Introduction
Rubidium-82 (
82Rb) is a widely used positron emission tomography (PET) tracer with a short half-life of ~ 75 seconds.
1 Stress perfusion imaging with
82Rb is used widely in cardiac PET centers for the assessment of known or suspected coronary artery disease.
2 The short half-life of
82Rb allows rapid sequential rest and stress imaging with low-radiation exposure to patients and medical staff.
3,
4 A mobile generator system is used to produce the
82Rb-chloride tracer on-demand from the parent radioisotope strontium-82 (
82Sr) which has a half-life of ~ 25 days.
5 The generator ion-exchange column is typically made of a hydrated tin-oxide, and loaded with 100 mCi (3.7 GBq) of
82Sr activity that decays continuously to produce
82Rb. The strong adsorption of
82Sr to tin-oxide compared to the weak binding of
82Rb enables the selective elution of
82Rb-chloride solution with standard physiologic saline.
6 The
82Rb activity is then replenished quickly on the generator column (93.8-99.6% within 5-10 minutes, respectively) according to the properties of parent–daughter secular equilibrium.
7
While the main components of
82Sr/
82Rb generators are similar across different commercial systems, e.g. ion exchange column, saline supply and pump, generator and patient intravenous lines,
8 the shelf-life (approved period for clinical use) can be different in clinical practice.
9‐
11 The factors that influence generator shelf-life are the available daily
82Rb isotope yield and the total volume of saline eluted through the column, both of which must be within the approved limits for clinical perfusion imaging.
12 Furthermore, the exact amount of delivered
82Rb activity and its infusion profile (e.g. constant flow-rate vs constant activity-rate) are important parameters for optimizing dynamic imaging accuracy and repeatability.
13,
14 In this context, the aims of the current study were to assess the isotope production efficiency (
82Rb yield/
82Sr parent activity) of two FDA-approved
82Sr/
82Rb generators during their clinical shelf-life and to characterize the effect of different daily quality assurance (QA) procedures on measured
82Rb activity.
Discussion
82Rb stress PET has become a routine method for myocardial perfusion imaging including the quantification of absolute myocardial blood flow (mL·min
−1·g
−1). The utility of
82Rb PET to diagnose coronary artery disease and to guide appropriate treatment for CAD patients has been well-established previously.
17 The documented prognostic value of
82Rb PET perfusion scanning,
2,
18,
19 its low-radiation exposure to patients and medical staff,
1 and the availability of approved on-site
82Sr/
82Rb generators
20 plays a key role in the increasing use of this modality. The number of imaging centers using
82Rb PET has risen dramatically since the initial commercial approval (CardioGen-82
®) and subsequent addition (RUBY-FILL
®) of these
82Rb generator systems.
21 With this increase in clinical capacity and diversity of generator systems, the need for standardized patient dosing is critical, as the net amount of injected activity ultimately determines the PET image quality and absorbed radiation dose to the patient.
22
This study compared the 82Rb production efficiency of the two-generator systems available for cardiac PET perfusion imaging, and characterized the daily QA procedure effects on measured isotope yield. The main findings indicated: (1) 82Rb yield of the CardioGen-82® generator was under-estimated by ~ 24% using dose-calibrator maximum values for daily QA; (2) a correction-factor of ~ 1.3 may be used to convert CardioGen-82® daily QA values to equivalent integrated total 82Rb activity as measured by the RUBY-FILL® system; (3) 82Rb isotope yield and production efficiency were estimated to be ~ 7% higher for the RUBY-FILL® vs CardioGen-82® system, as measured over a clinical shelf-life that was ~ 30% longer.
A novel method was developed to bring the
82Rb isotope yield measured by two different generators to a common scale for accurate comparison. The CardioGen-82
® elution protocol recommends the patient maximum volume to be set at 50 mL and the patient dose at 1480 MBq (40 mCi), with a range of 1110-2220 MBq (30-60 mCi).
11,
23 The RUBY-FILL
® system is user-adjustable for dose (e.g. 10-30 MBq·kg
−1 [0.27-0.81 mCi·kg
−1] body weight) and flow-rate (15-30 mL·min
−1) or infusion time (10-120 seconds).
9,
24 The infused patient activity is also reported differently by the two systems. RUBY-FILL
® reports only the integrated total activity for both daily QA and patient studies. The CardioGen-82
® system reports the integrated ‘Patient Dose’ and the ‘End-of-Infusion’ activity values, which must be calibrated at the same activity used for patient elutions. Neither of these activity values appear to be accurate on the daily QA breakthrough elution reports, therefore, they should not be used as a measure of CardioGen-82
® 82Rb isotope yield.
Both generator systems showed a small increase in
82Rb isotope production efficiency over time (Figure
5). The mechanism for this observed improvement is not completely understood but may be due to migration of
82Sr activity away from the column inlet, where it can be eluted from the column more completely. The difference in measured efficiency between vendors may be due to alternative chemical preparation/conditioning of the column resin and/or physical geometry/volume of exchanger in the generator columns, but these are proprietary data and difficult to confirm.
Three-dimensional (3D) PET systems are the current commercial standard which provides higher sensitivity for data acquisition, compared to previous generation 2D scanners.
20 However, highly sensitive signal acquisition carries a higher risk for detector saturation which can result in the over-estimation of myocardial blood flow. The saturation effect on dynamic imaging can be prevented by adjusting the injected activity for body weight to optimize the amount of tracer in the bolus first-pass transit.
13 To optimize the injected
82Rb activity in regard to the camera sensitivity, the difference between requested vs injected activity (and volume), maximum vs integral activity, and constant vs variable infusion flow rate are important parameters that need to be taken into consideration. In this context, the current study provides insights to the technical aspects of
82Sr/
82Rb generators that can affect the image quality and myocardial blood flow quantification.
14 Our results indicated that the RUBY-FILL
® elution activity and flow errors (in proportion to the requested values) are exceedingly small, which confirms the high precision of this infusion system. Moody et al. have shown previously that the error in actual measured/requested dose for the CardioGen-82
® generator was up to 50% for a requested activity of 370 MBq (10 mCi) and 33% for a requested activity of 550 MBq (15 mCi), and decreased as the requested activity approached 30 mCi (1110 MBq).
25 In our test elutions with RUBY-FILL
®, the range of requested elution activity was ~ 400 to 600 MBq (11 to 16 mCi), and the elution activity error was negligible (less than 0.2%) due to the accurate modeling of eluate transport delay and isotope decay en-route from the generator to the outlet of the patient infusion line.
6
Secondary analysis of the integrated total activity from the RUBY-FILL
® test elutions (Supplemental Figure S5) demonstrated similar trends compared to the primary analysis of the dose-calibrator maximum values in Figure
4. The integrated total activity increased consistently as a function of elution flow-rate, up to a peak value 95% of the daily QA calibration activity, using 30 mL at 30 mL·min
−1 elution. These data also suggest that if the RUBY-FILL
® generator was eluted using the CardioGen-82
® protocol (50 mL at 50 mL·min
−1), then the
82Rb yield could increase to ~ 120% of the daily QA calibration activity.
Clinical procedure guidelines published by the Society of Nuclear Medicine
26 recommend that radiopharmaceutical doses should be dispensed and subsequently administered to patients within 10% and 20% of those indicated by the prescribing physician. The U.S. NRC Regulations (10 CFR 35.63) on the use of medical isotopes state “… a licensee may not use a dosage if the dosage does not fall within the prescribed dosage range or if the dosage differs from the prescribed dosage by more than 20 percent”.
27 The most recent EANM guidelines for PET tumor imaging indicate that for automated administration “actual administered activity may not deviate by more than 3% from that indicated by the device”.
28 The activity error data from the RUBY-FILL
® generator (Figure S4A) show that this system can meet even these most strict criteria, with actual vs requested activity errors of < 0.11% over all test elutions performed. In comparison, the printed ‘End-of-Infusion’ activity values from the CardioGen-82
® breakthrough elution reports appeared to under-estimate the measured dose-calibrator maximum values by ~ 30%, even when using the recommended elution settings of 50 mL at 50 mL·min
−1. This underestimation may be due, in part, to the fact that this generator is typically calibrated using a requested
82Rb activity which is less than the total yield available on a given day, and subsequent elutions are only recommended using the same calibration activity.
The accuracy of
82Rb PET dosing is of clinical significance when the imaging results of different PET centers are compared or pooled together as part of multicenter imaging trials.
24 For example, these results may have important implications for the calculation of patient absorbed radiation doses. Effective dose (ED) estimates have been published recently for
82Rb PET in the range of 0.80 and 1.1 mSv·GBq
−1 based on PET studies using the RUBY-FILL
® and CardioGen-82
® generators, respectively.
3,
4 These radiation ED values are calculated using PET image-derived activity values divided (normalized) by the total injected activity, as reported by the respective generator systems. Because of this inverse relationship,
higher ED values would be expected from PET studies performed using a generator system which reports
lower administered
82Rb activity. In fact, the ratio of ED estimates from the RUBY-FILL
®/CardioGen-82
® dosimetry papers (0.73) is similar to the CardioGen-82
® under-estimation of activity reported in the present study (0.76), providing a possible explanation for the discrepancy between these independently reported ED values. Our results suggest that the lower ED values reported by Hunter et al. using PET studies performed with the RUBY-FILL
® system may be more accurate, because the same integrated total activity method of measurement is used both during daily QA calibration and during patient elutions.
3
Limitations
The CardioGen-82
® correction-factor was calculated using the RUBY-FILL
® daily QA protocol (integrated total activity of 35 mL elution @ 20 mL·min
−1) as the reference standard to enable accurate comparison of
82Rb isotope yield and production efficiency between the two systems. The results of this study do not suggest that all generator systems should use the same RUBY-FILL
® elution settings above for daily QA, but rather that the dose-calibrator maximum or ‘steady-state’ value measured in a calibration vial consistently under-estimates the integrated total activity delivered to the patient (Figure
2), and that this under-estimation is a function of the elution volume and flow-rate (Figure
4B). The integrated total activity could have been referenced instead to the CardioGen-82
® calibration elution settings (50 mL @ 50 mL·min
−1) in which case the RUBY-FILL
® calibration values would have been multiplied by the extrapolated value of 0.761 for inter-comparison, as shown in Table
2.
Direct comparison of the activity profiles using the same elution flow-rate on both systems was not feasible in this study. The RUBY-FILL
® maximum flow-rate is 30 mL·min
−1, whereas CardioGen-82
® uses a constant 50 mL·min
−1 elution flow-rate. The CardioGen-82
® system does include a switch to use lower flow-rates (20 or 35 mL·min
−1), but these are not part of the approved clinical labeling.
11 Therefore, direct comparison could be performed in future studies if CardioGen-82
® daily QA was also measured at a flow-rate of 20 mL·min
−1. In the current study, extrapolation of the RUBY-FILL
® test elution data from 30 to 50 mL·min
−1 was associated with a certain margin of error in estimation of the CardioGen-82
® correction-factor. Although the logarithmic and power functions provided excellent fits to these data (
R2 > 0.98), the extrapolated values should be interpreted with a degree of caution.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.