MedicaMundi

J.F. Verzijlbergen, B.E. Backus, R.L. Romijn - Department of Nuclear Medicine, St. Antonius Hospital, Nieuwegein (Utrecht), the Netherlands.

H. Wieczorek, T. Dey - Philips Research Laboratories, Aachen, Germany.

R. Bippus - Philips Research Laboratories, Hamburg, Germany.

H.H. Hines  - Philips Medical Systems, San Jose, CA, USA.

Myocardial Single Photon Emission Computed Tomography (SPECT) is the most widely used imaging modality for the diagnosis of functionally obstructive atherosclerotic coronary artery disease (CAD). CAD is caused by obstruction of coronary arteries and is the leading cause of death in Western countries.

Myocardial SPECT has now become standard for the evaluation of regional myocardial blood flow, called myocardial perfusion. Myocardial SPECT enables the identification of perfusion defects and risk stratification of patients with proven CAD. SPECT and gated SPECT are performed simultaneously and enable assessment of myocardial perfusion as well as evaluation of post-stress functional parameters like left ventricular volumes and ejection fraction.

In most cardiac SPECT protocols two acquisitions are performed, one study at rest and a second study after stress in order to detect and differentiate between ischemic lesions, which are regions affected by reduced blood flow, and scarred tissue due to a previous myocardial infarction.

There is an ongoing clinical study at St. Antonius Hospital, Nieuwegein (Utrecht), the Netherlands, in cooperation with Covidien Nederland B.V., Zaltbommel, the Netherlands, and Philips Research Laboratories, Aachen, Germany. The aim of this study is the development of a new myocardial SPECT protocol which will provide an improved clinical workflow, and better diagnosis, based on appropriate SPECT image quality. Technically, this protocol is performed with simultaneous technetium Tc-99m and thallium Tl-201 imaging, enabled by the use of attenuation and scatter correction in myocardial SPECT. This paper gives an overview of current myocardial SPECT and shows first clinical images using the new protocol.

Cardiac SPECT and competing modalities

Myocardial SPECT is the standard imaging modality for coronary artery disease in the United States, with an installed base of more than 13,000 SPECT systems and 9 million patients undergoing myocardial perfusion scans per year. The advantage of this non-invasive examination lies in the clinical experience with data based on a huge number of patients, laid down in databases of normal and abnormal cardiac SPECT scans, and the availability of well established and inexpensive tracers. In Europe, cardiac SPECT is not as widely spread as in the US, with five times fewer examinations per capita, but is likewise considered the standard for a five year prognosis of annual mortality due to CAD. This imaging method has a high sensitivity and specificity for coronary artery disease.

Among competing modalities in cardiology, positron emission tomography (PET) is promoted for superior image quality, optimized workflow, ability to quantitate myocardial flow, and higher sensitivity and specificity for CAD. However, cardiac PET is not considered a first line modality for myocardial perfusion imaging due to the high cost of tracers and equipment. While rubidium Rb-82 is available from generators but expensive, ammonia and PET imaging needs a cyclotron at the clinical site due to the short (10 min) half-life of the N-13 isotope. H2O-15 labeled water cardiac PET imaging needs a cyclotron on-site as well. Other and potentially less expensive tracers based on fluorine F-18 (Flurpiridaz) are still in the clinical test phase.

Computed Tomography Coronary Angiography (CTA) is feasible with modern multi-slice CT detectors, applying a contrast agent intravenously, and delivers high-quality images of the coronary arteries. It is unique in being the only non-invasive method that looks directly at coronary artery morphology. The diagnosis, however, is potentially different from that of functional imaging due to the mismatch between function and anatomy. Furthermore, the specificity of coronary lesions is low in the presence of calcified plaques, whilst a normal coronary angiogram has a very high negative predictive value. The method has recently come into discussion because of the high radiation dose applied. CTA may become a standard examination for coronary artery disease after development of appropriate low dose protocols and extensive clinical testing.

Calcium scoring is a CT-based method to detect coronary calcification. For patients with known coronary disease the calcium score has an additional prognostic value for long term survival. However, there is no short-term correlation of calcium scoring and myocardial perfusion imaging.

Echocardiography is the only available inexpensive imaging procedure besides SPECT. Inter-observer variability in interpretation of echo images is very high, and image quality and sensitivity in the detection of cardiac disease are lower than for SPECT. The main advantage of ultrasound is the high negative predictive value, comparable to SPECT and giving excellent risk stratification at relatively low cost and no radiation dose.

Cardiac magnetic resonance imaging (MR) accurately depicts morphological and functional data, ventricular and valvular function, wall thickening, perfusion, and myocardial viability. Despite the accuracy of this modality, MR is much less frequently used than echo due to lower availability, higher cost and long examination times. The special patient situation in the MR gantry and the need of repeated breath holds make this exam highly discomforting. MR can therefore not compete with myocardial perfusion SPECT as an established technology.

SPECT tracers and protocols

The SPECT tracer market is well established and consolidated. Thallium-201 chloride has been the first tracer used for cardiac imaging, available since 1977. The extraction of this tracer into the myocardium is high and shows a high linearity with the rate of myocardial blood flow, with 1-2% myocardial tracer uptake at rest and 3-5% under stress. Imaging must start 10 minutes after injection. Besides myocardial perfusion assessment, viability studies are used to differentiate between ischemia and myocardial infarct estimating the amount of redistribution of thallium in the myocardium. The main disadvantages of thallium are its long physical and biological half-life, requiring a low injected dose because of radiation protection considerations, and the low energy of the main emission lines at 70-80 keV energy leading to strong attenuation artifacts if no attenuation correction is applied. The resulting non-optimal image quality has led to a decline in the use of thallium chloride as a tracer for myocardial SPECT. Recently, it became evident that the radiation dose caused by thallium-201 has been overestimated for several years, but this is not yet documented in the procedural guidelines for myocardial perfusion imaging [2].

Technetium-based tracers such as Tc-99m-sestamibi and Tc-99m-tetrofosmine have been available from 1989 onwards. They show fast blood clearance but lower cardiac uptake compared to thallium. Technetium has a non-linear extraction-flow relationship with approximately 1.2% uptake at rest and 1.5% under stress, resulting in a relatively low lesion contrast. Imaging regularly starts 30-60 minutes after injection under stress and 45-90 minutes at rest, with the long waiting time required by slow excretion through the hepatobiliary system. The short physical half-life of six hours for Tc-99m allows for a higher injected dose. Additionally, self-absorption of the 140 keV-emission line in the patient is low so that the image quality obtained with technetium based tracers has been considered superior to thallium.

Most cardiac SPECT protocols use two image acquisitions, one at rest and one after stress. When technetium-based tracers are used in a one-day protocol, at least four hours waiting time between the two scans are required to reduce crosstalk from the first into the second SPECT image. According to EANM procedural guidelines [3] a three times higher dose is injected in the second examination, resulting in 13-16 mSv effective patient dose.

Alternatively, a two-day protocol allows using the same injected dose for both exams so that the effective dose is in the range of 10-15 mSv. This protocol is better in terms of defect contrast but is inconvenient and causes a delay in clinical workflow.

A separate dual-isotope imaging protocol using Tl-201 for the rest study and Tc-99m for the stress study has been developed at Cedars Sinai Medical Center [1], taking advantage of the Anger camera’s ability to collect data in different energy windows. In this procedure the thallium rest scan is followed by the technetium stress scan so that contamination of the thallium image at 70-80 keV energy by down-scatter from the technetium emission line at 140 keV is prevented. The gated stress image provides the necessary count statistics to be used for evaluation of functional parameters. Reduction of waiting time between rest and stress imaging results in completion of the entire procedure in less than two hours. The effective dose received by the patient is in the range of 22-27 mSv for this protocol under ASNC guidelines [4].

Simultaneous instead of sequential dual isotope imaging would be highly advantageous because of an optimized clinical workflow, and the camera acquisition time is cut in half. This approach was suggested earlier but shown to be unfeasible since at that time the software for correction of attenuation, scatter and cross-contamination of different isotopes necessary for simultaneous imaging was not yet available [12]. Meanwhile, correction for attenuation is supported both by ASNC and SNM [5], while correction for attenuation, scatter and collimator resolution is standard in state-of-the-art reconstruction software [6]. New developments in iterative reconstruction have shown that reconstruction can even account for cross-contamination of different isotopes, as required for the simultaneous dual isotope protocol [7, 8].

Technical feasibility

We have evaluated the technical feasibility of simultaneous dual isotope imaging on a Philips CardioMD gamma camera at St. Antonius Hospital, Nieuwegein. The camera is equipped with a Vantage^TM Gadolinium-153 scanning line source system that determines the attenuation map of the patient during the SPECT exam [9]. Simultaneously, Tc-99m and Tl-201 data are acquired in three different energy windows centered at 140.5 keV for Tc-99m, 167.4 keV for the upper Tl-201 line and 68.9-80.3 keV for the main Tl-201 emission. Data from these four energy windows are concurrently obtained by the gamma camera acquisition system.

Since the emission spectra of Tc-99m and Tl-201 are mutually contaminated mainly through down-scatter of gamma quanta within the patient, the reconstruction algorithm has to include corrections in order to obtain non-contaminated images of both isotopes. For this purpose an iterative dual-matrix OSEM reconstruction method has been developed at Philips Research Laboratories, Aachen, which performs a simultaneous reconstruction of thallium and technetium images using all available projection data from all energy windows. A fast Monte Carlo simulation is used to correctly quantify patient scatter and cross-contamination of different isotopes on basis of the simultaneously acquired attenuation map. Additionally, the collimator response function is included in the system model and corrected using resolution recovery so that virtually all physical effects affecting the SPECT images are taken into account [10].

We have assessed the performance of our reconstruction software by phantom experiments using an anthropomorphic torso phantom with a cardiac insert [11]. This phantom has a warm background with separate compartments for liver and lungs and a Teflon cylinder to simulate attenuation of the spinal cord. The heart insert includes an inner chamber for the left ventricle and an outer chamber for the myocardium. Two separate lesions, one with 5.6 ml volume in anterior position near the apex, another with 10 ml volume in inferior position near the base of the heart, were filled with Tc-99m activity to simulate lesions at stress in the Tl-201 image as expected in the proposed clinical protocol described below.

Images were reconstructed with 5 iterations and 8 subsets and post-filtered using a Butterworth filter of 5th order with 0.5 cycles per pixel cutoff frequency. Clinical evaluation was done on images reoriented by the AutoQuant software on a Philips JetStream workstation.

Figure 1 shows reconstructed Tc-99m rest and Tl-201 stress data for two phantom images acquired during ten hours of stepwise acquisition, simulating different Tc : Tl ratios on the same phantom. Thallium lesions are clearly seen on both stress images but definitely better visible when the concentration of both isotopes is comparable. This difference is caused by high energy contamination due to unavoidable Tl-200 and Tl-202 isotopes contamination that is not easily compensated for. The lesions are also hazily seen in the rest images due to the plastic lesion walls, which is inevitable. The quality of the images has been evaluated as fully sufficient for medical application, with a clear preference for the second image.

The SDI protocol

With the available correction software there is no reason to stick to the sequence of thallium and technetium scans, as with the separate dual isotope protocol. On the contrary, the use of technetium for the first scan is preferable because Tl-201 stress can be performed during the 45-60 minute waiting time required for hepato-biliary clearance of Tc-99m. When thallium has been injected the simultaneous acquisition of both images can be done with only ten minutes waiting time after peak stress. The timeline of this protocol, depicted in Figure 2,  shows that the combined rest and stress studies can be completed in slightly more than one hour.

Figure 2. Clinical timeline for Tc-rest / Tl-stress simultaneous dual isotope protocol.

This sequence is optimal for both isotopes, not only for the recommended waiting times, but for several other reasons. Technetium does not show any redistribution in the myocardium, as is known for thallium, so that the post-stress acquisition of the rest image does not pose any problems. The myocardial extraction of thallium shows better linearity than that of technetium. Under stress, 3-5% uptake is estimated in the myocardium, which is much higher than the 1-2% uptake after injection at rest. The use of thallium under stress instead of rest is in fact a necessary precondition to keep the radiation burden of the dual isotope protocol low.  With 250 MBq (6.8 mCi) Tc-99m injected at rest and 74 MBq (2 mCi) Tl-201 injected at stress we calculate an effective dose of only 12.5 mSv.

The use of thallium for the stress image does not rule out gated cardiac imaging. In fact, gated technetium images have a higher chance of demonstrating an intact LV-wall because of the injection at rest while the ability to demonstrate ischemia- or infarct-related wall motion abnormalities remains intact. Functional parameters are therefore evaluated in the same way as with any gated technetium protocol.

Clinical feasibility

A clinical feasibility study has been started at St. Antonius Hospital, Nieuwegein, to evaluate the image quality of the simultaneous dual isotope protocol. In this validation protocol, a 250 MBq Tc-99m-sestamibi rest injection is followed by 45 minute waiting time and 20 minute rest reference measurement (first scan). After 10 minute transfer and preparation, treadmill stress is applied with 74 MBq Tl-201-chloride injection at peak stress, followed by 10 minute waiting and 20 minute SDI imaging (second scan). A stress reference measurement after 350 MBq Tc-99m-sestamibi injection is added for validation of the Tl-201 SDI-stress scan (Figure 3).

Figure 3. Timeline for Tc-rest / Tl-stress simultaneous dual isotope validation protocol

Figures 4-6 show examples of patient images, with the stress measurements shown on the left side and the rest measurements on the right side. All images are shown in the short axis, horizontal long axis and vertical long axis view. The left side shows a comparison of the Tl-201 SDI stress measurement (upper row) with the Tc-99m stress reference (lower row), while on the right side there is a comparison of the Tc-99m SDI rest measurement with the Tc-99m rest reference in the upper and lower rows, respectively.

Figure 4. Patient A.

Patient A is a 66-year-old female patient with a mixture of typical and atypical chest pain and cardiovascular risk factors like hypertension, hypercholesterolemia and a family history of cardiac events.

The patient stressed 100 W (106% of the predicted exercise capacity) and reached a heartrate of 153 bpm (105%). During the stress test the patient did not suffer from chest pain or dyspnea and the abnormal repolarization at rest remained unchanged during exercise.

Stress and rest myocardial SPECT are of excellent quality and reveal no perfusion defects. The left ventricle is not dilated and post-stress LV-EF is 67%. LV-EF at rest is 65%.

Fig.  5: Patient B.

Patient B is a 66-year-old female patient who suffered an anterior infarct in 1978 and CABG in 2000. She was treated in the outpatient clinic because of mitral valve insufficiency and decreased LV-function. She was considered a ICD candidate. Although the patient did not suffer from chest pain, ischemia was considered because of decreased LV-function. The patient exercised 80 W (83%) and finished with a heart rate of 167 bpm (107%) because of fatigue and “heavy legs”. There were no signs of chest pain, but only mild dyspnea during recovery. There were no repolarization changes during exercise.

The stress and rest images are of good quality and reveal an anteroseptal-apical infarct. No ischemia is detected, but the LV-cavity looks dilated, which is confirmed on the post-stress gated SPECT. EDV is increased to 251 ml and ESV to 178 ml. LV-EF is 29%. Wall motion abnormalities are noted in a quite diffuse pattern with akinesia anteroseptal/apical. Also in this patient a conservative treatment regimen was chosen.

Fig.  6: Patient C.

Patient C is a 71-year old female patient with a previous history of CAD and PCI of the LAD. During coronary angiography a 50% ostium stenosis of the RCA was noted but regarded as functionally not significant.

The patient stressed 63 W (67%) and reached a heart rate of 148 bpm (99%). The stress test was stopped because of chest pain and dyspnea which she recognized from stressful situations during daily life. The ECG showed 1.5 mm horizontal ST-depression in leads V5 and V6.

The overall quality of the stress and rest myocardial SPECT images was good and revealed no stress-induced perfusion defects. The images remained unchanged after comparison with the resting situation. Post-stress gated SPECT showed a normal wall motion pattern and LV-EF above 70%. EDV was 65 ml (normal). LV-EF at rest was also larger than 70% and EDV at rest was 72 ml. A conservative treatment plan was chosen.

Discussion 

These first results of the clinical feasibility study at St. Antonius Hospital show that the simultaneous dual isotope protocol we propose is a viable alternative to one- or two-day technetium protocols and the sequential dual isotope protocol currently used. The interpretation of thallium stress SDI images is equivalent to that of separately acquired technetium images. The excellent thallium image quality is mainly due to the high myocardial uptake of Tl-201 in the stress study and improved iterative reconstruction methods including attenuation correction. When we compare the concentration of tracers in the myocardium during measurement we calculate 2.4 MBq of Tc-99m (250 MBq injection, 1.2% myocardial uptake at rest, 20% decay after 2 hours waiting time in the clinical evaluation) versus 3.0 MBq of Tl-201 (74 MBq injection, 4% myocardial uptake at stress) and a ratio of Tc : Tl = 0.8 : 1. Accounting for the different isotope branching ratios and stronger absorption of the thallium radiation we see that the number of detected thallium quanta is at least as high as for technetium.

Impact on clinical workflow

The proposed protocol will have a significant effect on the clinical workflow, both concerning the patient imaging and the evaluation step. On the patient side, there is only one acquisition, meaning a reduction in the time required for the study as a whole, or an increase in patient throughput with more patients being imaged on the same equipment per working day. Since there is no second imaging occasion there will be no “lost” studies due to patients not returning for the second study. For the evaluation of images, the perfect alignment of rest and stress studies will result in improved assessment of the images. On the other hand, there are disadvantages, especially the lack of the possibility to choose a “stress-only” study, and the missing opportunity to compare between stress and rest gated left ventricular ejection fractions.

Conclusions

The clinical dual-isotope protocol based on technetium rest and thallium stress imaging has several benefits besides a shorter overall examination time. First of all the quality of thallium stress images is comparable to that of technetium rest images due to the high myocardial uptake and linearity of thallium chloride, and the two images are naturally well registered. Furthermore the rest distribution is preferable for evaluation of functional parameters since it is less affected by lesions in the stress distribution of technetium. The protocol is therefore well suited for gated cardiac imaging. We are aiming at an overall radiation exposure of 12.5-15 mSv for this dual-isotope protocol which is well within the recommended range of the EANM procedural guidelines and far below the radiation exposure of currently used sequential dual-isotope protocols.

The first clinical images have confirmed our rating of this new myocardial SPECT protocol and will be followed by further assessment of clinical cases.

References

[1]  Berman DS, Kiat H, Friedman JD, et al. Separate Acquisition Rest Thallium-201 / Stress Technetium-99m Sestamibi Dual-Isotope Myocardial Perfusion Single-Photon Emission Computed Tomography: A Clinical Validation Study. J Am Coll Cardiol. 1993; 22(5): 1455-1464.

[2]  The biokinetic table data for Tl-201 published in ICRP No. 53 had a typographical error resulting in a 50% overestimated radiation dose for this isotope. See ICRP53 – Addendum 5, 6 and 7, p. 14, http://www.icrp.org/docs/add_5-7_to_P53.pdf.

[3]  EANM/ESC Procedural Guidelines for Myocardial Perfusion Imaging in Nuclear Cardiology. Eur J Nucl Med Mol Imaging. 2005; 32(7): 855-897.

[4]  American Society of Nuclear Cardiology. Imaging Guidelines for Nuclear Cardiology Procedures. J Nucl Cardiol. 2007; 14: e39-60.

[5]  American Society of Nuclear Cardiology and Society of Nuclear Medicine Joint Position statement. Attenuation Correction of Myocardial Perfusion SPECT Scintigraphy. J Nucl Cardiol. 2004; 11: 229-230.

[6]  Heller GV, Bateman TM, Cullom SJ, Hines HH, Da Silva AJ. Improved Clinical Performance of Myocardial Perfusion SPECT Imaging using Astonish Iterative Reconstruction. Medicamundi 2009; 53(3): 43-49.

[7]  Frey EC, Tsui BMW. A New Method for Modeling the Spatially-Variant, Object-Dependent Scatter Response Function in SPECT. Conference Record. IEEE 1996; 2: 1082-1086.

[8]  Beekman FJ, De Jong HW, van Geloven S. Efficient Fully 3-D Iterative SPECT Reconstruction with Monte Carlo-Based Scatter Compensation. IEEE Trans Med Imaging. 2002; 21: 867-877

[9]  Case AJ, Hsu BL, Bateman TM, Cullom SJ. A Bayesian Iterative Transmission Gradient Reconstruction Algorithm for Cardiac SPECT Attenuation Correction. J. Nucl Cardiol. 2007; 14(3): 324-333.

[10]  Botterweck H, Bippus R, Goedicke A, Salomon A, Wieczorek H. Quantitative Simultaneous Multiple Isotope SPECT Imaging with Iterative Monte-Carlo Reconstruction, Proc. 9th Fully 3D Image Reconstruction Meeting, Lindau. 2007: 221-224

[11]  Anthropomorphic Torso Phantom with Cardiac Insert. Data Spectrum Corporation, Hillsborough NC, USA.

[12]  Kiat H, Germano G, Friedman J, et al. Comparative Feasibility of Separate or Simultaneous Rest Thallium-201/Stress Technetium-99m-Sestamibi Dual-Isotope Myocardial Perfusion SPECT. J Nucl Med. 1994; 35: 542-548.