Myocardial perfusion imaging (MPI) using combined computed tomography (CT) and single-photon emission CT (SPECT) systems plays an important role in the management of patients with coronary artery disease (CAD).1–4 The method can be used to assess myocardial perfusion and left ventricular function simultaneously. It is an especially valuable tool for assessing short-term risk of CAD, thus effectively guiding decision-making regarding revascularisation.3,4
The system usually consists of a dual-headed, large-detector gamma camera united with a multislice diagnostic CT, and is designed for sequential SPECT and CT imaging. MPI with CT-based attenuation correction (AC) shows a consistent improvement in image quality and in apparent diagnostic accuracy for the identification of CAD compared with non-attenuation corrected (NAC) MPI. However, MPI is susceptible to several complex methodological, biochemical and physiological factors that expose it to several potential artefacts and pitfalls, potentially limiting its diagnostic value (see Figure 1).
CT can be used for enhanced object localisation and tissue AC in SPECT; in addition, by itself it provides assessment of the prognostic significance of coronary calcium score (CCS) and coronary CT angiography (CTA), at least with a 64-slice CT. As an atherosclerosis imaging method, CCS is likely to provide greater long-term risk assessment, and is therefore more useful in determining the need for aggressive medical prevention measures.3 The key advantage of CTA over MPI is that the results are highly unlikely to be normal in patients in whom revascularisation would be warranted; this is in contrast to MPI, where a balanced reduction in perfusion can occasionally result in a normal finding (three-vessel disease) despite the presence of severe and extensive CAD.
This article will briefly review the benefits, artefacts and pitfalls of combined SPECT and multislice CT that may compromise the performance and interpretation of MPI.
Physical Performance of Single-photon Emission Computed Tomography and Computed Tomography
Single-photon Emission Computed Tomography
SPECT images suffer from noise due to low count statistics and poor spatial resolution. New detector technologies and iterative reconstruction packages have been developed to improve SPECT resolution down to 7mm and to decrease noise by simultaneously correcting for attenuation, scatter and collimator response.5,6 These corrections almost halve the acquisition time, and gated MPI examinations can be performed in 15 minutes while maintaining the same signal-to-noise statistics and image quality that previously would have required twice the duration.
Computed Tomography
Combined systems are usually available in six-, 16- or 64-slice CT. A 64-slice system allows the capture of the whole heart in about a six-second breath-hold, while the 16-slice system requires a breath-hold of about 15 seconds in CTA studies. The typical tube voltage is 120–140kV and the tube current ranges from 250 to 700mAm, resulting in a spatial resolution of 0.4mm with relatively low noise of 0.3%. Several excellent reviews provide a better understanding of CCS and CTA.1,8–10
Single-photon Emission Computed Tomography and Computed Tomography Co-registration
CT-based AC is now a recommended technique for improving image quality.7 The attenuation map required for AC-MPI can be acquired using a low-dose CT mode (tube current of 20–30mA).The co-registration of SPECT and the attenuation map need to be verified for every patient, even when a semi-automatic method for detection and correction of SPECT-CT emission–transmission misalignment is used. The misalignment in any direction has to be less than ±3mm.
Electrocardiogram-gated Myocardial Perfusion Imaging
Technetium-99m-labelled myocardial perfusion tracers allow simultaneous assessment of perfusion and left ventricular function using electrocardiogram (ECG)-gated SPECT (see Figure 2). A typical dose in the one-day protocol is 300MBq during symptom-limited bicycle exercise or pharmacological stress and 700–900MBq at rest (three to four hours after exercise). Imaging is usually initiated 30–60 minutes after injection. Since these tracers do not show a significant redistribution phenomenon, SPECT images reflect myocardial perfusion at injection time.
Sixteen frames per RR interval – the time between two consecutive R waves in the ECG – are acquired in a 128x128 matrix over 180° (or 360°). In total, 30–32 projections/head are obtained using a body-contour or circular acquisition. The energy window is centred on the 140keV photopeak. Typical CT tube voltage for AC is 140kV, typical current is 20mA and the CT slice thickness is usually 5mm for a 15cm field of view. The patient holds her/his normal breath during the CT scan.
Image Analysis and Interpretation
Transaxial, short-axial and long-axial slices are reconstructed with carefully performed AC. The total slice thickness used is typically 6–7mm (two slices summarised). Visual interpretation of the images is then performed. Perfusion defects scoring and evaluation of left ventricular function can be performed using programs such as QGS™ (Cedars-Sinai Medical Center).11 This software provides operator-independent and reproducible results.
Artefacts and Pitfalls
Even in dedicated imaging systems, MPI is influenced by complex methodological, biochemical and physiological processes; as a result, it is vulnerable to artefacts and pitfalls in clinical routine.2,12 The following discussion of artefacts and pitfalls is primarily based on an excellent review by Burrell and MacDonald.12 Artefacts and pitfalls can arise at any step in the MPI study and can be grouped into issues related to the patient, the technologist/physician or the equipment (see Table 1).
Burrell and MacDonald12 summarise: “It is essential for both the technologist and the interpreting physician to be aware of these potential sources of error, take appropriate steps to limit them beforehand, where possible correct them if they occur, and, when they can not be eliminated, recognise their potential impact on the interpretation of the study.”
Role of Computed-tomography-based Attenuation Correction in Myocardial Perfusion Imaging
There are very limited data available on the performance of multislice CT-based AC-MPI. Previous cardiac SPECT/CT systems provide only a low-power (2.5mAs) non-diagnostic CT scan, and the gantry rotation is slow, necessitating a relatively long time in order for the CT scan to be acquired.13 A short report by Dey et al.14 on the use of multislice CT in 31 patients concludes that SPECT imaging with CT-based AC showed a consistent improvement in image quality and apparent diagnostic accuracy for the identification of CAD compared with conventional NAC studies.
We performed quantitative analysis of the MPI data for every 10th of the first 132 patients studied with our SPECT/multislice CT. Visual and region of interest (ROI) analysis was performed. SPECT slices were reconstructed with AC and NAC. The effect of AC on image quality was classified by two nuclear medicine experts as either 1 or 0, where 1 = superior and 0 = equivalent or worse image quality due to the appearance of artefacts. Fifteen of 26 AC image sets (58%) were ranked to be superior compared with NAC images. ROI analysis revealed that the inferior–anterior ratio markedly increased with AC from 0.85 to 1.01. The accuracy of co-registration of SPECT and CT images was on average ±1.7mm. We concluded that the image quality was indeed improved with AC when there was no critical misalignment between SPECT and CT.
Discussion and Conclusion
MPI is a well-established and commonly used technique for the evaluation of CAD. A normal MPI study defines patients at low risk of subsequent cardiac events,4,15 and the risk increases exponentially with worsening perfusion abnormality.15,16 However, recent findings suggest a new paradigm: rather than identifying only those patients at risk, MPI in a testing strategy can identify which patients may benefit from revascularisation.3
One limitation of MPI is under-recognition of multivessel disease. SPECT can measure only relative uptake, not absolute myocardial perfusion.17 If there is decreased perfusion to the whole myocardium, the abnormalities may not be recognised,18 and CTA is probably the method of choice. New techniques for 3D co-registration of CTA and MPI will be developed,19,20 which will further improve diagnostic accuracy for identifying early CAD.
Schuijf et al.21 recently asked a relevant question: “What are the implications of the observations for the clinical use of multislice CT in addition to SPECT?” They addressed the question as follows: “Rather than competing, multislice CT and SPECT appear to be complementary techniques, each having a valuable position in the diagnostic work-up of patients with suspected CAD. Still, before adopting these algorithms in clinical practice, more studies are needed, particularly focusing on long-term prognosis.”
In conclusion, MPI retains its important role in the management of patients with CAD, and AC significantly improves the specificity of MPI in clinical settings. The future goal is that imaging of morphology, function and metabolism can be performed in a single session.