Widely used single-photon emission computed tomography-myocardial perfusion imaging (SPECT-MPI) lacks spatial resolution to identify myocardial layers as well as the ability to measure absolute myocardial blood flow (MBF).1 The pitfalls of relative MPI using SPECT are well known in terms of underestimating the extent of coronary artery disease (CAD).2,3 These shortcomings have been overcome by cardiac positron emission tomography (PET) which quantifies absolute global MBF. Quantification of absolute MBF at stress and rest, and the derivation of myocardial flow reserve (MFR) is more desirable than relative flow assessment due to its diagnostic utility, powerful prognostic nature, improved accuracy, reproducibility, and simplified post-processing.1,4,5,6,7,8 Recent consensus statement from the American Society of Nuclear Cardiology and the Society of Nuclear Medicine and Molecular Imaging Cardiovascular Council indicates that these measures are at the cusp of translation into clinical practice.7 While it is well known that there is a transmural variation in MBF in CAD, wherein the subendocardium (SEN) is more vulnerable to ischemia than the subepicardium (SEP)9,10, the majority of cardiac PET literature has focused on measurement of average blood flow across the entire myocardial thickness rather than in layers. This is likely due to the complexities involved in making those measurements. However, if performed accurately, these measurements will reflect the true heterogeneous nature of ischemia and may have incremental diagnostic and prognostic value.

Over the past several decades, various techniques have been used to evaluate MBF in layers. The earliest studies, using radionuclide-labeled microspheres to measure regional MBF in animals, suggested a preferential increase in the blood flow to the SEP layer relative to the SEN layer in case of stenosis.9,10,11 Several studies utilizing different modalities including cardiac PET, magnetic resonance (MR), and computed tomography (CT) perfusion have demonstrated transmural variation in MBF in patients with CAD (Table 1).12,13,14 While it has been shown in these studies that it may be feasible to make these measurements using different modalities, the diagnostic utility of these measurements has not been well established. Current literature is conflictive in the added value of measuring MBF in layers: while MR assessment of MBF measurements in layers suggests a higher accuracy for CAD detection15,16, CT and [15O]H2O PET studies do not show a similar advantage over global MBF measurements.17,18

Table 1 Studies measuring myocardial blood flow in layers in patients with CAD
Full size table

Although transmural variation in MBF has been demonstrated by cardiac PET using [15O]H2O in patients with CAD17, in this issue of the journal, Sciagrà et al.19 for the first time explored the feasibility and utility of [13N]NH3 cardiac PET imaging for quantifying variations in MBF in separate layers of the myocardium using an automated procedure in patients with known or suspected CAD.

Sciagrà et al.19 studied 70 patients who underwent [13N]NH3 cardiac PET imaging, of whom 36 patients had significant CAD (defined as stenosis ≥ 70% on invasive coronary angiography) with 60 ischemic territories. Those authors did not observe any difference in resting global, SEN, or SEP MBF among patients with or without CAD. Whereas during hyperemia, there was a significant decrease in global, SEN, and SEP MBF in patients with versus those without CAD. Similar results were obtained when data were analyzed on a vascular territory basis. However, the measurement of MBF in layers did not show improved diagnostic accuracy above whole thickness measures. Whole thickness MFR and SEN MFR were significantly lower among patients with CAD, whereas hyperemic transmural perfusion gradient (TPG: the ratio of SEN MBF/SEP MBF) was not different between patients with or without CAD. They also demonstrate acceptable intraobserver correlation for ten randomly selected cases.

The current study differs in some respects from another cardiac PET study done by Danad et al.17, where [15O] H2O was used for quantification of MBF in layers in patients with an intermediate likelihood of CAD (defined as fractional flow reserve ≤0.80). In the study by Danad et al.17 MBF measures had similar directional results as observed by Sciagrà et al.19 However, they observed a lower hyperemic TPG in patients with significant CAD versus those without, which was not observed by Sciagrà et al.19 They also reported not only good intraobserver (like Sciagrà et al.19) but also reasonable interobserver correlation while making these measurements.

Taken together, both these studies suggest that measurement of MBF in layers in patients with CAD using cardiac PET is feasible. Both studies were done at experienced cardiac PET centers and demonstrate that more physiological MBF measurements are possible. However, the implementation of this knowledge beyond research realms may not have reached primetime yet. The amount of activity in the blood pool at the time of MBF measurement as well as differences in resolution related to positron ranges of different PET radionuclides will likely be the weighing factors in implementing these approaches. Furthermore, MBF measurements in layers are subject to similar pitfalls as global MBF measurements, including variability related to use of different software, lack of standardization in image acquisition/reconstruction protocols, and use of different radionuclides and scanners.20 There is also lack of data for normal MBF in layers using different radiotracers and the incremental diagnostic and prognostic utility of these measures.

This study paves the way to quantify MBF in layers using a cardiac PET with several caveats and questions which require further exploration:

  1. 1.

    Demonstration of intra- and interobserver variability of these measurements or lack thereof in multicenter studies.

  2. 2.

    Standardization of radiotracers, acquisition, reconstruction, kinetic modeling, and post-processing tools for making these measurements to reduce variability.

  3. 3.

    Assessment of variability introduced by the spatial resolution of cardiac PET, partial volume averaging, spillover, and other artifacts while making these measurements

  4. 4.

    Demonstration of the incremental value of transmural MBF variation in CAD patients beyond whole thickness measurements and possible ways to incorporate this information in making clinical decisions.

  5. 5.

    Demonstration of the diagnostic and prognostic values of these measurements in non-CAD diseases, including left ventricular hypertrophy, aortic stenosis, hypertrophic cardiomyopathy, and coronary microvascular disease.21

  6. 6.

    Comparison with MR and CT perfusion techniques in terms of accuracy, feasibility, radiation dose, cost, and incremental diagnostic value.

In conclusion, this innovative study should encourage further investigations to make a measurement of MBF in layers more reliable and clinically useful.