irradiation of neon. J Nucl Med 2000; 41: 243P.
60. Saitoh K, Inoue K, Kawamura M, Ishiwata K, Ido T, Kimura S. Studies
on the relationship between the biodistribution of 45Ti–AsA in the rat
kidney and renal function. Cyric Annual Report, 1987; 182.
61. Daube ME, Nickles RJ. Comparison of positron-emitting myocardical
perfusion agents. J Nucl Med 1983; 24: 119.
62. Hehrlein C, Gollan C, Donges K, et al. Low-dose radioactive endovascular
stents prevent smooth muscle cell proliferation and neointimmal hyper-
plasia in rabbits. Circ 1995; 92: 1570.
63. Jansen HML, Pruim J, van de Vliet AM et al. Visualization of damaged
brain tissue after ischemic stroke with cobalt-55 positron emission
tomography. J Nucl Med 1994; 35: 456.
64. Green MA, Mathias CJ, Welch MJ, et al. Copper-62-labeled pyruvalde-
hyde Bis(N4-methylthiosemicarbazonato) copper (II): synthesis and
evaluation as a positron emission tomography tracer for cerebral and
myocardial perfusion.
65. Martin CC, Oakes TR, Nickles RJ. Small cyclotron production of Cu-60
PTSM for PET blood flow measurements. J Nucl Medi 1990; 31: 815.
66. Koziorowski J, Finn RD, Pentlow KS, et al. Preliminary results on the
production of 66Ga, a non-conventional positron emitter. Eighth Work-
shop on Targetry and Target Chemistry, St. Louis, 1999.
67. Brockman J, Roesch F, Herzog H, Stolz B, Bruns C, Stocklin G. In vivo
uptake kinetics and dosimetric calculations of 86Y-DTPA-octreotide with
PET as a model for potential entotherapeutic octreotids labelled with 90Y.
J Lab Compd Radiopharm 1995; 37: 519.
68. Link JM, Krohn KA, Eary JF, et al. 89Zr for antibody labelling and
positron tomography. J Label Compd Radiopharm 1986; 23: 1296.
69. Qaim, SM. Production of high purity 94mTc for positorn emission
tomography studies. N Med Biol 2000; 27: 323.
Copyright # 2002 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2003; 46: 1–27