With the advent of fast scanning capability through the development of slip-ring technology, CT has become the imaging modality of choice in the assessment of coronary artery disease (CAD), particularly in the emergent situation of acute chest pain. Coronary computed tomography angiography (CCTA) is now widely used to image the morphology of coronary vessels to detect severe stenosis that requires revascularization.¹,² and ³ Even with its prevalence in the diagnostic armamentarium, CCTA remains an anatomical technique rather than a functional/physiologic technique for the sensitive detection of functional myocardial ischemia downstream to a stenosis.⁴,⁵ and ⁶ To address this shortcoming of CCTA, myocardial perfusion imaging (MPI) lately has become a focus of research for cardiac CT. In the triage of acute chest pain, the combination of CCTA and MPI promises to be the ideal one-stop shop for the comprehensive anatomical and physiologic evaluation of CAD⁶ before directing the appropriate patient to revascularization.

However, one main limitation of MPI with CT is the high radiation dose (approximately 20 mSv) from the necessity of repeatedly scanning the heart following injection of contrast (see next section). As such, recently a variety of methods have been developed to reduce the radiation dose of CT MPI. This chapter discusses these methods and illustrates them with examples.

To understand better how the new radiation dose reduction methods work, it is useful to describe the three different scanning protocols/methods for CT MPI. First, perfusion in ischemic myocardium is visually assessed from source CCTA images as the relative contrast enhancement with respect to normal myocardium⁷,⁸ (Fig. 3-1). Hypoattenuation relative to other parts of the myocardium is taken as ischemia. It is a qualitative assessment and depends on the time of arrival of contrast in the ischemic relative to that in the normal myocardium as well as the timing of the CCTA acquisition relative to the arterial (aorta) concentration. Second, the initial phase of time-density curves (TDCs) is acquired by scanning the heart with prospective ECG gating⁹,¹⁰ (Fig. 3-2). Myocardial perfusion is calculated as the front slope of the myocardial TDC normalized by the maximum value of the aortic TDC.¹⁰ Figure 3-2A-C shows an example of MPI in a dog model of acute heart attack. As shown, the method is able to detect stress- (adenosine) induced ischemia. One fundamental assumption of the upslope method is that during the time interval (10 seconds in Fig. 3-2) of its measurement, there is no outflow of contrast from the myocardium—no venous outflow. This is a rather stringent condition, and in general, it is not known whether it will be satisfied in all ischemic conditions. Third, entire TDCs are acquired by either cine-retrospective¹¹ or prospective ECG gating,¹⁰,¹² and kinetics modeling is applied to derive myocardial perfusion as shown in Figure 3-3. The application of kinetics modeling obviates the need of the no venous outflow assumption, and myocardial perfusion measurements obtained with this last method are the most accurate resulting in one-to-one correspondence with the gold standard microsphere measurements.¹³ The main disadvantage is the high radiation dose imparted.

As a comparison, the effective dose for a prospectively gated CCTA is about 3.7 mSv,¹⁴ while the effective dose of the kinetics modeling method for MPI of the whole heart using either cine-retrospective gating¹¹ or prospective gating¹⁰ is 19.4 and 6 to 17.7 mSv, respectively. The upslope method, because of its shorter acquisition window compared to the kinetics modeling method (Fig. 3-2 vs. Fig. 3-3), has an effective dose that is 40% to 50% less. Given the clinical utility of accurate MPI,⁶ there is an impetus to develop methods to reduce the higher effective dose required by the kinetics modeling. In the following sections, a number of recently developed methods for reduction of radiation dose are presented.

Besides prospective gating that can reduce the radiation dose of MPI by two to three times relative to retrospective gating, and currently, there are three methods under investigation: (a) iterative image reconstruction; (b) principal component analysis (PCA) of TDCs; and (c) compressed sensing (CS) image reconstruction.