Applied
`Radiation and
`Isotopes
`
`PERGAMON
`
`Applied Radiation and Isotopes 50 (1999) 1015–1023
`
`Manufacture of strontium-82/rubidium-82 generators and
`quality control of rubidium-82 chloride for myocardial
`perfusion imaging in patients using positron emission
`tomography
`
`Teresa M. Alvarez-Diez a, b, *, Robert deKemp b, Robert Beanlands b,
`John Vincent c
`
`aNuclear Medicine Division, Ottawa Civic Hospital, 1053 Carling Ave., Ottowa, Ont., Canada K1Y 4E9
`bOttawa Heart Institute, 1053 Carling Ave., Ottowa, Ont., Canada K1Y 4E9
`cTRIUMF, 4004 Westbrook Mall, Vancouver, B.C., Canada V6T 2A3
`
`Received 13 July 1998; received in revised form 5 September 1998; accepted 8 September 1998
`
`Abstract
`
`We describe a protocol to manufacture 82Sr/82Rb generators and 82RbCl for myocardial imaging with PET. The
`generators are manufactured in 3 stages: (1) preparation of a tin oxide column, (2) leak test of the generator column
`and (3) loading of the generator with 82Sr. The generators produced sterile and non-pyrogenic 82RbCl for i.v.
`injection. No significant 82Sr/85Sr breakthroughs were observed after elution with 20 l of saline. The automated
`system delivered human doses of 82RbCl accurately. # 1999 Elsevier Science Ltd. All rights reserved.
`
`1. Introduction
`
`Conventional diagnostic techniques used to assess
`coronary artery disease and its severity are of limited
`sensitivity. Stress SPECT (single photon emission com-
`puted tomography) using thallium-201 (201Tl) or tech-
`netium-99m-sestamibi (99mTc-sestamibi) has been used
`extensively in clinical practice to determine myocardial
`perfusion. However, this technique is limited by at-
`tenuation artifacts and does not permit in some cases
`the accurate distinction between hypoperfused but
`viable myocardium and infarcted tissues, thus underes-
`timating the areas of viable tissue (Wackers et al.,
`1976; Cloninger et al., 1988; Galli et al., 1988).
`Cardiac positron emission tomography (PET) using
`several short half-life (t1/2) radionuclides has been used
`
`* Corresponding author. Present address: 495 Canotia
`Drive, Orleans, Ont., Canada K4A 2J4. Tel.: +1-613-830-
`4733; e-mail: teresa.ad@sympatico.ca
`
`to characterize myocardial perfusion and metabolism
`non-invasively (Bergmann et al., 1985; Goldstein et al.,
`1986; Brunken et al., 1987; Camici et al., 1989; Demer
`et al., 1989; Saha et al., 1992). It has been demon-
`strated that rubidium-82 (82Rb), a positron emitter
`(t1/2=75 s)
`radionuclide with ultra-short half-life
`(Woods et al., 1987), permits the assessment of myo-
`cardial perfusion with high sensitivity and specificity
`(Goldstein et al., 1983; Gould et al., 1986, 1988; Go et
`al., 1990; Stewart et al., 1991; Grover-McKay et al.,
`1992). The accuracy of 82Rb PET has been shown to
`be superior to 201Tl SPECT imaging (Go et al., 1990;
`Stewart et al., 1991). This is especially important for
`the detection of early coronary artery disease and for
`the evaluation of therapy designed to protect or sal-
`vage myocardium. Some clinical studies have also
`demonstrated the utility of 82Rb as a quantitative mar-
`ker of myocardial necrosis/viability (Gould et al.,
`1991; Dahl et al., 1996).
`There are two other advantages of 82Rb. It has a
`ultra-short t1/2 which allows the sequential perform-
`
`0969-8043/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
`PII: S 0 9 6 9 - 8 0 4 3 ( 9 8 ) 0 0 1 7 0 - 5
`
`JUBILANT EXHIBIT 1012
`Jubilant v. Bracco, IPR2018-01449
`
`

`

`1016
`
`T.M. Alvarez-Diez et al. / Applied Radiation and Isotopes 50 (1999) 1015–1023
`
`ance of scans every 10 min and reduces the exposure
`of the patients to radiation. Secondly, 82Rb is a PET
`radiopharmaceutical which is generator-produced from
`its parent radionuclide strontium-82 (82Sr). This makes
`it feasible for institutions to participate in investiga-
`tional and clinical PET studies without the need to
`have expensive on-site cyclotrons.
`Several studies have been reported in the literature
`about
`the manufacture of strontium-82/rubidium-82
`(82Sr/82Rb) generators. Neirinckx et al., 1982 demon-
`strated the e(cid:129)ciency of hydrous tin oxide, a cationic
`exchanger material, for the separation of 82Rb from
`82Sr in a column. A special
`issue of
`this journal
`(Waters and Coursey, 1987) contained 11 papers deal-
`ing with 82Sr production and the physics and chemistry
`of the 82Sr/82Rb generators and 2 papers dealing with
`the clinical applications of 82Rb. However, much of
`these works were taken from on-going projects and
`there are in several cases discrepancies reported. Also,
`our group had previously described the development
`of a 82Rb generator system while focusing on the safe
`82Sr from metallic rubidium targets
`production of
`(Cackette et al., 1993). However, the 82Rb eluted from
`these generators is not suitable for clinical studies but
`only for research applications. In addition, a detailed
`protocol for the manufacture of 82Sr/82Rb generators
`from simple components that include quality control
`procedures and specifications of 82Rb for clinical ima-
`ging has not yet been reported in the literature.
`Thus, we describe in this manuscript a novel and
`simple manufacturing protocol which include the qual-
`ity control procedures for the production of 82Sr/82Rb
`generators and 82Rb chloride doses for i.v. adminis-
`tration in patients. Other nuclear pharmacies could
`produce generators in a similar fashion and participate
`in clinical PET studies. We discuss some of the factors
`that determine the performance of 82Sr/82Rb genera-
`tors for clinical imaging: (1) the type and properties of
`the cationic exchanger material used in the manufac-
`ture of the generator column; (2) the tightness of the
`cationic exchanger packing and (3) the type of eluent.
`We also report the quality control tests and specifica-
`tions of 82Rb chloride suitable for clinical myocardial
`imaging.
`
`2. Materials and methods
`
`All components of the generator column and the
`wrenches used during manufacturing were thoroughly
`washed with laboratory soap and water, rinsed with
`sterile water and then autoclaved. Soaking of the gen-
`erator components in soap and water for at least 20
`min removes the traces of lubricant used during its
`manufacturing. Rinsing the stainless steel components
`with a diluted solution of hydrochloric acid before
`
`autoclaving is not recommended because we observed
`that HCl with repeated autoclaving accelerates the oxi-
`dation of the components. The bu€ers and solutions
`used during the manufacturing of the generator col-
`umn are sterile and pyrogen free. 82Sr/82Rb generators
`are manufactured under aseptic conditions in 3 stages.
`
`2.1. Manufacturing of the generator column
`
`2.1.1. Preparation of hydrous tin oxide
`The generator column is prepared in a clean manu-
`facturing room dedicated to the compounding of
`radiopharmaceuticals. An excess of hydrous tin oxide
`(International Tin Research Institute, Middlesex,
`England) was sieved with a 150 mm stainless steel sieve
`for 10 min. The sieved tin oxide (with particle size
`<150 mm) was then sieved thoroughly with a 75 mm
`sieve in order to separate the fines (tin particles of
`<75 mm) from the tin oxide particles with sizes in the
`interval of interest, 150 mm>x>75 mm. The fines were
`discarded. The tin oxide, approximately 3.5 g, was
`washed with 30 ml of 0.1 N NH4OH/NH4Cl, pH 10
`and incubated overnight with 10 ml of this bu€er in
`order to activate its cationic exchanger properties. The
`tin oxide was kept in the bu€er of incubation until
`loaded in the column subassembly.
`
`2.1.2. Preparation of the generator column assembly
`The generator column assembly consists of two 9.5–
`1.5 mm Swagelok reducing adaptors with nuts and fer-
`rules, one column and two 25 mm filters (frits) (Fig. 1).
`The dimensions of the generator column are: 2.6 cm
`length, 6 mm internal diameter and 0.5 mm wall thick-
`ness. All components are made of stainless steel type
`316. A generator column subassembly is first prepared
`by attaching the column to a reducing adaptor con-
`taining a 25 mm filter. This column subassembly is
`then loaded with about 3.5 g of the wetted a-hydrous
`tin oxide (Sn2O(cid:1)xH2O where x = 1–2) in 10 ml of 0.1
`N NH4OH/NH4Cl bu€er. A 3-cm Teflon reservoir
`containing 0.1 N NH4OH/NH4Cl is connected to the
`upper end of the generator column and a vacuum
`aspirator to the lower end to facilitate the packing of
`the tin oxide into the column. It is important not to
`draw air through the column. An alternative method
`that facilitates the packing of the tin oxide column is
`to gently shake the reducing adaptor of the loaded col-
`umn with a small electrical vibrator (i.e. an engraver).
`When the packing is finished (the resin is leveled with
`the top of the column), a 7.9 mm Swagelok capping
`nut (plug) is screwed onto the column outlet to prevent
`fluid drainage. The column assembly is then completed
`by attaching a second Swagelok reducing adapter con-
`taining a 25 mm filter to the upper end of the column.
`The Swagelok nut is tightened one full turn past finger
`tight.
`
`

`

`T.M. Alvarez-Diez et al. / Applied Radiation and Isotopes 50 (1999) 1015–1023
`
`1017
`
`passing a few ml of the eluent, will indicate that the
`tin oxide of the generator column is in the Na+ form
`and that it is successfully working as a cationic exchan-
`ger material, exchanging some of its Na+ for the H+
`of the eluent. If the pH of the eluate is outside this pH
`interval, the generator column is discarded.
`After this test has been performed the hydrous tin
`oxide is replenished with sodium cations by repeating
`the 2 M NaCl flush.
`
`2.2. Generator assembly and leak test
`
`The generator inlet and outlet lines used are pre-
`formed 1.5 mm stainless steel tubes of 7 and 15.5 cm,
`respectively (Triumf, Vancouver, Canada) (Fig. 1). The
`outlet line is filled with 0.9% NaCl and attached to the
`column outlet using a 7.9 mm Swagelok nut and fer-
`rules. The nut is tightened 3/4 of a turn past finger
`tight. The inlet line is attached to the generator column
`inlet in the same way. This assembly is then attached
`to the lid of the shielding body (Fig. 2). The generator
`
`Fig. 1. Physical components of the strontium-82/rubidium-82
`generator column and infusion line used during generator’s
`manufacture. The components (1–11) are made of stainless
`steel type 316. The generator column assembly consistent of
`two 9.5–1.5 mm Swagelok reducing adaptors (4) with nuts
`and ferrules (1–3), one column (6), two 25 mm filters (frits) (5)
`and the generator inlet (7) and outlet (11) lines which are
`attached to the column using 7.9 mm Swagelok nuts (8) and
`ferrules (9–10). The Teflon infusion line (12) is attached to
`the syringe pump (luer end) and to the column inlet in order
`to infuse the bu€ers through the hydrous tin oxide column.
`
`The column backpressure is measured by passing ap-
`proximately 120 ml of 0.1 N NH4OH through the col-
`umn at flow rates of 10 and 20 ml/m using a syringe
`connected to a syringe pump and an in-line pressure
`gauge.
`The tin oxide exchanger is then saturated with
`sodium cations by passing 120 ml of 2 M NaCl
`through the column at a flow rate of 0.5 ml/m fol-
`lowed by 500 ml of 0.9% saline at a flow rate of 10
`ml/m.
`The cationic exchanger properties of hydrous tin
`oxide were tested by passing through the column
`about 10 ml of sterile water for irrigation USP that
`was acidified with 0.1 N HCl to pH 4. A rapid shift on
`the pH of the eluate from pH 6 to pH 9 to 10, after
`
`Fig. 2. This figure shows the generator column assembled,
`packed with hydrous tin oxide and attached to the lid of the
`shielding body. Both shielding lid and body (not shown) con-
`tain depleted uranium and weigh approximately 17 kg.
`
`

`

`1018
`
`T.M. Alvarez-Diez et al. / Applied Radiation and Isotopes 50 (1999) 1015–1023
`
`lines are passed through the hexagonal holes in the
`shielding lid until the lines protrude at least 1 cm
`above the surface of the lid. Female threaded luer fit-
`tings are attached to the inlet and outlet lines with 7.9
`mm Swagelok nuts and ferrules. The nut is tightened
`1/4 of a turn. The outlet line is then capped with a
`male luer plug. The inlet
`line of
`the generator is
`attached to a research purity argon gas cylinder using
`a 7.9 mm Swagelok nut. The generator column is then
`leak tested by passing a static pressure of 50 psi of
`argon. Leaks are detected by immersing the generator
`column into a 1 l beaker containing a su(cid:129)cient volume
`of sterile water to cover the column assembly (Fig. 3).
`If no leaks are observed, this assembly is detached
`from the argon gas cylinder and the generator column
`inserted into the shielding body. The shielding body is
`a cylinder of 2.8 cm internal diameter, 3 cm wall thick-
`ness and 15.5 cm length (not shown). Both shielding
`lid and body contain depleted uranium and weigh ap-
`
`Fig. 3. Leak test of the generator column. The generator col-
`umn is leak tested by capping the outlet line of the column
`and passing a static pressure of 50 psi of argon through the
`column. The generator column is immersed in a beaker con-
`taining 1 liter of sterile water. If a leak occurs, argon bubbles
`will be observed in the water.
`
`proximately 17 kg in total. The shielding lid is secured
`to the body using lever clamps.
`Spent generators are disassembled and the columns
`stored for decay. Generator shielding assemblies are
`reused to manufacture subsequent generators.
`
`2.3. Loading 82Sr into the generator column
`
`Loading of the generator with 82Sr is performed
`inside a lead castle in the Radiopharmacy laboratory.
`This castle has a front and back wall thickness of 7.5
`cm. Mirrors placed on the inside front and back walls
`of the castle provide an inside view. Lead bricks at the
`top of the side walls can be removed to enable manual
`access during the manufacturing procedure.
`The arrangement for loading the generator with 82Sr
`is shown in Fig. 4. It consists of the assembled genera-
`tor column, a 30 ml vial containing the 82Sr stock sol-
`ution (MDS Nordion, Vancouver, Canada), a syringe
`isolator and a waste container. The stock vial, syringe
`isolator and the waste container are shielded with lead.
`The syringe isolator (Triumf, Vancouver, Canada) con-
`sists of a 20 ml syringe attached by its plunger to the
`plunger of a 60 ml syringe. The plunger of the 60 ml
`syringe was modified to permit attachment of a 20 ml
`syringe. The 60 ml syringe of the isolator is connected
`to a remote 60 ml syringe which can be operated using
`a syringe pump.
`50 mCi of 82Sr in 0.1 N HCL with a concentration
`of <50 mCi/ml (MDS Nordion, Vancouver, Canada)
`was mixed with 15 ml of 0.5 M Tris bu€er pH 7.5.
`Using the remote syringe the 82Sr was withdrawn into
`the shielded syringe of the isolator with a 19G sterile
`spinal needle. The 82Sr solution is then pumped from
`the isolator through the generator column at a flow
`rate of 2 ml/h. After the generator column has been
`loaded with 82Sr, the column is purged with 500 ml of
`0.9% NaCl at a moderate flow rate of 0.5 ml/m.
`Washing the column with a large volume of saline is
`intended to remove any possible impurities contained
`in the 82Sr stock solution.
`
`2.4. Quality control of the 82Sr/82Rb generators
`
`The 82Sr/82Rb generators were eluted with sterile
`and pyrogen free 0.9% NaCl. The initial quality con-
`tests performed on the 82Rb chloride eluate
`trol
`included: visual
`inspection, pH measurement, radio-
`nuclide purity (82Sr and 85Sr breakthrough), chemical
`purity, sterility and pyrogen tests.
`The radionuclide purity of a sample of 82Rb was
`measured by quantitative gamma spectrometry using a
`calibrated and certified multi-channel analyzer (Tracor
`Northen, Middleton, Wisconsin, USA) with an intrin-
`sic germanium lithium detector (Ge(Li)) and computer
`analysis (AccuSpec). (The multi-channel analyzer was
`
`

`

`T.M. Alvarez-Diez et al. / Applied Radiation and Isotopes 50 (1999) 1015–1023
`
`1019
`
`Fig. 4. Arrangement for loading strontium-82 into the generator column. It consists of the assembled generator column (1), a vial
`containing the strontium-82 chloride solution (2), a syringe isolator (3) and a waste container (4).
`
`certified by the Institute for National Measurements
`Standards; National Research Council, Ottawa,
`Canada). The amounts of 82Sr and 85Sr in the eluate
`were measured relative to eluted yield of 82Rb. The
`82Sr/82Rb ratio limit established was 0.02 mCi/mCi.
`The limit of the 85Sr/82Rb ratio was 0.2 mCi/mCi. The
`elution yield of 82Rb was also measured with a dose
`calibrator and corrected for decay of the measured
`82Rb activity to the time of loading.
`A sample of 82Rb was sent to an independent lab-
`oratory for trace metal analysis by inductively coupled
`plasma/atomic emission spectroscopy (ICP/AES).
`A sterility test of 82Rb was performed by inoculation
`in trypticase soy broth and thioglycolate media for the
`detection of aerobic and anaerobic microorganisms.
`The absence in the eluate of pyrogenic substances
`was established using the limulus amebocyte lysate
`(LAL) test.
`82Sr and 85Sr breakthrough measurements using a
`dose calibrator and pyrogen tests of 82Rb were also
`performed daily before administration to the patients.
`The 82Sr and 85Sr breakthroughs relative to the peak
`of 82Rb radioactivity eluted were measured in 20-ml
`samples of 82Rb eluted at a rate of 20 ml/m and which
`were allowed to decay for 1 h after elution. 30 ml of
`the first elution of 82Rb is discarded every day. 82Sr
`breakthrough was
`calculated using the
`following
`
`equations:
`
`82Sr breakthrough (cid:136) 82Sr
`82Rb
`
`(cid:133)1(cid:134)
`
`where 82Sr is the amount of 82Sr in the sample in mCi
`calculated from Eq. (2); 82Rb is the peak of 82Rb
`radioactivity in mCi measured at
`the end of
`the
`elution.
`
`82Sr (cid:136) Radioactivity at one hour
`1 (cid:135) 0:48 (cid:2) R0
`
`:
`
`(cid:133)2(cid:134)
`
`The numerator of Eq. (2) is the radioactivity of the
`sample in mCi measured in a dose calibrator (on the
`82Rb and/or 82Sr setting) at least 1 h after elution (at
`complete decay of 82Rb) and R0 is the 85Sr/82Sr ratio
`on the date of the measurement, calculated from the
`manufacturer specifications. The correction factor 0.48
`is used to compensate for the contribution of the 85Sr
`to the reading. The 85Sr breakthrough was calculated
`by multiplying the 82Sr (mCi) by R0.
`85Sr breakthrough (cid:136)82 Sr Breakthrough (cid:2) R0:
`
`(cid:133)3(cid:134)
`
`A sterility test of a sample of 82Rb was also performed
`every day retrospectively.
`
`

`

`1020
`
`T.M. Alvarez-Diez et al. / Applied Radiation and Isotopes 50 (1999) 1015–1023
`
`2.5. 82Rb automated infusion system
`
`When the initial quality control of the 82Sr/82Rb
`generator is finalized, the generator is connected to an
`automated delivery system (Fig. 5). Automated intra-
`venous (i.v.) infusion of 82Rb chloride eluted from the
`generator into the patients is required due to the short
`half-life of 82Rb (75 s). The delivery system is a com-
`puter controlled peristaltic pump with direct monitor-
`ing of the 82Rb radioactivity eluted from the 82Sr/82Rb
`generator. The primary function of the delivery system
`is to accurately deliver a patient dose of 82Rb during
`the infusion time. This is achieved by varying the infu-
`sion rate of 82Rb to a maximum of 50 ml/m and by
`monitoring the 82Rb radioactivity eluted from the 82Sr/
`82Rb generators. The infusion system also records the
`actual volume of 82Rb administered (<100 ml) as well
`as the infusion backpressure and administered dose
`profiles over time. The backpressure from the genera-
`tor column is limited to 30 psi. An emergency stop
`button can be depressed in the event of patient distress
`or a mechanical malfunction. Sterile tubing and fittings
`are used throughout. A disposable sterile i.v. infusion
`line with a one-way valve on each end is used for each
`
`patient. A sterilizing 0.22 mm Millipore filter is also
`attached to the patient valve outlet line. A bypass line
`permits the flushing of the remaining 82Rb radioac-
`tivity in the patient infusion line, reducing background
`radiation.
`
`3. Results and discussion
`
`The adsorption of 82Sr into the generator column
`was greater than 99.99%.
`82Sr and 85Sr breakthroughs measured by gamma
`spectroscopy were lower than the established limits;
`82Sr/82Rb ratio was <0.02 mCi/mCi and the 85Sr/82Rb
`ratio <0.2 mCi/mCi. The minimum detectable activity
`of the multi-channel analyzer used was 2 (cid:2) 10(cid:255)3 mCi of
`82Sr and 2 (cid:2) 10(cid:255)4 of 85Sr. The 82Sr and 85Sr break-
`through measurements performed daily in a dose cali-
`brator using Eqs. (1) and (3) enabled us to correct the
`contribution of 85Sr to the 82Sr breakthrough reading,
`which could have otherwise resulted in overestimates
`of the 82Sr breakthrough. No significant breakthrough
`was observed either by gamma spectrometry or by
`using a dose calibrator after elution of the generators
`
`Fig. 5. Positron emission tomograph (T) and strontium-82/rubidium-82 generator shielded with depleted uranium and lead rings
`(inside the stainless steel cart) (G) connected to the automated patient’s delivery system (D). The system delivers accurate intrave-
`nous doses of 82RbCl to the patients by pumping 0.9% NaCl (S) through the 82Sr/82Rb generator. The delivery system is connected
`to a controlling computer with a touch screen (C).
`
`

`

`T.M. Alvarez-Diez et al. / Applied Radiation and Isotopes 50 (1999) 1015–1023
`
`1021
`
`with more than 20 l of 0.9% NaCl. No other radio-
`nuclide impurities were detected by gamma spec-
`troscopy.
`The breakthrough of radioactive Sr is a function of
`the type and volume of the adsorbent of the generator
`column and the type and flow rate of the eluent. We
`and others (Brihaye et al., 1987; Kensett et al., 1987;
`Cackette et al., 1993) have previously demonstrated
`that hydrous tin oxide is an e(cid:129)cient cation exchanger
`material for the preparation of reliable 82Sr/82Rb gen-
`erators. Donaldson and Fuller, 1970 demonstrated that
`+ form is an e(cid:129)cient ma-
`hydrous tin oxide in the NH4
`terial for the separation of alkali metals from alkaline
`earth elements in a column. Sr2+ and Ca2+ were com-
`pletely retained on the column. However, the mono-
`valent cations tested (Na+, K+ and Cs+) were easily
`eluted without alkaline earth breakthrough using a
`mobile phase of NH3/NH4NO3. Neirinckx et al., 1982
`for the
`measured the distribution coe(cid:129)cients (KD)
`equilibration of 82Sr (II) and 82Rb (I) between hydrous
`tin oxide and NaCl solutions of di€erent pH and con-
`centrations (mobile phases). They showed that the KD
`values for 82Sr (II) were high and essentially constant
`between pH 7.2 and 11.4 (KD=53,000–56,000). The
`corresponding KD values for 82Rb (I) were 4 orders of
`magnitude lower. The 82Sr breakthroughs reported in
`this study were 5 (cid:2) 10(cid:255)8 and <10(cid:255)9/ml using hydrous
`tin oxide columns of 1- and 2-cc bed sizes, respectively.
`These columns were eluted with 0.9% NaCl at 10 ml/
`m.
`Other adsorbents previously used in the manufactur-
`ing of the 82Sr/82Rb generator column include Al2O3,
`the 82Sr
`ZrO2 and organic adsorbents. However,
`breakthrough values reported using these materials
`(Yano et al., 1979) were higher when compared with
`the breakthrough measurements observed when hy-
`drous tin oxide was employed. In addition, 82Sr/82Rb
`generators manufactured
`using Al2O3
`adsorbent
`require 2% NaCl eluent with pH 8–9 in order to mini-
`mize 82Sr breakthrough. This hypertonic eluent is not
`adequate for clinical use. It needs to be diluted prior
`to i.v. administration which further dilutes the 82Rb
`bolus, thus requiring higher patient infusion rates.
`However, when hydrous tin oxide is used as the
`adsorbent to manufacture 82Sr/82Rb generators,
`it is
`possible to elute 82Rb with physiological saline sol-
`utions of pHr 5.5 if the exchanger in the NH4
`+ form
`has been previously saturated with sodium ions. There
`is no need to neutralize the 82Rb chloride eluted. This
`was done following the method previously reported by
`our group (Cackette et al., 1993) and others (Brihaye
`et al., 1987). An exchange between the H+ ions of
`NaCl solution and the Na+ of the hydrous tin oxide
`occurs resulting in an increase of the pH value of the
`eluate to at least pH 6. At this pH, the separation of
`82Sr and 82Rb is maximized, thus minimizing the Sr
`
`breakthrough in the 82Rb eluted. Organic adsorbents
`have not been extensively utilized because they are un-
`stable to radiation.
`Another factor that determines the performance of
`the generator column is the tightness of the tin oxide
`packing. This factor was assessed by measuring the
`backpressure of the column when passing a volume of
`NH4OH at flow rates of 10 and 20 ml/m. The column
`backpressure ranged from 20 to 30 kPa and from 50
`to 60 kPa for these flow rates, respectively.
`The minimum backpressure established were 20 and
`50 kPa for 10 and 20 ml/m. Backpressures lower than
`20 kPa are the result of loose packing of the tin oxide
`into the column, which may result in low 82Rb yield
`and high 82Sr breakthrough. Columns with backpres-
`sures higher than 150 kPa were not accepted because
`the pressure limit of the infusion pump used is reached
`at approximately this pressure.
`The 82Rb elution yield obtained was at least 80%
`using an elution rate of 20 ml/m.
`The 82Rb infusion system permits the accurate deliv-
`ery of a 82Rb patient dose of 5 to 10 mCi over 1 min.
`This is achieved by setting the dose of 82Rb to be
`administered and the infusion time in the automated
`system and using variable infusion rates of 82Rb (up to
`50 ml/m)
`and
`infusion
`volumes
`(maximum
`volume = 30 ml). However, commercially available
`82Rb delivery systems (Squibb Diagnostic, 1995) do
`not permit slow infusions of 82Rb because the elution
`rate must be maintained constant at 50 ml/m. These
`doses of 82Rb are delivered instead by bolus adminis-
`tration. The infusion of 82Rb versus a bolus injection
`is advantageous for quantification studies in order to
`measure the arterial input function accurately and to
`assess necrosis/viability in patients.
`The chemical purity of samples of 82Rb eluted from
`3 di€erent 82Sr/82Rb generators was measured by ICP/
`AES, (Table 1). Stannous (Sn) was not detected in the
`samples. The detection limit of Sn using this method is
`5 (cid:2) 10(cid:255)2 ppm. 82Rb chloride samples contained 0.12–
`2.67 ppm of NH3, 0–2 (cid:2) 10(cid:255)3 ppm Pb, 0–0.11 ppm of
`Cu, 4 (cid:2) 10(cid:255)2–7 (cid:2) 10(cid:255)2 ppm of Fe and 0–1 ppm of Ca
`and Mg, respectively. However, the amounts of metal
`impurities contained in the maximum patient volume
`of 82Rb administered are lower than the toxic levels
`for i.v. injections (data not shown).
`Titration of Sn and trace metal analysis of samples
`of two batches of hydrous tin oxide were also per-
`formed by ICP/AES and compared with the data pro-
`vided by the manufacturer (Table 2). The samples
`analyzed contained more than 68% of Sn. The first
`column shows a typical analysis of hydrous tin oxide.
`The second and third columns show the actual trace
`metal analysis of the two batches of hydrous tin oxide
`analyzed. We have also indicated on column 5, the
`method detection limit
`(MDL)
`for
`each metal
`
`

`

`1022
`
`T.M. Alvarez-Diez et al. / Applied Radiation and Isotopes 50 (1999) 1015–1023
`
`Table 1
`Trace metal analysis of 82Rb chloride eluted from the 82Sr/
`82Rb generators
`
`During the entire shelf life (2–3 months) of the 82Sr/
`82Rb generator, all samples of 82Rb chloride eluted
`were sterile and pyrogen free.
`
`Trace metals
`
`Concentration (ppm)
`
`Sn
`N–NH3
`Pb
`Cu
`Fe
`Ca
`Mg
`
`NDb
`0.12–2.67
`ND–2 (cid:2) 10(cid:255)3
`ND–0.11
`4 (cid:2) 10(cid:255)2–7 (cid:2) 10(cid:255)2
`ND–1
`ND–1
`
`MDL (ppm)a
`5 (cid:2) 10(cid:255)2
`2 (cid:2) 10(cid:255)2
`2 (cid:2) 10(cid:255)3
`5 (cid:2) 10(cid:255)2
`5 (cid:2) 10(cid:255)2
`1
`1
`
`The second column of this table indicates the minimum
`and maximum concentrations of metals in ppm observed in
`the 82Rb chloride samples analyzed (n = 3). aMinimum detec-
`tion limit (MDL) of inductively coupled plasma/atomic emis-
`sion spectroscopy (ICP/AES). bNot detected (ND).
`
`measured. The MDL reported here are di€erent than
`those reported on Table 1 because these parameters
`varied with the sample matrix. 82Rb chloride is a sol-
`ution, however hydrous tin oxide is a solid. The levels
`of Pb and Cu, measured in the second batch of hy-
`drous tin oxide samples were higher than the typical
`levels reported by the manufacturer. The amounts of
`Ca and Mg on hydrous tin oxide were relatively high
`compared with other metals. This may be due to the
`absorption of these metals during the manufacturing
`process. Hydrous tin oxide is a good adsorbent of
`divalent cations. However, Ca and Mg were not
`detected in the 82Rb eluate. Even when the batch of tin
`oxide that contained relatively high levels of
`trace
`metals was used to manufacture the generator column,
`we did not observe a reduction on its cationic
`exchange e(cid:129)ciency.
`
`Table 2
`Sn assay and trace metal analysis of hydrous tin oxide
`
`Trace
`metals
`
`Typical
`analysisa
`
`Actual analysisb
`
`MDLc
`
`batch 1
`
`batch 2
`
`Sn
`Pb
`Cu
`Fe
`Sb
`Ca
`Mg
`
`66%
`0.1 ppm
`10
`20
`40
`
`68.5%
`NDd
`ND
`4
`ND
`ND
`ND
`
`68.8%
`34 ppm
`70
`100
`1
`458
`159
`
`0.01%
`3 ppm
`1
`1
`1
`25
`25
`
`a Analysis provided by Tin Research Institute, Kingston
`Lane, Middlesex, England. bTrace metal analysis of
`two
`batches of hydrous
`tin oxide performed by inductively
`coupled plasma (ICP)/atomic emission spectroscopy (AES).
`cMethod detection limit.dNot detected.
`
`4. Conclusion
`
`82Sr/82Rb generators manufactured following this
`protocol provide a means to obtain sterile and pyrogen
`free solutions of 82Rb chloride injection. This protocol
`uses simple components and quality control tests com-
`monly used in radiopharmacy.
`Our delivery system permits the infusion of small
`doses of 82Rb into the patients which is suitable for
`quantification studies of myocardial blood flow.
`Other nuclear medicine facilities could use this tech-
`nique to produce on-site 82Sr/82Rb generators and
`82Rb chloride for clinical studies.
`
`References
`
`Bergmann, S.R., Fox, K.A.A., Geltman, E.M., Sobel, B.E.,
`1985. Positron emission tomography of the heart. Prog.
`Cardiovasc. Dis. 28, 165–194.
`Brihaye, Cl., Guillaume, M., O’Brien, H.A., Raets, D., De
`Landsheere, CH., Rigo, P., 1987. Preparation and evalu-
`ation of a hydrous tin (IV) oxide 82Sr/82Rb medical genera-
`tor system for continuous elution. Appl. Radiat. Isot. 38,
`213–217.
`Brunken, R., Schwaiger, M., Grover-McKay, M., Phelps, M.,
`Tillisch, J., Schelbert, H.R., 1987. Positron emission tom-
`ography detects tissue metabolic activity in myocardial seg-
`ments with persistent thallium perfusion defects. JACC 10,
`557–567.
`Cackette, M.R., Ruth, T.J., Vincent, J.S., 1993. Sr production
`from metallic Rb targets and development of an Rb genera-
`tor system. Appl. Radiat. Isot. 44, 917–922.
`Camici, P., Ferrannini, E., Opie, L.H., 1989. Myocardial
`metabolism in ischemic heart disease: basic principles and
`applications to imaging by positron emission tomography.
`Prog. Cardiovasc. Dis. 32, 217–238.
`Cloninger, K.G., DePuey, E.G., Garcia, E.V., Roubin, G.S.,
`Robbins, W.L., Nody, A., DePasquale, E.E., Berger, H.J.,
`1988. Incomplete redistribution in delayed thallium-201
`single photon emission computed tomographic (SPECT)
`images: an overestimation of myocardial scarring. J. Am.
`Coll. Cardiol. 12, 955–963.
`Dahl, J.V., Muzik, O., Wolfe, E.R., Allman, C., Hutchins,
`G., Schwaiger, M., 1996. Myocardial rubidium-82 tissue
`kinetics assessed by dynamic positron emission tomography
`as a marker of myocardial cell membrane integrity and via-
`bility. Circulation 93, 238–245.
`Demer, L.L., Gould, K.L., Goldstein, R.A., Kirkeeide, R.L.,
`Mullani, N.A., Smalling, R.W., Nishikawa, A., Merhige,
`M.E., 1989. Assessment of coronary artery disease severity
`by positron emission tomography. Circulation 79, 825–835.
`
`

`

`T.M. Alvarez-Diez et al. / Applied Radiation and Isotopes 50 (1999) 1015–1023
`
`1023
`
`Donaldson, J.D., Fuller, M.J., 1970. Ion exchange properties
`of tin (IV) materials-III. J. Inorg. Nucl. Chem. 32, 1703–
`1710.
`Galli, M., Bencivelli, W., Pardo, N.F., Tavazzi, L., 1988.
`Underestimation of residual ischemia by 201-Thallium scin-
`tigraphy after myocardial infarction. Chest 94, 876–878.
`Go, R.T., Marwick, T.H., MaIntyre, W.J., Saha, G.B.,
`Neumann, D.R., Underwood, D.A., Simpfendorfer, C.C.,
`1990. A prospective comparison of rubidium-82 PET and
`thallium-201 SPECT myocardial perfusion imaging utilizing
`a single dipyridamole stress in the diagnosis of coronary
`artery disease. J. Nucl. Med. 31, 1899–1905.
`Goldstein, R.A., Mullani, N.A., Marani, S.K., Fisher, D.J.,
`Gould, L.K., O’Brien, H.A., 1983. Myocardial perfusion
`with rubidium-82. II. E€ects of metabolic and pharmaco-
`logic interventions. J. Nucl. Med. 24, 907–915.
`Goldstein, R.A., Mullani, N.A., Wong, W.-H., Hartz, R.K.,
`Hicks, C.H., Fuentes, F., Smalling, R.W., Gould, K.L.,
`1986. Positron imaging of myocardial infarction with rubi-
`dium-82. J. Nucl. Med. 27, 1824–1829.
`Gould, K.L., Goldstein, R.A.