European Journal of Nuclear Medicine and Molecular Imaging

, Volume 41, Issue 3, pp 536–547

Low-dose single acquisition rest 99mTc/stress 201Tl myocardial perfusion SPECT protocol: phantom studies and clinical validation


    • Institute of Imaging & Computer VisionRWTH Aachen University
  • Barbra E. Backus
    • Department of Nuclear MedicineSt. Antonius Hospital
  • R. Leo Romijn
    • Department of Nuclear MedicineSt. Antonius Hospital
  • Herfried Wieczorek
    • Philips Research
  • J. Fred Verzijlbergen
    • Department of Nuclear MedicineSt. Antonius Hospital
    • Department of Nuclear MedicineErasmus Medical Center
Original Article

DOI: 10.1007/s00259-013-2551-3

Cite this article as:
Dey, T., Backus, B.E., Romijn, R.L. et al. Eur J Nucl Med Mol Imaging (2014) 41: 536. doi:10.1007/s00259-013-2551-3



We developed and tested a single acquisition rest 99mTc-sestamibi/stress 201Tl dual isotope protocol (SDI) with the intention of improving the clinical workflow and patient comfort of myocardial perfusion single photon emission computed tomography (SPECT).


The technical feasibility of SDI was evaluated by a series of anthropomorphic phantom studies on a standard SPECT camera. The attenuation map was created by a moving transmission line source. Iterative reconstruction including attenuation correction, resolution recovery and Monte Carlo simulation of scatter was used for simultaneous reconstruction of dual tracer distribution. For clinical evaluation, patient studies were compared to stress 99mTc and rest 99mTc reference images acquired in a 2-day protocol. Clinical follow-up examinations like coronary angiography (CAG) and fractional flow reserve (FFR) were included in the assessment if available.


Phantom studies demonstrated the technical feasibility of SDI. Artificial lesions inserted in the phantom mimicking ischaemia could be clearly identified. In 51/53 patients, the image quality was adequate for clinical evaluation. For the remaining two obese patients with body mass index > 32 the injected 201Tl dose of 74 MBq was insufficient for clinical assessment. In answer to this the 201Tl dose was adapted for obese patients in the rest of the study. In 31 patients, SDI and 99mTc reference images resulted in equivalent clinical assessment. Significant differences were found in 20 patients. In 18 of these 20 patients additional examinations were available. In 15 patients the diagnosis based on the SDI images was confirmed by the results of CAG or FFR. In these patients the SDI images were more accurate than the 99mTc reference study. In three patients minor ischaemic lesions were detected by SDI but were not confirmed by CAG. In one of these cases this was probably caused by pronounced apical thinning. For two patients no relevant clinical follow-up information was available for evaluation.


The proposed SDI protocol has the potential to improve clinical workflow and patient comfort and suggests improved accuracy as demonstrated in the clinical feasibility study.


Dual isotope protocolRest-stress myocardial perfusion imagingSingle photon emission computed tomographyCoronary artery diseaseSPECT


Myocardial perfusion imaging (MPI) with single photon emission computed tomography (SPECT) is widely used the for diagnosis of coronary artery disease (CAD). In the past 20 years several innovations like iterative reconstruction and new tracers have improved the diagnostic and prognostic value of SPECT imaging. Progress in image quality and ECG-gated imaging has been enabled by technetium-based myocardial tracers with better emission properties and higher applicable doses than the previously used thallium chloride. There is, however, still room for improvement regarding accuracy. Furthermore, new challenges are induced by the demand for dose and cost reductions as well as improved patient comfort.

Standard MPI is done in 1 or 2-day protocols. Two-day protocols exhibit ideal acquisition properties since contamination by prior tracer injection can be avoided. Moreover, rest studies can be skipped if stress studies are normal and the likelihood of CAD is low. The drawbacks of a 2-day protocol are the delayed diagnosis and the inconvenient scheduling on 2 days for patients and medical staff. One-day protocols overcome these disadvantages but require either a long waiting time of 3–4 h between rest and stress studies for radioactive decay or high doses for the second acquisition.

Dual isotope protocols with rest 201Tl and stress 99mTc have also been used for improved clinical workflow [1, 2]. Dual isotope imaging with separate acquisitions reduces the time of 1-day protocols from about 4–6 to 2–2.5 h. The thallium scan has to be performed prior to 99mTc injection. Otherwise, Compton scatter of the 99mTc emission at 140 keV would disturb the signal in the lower 201Tl energy window at 70–80 keV. The first scan has to be the rest study since recovery after physical or pharmacological stress would prolong the protocol.

For further improvement of workflow, single acquisition dual isotope (SDI) protocols have been proposed. Besides a higher patient throughput, simultaneous dual isotope imaging provides more patient comfort and perfect alignment of rest and stress studies. In conventional rest and stress imaging misregistration can lead to misinterpretation of infarct lesions as ischaemic areas [3]. Several methods have been discussed to overcome the cross-contamination of 99mTc in the lower 201Tl window to enable simultaneous dual isotope imaging. Monte Carlo simulations are the most promising approach for accurate scatter estimation and correction of cross-contamination of multiple isotopes [4, 5].

First attempts to implement SDI protocols have been based on the standard sequence of rest 201Tl and stress 99mTc injections. This protocol requires that the distribution of 201Tl injected at rest is not influenced by physical stress during the next phase of the cardiac study. Prior studies show that the distribution of 201Tl injected at rest remains unchanged after vigorous exercise in normal and ischaemic tissue [6]. Hence, 201Tl redistribution does not hamper the use of a simultaneous dual isotope protocol with rest 201Tl and stress 99mTc injection.

However, this protocol is not optimized for 201Tl imaging since the waiting time from 201Tl injection to acquisition is needlessly long. Therefore, we propose a single scan dual isotope protocol using 99mTc at rest and 201Tl at stress. This enables the acquisition of rest and stress images within 1 h on standard SPECT cameras. In this study we present a technical feasibility assessment of the proposed protocol by phantom studies and first clinical results in 53 patients.

Materials and methods

Rest Tc and stress Tl SDI protocol

For simultaneous dual isotope imaging, two isotopes with distinct emission energy are required. 99mTc-Sestamibi and 99mTc-tetrofosmin are known as excellent tracers for MPI. The gamma emission at 140 keV can be well detected by high-resolution collimated gamma cameras and is not as prone to scatter and attenuation as e.g. the lower thallium emission at 70–80 keV. Moreover, the relatively short half-life of 6 h offers good count statistics with moderate to low radiation burden. The minimal redistribution of these tracers allows some flexibility in protocol timing. A certain limitation of sestamibi is the slow liver clearance, requiring a waiting time of 15–30 min at stress and 45–60 min at rest between injection and acquisition. 99mTc-Tetrofosmin is known for a faster liver clearance, but image quality and clinical outcome are reported to be equivalent for both tracers [7, 8]. In this study we use 99mTc-sestamibi since the hospital has a long-lasting experience with this tracer and the protocol provides sufficient waiting time for liver clearance.

201Tl-Chloride is one of the first tracers that were used for MPI. It provides better linearity and higher uptake, 1–2 % at rest and 3–5 % at stress compared to1.0 ± 0.4 % at rest and 1.4 ± 0.3 % at stress for 99mTc-sestamibi [9]. The main drawback of 201Tl is its long half-life of 73 h causing a relatively high radiation burden for patients, limiting the applicable dose and count statistics. Furthermore, the main emission energy of 70–80 keV is highly susceptible to attenuation and Compton scatter. Hence, the image quality of 201Tl images has generally been seen as inferior compared to 99mTc images. The redistribution of 201Tl is routinely used for viability studies of the myocardium but may be unfavourable for standard MPI. According to the guidelines, acquisition should begin 5–10 min after peak stress injection of thallium. This offers a stable tracer distribution and reduces possible artefacts due to post-stress heart motion.

Using 99mTc at rest and 201Tl at stress we provide the necessary long waiting time for 99mTc and short waiting time for 201Tl. Additionally, the relatively high myocardial uptake of 201Tl at stress provides at least twofold better utilization of the injected dose. 99mTc-Sestamibi application at rest is beneficial since no significant redistribution is expected during the subsequent phase of physical or pharmacological stress. Figure 1 shows the timeline of an optimized protocol based on the feasibility of a simultaneous rest 99mTc-sestamibi/stress 201Tl dual isotope acquisition protocol.
Fig. 1

Timeline and doses of an optimized rest 99mTc/stress 201Tl simultaneous dual isotope protocol. Following 99mTc-sestamibi injection a waiting time of 20 min prior to stress is applied to provide about 50 min between injection and image acquisition for liver clearance. 201Tl is injected at peak stress. Waiting time prior to acquisition is needed to prepare the patient for acquisition and is beneficial to reduce upward creep

The 99mTc window can be used for ECG-gated cardiac imaging to determine functional parameters. This provides a unique combination of rest tracer distribution and post-stress wall motion. The 99mTc uptake at rest is unaffected by ischaemia and therefore the detection of myocardial boundaries in the gated images might be improved compared to conventional stress 99mTc images with ischaemia and infarct-related defects.

Waiting time between peak stress and acquisition is only 10 min in the new SDI protocol and notably shorter than the 30 min minimally recommended for standard stress 99mTc imaging. Therefore, the gated images acquired by the SDI protocol are supposed to demonstrate myocardial wall motion and functional parameters at stress potentially better than in standard stress 99mTc protocols.

Radiation burden

To confine the radiation burden of the dual isotope protocol, the injected dose is limited to a standard value of 74 MBq for 201Tl and 250 MBq for 99mTc. With these injected activities the effective dose of the protocol is 12.5 mSv and well within American Society of Nuclear Cardiology (ASNC) and European Association of Nuclear Medicine (EANM) guidelines.

Acquisition parameters

All acquisitions were made on a conventional dedicated cardiac SPECT camera (CardioMD, Philips Healthcare Systems, Andover, MA, USA) with a low-energy high-resolution collimator (LEHR). Simultaneously, 99mTc and 201Tl data are acquired in three different energy windows centred at 140.5 keV (width 20 %) for 99mTc, 167 keV (width 15 %) for the upper 201Tl line and 74.6 keV (width 15 %) keV for the main 201Tl emission.

During 1 scan 64 projections of 64 × 64 pixels with a pitch of 6.4 mm are acquired by 2 camera heads with an acquisition time of 40 s per projection. Transmission projections for creation of attenuation maps are determined by the Vantage system, using a collimated and moving 153Gd line source. Since transmission and emission data are acquired simultaneously, down-scatter into the 153Gd energy window at 100 keV must be corrected. For this purpose electronic windowing of this energy window is used. Contamination of the lower 201Tl window by collimated 153Gd line source emission is known to be negligible. The radiation dose of 50 μGy is also negligible in comparison to the radiation burden caused by tracer injection.

Reconstruction and scatter correction

Cross-contamination has been identified as the limiting factor for simultaneous dual isotope protocols. Most important is the down-scatter of the 140 keV 99mTc emission line into the lower 201Tl energy window at 70–80 keV. In patients, approximately 27 % of the primary 99mTc counts are scattered into the lower 201Tl window. Previous clinical studies have shown that uncorrected 99mTc down-scatter leads to a systematic underestimation of 201Tl lesion severity [1]. On the other hand, due to the lower abundance of the high energy lines of 201Tl, down-scatter of the upper 201Tl window at 167 keV into the 99mTc window is only about 2.9 % and therefore negligible.

Several correction methods for cross-contamination of isotopes in simultaneous acquisitions have been proposed. The triple energy window (TEW) method for (down-) scatter correction is based on the measurement of two additional, narrow energy windows above and below the actual emission window. Using these two energy windows the amount of scatter in the emission window is estimated and corrected by subtraction [10]. This method is straightforward to implement and has been validated in simulation and clinical studies also for dual isotope imaging. It is also capable of handling scatter originating from outside the field of view (FOV) [11, 12]. We have chosen a fully 3D Monte Carlo-based method for scatter estimation and correction. In order to reduce radiation burden we aim at low doses of Tl, and Monte Carlo-based scatter correction has shown accurate results and improved contrast and noise properties [5, 1315]. Besides correction methods, the utilization of a low 99mTc dose of 250 MBq reduces down-scatter into the lower 201Tl window. Based on the uptake ratios of 99mTc-sestamibi and 201Tl-chloride we expect a 99mTc to 201Tl ratio of 0.5:1 to 1.4:1 in the myocardium.

The scatter estimation using Monte Carlo simulation is implemented in the ordered subset expectation maximization (OSEM) reconstruction algorithm [16]. Additionally, attenuation correction and resolution recovery are applied, the latter by means of an analytically calculated collimator point response function.

For visual interpretation, 5 full iterations with 8 subsets are performed and images are post-filtered with a Butterworth filter (fifth order, cut-off frequency 0.5 voxels/pitch) and reorientated in long and short axis views.

Phantom measurements

The technical feasibility of the SDI protocol and performance of the reconstruction method are assessed by measurements of an anthropomorphic torso phantom. The phantom contains separate compartments for the lungs, partly filled with Styrofoam, the liver and the spinal cord simulated by a Teflon cylinder. The heart insert consists of two interleaved chambers, the inner chamber defining the left ventricle and the outer one mimicking the myocardium. The myocardium is equipped with two fillable lesions. One lesion is positioned anterior near the apex and covers one eighth of the myocardial circumference. The second, larger lesion in the inferior wall covers one fourth of the circumference and is positioned near the myocardial base. Both lesions are 2 cm high and filled with 99mTc only to simulate ischaemic regions. We take advantage of the different half-lives of 99mTc and 201Tl to create the expected 99mTc to 201Tl ratios by a series of three measurements resulting in ratios of 2:1, 0.7:1 and 0.5:1.

Table 1 summarizes the absolute tracer concentrations. These are higher than the concentrations expected in patients, which are about 30 kBq/cc for both 99mTc and 201Tl in the myocardium. Comparable count numbers are achieved by cutting the acquisition time to 20 s, which is one half of the value used in clinical patient studies.
Table 1

Concentrations of isotopes for different anthropomorphic torso phantom acquisitions

Concentration in kBq/cc


1st acquisition

2nd acquisition

3rd acquisition



































Decreasing concentrations are caused by radioactive decay. Time span between 1st and 2nd acquisition is 13 h and about 15 h between 1st and 3rd acquisition. Tracer concentration in lesion is equal to myocardial concentration for 99mTc and zero for 201Tl

Clinical feasibility

Validation protocol

A clinical feasibility study was done to evaluate the image quality of the simultaneous dual isotope protocol. All human and animal studies were approved by the Ethics Committee of St. Antonius Hospital, Nieuwegein, and were therefore performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. All persons gave their informed consent prior to their inclusion in the study. The timeline of the validation protocol is depicted in Fig. 2. 99mTc stress and rest reference scans are taken from a 2-day protocol.
Fig. 2

Timeline of the validation protocol to test rest 99mTc/stress 201Tl simultaneous dual isotope study. Stress reference 99mTc is performed at least 1 day before rest reference 99mTc and SDI imaging. On the second examination day a single 99mTc-sestamibi injection is used both for rest reference and SDI imaging. In contrast to the proposed protocol the time span between injection and SDI imaging is increased from 50 min to about 2 h

The stress reference acquisition is performed with a dose of 500 MBq 99mTc. One week later, when the patient has approved participation in the study, the rest 99mTc and simultaneous dual isotope measurement are conducted. Injection at rest of 250 MBq 99mTc-sestamibi is followed by 45 min waiting time and 20 min rest reference measurement (first scan). After 10 min transfer and preparation, exercise or pharmacological stress is applied with 74 MBq 201Tl-chloride injection at peak stress, followed by 10 min waiting and 20 min SDI imaging (second scan). During the study it was recognized that in two obese patients [body mass index (BMI) > 32] the injected 201Tl dose was not sufficient for adequate image quality. Hence, the dose for obese patients was adapted to 110 MBq and in one case to 131 MBq.

Reference images are reconstructed with a maximum likelihood expectation maximization (MLEM) algorithm with 30 iterations and attenuation correction. For the SDI acquisitions the same acquisition and reconstruction parameters as for the phantom study are used. Since our research reconstruction algorithm is not capable of gated reconstruction, functional parameters are determined by QGS based on MLEM reconstruction. Down-scatter of 201Tl high energy emission into the 99mTc energy window is negligible. Thus gated reconstruction of 99mTc distribution is feasible without down-scatter correction. End-systolic and end-diastolic volumes as well as the ejection fraction of the left ventricle are determined as functional parameters from the gated acquisitions. Ejection fractions of 45 % or less are regarded as significantly lowered.

For visual interpretation the images are blinded and reorientated in short and long axis views. The evaluation itself is done by an experienced nuclear medicine physician. We evaluate images on visual interpretation and semi-quantitative 17-segment, 5-point scoring to compare SDI and 99mTc reference imaging. Comparison by automatic quantitative assessment of myocardial perfusion via e.g. QPS would be appealing but is not feasible since as yet no normal database has been created for this particular protocol and reconstruction.


For the study we selected patients for whom, after the initial stress scan, an additional rest scan was requested. Patient demographics are given in Table 2. In total 53 (14 female and 39 male) patients were prospectively included in the study with most of them (70 %) suffering from documented CAD. About 38 % had a history of prior infarction and were referred for MPI, because of recurrent chest pain complaints. In 37 patients 2 or more risk factors (diabetes mellitus, hypertension, hypercholesterolaemia, smoking, obesity and family history) were present. Patients who were not able or unwilling to undergo a second stress test were excluded from the study.
Table 2

Demographic data of patients included in study


Patient group, 53 patients


14 female, 39 male


61.7 ± 9.5 years (37–81)


28.2 ± 4.8 (18–39)

History of myocardial infarction

20 (38 %)

Coronary artery disease

37 (70 %)

Risk factors

 Diabetes mellitus

13 (25 %)


28 (53 %)


25 (47 %)


6 (11 %)


21 (40 %)

 Family history

34 (64 %)

In 14 patients pharmacological stress with adenosine was applied. The remaining 39 patients were stressed physically on a bicycle ergometer. The mean workload during reference stress was 131 W and 137 W during SDI stress. The mean maximum heart rate during stress was 154 bpm at reference and 152 bpm at SDI stress. The difference in workload is small but significant (p = 0.0074), while the difference in heart rate is not (p = 0.146). Since SDI as well as reference stress were both symptom limited and differences are small we regard the stress level as equivalent.

The clinical follow-up of the patients participating in the study was tracked and later cardiac events, examinations or treatment were used to support the diagnostic value of the SDI myocardial perfusion test.

Statistical analysis

The analysis of the differences between reference and SDI imaging for continuous parameters like heart rate, workload and ejection fraction were performed using the paired t test (two-tailed). For statistical analysis the summed stress score (SSS) was also regarded as a continuous parameter. A probability value (p value) <0.05 was considered statistically significant.


Phantom measurements

Two sets of reconstructed and reorientated images with 99mTc to 201Tl ratios of 2:1 and 0.5:1 of the anthropomorphic torso phantom are shown in Fig. 3.
Fig. 3

Reconstructed anthropomorphic phantom data mimicking stress 201Tl and rest 99mTc. Upper image (a) is acquired with a ratio of 99mTc to 201Tl of 2:1. Lower image (b) shows the third acquisition obtaining a 99mTc to 201Tl ratio of 0.5:1. In both images the upper rows show 201Tl images, representing stress in the proposed dual isotope protocol. The lower rows contain 99mTc images, which simulate rest imaging. Anterior and inferior lesions are clearly visible in 201Tl images and also hazily in 99mTc images, due to the plastic walls of lesion containers

Both 201Tl lesions are clearly seen on both stress images. In particular the larger lesion in the inferior wall is reported to be better visible when the concentration of 99mTc is reduced. The lesions are also hazily seen in the rest images due to the plastic lesion walls, which is inevitable. Image quality was assessed by an experienced nuclear medicine specialist and reported to be fully sufficient for medical application with a clear preference for the second image.

Clinical feasibility

In 51 of 53 patients the image quality of the SDI acquisitions was assessed as fully suitable for clinical evaluation. In the remaining two obese patients with a BMI > 32 the 201Tl dose of 74 MBq was not sufficient for appropriate imaging and these two patients were excluded from the study. As mentioned in the “Materials and methods” section, the 201Tl dose was adapted for obese patients during the study. After this adaptation the image quality of later obese patients was acceptable for clinical assessment. However, the adjustment increases the radiation burden for these patients to a maximum value of 17.5 mSv.

In some patients the SDI rest images show a high 99mTc uptake in the gall bladder and the intestinal tract. This is caused by the set-up of the validation protocol, which differs from the suggested, optimized protocol by an extra hour waiting time between 99mTc injection and SDI acquisition. However, this uptake pattern has not hampered the clinical value of the SDI rest images.

Ischaemia or infarction was found in 39/53 patients (6/14 female and 33/39 male); this high portion of positive tests is caused by selection bias. In 31 patients (58 % of all patients) SDI and reference images led to equivalent assessments, though in general the lesions appeared clearer and more severe in SDI images. In 20 patients (38 %) significant differences were found between SDI and 99mTc reference studies. In 18 of these patients coronary angiography (CAG) partly with fractional flow reserve (FFR) assessment was performed. In 15 of these cases (28 % of all patients) angiography proved that the SDI result was more accurate than the 99mTc reference. In three cases (6 %) minor ischaemic lesions detected by the SDI study were not supported by CAG; we regard SDI imaging as less accurate in these patients. In one of these cases a small apical reversible defect (SSS = 3) was found that must be seen as pronounced apical thinning. The results are presented in Fig. 4.
Fig. 4

Comparison of reference and SDI imaging. Image quality of SDI was sufficient in all patients except two obese patients. In the course of the study the 201Tl dose was adapted to BMI and no more patients had to be excluded. If significant differences were found, further tests like CAG were used to check SDI and reference imaging results. In 15 patients SDI imaging was more consistent with the additional test. In three patients SDI found minor ischaemic lesions, which could not be supported by follow-up. In one of these cases this was probably due to pronounced apical thinning

Figure 5a shows the SSS of SDI and 99mTc reference separated in the categories ‘clinically equivalent’, ‘SDI more accurate’, ‘SDI less accurate’ and ‘no follow-up available’. For patients in the category clinically equivalent the correlation between the SSS of SDI and 99mTc reference is highly significant (r = 0.9447, p <0.0001). The mean difference (between 99mTc reference and SDI) of SSS for those patients is −1.55 points and significant (p < 0.0001). This relates to the aforementioned observation that lesions appear clearer and more severe in SDI imaging. Also more vascular territories exhibit defects in SDI than in the 99mTc reference. The differences in SSS between 99mTc reference and SDI are illustrated by a Bland-Altman plot (Fig. 5b).
Fig. 5

Results of summed stress scoring (SSS) of the validation protocol. The correlation between SSS of SDI and 99mTc reference imaging is depicted in a. The differences in SSS between 99mTc reference and SDI are illustrated by a Bland-Altman plot (b). Correlation coefficient (r = 0.9447, p < 0.0001) and mean difference and standard deviation are calculated for patients with a clinically equivalent outcome in SDI and 99mTc reference

For comparison we selected in all patients those territories which showed a defect in the respective segments. For the left anterior descending artery (LAD) area the 99mTc reference demonstrated a defect in 28 patients compared to 39 in SDI imaging. In the right coronary artery (RCA) and circumflex (CX) territories the 99mTc reference image showed significant lesions in 21 and 22 patients, respectively, whereas SDI imaging depicted defects in 29 patients for each of these two coronary territories.

Functional parameters

In Table 3 the average heart rate (HR), left ventricular ejection fraction (LVEF), left ventricular end-diastolic volume (LVEDV) and left ventricular end-systolic volume (LVESV) are shown for 99mTc reference and SDI imaging. SDI imaging obtained the highest heart rate due to the short waiting time and the highest LVEF was determined by stress 99mTc imaging. The differences in HR, LVEF and LVEDV between reference stress and SDI are significant. The averaged ejection fractions of rest 99mTc and SDI show very good concordance.
Table 3

Averaged functional parameters in reference (stress 99mTc and rest 99mTc) and SDI imaging


HR (bpm)


LVEDV (ml)

LVESV (ml)

Stress 99mTc (ref.)


57.8 %



Rest 99mTc (ref.)


55.0 %



SDI imaging


55.0 %



p (SDI-stress 99mTc ref.)





p (SDI-rest 99mTc ref.)





p (stress 99mTc ref.- rest99mTc ref.)





The LVEF was also compared for each patient individually. Since it is known that determination of functional parameters is difficult in cases of small hearts and high ejection fractions we excluded ten patients with an LVEF of more than 65 % in all three acquisitions. The differences (EFTc-stress-EFSDI) and (EFTc-rest-EFSDI) are shown in Bland-Altman plots (Fig. 6). On average the difference between stress 99mTc and SDI (EFTc-stress-EFSDI) was 2.3 ± 5.7 % and between reference rest and SDI the difference was (EFTc-rest-EFSDI) = −0.3 ± 5.6 %. For comparison the difference (EFTc-stress-EFTc-rest) was in the same order of magnitude, 2.6 ± 5.6 %.
Fig. 6

Bland-Altman plots of ejection fraction differences betweenSDI and Tc-stress (a) and Tc-rest (b). Patients with an ejection fraction above 65 % in both reference and SDI studies were excluded since the low spatial resolution of SPECT results in high variance and no clinical value

Figure 7 shows one patient data set in which SDI imaging showed more accurate diagnostic results than conventional imaging.
Fig. 7

Patient A is a 56-year-old man (BMI 25.5) suffering from typical angina. Heart rate and workload reach the targeted values of 164 bpm and 185 W, respectively. He complains about chest pain at stress with abatement after exercise. The reference scan (a) shows severe ischaemia in the anterior wall. Gated study depicts motion abnormality suggesting a light infarction in the inferior wall. On the SDI images (b) the anterior lesion appears more severe and additionally inferoseptal ischaemia is clearly visible. A bullseye plot of the scoring is shown in c. A CAG performed later detected significant stenoses in RCA, distal LAD, MO1 and MO2. For treatment the stenoses were dilated by percutaneous coronary intervention and stenting


In this study we tested the feasibility of simultaneous, dual isotope acquisition of rest 99mTc and stress 201Tl distribution and developed a new optimized, dual isotope protocol for improved patient comfort. The known issue of down-scatter of 99mTc emission at 140 keV into the lower 201Tl windows was addressed by scatter correction based on Monte Carlo simulation. High 201Tl uptake at stress and a low 99mTc dose were utilized to reduce the 99mTc to 201Tl ratio and limit the effect of down-scatter on 201Tl images.

Measurements of an anthropomorphic torso phantom covering the expected 99mTc to 201Tl ratios of 0.5:1.4 showed adequate image quality and detectability of 201Tl lesions.

Clinical feasibility was shown by comparison of SDI imaging to 99mTc rest and stress reference acquisitions. The image quality of the SDI images was reported to be sufficient for clinical assessment with the exception of two obese patients, which led to weight-dependent dose adaptation. With this dose adaptation to a maximum of 131 MBq of 201Tl-chloride the SDI images of the subsequent patients were of good quality also for patients with a high BMI up to 36. Hence, obesity is not an exclusion criterion for SDI patients, but 201Tl-chloride dose has to be adapted and increased radiation burden has to be taken into account particularly for younger, obese patients.

The majority of patients showed equivalent image quality and clinical results in reference and SDI images, and in most patients with different findings and available clinical follow-up SDI imaging was regarded to be more accurate.

The dual isotope approach in MPI imaging intends to reduce overall examination time since waiting times for nuclear decay or recovery after stress can be avoided. Simultaneous reconstruction spares one imaging procedure and therefore reduces protocol time further and improves patient comfort. Moreover, simultaneous reconstruction allows one to take advantage of the tracers’ clinical properties since acquisition is independent of the sequence of tracer injection. Accordingly the proposed protocol exploits relatively high 201Tl uptake at stress and stability of myocardial 99mTc-sestamibi distribution on the rest image. Furthermore, the waiting time required for liver clearance of 99mTc is partly utilized for patient stress leading to an overall examination time of only 1 h. For 201Tl no waiting time is required, but it is recommended to start acquisition 10 min post-stress to reduce upward creep. In contrast to pure 201Tl imaging the SDI protocol allows acquisition of gated images based on the rest 99mTc distribution, and in contrast to conventional gated stress 99mTc studies, for which a waiting time of at least 30 min is recommended, the acquisition can start shortly after stress. It is supposed that this early gated imaging potentially demonstrates wall motion defects, and therefore functional imaging at stress may be more valuable than acquisitions conducted with a longer waiting time. For segmentation of the myocardium, which is a prerequisite for determination of functional parameters like ejection fraction, the rest distribution is beneficial since it is unaffected by ischaemia.

As expected, the heart rate during SDI acquisition is significantly higher than at reference rest and at stress due to the smaller time span between maximum stress and acquisition. On average, the ejection fraction at SDI is slightly lower than stress 99mTc (55 vs 57.8 %), but well comparable to rest 99mTc. However, quite high differences of 2.3 ± 5.7 % in ejection fractions between stress 99mTc and SDI are found when we compare patients individually. The comparison of rest 99mTc and SDI shows a small average difference but still a high variation (EFTc-rest-EFSDI) = −0.3 ± 5.6 %. For ejection fraction a day-to-day variability of about 3 % is reported for rest 99mTc protocols [17]. Slightly smaller variability (2.3 %) for 99mTc acquisition is reported if the patient is not repositioned [18]. For a 201Tl protocol the same study reports a variability of 5.2 % which is similar to the one we found in the present study. Although we use 99mTc for gated acquisition in the SDI protocol, count rates of 99mTc in our study are comparable to standard 201Tl protocols since we use a low 99mTc dose to reduce down-scatter. Thus, the low count statistic is supposed to be responsible for the variation of ejection fraction. A possible next step is to introduce resolution recovery and Monte Carlo-based scatter correction also for gated images to assess functional parameters of low-dose images.

The Bland-Altman plots also indicate that the higher variability is present at high ejection fraction. This might also be linked to limited spatial resolution, since high LVEF are often linked to small LVESV, which are difficult to determine with typical SPECT resolution. However, classification of patients with a lower ejection fraction (<46 %) found 14 patients and is consistent for reference and SDI acquisition, except for one patient, who showed an LVEF of 51 % at SDI and 43 % at stress 99mTc and 45 % at rest 99mTc. In summary, gated post-stress imaging utilizing rest tracer distribution is promising, but low-dose imaging is suspected of inducing variability in particular at high ejection fractions.

A certain benefit of simultaneous acquisition is inherent coregistration of rest and stress images, and this avoids artefacts induced by inconsistent reorientation of rest and stress images. In addition, this also eases reorientation, since reorientation parameters have to be determined only once and on the image that shows myocardial orientation more clearly, generally the rest image.

The need for improvement of workflow and patient comfort has also been addressed by sequential dual isotope and stress-only protocols. Stress-only imaging provides low radiation burden (4 mSv) and can be performed within 1 h. Based on the assessment of stress acquisition, it is decided whether an additional rest scan is required. Patients who are scheduled for the rest scan cannot take advantage of the short examination time. For those who definitely need a comparison of rest and stress images, e.g. patients with a history of CAD, the stress-only approach is not reasonable. Moreover, if stress and optional rest scans are planned to be performed on the same day a high flexibility in department management is required, since the number of rest scans is unknown a priori.

Stress-only and SDI imaging can be considered as complementary approaches. For those with a low likelihood of rest scan requirement, stress-only imaging can reduce the radiation burden. Others who probably need rest and stress scans can benefit from the single acquisition and short protocol time of the SDI protocol but have to accept a higher but still limited radiation burden of 12.5 mSv.

Dedicated cardiac CZT cameras can shorten the imaging time to 3–10 min. This improves overall examination time, but the effect is limited since waiting times cannot be reduced. The SDI protocol can be used on conventional SPECT cameras and would only require adaptation of the reconstruction software.

The focus of developing the SDI protocol is on workflow improvement; hence, SDI images are compared to conventional 99mTc reference imaging with standard MLEM reconstruction. For the SDI reconstruction additional scatter correction and resolution recovery are essential components. It is a limitation of the current study that we cannot determine whether the improved accuracy of SDI images found in 15 patients is caused by a different reconstruction method or by the use of 201Tl as stress tracer. This has to be addressed by further investigation.

A second limitation is that clinical follow-up is not available for all patients. This might bias the assessment of the SDI protocol, although it is encouraging that a defect was found that was not verified by CAG in only 3 of 18 patients with clinical follow-up, and in 1 case this was probably caused by exaggerated apical thinning.

The creation of attenuation maps by 153Gd line source imaging can be seen as a further limitation since CT-based attenuation maps are expected to be of higher quality, but are also prone to misregistration of attenuation and emission images.

Utilization of 99mTc-tetrofosmin instead of 99mTc-sestamibi would be an interesting extension of the study since the faster liver clearance promises a further shortening of the optimized protocol (shown in Fig. 1) and the lower gall bladder uptake might reduce extra cardiac activity, which was observed in some patients during the clinical feasibility study [19, 20].

It must be considered as a drawback of the SDI protocol itself that gated images are only available post-stress and no gated rest images can be acquired. Therefore, a comparison of stress and rest wall motion and functional parameters is not possible with SDI imaging. In addition the lack of rest function and wall motion impedes the verification of rest perfusion lesions by wall motion assessment. However, in principle the proposed protocol allows an additional gated pure rest 99mTc scan without tracer injection after the patient has recovered from stress, if necessary.

Due to different attenuation coefficients of 99mTc and 201Tl emissions, image artefacts by inadequate attenuation correction are potentially different at rest and stress acquisitions. This might hamper the identification of these artefacts.

Although the effective dose of 201Tl has been overestimated in the past, the use of 201Tl causes a relatively high radiation burden compared to low-dose 99mTc protocols [21]. Hence, the proposed SDI protocol is especially recommended for elderly patients or those who will probably need rest and stress scans. Patients with a high probability of ischaemia or those who have suffered from prior infarction are typical examples. Elderly patients also probably benefit most from sparing one straining imaging procedure by a single scan examination.

The optional possibility of viability studies using 201Tl redistrinbution is also a benefit of the proposed SDI protocol. In the present study we focused on myocardial perfusion; therefore, we did not schedule an additional acquisition to examine viability. The clinical feasibility protocol shown in Fig. 2 is already stressful for the participating patients; hence, a further viability scan including waiting times of 3–4 h was not performed. This is an appealing topic for subsequent studies using the optimized protocol (see Fig. 1) with an overall protocol time of about 1 h. The benefits and drawbacks of the proposed SDI protocol discussed in this section are summarized in Table 4.
Table 4

Summary of benefits and drawbacks of the proposed SDI protocol



- Improved workflow due to stress and rest acquisition in about 1 h

- No stress-only imaging possible

- Improved patient comfort, by single acquisition imaging

- Only one post-stress gated study, no determination of functional parameters at rest; rest perfusion findings cannot be verified by functional imaging

- Inherent coregistration of stress and rest images

- Different attenuation coefficients might induce different image artefacts in stress and rest images

- Post-stress gated imaging using tracer distribution at rest

- Absolute radiation burden is higher than for 99mTc low-dose or stress-only protocols (still well within EANM limits)

- Optional 201Tl redistribution study for viability assessment without extra dose


- Optional gated rest 99mTc acquisition without extra dose after patient recovered from stress

- Low radiation burden for a dual isotope protocol (12.5 mSv)


Phantom and patient studies indicate the technical and clinical feasibility of rest 99mTc/stress 201Tl dual isotope single acquisition imaging. The optimized SDI protocol, basing on this feasibility, has the potential to improve clinical workflow and patient comfort. Furthermore, an improved accuracy is suggested as demonstrated in patient examples.


The nuclear medicine team of St. Antonius Hospital, Nieuwegein and in particular Naomi Mul are gratefully acknowledged for their abundant help and support. Likewise the authors would like to thank Horace Hines and Rolf Bippus for their long-lasting support of this project. This work has been funded by Covidien and by Philips Research.

Conflicts of interests

The authors declare that they have no conflicts of interest. One of the authors (H. Wieczorek) is an employee of Philips Research, Eindhoven.

Copyright information

© Springer-Verlag Berlin Heidelberg 2013