Introduction

Ischaemic heart disease is the leading cause of morbidity and mortality in developed countries. Diagnosis and treatment of ischaemic heart disease often require the use of techniques and procedures that employ some amount of radiation. Consequently, a significant increase in exposure of the population to ionizing radiation due to medical imaging has been observed since the 1980s. [1–3]

Recent retrospective studies on a cohort of about one million US people report that computed tomography and nuclear imaging together represent only 20 % of the total number of procedures but they contribute more than 75 % of the total effective dose to the population [4]. Other authors reported that cardiac imaging is responsible for 20 % of the collective effective dose [5].

Despite current risk estimates for cardiac imaging procedures which are just projections from the available epidemiological evidence, due to the absence of direct evidence from cardiac imaging, recently, the growing use of imaging procedures has raised concerns about the risks of exposure to low-dose ionizing radiation in the general population, in particular of malignancy [6, 7].

Thus, scientific societies focused their attention on producing recommendations aimed at reducing radiation exposure in cardiac imaging. Concerning myocardial perfusion imaging (MPI), imaging protocols that will achieve a radiation exposure of no more than 9 mSv in 50 % of the studies by 2014 have been recommended [8]. Several strategies can be adopted to contain patients’ exposure [9, 10]. New iterative reconstruction algorithms have been recently proposed for cardiac gated single photon emission computed tomography (SPECT) with 99mTc tracers, allowing for studies with reduced count statistics. This allows for half-dose, or even lower, cardiac gated SPECT, with image quality and quantitative data comparable to standard-dose studies [11–14].

The aim of the present study was twofold: to analyse the temporal evolution in the administered activities in patients scheduled for a dual-day 99mTc-sestamibi gated SPECT study and to compare different activity administration protocols in terms of patients’ and staff’s exposure.

Materials and methods

Data related to gated SPECT studies consecutively performed in our institution from 1 July 2002 to 31 January 2012 were retrospectively retrieved from our institutional database. Patients scheduled for a dual-day stress-rest 99mTc-sestamibi gated SPECT were allocated according to the study protocol adopted into group 1: patients injected with a fixed activity of tracer according to the diagnostic reference level suggested by the Italian Association of Nuclear Medicine (740 MBq up to 80 kg, proportionally adapted for weight <60 kg, 900 MBq 80–100 kg and 1,110 MBq more than 100 kg); group 2: patients injected with a weight-adjusted activity: 8 MBq/kg up to a maximum of 1,110 MBq; and group 3: patients injected with 4 MBq/kg and studies reconstructed with a new iterative reconstruction algorithm (wide beam reconstruction, WBR, by UltraSPECT).

A conventional dual-head Anger camera (Millennium VG, GE), equipped with a low-energy high-resolution collimator, was used in all studies, with the following acquisition parameters: 180° acquisition, 60 projections, 64 × 64 matrix, pixel size 6.3 mm and 8 frames/cycle gated acquisition. Group 1 and group 2 studies were reconstructed with a conventional filtered back-projection (FBP) algorithm (Butterworth 0.5–10). Group 3 studies were reconstructed with an ordered subset expectation maximization (OSEM)-modified iterative reconstruction algorithm (WBR), as previously described [13]. No attenuation or scatter correction was performed.

Patients’ exposure and detrimental effects

According to Italian and European regulations, the injected activities are recorded. Patients’ effective dose was estimated according to published data as follows: 0.0079 mSv/MBq for a stress study and 0.009 mSv/MBq for a rest study [15, 16]. The probability of detrimental effect was estimated, according to published data, on the basis of the approximated overall fatal risk coefficient (0.05×Sv−1) as recommended by the International Commission on Radiological Protection (ICRP) [17]

Staff exposure

The evaluation of the effective dose of the physician dedicated to performing the stress test and injecting the tracer was monitored with film badges for whole-body dose equivalent; monthly badge reports were reviewed retrospectively over the entire study period.

Statistical analysis

Data are reported as mean ± SD. When appropriate, 95 % confidence intervals (CI) are reported. Analysis of variance (ANOVA) was used to compare differences between groups and trends over time. Categories were compared with the chi-square test. A p value <0.05 (two-tailed) was considered significant.

Results

Study population

The study cohort consisted of 9,060 patients consecutively scheduled for a dual-day stress-rest sestamibi gated SPECT study from 01 July 2002 to 31 January 2012. The study cohort was allocated according to the protocol adopted: group 1: from 1 July 2002 to 1 May 2007 (period 1), 4,751 patients; group 2: from 2 May 2007 to 30 April 2010 (period 2), 2,844 patients; and group 3: from 01 May 2010 to 31 January 2012 (period 3), 1,465 patients. The patients in groups 2 and 3 were slightly older than group 1, according to the aging of the general population, and less frequently were evaluated after a recent acute myocardial infarction (Table 1).

Table 1 Clinical characteristics of the three study cohorts
Full size table

Effective dose of radiation

The average activity injected at stress and rest in the three groups is reported in Table 2. The average cumulative effective dose significantly reduced from 13.7 ± 2.7 mSv in group 1 to 9.5 ± 2.4 mSv in group 2 (p < 0.001) and a further reduction was documented in group 3 (5.7 ± 1.2 mSv, p < 0.001 vs groups 2 and 3) (Figs. 1 and 2). The percentile curves of the cumulative estimated patients’ effective dose in the three groups are reported in Fig. 3 and the trend in the 50th percentile at stress, rest and cumulative is reported in Table 3. A significant reduction over time in the cumulative (stress + rest) 50th percentile was documented and the suggested target of 9 mSv in at least 50 % of the study for a complete stress-rest evaluation was already obtained in group 2 patients, studied from 02 May 2007 to 30 April 2010 (Fig. 3).

Table 2 99mTc-Sestamibi average activities injected at stress and rest studies in the three study groups
Full size table
Fig. 1
figure 1

Estimated effective dose at stress, rest and cumulative in group 1 (black bars), group 2 (white bars) and group 3 (grey bars)

Full size image
Fig. 2
figure 2

Gaussian-fitted frequency distribution of stress studies’ effective dose in the three groups. A significant leftward shift was observed from group 1 (grey bars) to group 2 (white bars) and group 3 (light grey bars)

Full size image
Fig. 3
figure 3

Percentile curves of the effective cumulative dose in group 1 (blue), group 2 (red) and group 3 (green). Dotted lines indicate the target suggested by the American Society of Nuclear Cardiology of 50 % of studies with a cumulative effective dose less than 9 mSv [8]

Full size image
Table 3 Trend in the 50th percentile for stress and rest studies and cumulative
Full size table

The probability of detrimental effects was estimated in group 1 as 0.685×10-3 (95 % CI 0–1.3), in group 2 as 0.485×10−3 (95 % CI 0–1.3) and in group 3 as 0.285×10−3 (95 % CI 0–1.1), corresponding to 1:1,460 patients, 1:2,105 patients and 1:3,509 patients, in group 1, group 2 and group 3, respectively. Thus, a 1.4-fold and 2.4-fold reduction in groups 2 and 3, respectively, as compared to group 1 was documented.

When the exposure of the physician dedicated to performing the stress test and injecting the tracer was analysed, a significant reduction over time was observed in the exposure (normalized by the number of studies performed) from 2.1 μSv/study in group 1 to 1.5 μSv/study in group 2 and to 1.0 μSv/study in group 3 (all p < 0.001), which is more than 50 % reduction (Fig. 4).

Fig. 4
figure 4

Trend in average physician’s exposure. A significant reduction (>50 %) was documented from period 1 to period 3

Full size image

Discussion

In the present study, a significant reduction over time in patients’ dosimetry was observed. A simple weight-adjusted activity of injected tracer was able to reduce the estimated exposure by 30 % (from 13.7 mSv in group 1 to 9.5 mSv in group 2, p < 0.001). The additional use of a new reconstruction algorithm (WBR) was able to further reduce patients’ effective dose by an additional 40 % (from 9.5 mSv in group 2 to 5.7 mSv in group 3, p < 0.001); the total reduction in patients’ exposure from group 1 to group 3 was almost 60 %. At the same time, a significant reduction, greater than 50 %, in the staff exposure was also documented. Finally, in select group 3 patients, a stress-only strategy would be able to reduce patients’ exposure to as low as <3 mSv.

As a matter of fact, the combination of weight-adjusted tracer activity and the use of WBR, a new OSEM-derived iterative reconstruction algorithm, significantly reduced both patients’ and staff’s exposure and was able to reach the suggested target of a patient’s exposure of 9 mSv or less in more than 50 % of our population. Several softwares are now available from different vendors to reconstruct gated SPECT studies acquired with fast protocol (half-time or less), with image quality comparable to that of standard time acquisition. Several studies demonstrated that these softwares are able to provide good quality images also in studies obtained with reduced tracer activity. In particular, with WBR we demonstrated that performance and image quality was comparable to that of conventional FBP, allowing for either half-time or half-dose studies [11–14]

New SPECT camera technologies, with high sensitivity solid-state detectors and utilizing advanced reconstruction algorithms, also allow lower doses of injected activity for diagnostic studies as compared to conventional Anger cameras [18–21]. These systems, however, are more expensive and almost exclusively dedicated to cardiac studies.

Imaging protocols, tracer selection and hardware are only parts of the principle of “as low as reasonably achievable” (ALARA), and in line with the importance of justification of all studies involving ionizing radiation, an appropriate patient selection according to appropriateness criteria and clinical guidelines is the initial and equally important component of a programme of radiation exposure reduction. Moreover, in selected patients, a stress-only approach has been documented to provide accurate diagnostic and prognostic evaluation, with a reduced radiation exposure [22]

Radiation-related side effects

Great emphasis has recently been placed on the issue of patients’ radiation safety by physicians, patients and by the media. The risk of cancer in the range of a “low-dose” level of ionizing radiation (<100 mSv) is largely based on the “linear no-threshold” model extrapolated from follow-up of survivors of the atomic bomb explosions in Hiroshima and Nagasaki, Japan, in 1945 [6, 7, 9]. According to these assumptions, the lifetime risk of malignancy related to MPI is in the range of 1:2,000 patients for a 10 mSv effective dose. Results from the present study suggest that the combination of a weight-adjusted tracer activity with WBR is able to reduce the estimated risk up to 1:3,509, with a 2.4-fold reduction as compared to a fixed-dose protocol. In Italy the MPI rate is in the range of 1,700 stress-rest study/million inhabitants; with a low-dose approach, approximately more than 400 MPI-related malignancies over 10 years could be avoided. Furthermore, the majority of patients undergoing nuclear stress tests, in our population, are above 65 years of age. The lifetime risk of a radiation-related malignancy is much lower in older adults than children and fertile women. Thus, the risk of adverse events due to a serious heart disease or the risk of missing a serious cardiovascular diagnosis is much greater than the theoretical risk of radiation-related harm.

Limitations

The present report is based on a retrospective analysis. A prospective study would be required to verify the impact of low-dose protocols on patients’ long-term incidence of radiation-induced side effects. However, this would require a very large sample size (in the range of hundreds of thousands of subjects) with a very long-term follow-up. Results of the present study are based on tracer activities optimized for our gated SPECT camera, study protocols and the iterative algorithm employed (WBR) and do not necessarily apply to other camera systems or softwares. Each centre should define the optimal tracer activity according to its own protocols, reconstruction software and camera system in order to match image quality and radiation protection. Finally, attenuation correction was not performed in our population, so the impact of CT scan on global patients’ exposure was not assessed. Recent contributions, however, suggest that also for the use of CT in a hybrid imaging protocol, either for attenuation correction or non-invasive CT coronary angiography, low-dose protocols can be adopted [23–25]. With low-dose protocols, attenuation correction is expected to increase patients’ exposure by approximately 13 % for two correction maps (stress and rest); however, CT coronary angiography implies at least a 36 % additional exposure also with the newer low-dose protocol (Fig. 5)

Fig. 5
figure 5

Expected patients’ exposure according to the tracer dose and imaging protocol adopted. Gray bars indicate the expected patients’ exposure related to gated SPECT, black bars indicate the contribution of CT source in hybrid imaging. AC attenuation correction, CT computed tomography, CTA computed tomography angiography, DRL diagnostic reference limits. **Estimated exposure based on a dual slice CT system with default AC acquisition parameters (130 kVp; 60 mAs; pitch 1.5) for an average 20-cm scan length. Stress plus rest AC maps. *Stress plus rest AC estimated exposure with low-dose systems, based on data from [23. $Estimated hypothetical exposure with a low-dose acquisition protocol using an ECG-driven tube current modulation, a body mass index-adapted tube voltage modulation and prospective ECG-triggered sequential scanning. Data from [25]

Full size image

Conclusion

A simple change in the administered dose protocols can result in a significant reduction in patients’ exposure and allows the recommendations from scientific societies to be fulfilled. Further reductions can be obtained with new iterative reconstruction algorithms, with a 240 % concomitant reduction in the probability of malignancy. This could be particularly important in patients undergoing multiple evaluation over time. In selected patients, a stress-only strategy allows for studies with <3 mSv exposure. Thus, according to our results, to reduce patients’ and staff’s exposure, at least a software implementation should be encouraged and, accordingly, activities to be used for cardiac gated SPECT are to be revised. Finally, an update of the diagnostic reference limits should be considered.