Comparison of Microsphere-Equivalent Blood Flow (¹⁵O-Water PET) and Relative Perfusion (^99mTc-Tetrofosmin SPECT) in Myocardium Showing Metabolism-Perfusion Mismatch

MATERIALS AND METHODS

Data Analysis

Reorientation and Overlay.

¹⁸F-FDG PET and ¹⁵O-water PET studies were acquired primarily using ECAT 7.1 software (CTI/Siemens). Transaxial SPECT slices were imported into the ECAT 7.1 software. Next, all PET and SPECT images were reoriented in left ventricular short-axis slices of 1.2-cm thickness. For reorientation of the dynamic ¹⁵O-water PET studies, the reorientation axes of the respective ¹⁸F-FDG PET studies were transferred to the ¹⁵O-water PET studies and controlled for accuracy on early ¹⁵O-water study frames delineating the ¹⁵O-water-filled left ventricle chamber.

Each of 3 ¹⁸F-FDG PET short-axis slices (apical, midventricular, basal) was divided into 6 ROIs (segments) measuring 7 mm in radial diameter (anteroseptal, septal, posteroseptal, posterolateral, lateral, and anterolateral), yielding a total of 18 segments per dataset (Fig. 1). These segments were projected to both the ¹⁵O-water PET images and the ^99mTc-tetrofosmin SPECT images.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Definition of 18 myocardial tissue ROIs on 3 representative short-axis slices (apex, midventricular, and base). Left ventricular input (LV-Inp.) ROI is located in basal slice.

Quantification.

Myocardial ^99mTc-tetrofosmin and ¹⁸F-FDG PET uptake was assessed by calculating the count densities (counts per pixel) in every segment. The reference region for both ^99mTc-tetrofosmin and ¹⁸F-FDG uptake was the region with the maximum ^99mTc-tetrofosmin uptake. Segmental ^99mTc-tetrofosmin and ¹⁸F-FDG uptake values were then expressed as percentages of the respective reference region.

Myocardium was classified as normal if ^99mTc-tetrofosmin was >70% and as MPMM if ^99mTc-tetrofosmin was ≤70%, if ¹⁸F-FDG minus ^99mTc-tetrofosmin was ≥20%, and if ¹⁸F-FDG PET was >70%. This classification was derived from previously reported studies on the same type of patient and using identical methods that showed that MPMM myocardium often reflects altered, dysfunctional myocardium capable of functional recovery after revascularization (18,19). Relative ^99mTc-tetrofosmin uptake was averaged by segment for the 6 segments and for all ROIs for both normal and MPMM myocardium.

MBF_kin and MBF_micr.

MBF and water-perfusable tissue fraction (PTF) were calculated for each myocardial segment with a previously validated fitting routine based on a single-tissue-compartment model (11,12). The arterial input function was obtained from an ROI placed in the left ventricular chamber close to the base (26). MBF_kin was calculated as flow rate per mass of myocardium that is capable of rapidly exchanging radiolabeled ¹⁵O-water (mL·min⁻¹·g⁻¹). MBF_kin is free of the so-called partial-volume effect (PVE) due to the PTF, which is defined as the ratio of the mass of perfusable tissue within a given ROI to the volume of that ROI (g/mL) (13,15,16). The product of MBF_kin and PTF reflects the non-PVE-corrected blood flow in this ROI volume (13,15,16) in mL·min⁻¹·mL⁻¹ and was therefore termed the MBF_micr (15).

To prove that the inversion of the PVE correction of MBF_kin works well, MBF_kin, PTF, MBF_micr, ^99mTc-tetrofosmin, and ¹⁸F-FDG uptake values were quantified in ROIs of different sizes (Fig. 2A), all containing approximately the same myocardial mass, as described by Iida et al. (13). The dependence of the quantities on the ROI volumes is shown in Figure 2B.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

(A) ROIs with increased volumes but all containing approximately same myocardial mass. (B) ROI volume dependencies of MBF_kin, ^99mTc-tetrofosmin uptake, MBF_micr, tissue fraction, and ¹⁸F-FDG uptake. All values are expressed as ratios of respective value with A (smallest ROI).

MBF_micr (mL·min⁻¹·mL⁻¹) values were calculated for each segment, and the results of both runs were averaged. The mean MBF_micr of all 6 segments (anteroseptal, septal, posteroseptal, posterolateral, lateral, and anterolateral) and of all ROIs was calculated for both normal and MPMM myocardium.

Segmental MPMM-to-Normal Percentages.

For each of the 6 segments, both the MBF_micr and the relative ^99mTc-tetrofosmin uptake values in MPMM myocardium were divided by the segmental averaged corresponding values in normal myocardium, yielding the segmental MPMM-to-normal percentages.

RESULTS

MBF_micr and ^99mTc-Tetrofosmin Uptake in “Normal” Myocardium

The mean MBF_micr of patients with severe CAD in myocardial segments classified as normal (n = 493) was 0.64 ± 0.01 mL·min⁻¹·mL⁻¹, and the respective mean relative ^99mTc-tetrofosmin uptake was 86% ± 1%.

The segmental distribution of the MBF_micr values (Fig. 3) shows only a small variability within a range from 0.58 ± 0.02 mL·min⁻¹·mL⁻¹ (anteroseptal regions) to 0.68 ± 0.02 mL·min⁻¹·mL⁻¹ (lateral regions). Similarly, the ^99mTc-tetrofosmin uptake shows a small variation over the segments from 83% ± 1% (posteroseptal regions) to 91% ± 1% (lateral regions). Figure 3 shows that both MBF_micr and ^99mTc-tetrofosmin uptake tend toward higher values in the lateral than in the septal regions.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Mean MBF_micr (mL·min⁻¹· mL⁻¹) and relative ^99mTc-tetrofosmin uptake of investigated 6 segments and of all segments classified as normal (abscissa). ^99mTc-tetrofosmin uptake values are normalized to respective patient’s maximum uptake region, set as 100%. Ant = anterior; lat = lateral; post = posterior; seg. = segments; sep = septal.

MBF_micr and ^99mTc-Tetrofosmin Uptake in MPMM

MBF_micr, at 0.60 ± 0.03 mL·min⁻¹·mL⁻¹ (−6%), was not significantly reduced in those segments classified as MPMM (n = 82), whereas mean relative ^99mTc-tetrofosmin uptake was significantly reduced to 56% ± 1% in MPMM versus normal myocardium (−35% [relative]; P < 0.001) (Fig. 4).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Mean MBF_micr (mL·min⁻¹·mL⁻¹) and relative ^99mTc-tetrofosmin uptake in normal myocardium and MPMM myocardium. ^99mTc-tetrofosmin uptake values are normalized to respective patient’s maximum uptake region, set as 100%. n. s. = not statistically significant.

Segmentwise Comparison of MPMM-to-Normal Percentages for MBF_micr and ^99mTc-Tetrofosmin Uptake

The segmental MPMM-to-normal percentage describes the relationship between MBF_micr and ^99mTc-tetrofosmin uptake in MPMM compared with normal myocardium within the same segment. These calculations yielded values from 81% ± 9% (lateral regions) to 106% ± 7% (posteroseptal regions) for MBF_micr and from 56% ± 6% (anterolateral regions) to 72% ± 3% (anteroseptal regions) for ^99mTc-tetrofosmin uptake (Fig. 5). In all but the anteroseptal region, ^99mTc-tetrofosmin uptake was significantly more reduced than was MBF_micr. For the septal (n = 24), posteroseptal (n = 16), and posterolateral regions (n = 17), the relative differences were −22%, −37%, and −36%, respectively, and reached significance at P < 0.005. For the lateral (n = 8) and anterolateral (n = 8) regions, the relative differences were −25% and −39%, respectively, and reached significance at P < 0.05.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Segmental MPMM-to-normal percentages of mean MBF_micr and relative ^99mTc-tetrofosmin uptake. Mean normal values are set as 100% for each segment for either parameter. n = number of MPMM segments. *P < 0.05. ***P < 0.005. Ant = anterior; lat = lateral; post = posterior; sep = septal.

DISCUSSION

Many studies have shown that myocardial ^99mTc-tetrofosmin uptake correlates closely with blood flow (2,3), but several authors have pointed out that in animal models, ^99mTc-tetrofosmin uptake is more reduced than reperfusion blood flow after acute ischemia (4,5). These findings suggest that myocardial alterations affect ^99mTc-tetrofosmin uptake independently of blood flow. To overcome the limitations of a perfusion tracer whose uptake is related to energy-dependent processes (Na⁺/H⁺ ion channel activity, cellular and mitochondrial membrane potentials (6–10)), we used ¹⁵O-water PET as a reference method in assessing MBF. In this approach, relative ^99mTc-tetrofosmin uptake and MBF_micr were compared in normal and MPMM myocardium, and the flow-independent influence of cellular alterations on ^99mTc-tetrofosmin uptake in MPMM myocardium was quantified.

The mean MBF_micr (n = 493) in normal myocardium was 0.64 ± 0.01 mL·min⁻¹·mL⁻¹, which agrees with the findings of Iida et al. (15), who reported an average MBF_micr of 0.63 mL·min⁻¹·mL⁻¹ in 9 healthy volunteers. Estimation of MBF_micr therefore seems to work well. In addition, the unaffected MBF_micr in normal regions agrees well with the findings of Gerber et al. (27), who reported that in remote regions of dysfunctional myocardium, MBF_kin was similar to that in “normal” volunteers.

As expected from our classification, ^99mTc-tetrofosmin uptake in MPMM myocardium was significantly reduced, from 86% ± 1% (normal) to 56% ± 1%. Contrary to these findings, the MBF_micr showed only a nonsignificant reduction, from 0.64 ± 0.01 mL·min⁻¹·mL⁻¹ to 0.60 ± 0.03 mL·min⁻¹·mL⁻¹. Despite the assumption of a strictly flow-proportional perfusion marker, the discrepancy between reduced ^99mTc-tetrofosmin uptake and preserved MBF_micr in MPMM versus normal myocardium indicates that in MPMM myocardium, processes governing ^99mTc-tetrofosmin uptake or retention seem to be impaired. Contrary to these findings, ¹⁵O-water diffusion along the sarcolemmal membranes was unaffected by these alterations, since otherwise the MBF_micr would also have to be reduced.

Arbab et al. (6) reported that amiloride, an inhibitor of the Na⁺/H⁺-antiporter, reduced ^99mTc-tetrofosmin uptake in cultures of rat myocardium cells by about 35%. Arbab et al. also reported that ^99mTc-tetrofosmin-incubated cells released 38% of the accumulated tracer if they were treated with carbonyl cyanide m-chlorophenylhydrazone, an uncoupler of oxidative phosphorylation that depolarizes the mitochondrial membrane potential (6). Younès et al. (9) found that 2,4-dinitrophenol, another mitochondrial uncoupler, caused a 92% release of accumulated ^99mTc-tetrofosmin in incubated isolated adult rat heart mitochondria. Spadafora et al. (10) reported that the cytoprotective effects of trimetazidine, which are related to the mitochondrial function and not to coronary hemodynamic changes, increased myocardial ^99mTc-tetrofosmin uptake after 1 wk of administration in CAD patients. All these results indicate that ^99mTc-tetrofosmin uptake and retention are related to active ion transport systems and membrane potentials as well as to mitochondrial integrity.

Comparing the unchanged MBF_micr with reduced ^99mTc-tetrofosmin uptake and retention suggests that the discrepancy between these 2 findings may be caused by alterations of active ion transport systems (e.g., Na⁺/H⁺-antiporter), decreased membrane potentials, or ultrastructural alterations of mitochondria in MPMM versus normal myocardium. These assumptions agree well with data from Schwarz et al., who showed that ultrastructural alterations of mitochondria, particularly reduction in size and reduction in number of cristae, are present in reversibly dysfunctional MPMM myocardium (22,23). These data also agree well with data from Ausma et al., who showed that ultrastructural alterations of mitochondria are present in reversibly dysfunctional myocardium (28).

Reduced ^99mTc-tetrofosmin uptake and impaired retention cannot be differentiated on the basis of these data, since acquisition was performed 60 min after injection and the resulting ^99mTc-tetrofosmin values are determined by uptake and retention. Studies comparing early and late ^99mTc-tetrofosmin SPECT are needed to differentiate between impaired uptake and retention mechanisms in these patients.

Comparing the 6% MBF_micr flow reduction and the 35% reduction in ^99mTc-tetrofosmin uptake yields an approximate landmark value of about 29% for flow-independent ^99mTc-tetrofosmin uptake reduction between MPMM and normal myocardium. This 29% difference gives an idea of the effect of cellular alterations in MPMM myocardium on ^99mTc-tetrofosmin uptake independent of blood flow.

Regarding limitations of the study, the following points should be mentioned. First, the comparison of ^99mTc-tetrofosmin SPECT and ¹⁵O-water PET is a comparison of 2 different modeling techniques. The ¹⁵O-water model provides PVE-corrected MBF only in water-perfusable, not nonperfusable, tissue (13,15,16), whereas ^99mTc-tetrofosmin uptake values represent tracer accumulation within the ROI volume. However, the PTF was introduced to correct for the PVE. This PTF reflects the aspect of an ROI volume-based quantity in the ¹⁵O-water model. Therefore, multiplying MBF_kin and tissue fraction reverses the PVE correction, and the resulting MBF_micr (13,15) and ^99mTc-tetrofosmin uptake reflect both non-PVE-corrected ROI volume-based values and are therefore comparable. Evidence that these assumptions are correct lies in the dependency of the curves for MBF_kin, PTF, MBF_micr, ^99mTc-tetrofosmin, and ¹⁸F-FDG uptake on the ROI volume (Fig. 2B). The nearly parallel curves of ^99mTc-tetrofosmin uptake and MBF_micr validate the mathematics underlying the PVE correction inversion, as described by Iida et al. (13), and hence prove that ^99mTc-tetrofosmin uptake and MBF_micr are comparable.

Second, the comparison of ^99mTc-tetrofosmin SPECT and ¹⁵O-water PET is a comparison of different quantification (absolute vs. relative) methods. The difference between absolute and relative quantification could be eliminated by calculating the MPMM-to-normal percentages for both MBF_micr and ^99mTc-tetrofosmin uptake. Therefore, comparability is given because both values are expressed as percentages.

Third, for the PET studies, attenuation was corrected by transmission measurements using ⁶⁸Ge/⁶⁸Ga rod sources, whereas the SPECT studies were reconstructed using a validated attenuation and scatter correction (25) based on a triple-energy-window acquisition without a transmission scan. That the observed effects are not related to the different attenuation correction methods can be derived from the result that the segmental MPMM-to-normal ^99mTc-tetrofosmin uptake percentages are significantly more reduced than the MBF_micr percentage for all but the anteroseptal regions. Therefore, the attenuation correction applied in the SPECT studies seems to work as well as the attenuation correction by transmission scan in the PET studies for all segments.

Fourth, because the different intrinsic resolutions of the camera systems make a quantitative comparison more difficult, a flow-independent ^99mTc-tetrofosmin uptake reduction of about 29% can provide a landmark but not an accurate value.

CONCLUSION

As expected from the classification, mean relative ^99mTc-tetrofosmin uptake is significantly reduced in MPMM myocardium, whereas mean MBF_micr does not show a significant difference between normal and MPMM myocardium. The difference between reduced ^99mTc-tetrofosmin uptake and unchanged MBF_micr can be explained by cellular alterations in MPMM myocardium, which may affect uptake and retention by impaired transport mechanisms such as active ion channel systems and membrane potentials as well as by ultrastructural mitochondrial alterations independently of blood flow. Therefore, it seems that perfusion deficits in MPMM myocardium are greatly overestimated when ^99mTc-tetrofosmin is used and tend to give false-positive findings.