Author + information
- Received May 26, 2015
- Revision received July 21, 2015
- Accepted July 21, 2015
- Published online October 1, 2015.
- ∗Division of Cardiovascular Medicine, University of Maryland School of Medicine, Baltimore, Maryland
- †Department of Medicine, Greenville Health System, Greenville, South Carolina
- ‡Intersocietal Accreditation Commission, Ellicott City, Maryland
- §Division of Cardiology, Emory University School of Medicine, Atlanta, Georgia
- ↵∗Reprint requests and correspondence:
Dr. Scott D. Jerome, University of Maryland School of Medicine, 295 Stoner Avenue, #103, Westminster, Maryland 21157.
Objectives This study sought to examine current laboratory practices for radiation effective doses for myocardial perfusion imaging (MPI) and laboratory adherence to guideline-directed radiation reduction practices.
Background A recent focus on radiation dose reduction for cardiovascular imaging has led to several published guidelines and consensus statements detailing performance metrics for laboratory practices. We sought to examine laboratory adherence to optimized radiation dose protocol recommendations among 5,216 submitted cases from 1,074 MPI laboratories evaluated for Intersocietal Accreditation Commission accreditation.
Methods Eligible imaging centers included MPI laboratories enrolled in the Intersocietal Accreditation Commission data repository of accreditation applications from 2012 to 2013. Accreditation requires submission of 3 to 5 cases for evaluation of a range of representative cases. Based on standard dosimetry for rest and stress MPI, an effective dose (in millisieverts) was calculated. Model simulations were performed to estimate guideline-directed effective doses.
Results The average effective dose was 14.9 ± 5.8 mSv (range 1.4 to 42.4 mSv). A 1-day technetium Tc 99m protocol was used in 82.9% of cases, whereas a 2-day technetium Tc 99m and dual isotope protocol was used in 7.5% of submitted cases. Only 1.5% of participating imaging centers met current guidelines for an average laboratory radiation exposure ≤9 mSv, whereas 10.1% of patient effective doses were >20.0 mSv. A model simulation replacing the radiation exposure of dual isotope MPI with that of a 1-day technetium Tc 99m protocol reduced the proportion of patients receiving an effective dose >20 mSv to only 2.7% of cases (p < 0.0001).
Conclusions Mandatory laboratory accreditation for MPI allows for examination of current radiation dosimetry practices. Current guidelines for reduced patient-specific radiation exposure are rarely implemented, with few laboratories meeting recommendations of ≤9 mSv for 50% of patients. Increased educational efforts and the development of performance measures for laboratory accreditation may be required to meet current radiation dose-reduction standards.
Recently, considerable attention has been focused on the ever-increasing proportion of exposure to ionizing radiation that is attributable to medical imaging, leading many cardiovascular professional societies to scrutinize current practices and to issue radiation control directives (1–7). In 2010, the American Society of Nuclear Cardiology (ASNC) published an information statement that included a combination of radiation reduction measures that, if implemented correctly, would result in quality imaging at the lowest possible radiation dose (7). Current guidance documents recommend dose-reduction strategies whereby 50% or more of patients undergoing single photon emission computed tomography myocardial perfusion imaging (MPI) receive a total effective dose ≤9 mSv (7). Moreover, a recent symposium sponsored by the National Institutes of Health’s National Heart, Lung, and Blood Institute and National Cancer Institute proposed that cardiovascular diagnostic procedures with effective doses >20 mSv be rarely employed due to the potential projected cancer risk and that this threshold should be considered for laboratory tracking as a performance measure (5,8).
Current evidence is lacking as to nationwide laboratory practices of radiation dosing patterns, particularly for MPI due to its high rate of utilization (9). Recent survey findings representing only 20% of ASNC member laboratories revealed unnecessary and high radiation dose practices with minimal attention to dose (4). The Intersocietal Accreditation Commission (IAC) data repository provides a unique opportunity to evaluate current clinical practice and radiation dosing patterns from a large nationwide cohort of laboratories and patients. The aim of this analysis was to examine current laboratory practice for radiation effective doses for MPI and laboratory adherence to guideline-directed radiation reduction practices (5,7,8).
The IAC data repository enrolls prospective submission of 3 to 5 patient MPI cases as part of laboratory accreditation requirements. Accreditation requirements include submission of 5 MPI cases for nuclear cardiology and 3 cases for general nuclear medicine accreditation. Of the submitted cases, only 1 may be normal with the remaining cases demonstrating an abnormality. Specific cases included case 1—normal, case 2—ischemia, case 3—myocardial infarction, case 4— left ventricular wall motion abnormality, and case 5— an additional abnormality. A total of 1,074 laboratories submitted applications for accreditation during 2012 and 2013. This time frame was selected to reflect a stable period of isotope production and availability in the United States, without restrictions and potential need for use of higher radiation dose protocols.
Laboratory characteristics including state, U.S. region (e.g., south, northeast, west), laboratory type (i.e., hospital vs. private practice), annual volume of MPI studies, number of physicians in the practice, number of physicians certified by the Certification Board of Nuclear Cardiology, number of technologists, and number of locations providing MPI services were also documented.
Calculation of radiation dosimetry
For each case, a laboratory must submit a final report from which we extracted the amount of radioactivity injected at rest and stress in millicuries per patient. We also identified the standardized MPI protocol used (technetium Tc 99m [Tc99m] 1-day, Tc99m 2-day, thallium Tl 201 [Tl201]/Tc99m dual isotope, Tl201 or Tc99m stress-only) (10). The effective radiation exposure per patient was estimated by converting 0.3 mSv/mCi for Tc99m (both Tc99m sestamibi and Tc99m tetrofosmin were treated similarly) (7) and 6.3 mSv/mCi for Tl201 (11,12). The administered dose was converted to an effective dose (in millisieverts) based on ASNC guidance documents (7).
Comparisons were made between the average effective dose per protocol and the guideline average recommended dose by protocol: Tc99m 1-day (370 MBq [10 mCi] low dose/1,110 MBq [30 mCi] high dose); Tc99m 2-day (1,110 MBq [30 mCi] day 1/1,110 MBq [30 mCi] day 2); dual isotope (122 MBq [3.3 mCi] Tl201 rest dose/1,110 MBq [30 mCi] Tc99m stress dose); and Tl201 (122 MBq [3.3 mCi] stress dose/56 MBq [1.5 mCi] rest reinjection dose).
The IAC data repository was locked in 2014 following thorough cleaning and independent examination of implausible data ranges/outliers. Initial data analysis included descriptive statistics evaluating percentages for categorical variables or mean ± SD for continuous variables. Comparisons of categorical variables were calculated using chi-square statistics and Student t tests for comparisons of continuous measures. Across laboratory volume, effective doses, and other continuous measures, we categorized these variables using median and interquartile range values.
Comparisons were made between the years 2012 and 2013 for alterations in protocol usage, average effective doses, percentage of cases ≤9 mSv (7) and >20 mSv (5). As well, the frequency of use of the dual isotope protocol was evaluated, as this is the highest effective radiation exposure used for MPI. Correlation of the average observed exposure for each protocol to the average guideline-directed dosing recommendation from ASNC was performed by Spearman rho correlation. A model simulation was devised examining alterations in mean effective dose per laboratory following elimination of dual isotope imaging and replacement with a lower effective dose protocol per ASNC guidelines (13). The results were modeled by replacing the administered doses of each dual isotope MPI with the average guideline recommended doses for a 1-day Tc99m protocol. Statistical difference was determined using the McNemar test. A p value <0.05 was chosen to reflect statistical significance for all analyses. Statistical analysis was performed using IBM SPSS version 22.0 (IBM Corporation, Armonk, New York).
The majority of laboratories were from the southern states, with 30% located in the northeastern part of the United States (Table 1). Submitting laboratories were largely defined as private practice. The median of the annual volume of rest/stress MPI cases was 1,000 (interquartile range: 566 to 1,753). Submitting laboratories represented a median of 3 nuclear cardiologists and a median of 1 Certification Board of Nuclear Cardiology–certified physician per laboratory. The submitting laboratory had a median of 1 technologist and a median of 1.0 sites for MPI testing.
Laboratory effective doses across laboratories and patient cases
The IAC data repository included 5,216 reports from 1,074 laboratories applying for accreditation with a total of 10,405 rest and stress administered radiopharmaceutical doses for MPI. The calculated average effective dose per patient was 14.9 ± 5.8 mSv with a range of 1.4 to 42.4 mSv.
The Tc99m 1-day protocol was the most frequent protocol (82.9% of cases), followed by the Tc99m 2-day protocol and the rest Tl201/stress Tc99m dual isotope protocols, both at 7.5% (Table 2). Based on the protocol used, marked variation in radiation effective dose was observed. The mean effective dose was 12.9 mSv for the Tc99m MPI protocol and increased to an average of 17.9 mSv for the 2-day Tc99m MPI protocol. For the 2-day protocol, no significant difference was seen between the effective dose on day 1 versus day 2: 8.8 mSv and 8.9 mSv (p = 0.251). Higher average effective doses were observed for Tl201 at 23.9 mSv. The highest average effective dose protocol was for rest Tl201/stress Tc99m dual isotope protocol at 32.8 mSv. Importantly, the average effective dose was only 7.5 mSv for a Tc99m stress-only protocol. The lowest effective dose, 1.4 mSv, was from a stress-only MPI performed on a cadmium zinc telluride multidetector gamma camera.
Adherence to radiation dose-reduction guidelines
A cumulative plot of MPI effective doses is reported in Figure 1. Only 1.5% of laboratories reported a median effective dose for rest/stress MPI of ≤9 mSv. In total, cases with effective doses ≤9 mSv were submitted from only 44 laboratories. Nearly 86% of submitted cases had an effective dose of 10.1 to 20 mSv. By comparison, nearly 1 in 10 laboratories had a median reported rest/stress MPI effective dose of >20 mSv. Of those cases, 74.4% resulted from a dual isotope protocol, 13.8% were from a Tl201 protocol, and 11.6% were from the Tc99m 2-day protocol.
Of the submitted cases meeting low and high effective dose thresholds, only 1.1% of laboratories employed a dose ≤9 mSv in any given patient, whereas 9.8% had a dose >20 mSv (Figure 2). Moreover, nearly 1 in 11 cases used the rest Tl201/stress Tc99m dual isotope protocol.
Figure 3 plots the median (interquartile range) average effective doses for rest and stress MPI for all submitting laboratories and for the normal MPI case. For those with a normal MPI, use of the stress-only protocol could reduce the average effective dose by nearly 25%.
Comparison of observed to recommended average radiation exposure
The average observed exposure for each protocol was compared with the average guideline-directed dosing recommendation from ASNC (13). Although there was a strong correlation (r = 0.93) between observed and recommended exposure, the guideline-directed recommended exposure was significantly lower (mean difference: −0.9 ± −2.09, p < 0.0001), supporting opportunities for optimized dose-reduction practices across participating laboratories. Compared with the guideline-directed recommended dose, laboratories administered a higher effective dose for dual isotope and 1-day Tc99m protocols, +3.3 mSV and +0.9 mSv (13).
Our model simulation results revealed that replacement of a rest Tl201/stress Tc99m dual isotope protocol with the average guideline-directed recommended Tc99m 1-day protocol dose would yield a sizeable reduction in overall population radiation exposure. Based on this simulation, the guideline-directed average effective dose for a laboratory would be reduced to 13.4 mSv, with only 2.6% of doses being >20 mSv (p < 0.0001). From this model, an examination of radiation effective doses reveals that maximum exposure would be reduced from 42.4 mSv to 33.4 mSv (p < 0.0001).
Temporal changes in radiation dosimetry
Additional analyses compared temporal changes in radiation dosimetry from 2012 to 2013 (Table 3). We failed to observe any significant differences in the average effective doses over time (2012: 15.1 ± 5.7 mSv and 2013: 14.7 ± 5.1 mSv, p = 0.19). The percentage of exposures ≤9 mSv was 1.1% in 2012 and 2013 (p = 0.92). Additionally, there was no statistical difference in the frequency of effective doses >20 mSv, with 10.6% occurring in 2012 and 8.4% in 2013 (p = 0.22). However, we did observe a trend in reduced use of dual isotope imaging: from 10.1% in 2012 to 6.9% in 2013 (p = 0.04). Of the sustained users of dual isotope imaging, no difference in average radiation effective dose was observed from 2012 to 2013.
During the past decade, we have observed an increased rigor in evaluation of the appropriate use and justification of exposure to ionizing radiation for cardiovascular imaging procedures such as MPI and cardiovascular computed tomographic imaging (4). Concomitant to this are increasing requirements for MPI laboratories to be accredited in order to qualify for reimbursement from both public and private payers. The IAC data repository provides a unique opportunity to assess “real-world” MPI practice and radiation dosing patterns in the United States. Observations from the IAC data repository may provide a unique means to propose data-driven standards for MPI and other imaging modality practices and to track proposed performance standards of high-quality imaging practices in the United States. The current report includes a nationwide sample of 5,216 patients with 10,405 administered rest and stress MPI radiopharmaceutical doses evaluated from 1,074 laboratories.
Current laboratory effective dose for MPI
Few previous reports, largely using self-reported data, have been published revealing current MPI protocol use (4). As reported in a 2001 ASNC survey, dual isotope MPI use was reported in 72% of patients (14). By comparison, in a recent report from 2011, Einstein et al. (4) reported from an ASNC member survey (with a response rate of ∼20%) that dual isotope MPI was used in 15.6% of laboratories. Our study demonstrated further decline in dual isotope MPI use from 10.1% in 2012 to 6.9% of patients in 2013. This represents a positive trend toward reductions in high exposures, but additional findings from our IAC registry also support that further reductions (i.e., to 2.9%) may be realized. From a survey of MPI laboratories in Germany, a notable decrease in the use of Tl201 was observed from 2005 to 2012 (20% to 5%) (15).
Achieving population dose-reduction strategies
Our results from unselected laboratories revealed an average effective dose of 14.9 mSv. From a recent 2012 survey in Germany (15), the average effective dose per patient was estimated at 7.4 mSv. The lower German radiation doses may reflect tighter control based on the regulations of the German Office for Radiation Protection that were initiated in 2003. It remains plausible that additional U.S. regulations, particularly in the form of a diagnostic reference levels applied as performance metrics may be helpful to further reduce unnecessarily high effective dose utilization practices (4).
In 2010, ASNC proposed radiation dose-reduction recommendations targeting a total effective dose of ≤9 mSv for a minimum of 50% of MPI patients tested in a given laboratory by 2014 (7). Our results reveal that this goal has not been achieved; only 1.9% of all studies were ≤9 mSv, and only 1.4% laboratories administered ≤9 mSv in >50% of cases. Hindsight reveals that although laudable, this goal of ≤9 mSv was largely unrealistic, based on the IAC data repository findings. The current age of MPI equipment, lack of physician and technologist’s knowledge of contemporary dose-reduction practices, or the failure of positron emission tomography to be optimally used in the United States are all factors contributing to a failure to achieve this laboratory goal of ≤9 mSv for 50% of patients (7–14). As widespread laboratory dosing practices were not available when this targeted level of ≤9 mSv was devised, the use of the IAC data repositories and other databases may prove optimal to set realistic quality-based standards for radiation dose practices as well as other metrics, such as appropriate use criteria (16).
A recent National Institutes of Health–sponsored symposium recently made strict recommendations for limiting and tracking radiation effective doses of >20 mSv to a minimal number of patients with justified indications for MPI (5). This target was obviously focused on high-dose exposure, such as with dual isotope MPI and adoption of this recommendation, based on IAC data, could have a significant effect in lowering overall radiation exposure for ∼1 in 10 patients. As the average radiation exposure for laboratories using the dual isotope protocol was 32.8 mSv, adoption of a 20-mSv threshold would encourage laboratories to eliminate use of the dual isotope protocol and reduce exposures for the majority of patients now receiving high effective doses in the IAC data repository.
Reducing population exposure by targeting high effective dose cases
Importantly, the dual isotope protocol has the highest potential to affect projected cancer risk for patients undergoing MPI (5,7). Recent estimates reported that the use of dual isotope MPI can be expected to result in 25 (range 9 to 58) incident cancer cases for every 10,000 tested patients, a rate 2.5-fold higher than that of a 1-day Tc99m protocol (incident cancer cases: 10 per 10,000 tested patients) (8).
Elimination of the dual isotope protocol with substitution of a Tc99m 1-day protocol would result in a significant decrease in radiation exposure, based on our findings. The IAC data repository of 1,074 laboratories represents a total volume of nearly 1.5 million MPI studies performed each year. Replacing the dual isotope protocol with the Tc99m 1-day protocol would lower the radiation exposure by ∼2-fold and improve the safety of more than 100,000 patients undergoing MPI each year in the United States.
In a previous National Institutes of Health–sponsored symposium, the diagnostic reference level for tracking was proposed at >20 mSv (5). We observed that ∼10% of patients received an effective dose >20.0 mSv, which, based on our modeling, could be reduced to a sizeable extent and improve overall population exposure appreciably. It would seem reasonable that use of the IAC repository may be a means of tracking radiation effective doses and to document adherence to the use of high-dose exposures (i.e., >20 mSv) as part of the accreditation process. As many physicians are relying on databases, such as the IAC repository, for documentation of their performance improvement activities for their maintenance of certification, the development of a more rigorous educational program and tracking of laboratory radiation exposures may prove an optimal means to reduce patient effective doses.
Stress-only MPI protocol use
Stress-only MPI remains a valuable means to significantly reduce total radiation exposure (17–20). By applying stress-only imaging, the stress image is performed first and if normal, then the rest dose is not required. Unfortunately, the IAC data repository revealed a very low use of the stress-only protocol (1.4%). Considering data only from 2-day protocols (average radiation exposure of 8.8 mSv for the first dose and 8.9 mSv for the second dose), if stress imaging was performed first and the second (rest) dose was eliminated in patients with a normal study, patients would realize a sizeable reduction in radiation exposure (∼25%).
Important limitations should be discussed with regard to our current findings. First, we applied a large, nationwide data repository whose primary use is for accreditation. The submitted cases were for review of image quality and interpretive acumen and not for consideration of radiation dose practices. As such, some referral and selection bias may be operational, although our data reflect a broader representation of MPI laboratories than previous survey reports have (4). There was no mechanism to determine whether the protocol intent started as stress-only but rest images were subsequently acquired because the stress study was abnormal. We moreover did not include cardiac positron emission tomography MPI, which has lower radiation exposure. A further limitation is that the administered dose was extracted from the patient report and may not reflect the true amount of administered radioactivity. In addition, at the time of data collection, some patient characteristics such as height and weight were not mandatory reporting requirements and thus were not available. The IAC reporting requirements have been revised to require the addition of patient height, weight, and sex beginning 2016. The study is also limited in that radiation exposure for both Tc99m sestamibi and Tc99m tetrofosmin was assumed similar at 0.3 mSv/mCi, although small differences do exist and radiation dosimetry varies based on patient weight and sex. Finally, this study only evaluated laboratories applying for IAC accreditation, which represents a large percentage of outpatient laboratories, and the results may not be generalizable to other populations.
The results of our study suggest that strategies to lower radiation exposure such as stress-only imaging are underused. Use of the data repositories, such as the IAC, may prove an optimal means to incentivize physicians for performance improvement and to document radiation dose-reduction practices. Findings from our analyses reveal that there are significant opportunities for reducing radiation exposure in the United States, and increased educational efforts or proscriptive dosing regimens may be required for laboratories to optimally lower radiation exposure for the large sector of stable ischemic heart disease patients undergoing MPI each year.
COMPETENCY IN PRACTICE-BASED LEARNING: Although radiation dose-reduction recommendations have been available since 2010, this study supports that current strategies to lower radiation exposure are underused across a wide spectrum of nuclear cardiology laboratories. A combination of radiation dose-reduction techniques such as stress-only imaging, elimination of dual isotope imaging, or weight-based dosing should be evaluated and incorporated into daily practice to optimize image quality at the lowest possible radiation dose. More rigorous incorporation of performance measures and dose reference levels into accreditation and patient registries are an important means to improve the safety of populations at risk for coronary artery disease.
TRANSLATIONAL OUTLOOK: Additional comprehensive studies that evaluate radiation exposure with image quality and study appropriateness are needed. In addition, the inclusion of individual patient characteristics such as sex, age, height, and weight should be evaluated along with the accuracy of imaging.
This project was supported by a grant from the Intersocietal Accreditation Commission (IAC). Its contents are solely the responsibility of the authors and do not necessarily represent the views of the IAC. Dr. Jerome is on the IAC Nuclear/Positron Emission Tomography Board of Directors. Ms. Farrell is an employee of the Intersocietal Accreditation Commission. The other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Ami E. Iskandrian, MD, served as Guest Editor for this paper.
- Abbreviations and Acronyms
- American Society of Nuclear Cardiology
- Intersocietal Accreditation Commission
- myocardial perfusion imaging
- technetium Tc 99m
- thallium Tl 201
- Received May 26, 2015.
- Revision received July 21, 2015.
- Accepted July 21, 2015.
- American College of Cardiology Foundation
- Chen J.,
- Einstein A.J.,
- Fazel R.,
- et al.
- Einstein A.J.,
- Tilkemeier P.,
- Fazel R.,
- Rakotoarivelo H.,
- Shaw L.J.,
- American Society of Nuclear Cardiology
- Einstein A.J.,
- Berman D.S.,
- Min J.K.,
- et al.
- Fazel R.,
- Gerber T.C.,
- Balter S.,
- et al.,
- for the American Heart Association Council on Quality of Care and Outcomes Research, Council on Clinical Cardiology, Council on Cardiovascular Radiology and Intervention
- Berrington de Gonzalez A.,
- Kim K.P.,
- Smith-Bindman R.,
- McAreavey D.
- Shaw L.J.,
- Marwick T.H.,
- Zoghbi W.A.,
- et al.
- ↵Holly TA, Abbott BG, Al-Mallah M, et al. ASNC imaging guidelines for nuclear cardiology procedures: single photon-emisson computed tomography. Available at: http://www.asnc.org/files/SPECT%202010.pdf/. Accessed August 27, 2015.
- Henzlova MJ, Cerqueira MD, Hansen CL, Taillefer R, Yao S. ASNC imaging guidelines for nuclear cardiology procedures: stress protocols and tracers. Available at: http://www.asnc.org/files/Stress%20Protocols%20and%20Tracers%202009.pdf/. Accessed August 27, 2015.
- Lindner O.,
- Burchert W.,
- Bengel F.M.,
- et al.,
- for the Working Group Cardiovascular Nuclear Medicine of the German Society of Nuclear Medicine, Working Group Nuclear Cardiology of the German Cardiac Society
- Wolk M.J.,
- Bailey S.R.,
- Doherty J.U.,
- et al.
- Chang S.M.,
- Nabi F.,
- Xu J.,
- Raza U.,
- Mahmarian J.J.
- Iskandrian A.E.