- Email: [email protected]
Comparison Between Angiography and Fractional Flow Reserve Versus Single-Photon Emission Computed Tomographic Myocardial Perfusion Imaging for Determining Lesion Significance in Patients With Multivessel Coronary Disease
Comparison Between Angiography and Fractional Flow Reserve Versus Single-Photon Emission Computed Tomographic Myocardial Perfusion Imaging for Determining Lesion Significance in Patients With Multivessel Coronary Disease Michael Ragosta, MD*, Andrew H. Bishop, MD, Lewis C. Lipson, MD, Denny D. Watson, PhD, Lawrence W. Gimple, MD, Ian J. Sarembock, MB, ChB, MD, and Eric R. Powers, MD We hypothesized that myocardial perfusion imaging (MPI) would fail to identify all vascular zones with the potential for myocardial ischemia in patients with multivessel coronary disease (MVD). MPI is based on the concept of relative flow reserve. The ability of these techniques to determine the significance of a particular stenosis in the setting of MVD is questionable. Fractional flow reserve (FFR) can determine the significance of individual stenoses. Thirty-six patients with disease involving 88 arteries underwent angiography, FFR, and MPI. FFR was performed using a pressure wire with hyperemia from intracoronary adenosine. Myocardial perfusion images were analyzed quantitatively and segments assigned to a specific coronary artery. The relation between FFR and perfusion was determined for each vascular zone. Of the 88 vessels, the artery was occluded (n ⴝ 20) or had an abnormal FFR <0.75 (n ⴝ 34) in 54 of 88 (61%). MPI showed no defect in 51 zones (58%). Concordance between angiography, FFR, and MPI was seen in 61 of 88 zones (69%). Discordance was seen in the remaining 27 zones (31%) and was predominantly from the finding of a FFR <0.75 or total occlusion despite no defect on MPI. In conclusion, many patients with MVD show no perfusion defect in zones supplied by arteries with total occlusion or a FFR <0.75. Thus, MPI underestimates ischemic burden and FFR may be better at guiding revascularization decisions than perfusion imaging in patients with MVD. © 2007 Elsevier Inc. All rights reserved. (Am J Cardiol 2007;99:896 –902)
Myocardial perfusion imaging (MPI) techniques are based on the concept of relative flow reserve (i.e., hyperemic flow in a stenotic artery vs hyperemic flow in a nonstenotic artery) and require the presence of ⱖ1 normal vascular bed to demonstrate ischemia. These techniques are often applied to determine the significance of disease in an individual coronary artery. Although this is likely of use in patients with single-vessel coronary disease, the ability of MPI to determine the significance of disease in each artery in the setting of multivessel disease (MVD) has not been demonstrated and is in question.1,2 In a recent study, despite angiographically severe 3-vessel disease, myocardial perfusion showed no defect or a single-vessel disease pattern in 54% of patients.3 Explanations for this discrepancy include a lack of spatial resolution to discriminate individual vascular beds and/or the presence of “balanced” ischemia, with patients with MVD lacking a normal reference bed. Fractional flow reserve (FFR) is an invasive pressure-based methodology that determines the physiologic significance of disease within a specific coronary artery with FFR ⬍0.75, which is considered “ischemic.”4 –9 FFR may be better than Cardiovascular Division, Department of Medicine, University of Virginia Health System, Charlottesville, Virginia. Manuscript received August 23, 2006; revised manuscript received and accepted November 7, 2006. *Corresponding author: Tel: 434-924-2420; fax: 434-982-0901. E-mail address: [email protected] (M. Ragosta). 0002-9149/07/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2006.11.035
MPI at determining the significance of disease in individual arteries. Accordingly, we hypothesized that, in patients with MVD, MPI would fail to identify ischemia in all vascular zones subtended by arteries with total occlusion or a FFR ⬍0.75. Methods The study group was selected from 366 consecutive and prospectively identified patients who underwent stressgated technetium-99m sestamibi single-photon emission computed tomographic (SPECT) imaging followed by coronary arteriography. To incorporate all potentially flowlimiting lesions, we considered all arteries with ⱖ50% stenosis by angiography. MVD, defined as ⬎50% stenosis by visual analysis in the proximal to mid segment of ⱖ2 of the 3 major epicardial arteries or their major branches, was present in 56 patients. The remaining 310 patients had normal arteries, minimal atherosclerosis, or significant single-vessel disease. Of those with MVD, 20 patients were excluded because the operator declined (n ⫽ 5), technical difficulties in obtaining reliable FFR measurements (n ⫽ 2), or the patient required coronary bypass surgery to revascularize, and the operator felt it unsafe to instrument the artery (n ⫽ 13). Thus, the study cohort consisted of 36 patients with 88 diseased vessels. The study was approved by the institutional committee on human research. www.AJConline.org
Coronary Artery Disease/FFR Versus Nuclear Scintigraphy in MVD
897
Figure 1. Fourteen-segment model used to define vascular zones by perfusion imaging. Tracer uptake for each segment is determined relative to the segment with the greatest uptake and expressed as percent uptake. The upper number for each segment represents stress uptake and the lower number represents uptake at rest. Upper apical segments shown are averaged to represent segment 1 and lower apical segments are averaged to represent segment 2. A significant defect is defined as ⬍75% uptake and reversibility is present if stress uptake is ⬍75% and increases on imaging at rest.
Study protocol overview: For each patient, FFR was assessed in each artery with ⬎50% stenosis determined by visual estimation of the angiogram by the operator and later measured by offline quantitative angiography. Because it is not possible to measure FFR in a totally occluded artery, vessels with total occlusion were a priori considered “significant.” Quantitative SPECT perfusion images were analyzed using a 14-segment model. For each patient, the SPECT perfusion segments correlating with the artery with ⬎50% narrowing were determined based on the angiogram, taking into consideration coronary dominance and unique vascular territories supplied by each vessel. Concordance between angiographic occlusion, FFR and SPECT perfusion imaging was determined for each coronary artery with ⬎50% narrowing. Clinical characteristics were extracted from a prospectively collected, clinical database (Clinical Automated Office Solutions, Intelligent Business Solutions, Inc., Winston-Salem, North Carolina). Coronary angiography: Patients underwent coronary angiography within a mean of 7 ⫾ 8 days after SPECT imaging with no clinical event, change in status, or revascularization procedure occurring between examinations. Coronary angiography was performed in multiple oblique projections using standard techniques. The decision to perform FFR was based on the operator’s visual assessment of coronary artery narrowing ⬎50%. Off-line quantitative coronary angiography was performed to further characterize stenoses by a single observer blinded to the clinical, SPECT, and FFR data. For patients with total occlusion,
collaterals to the occluded artery were graded 0 to 3, where 0 represents absent collaterals, 1 represents minimal collaterals to small branches, 2 represents near complete filling of the occluded artery, and 3 represents complete filling of the occluded artery. Coronary angiographic images were analyzed with a computer-assisted program (DICOM, Heartlab, Inc., Westerly, Rhode Island). Single frames demonstrating the most severe lumen narrowing without foreshortening were selected for analysis. Angiographic catheter diameter was used as a calibration standard. Luminal diameter proximal or just distal to the moderate stenosis (reference lumen) and minimal lumen diameter at the site of the moderate lesion were determined for all lesions. Percent diameter stenosis was calculated as the ratio of minimal lumen diameter to reference lumen diameter. Determination of FFR: Using a 6Fr guide catheter without side holes and after administration of intravenous heparin (50 U/kg), a 0.014-inch sensor-tipped high-fidelity Pressure Wire (RADI Medical, Uppsala, Sweden) was set at 0, calibrated and aligned, and advanced through the guiding catheter. The pressure sensor was positioned beyond the stenosis in the distal portion of the artery. Phasic and mean aortic pressures and phasic and mean coronary pressures distal to the stenosis were measured under maximum coronary hyperemia induced by administration of intracoronary adenosine (30- to 40-g bolus in the right coronary artery or 80- to 100-g bolus in the left coronary artery) with measurements recorded continuously for 15 seconds after adenosine bolus. Hemodynamic data were digitally stored on a
898
The American Journal of Cardiology (www.AJConline.org)
computer system (WITT Biomedical, Melbourne, Florida), allowing offline analysis. During off-line analysis, FFR was calculated as the ratio of mean distal intracoronary pressure to mean aortic pressure at the time of peak hyperemia at the point of maximum trans-stenotic pressure gradient. FFR values ⬍0.75 and total artery occlusions were classified as significant lesions. SPECT perfusion imaging: SPECT perfusion imaging during stress and at rest was performed as previously described.10 –12 Briefly, a weight-adjusted dose of 240 to 300 MBq of technetium-99m sestamibi was administered at rest and 750 to 900 MBq was given at peak stress ⱖ3 hours after the first injection. Stress testing was performed by exercise using a standard Bruce protocol in 20 patients (with all 20 patients achieving adequate exercise, i.e., ⱖ85% maximum predicted heart rate) and by dipyridamole infusion in 16 patients. Images were obtained 60 minutes after tracer injection. Left ventricular ejection fraction and regional function were determined from gated images acquired on a Picker Prism 3000 3-headed gamma camera using a lowenergy high-resolution parallel-hole collimator (Picker, Cleveland, Ohio) and University of Virginia quantitatively gated SPECT software as previously described.11,12 For myocardial perfusion, quantitative tracer uptake in each segment was determined and relative tracer activity in each segment compared with a gender-specific normal database. Scintigraphic interpretation: First, the angiographer determined the SPECT scintigraphic segments corresponding to each artery in which FFR was assessed based on review of the coronary angiograms. Two reviewers blinded to the results of the angiograms and FFR measurements then interpreted the scintigraphic images. Figure 1 shows an example of the 14-segment scintigraphic model used. In general, the apical segments (segments 1 and 2) were assigned to the left anterior descending artery unless the coronary angiogram showed that these segments were supplied from the posterior descending artery of the right coronary artery or a dominant circumflex artery. Segments 3, 4, 9, and 10 were assigned to the left anterior descending artery. Dominant right coronary arteries were assigned segments 5, 6, 11, and 12, and the circumflex artery was assigned segments 7, 8, 13, and 14. Adjustments in these assignments were made depending on the presence of unique anatomic circumstances. Each vascular zone was classified as showing “no perfusion defect” if all segments in the zone had normal quantitative uptake. A vascular zone was classified as showing a “perfusion defect” if quantitative tracer uptake in any segment within the vascular zone on the stress image was ⬍75%. Each zone with a perfusion defect was further classified as “reversible” if any of the segments with a stress defect quantitatively improved tracer uptake on the image at rest or “fixed” if any of the segments with a stress defect showed no change or a further decrease in tracer uptake on the image at rest. Statistical analysis: Data were analyzed with RS/1 6.0.1 (Domain Manufacturing Corp., Burlington, Massachusetts). All normally distributed data were expressed as mean ⫾ 1 SD; data not normally distributed were expressed as median
Table 1 Patient characteristics (n ⫽ 36) No. of vascular beds Left anterior descending coronary artery Left circumflex coronary artery Right coronary artery Ramus/diagonal Total occlusion Age (yrs) Men Diabetes mellitus Hypertension Left ventricular hypertrophy Previous Q-wave myocardial infarction Previous coronary bypass surgery Ejection fraction (%) 2-Vessel coronary disease 3-Vessel coronary disease ⱖ1 Occluded coronary artery Exercise stress Dipyridamole stress
88 27 (31%) 27 (31%) 31 (35%) 3 (3%) 20 (23%) 62 ⫾ 11 27 (75%) 16 (44%) 28 (78%) 5 (14%) 7 (19%) 1 (3%) 51 ⫾ 10 21 (58%) 15 (42%) 19 (53%) 20 (56%) 16 (44%)
Figure 2. Relation between FFR and SPECT perfusion imaging in nonoccluded coronary arteries. The lower limit cut-off value (dashed line) represents ischemia of 0.75.
(25th, 75th percentile). Comparisons between groups were performed, and group differences of continuous factors were compared using Wilcoxon signed-rank tests. Group differences of categorical variables were compared using chi-square tests or, in cases of small cells, Fishers’ exact test. All p values are from 2-sided tests, and p values ⬍0.05 were considered statistically significant. Results Clinical characteristics: Clinical characteristics are listed in Table 1. Stress testing and subsequent catheterization were performed for evaluation of typical stable angina in 26 patients (72%), dyspnea without chest pain in 4 patients (11%), atypical chest pain in 3 patients (8%), and as a preoperative risk assessment in 3 asymptomatic patients (8%). Twenty-one patients (58%) had 2-vessel disease, and the remainder had 3-vessel disease (42%). Coronary angiographic and hemodynamic data: All 88 study arteries were narrowed ⱖ50% by quantitative
Coronary Artery Disease/FFR Versus Nuclear Scintigraphy in MVD
Figure 3. Chart shows number of vascular zones and relation between FFR and SPECT imaging. Concordant vascular zones are those with a SPECT defect and a FFR ⬍0.75 or total occlusion and those with no SPECT defect and a FFR ⬎0.75. Discordant vascular zones are those with no SPECT defect and a FFR ⬎0.75 or total occlusion and those with a SPECT defect but a FFR ⬎0.75.
Table 2 Comparison between concordant and discordant vascular territories Variable
Catheterization data Diameter stenosis (%) Left anterior descending coronary artery Right coronary artery Circumflex/ramus/diagonal Total coronary occlusion FFR ⬍0.75 Perfusion imaging data Exercise stress No perfusion defect Fixed defect Reversible defect Uptake at rest Stress uptake Delta uptake Wall motion abnormality
Concordant Zones (n ⫽ 61)
Discordant Zones (n ⫽ 27)
p Value
76 ⫾ 17 19 (31%)
74 ⫾ 15 8 (30%)
0.55 1.0
22 (36%) 20 (34%) 16 (26%) 16 (26%)
9 (33%) 10 (37%) 4 (15%) 18 (67%)
1.0 0.81 0.28 0.004
33 (54%) 29 (48%) 4 (7%) 28 (46%) 79 ⫾ 13% 74 ⫾ 14% 5 ⫾ 6% 32 (52%)
16 (59%) 22 (81%) 3 (11%) 2 (7%) 82 ⫾ 10% 81 ⫾ 9% 1 ⫾ 5% 12 (44%)
0.82 0.004 0.67 ⬍0.005 0.41 0.04 0.005 0.64
angiographic analysis (mean 76 ⫾ 16%). Total occlusion was observed in 20 of 88 lesions (23%) in 19 of 36 patients (53%). In the 68 arteries without total occlusion, the FFR was ⬍0.75 in 34 (50%) with a mean FFR of 0.75 ⫾ 0.14. Thus, FFR was ⬍0.75 or the artery was totally occluded in 54 of 88 arteries (61%). SPECT perfusion data and relation across angiography, FFR, and SPECT perfusion: No perfusion abnormality was seen in 51 of 88 vascular zones (58%). Of the 37 zones with a perfusion abnormality, reversibility was seen in 30 and fixed defects in 7. The relation between FFR and SPECT perfusion imaging for the 68 nonoccluded vascular zones is shown in Figure 2. In 47 zones with no perfusion abnormality, 18 (38%) had a FFR ⬍0.75. In 15 zones with reversible defects, FFR was ⬍0.75 in 13 (87%); in 6 zones
899
with fixed defects, FFR was ⬍0.75 in 3 (50%). Thus, combining reversible and fixed defects, concordance between FFR and SPECT perfusion imaging was observed in only 45 of 68 nonoccluded vascular zones (66%), and discordance was observed in 23 of 68 (34%). In 18 of these 23 discordant zones, discordance was due to a FFR ⬍0.75 despite no perfusion abnormality. The remaining 5 discordant vascular zones showed perfusion defects but a FFR ⬎0.75 (2 of these showing reversible and 3 showing fixed defects). For the 20 totally occluded vessels in which FFR could not be directly measured, 15 of 20 vascular zones (75%) showed a reversible perfusion defect, 1 of 20 (5%) showed a fixed defect, and only 4 of 20 (20%) had no perfusion defect. Comparison of the quantitative SPECT perfusion data between vascular territories subtended by an occluded artery and vascular territories subtended by a nonoccluded artery showed significantly worse quantitative perfusion during imaging under stress (64 ⫾ 11% vs 80 ⫾ 11%, p ⬍0.001) and at rest (72 ⫾ 12% vs 82 ⫾ 11%) than nonoccluded zones. There was no relation between extent of angiographic collateral and presence of a perfusion defect. In the 16 zones with a perfusion defect, 12 had good collaterals (grade 2 to 3) and 4 had poor collaterals (grade 0 to 1). For the 4 zones with no perfusion defect, 3 had good collaterals and 1 had no collateral. When all 88 vascular zones were included, concordance across angiographic findings, FFR, and SPECT perfusion imaging was present in 61 of 88 vascular zones (69%), and discordance was observed in 27 of 88 (31%; Figure 3). In 22 of the 27 vascular zones with discordance, discordance was due to a FFR ⬍0.75 or total coronary occlusion despite no perfusion abnormality. Analysis of concordant and discordant vascular zones: Vascular zones demonstrating concordance between FFR and MPI were not statistically significantly different from discordant vascular zones in terms of stenosis severity, presence or absence of total coronary occlusion, location of artery, or method of stress induction (exercise or pharmacologic; Table 2). The main discrepancy between concordant and discordant vascular zones was due to a larger proportion of discordant zones with a FFR ⬍0.75 despite no perfusion abnormality. There was no difference between concordant and discordant zones in terms of presence of wall motion abnormality within the vascular zone. Although quantitative uptake of sestamibi at rest was similar, stress uptake was significantly lower in concordant zones with a greater change in uptake from stress to uptake at rest consistent with the larger proportion of reversible defects in concordant zones. Comparison between patients with concordant and discordant vascular zones: Complete concordance among nuclear perfusion imaging, coronary angiographic, and FFR findings was observed in only 14 patients (39%). Thus, 22 of 38 patients (61%) with MVD had ⱖ1 discordant vascular zone. Patients with concordant and discordant zones were similar in terms of clinical characteristics, presence of typical angina, method of stress induction, left ventricular func-
900
The American Journal of Cardiology (www.AJConline.org)
Figure 4. Examples of nuclear perfusion imaging, coronary angiographic, and FFR findings in a 63-year-old man with increasing angina pectoris and 3-vessel coronary disease. (A) Perfusion images show a reversible defect only in the distribution of the left anterior descending coronary artery (LAD). (B) Coronary angiography shows a severe stenosis in the LAD with a FFR of 0.45 and thus concordant with perfusion images. However, there is also a moderate lesion in the left circumflex artery (LCX) with a FFR of 0.64, demonstrating discordance between nuclear imaging and FFR findings. (C) The right coronary artery has a moderate lesion in the proximal portion with a FFR of 0.78, concordant with nuclear imaging. This is an example of how perfusion images identify only the most severe stenosis in the setting of MVD.
tion, and presence of ⱖ1 totally occluded vessel. Importantly, a significantly larger proportion of patients with discordant zones had 3-vessel coronary disease (59% vs
21%, p ⫽ 0.04) and normal SPECT perfusion scans (41% vs 7%, p ⬍0.05) compared with patients with concordant zones.
Coronary Artery Disease/FFR Versus Nuclear Scintigraphy in MVD
Reasons for discordance between FFR and SPECT perfusion in MVD: The various reasons responsible for a discrepancy among the angiograms, FFR result, and MPI in the 22 patients with ⱖ1 discordant zone were explored. The most common reason for discordance, seen in 11 of 22 patients (50%), occurred when SPECT perfusion identified 1 stenosis and failed to identify a second or third significant stenosis (a FFR ⬍0.75 or total occlusion). In 8 of 11 of these cases, a perfusion defect was associated just with the “worst” stenosis, i.e., the artery with the lowest FFR; normal perfusion was present in other zones despite significant stenoses. In 7 of 22 patients (32%), SPECT perfusion imaging was completely normal in all zones despite a low FFR in ⱖ1 zone. In the remaining 4 patients, a perfusion defect was present despite a FFR ⬎0.75; in 3 of these patients the perfusion defect was fixed. Figure 4 shows an example of a patient with angina pectoris and discordance in whom nuclear perfusion imaging identified the zone supplied by the most severe stenosis and failed to identify a significant stenosis in another zone.
Discussion There are several important findings of the present study. First, in patients with MVD, 36% of vascular zones with no perfusion abnormalities on MPI were ischemic as defined by the presence of total coronary occlusion or a FFR ⬍0.75. Totally occluded arteries were more likely to be concordant and associated with perfusion defects than patent arteries (80% vs 31%), so most of the disparity was observed in patent arteries of which 18 of 47 (38%) had a FFR ⬍0.75 despite normal perfusion. This is important because it implies that significant stenoses would be overlooked if a clinician relied on perfusion scintigraphy to determine the significance of an ambiguous stenosis. Second, routine clinical variables could not discriminate zones more likely to show discordance between FFR and nuclear scintigraphy. Stenosis severity, vessel involved, and type of stress protocol used were similar in concordant and discordant zones. Patients with ⱖ1 discordant zone were similar to patients with complete concordance with the exception that patients with discordant zones were more likely to have 3-vessel disease. This is an important observation, because this is a particularly high-risk group in whom it is important to choose optimal treatment. Third, the reason for discordance supported our study hypothesis because most cases of discordance were due primarily to perfusion imaging correctly identifying the most severe stenosis but not identifying other zones subtended by significant lesions. The phenomenon of balanced ischemia in which perfusion appears normal because of severe flow limitation in all vascular zones was not observed in this study but has been demonstrated by other investigators using FFR methodology.13 There are several limitations to this study. Selection bias may have influenced the results because the study enrolled consecutive patients undergoing stress testing followed by angiography with no attempt made to stan-
901
dardize criteria for catheterization after stress testing, and an unknown number of patients underwent catheterization without a stress test. Ideally, FFR should have been measured in every patent vessel, not just those with moderate degrees of stenosis. Although unlikely, it is possible that a “normal”-appearing vessel had a FFR ⬍0.75. However, including these angiographic normal arteries could only have strengthened the data because no patient with 2-vessel disease had perfusion defects in the territory of the normal-appearing artery. Another limitation relates to our analysis combining fixed and reversible defects. Some fixed defects may represent severe ischemia, so we considered any defect as potentially ischemic. However, most defects were reversible, and the major results would not have changed if the few fixed defects were excluded. We chose high-dose intracoronary adenosine to induce hyperemia because this is the method most commonly used in the clinical setting. Other investigators use intravenous infusion (140 g/kg/min). Although comparison between these 2 routes has shown greater hyperemia with the intravenous route, the difference in FFR calculation was small, with only 10% of patients with a FFR ⬎0.75 by 60 g of intracoronary adenosine changing to a FFR ⬍0.75 with intravenous infusion.14 Similar results applied to our study would not have changed the conclusions and intravenous adenosine could only have strengthened our data and led to a further increase in disparity between nuclear scintigraphic and FFR assessments.
1. Daou D, Delahaye N, Vilain D, Lebtahi R, Faraggi M, Le Guludec D. Identification of extensive coronary artery disease: incremental value of exercise Tl-201 SPECT to clinical and stress test variables. J Nucl Cardiol 2002;9:161–168. 2. Zaacks SM, Ali A, Parrillo JE, Barron JT. How well does radionuclide dipyridamole stress testing detect three-vessel coronary artery disease and ischemia in the region supplied by the most stenotic vessel? Clin Nucl Med 1999;24:35– 41 3. Lima RSL, Watson DD, Goode AR, Siadaty MS, Ragosta M, Beller GA, Samady H. Incremental value of combined perfusion and function over perfusion alone by gated SPECT myocardial perfusion imaging for detection of severe three vessel coronary artery disease. J Am Coll Cardiol 2003;42:64 –70. 4. Pijls NHJ, van Son JAM, Kirkeeide RL, De Bruyne B, Gould KL. Experimental basis of determining maximum coronary, myocardial, and collateral blood flow by pressure measurements for assessing functional stenosis severity before and after percutaneous transluminal coronary angioplasty. Circulation 1993;87:1354 –1367. 5. De Bruyne B, Baudhuin T, Melin JA, Pijls NH, Sys SU, Bol A, Paulus WJ, Heyndrickx GR, Wijns W. Coronary flow reserve calculated from pressure measurements in humans: validation with positron emission tomography. Circulation 1994;89:1013–1022. 6. Pijls NHJ, Van Gelder B, Van der Voort P, Peels K, Bracke FALE, Bonnier HJRM, El Gamal MIH. Fractional flow reserve: a useful index to evaluate the influence of an epicardial coronary stenosis on myocardial blood flow. Circulation 1995;92:3183–3193. 7. De Bruyne B, Bartunek, J, Sys SU, Heyndrickx GR. Relation between myocardial fractional flow reserve calculated from coronary pressure measurements and exercise-induced myocardial ischemia. Circulation 1995;92:39 – 46. 8. Pijls NHJ, De Bruyne B, Peels K, van der Voort PH, Bonnier HJRM, Bartunek J, Koolen JJ. Measurement of fractional flow reserve to assess the functional severity of coronary-artery stenoses. N Engl J Med 1996;334:1703–1708.
902
The American Journal of Cardiology (www.AJConline.org)
9. Bishop AH, Samady H. Fractional flow reserve: critical review of an important physiologic adjunct to angiography. Am Heart J 2004;147: 792– 802. 10. Taillefer R, DePuey EG, Udelson JE, Beller GA, Latour Y, Reeves F. Comparative diagnostic accuracy of Tl-201 and Tc-99m sestamibi SPECT imaging (perfusion and ECG-gated SPECT) in detecting coronary artery disease in women. J Am Coll Cardiol 1997;29: 69 –77. 11. Calnon DA, Kastner RJ, Smith WH, Segalla D, Beller GA, Watson DD. Validation of a new counts-based gated single photon emission computed tomography method for quantifying left ventricular systolic function: comparison with equilibrium radionuclide angiography. J Nucl Cardiol 1997;4:464 – 471.
12. Calnon DA, Kastner RJ, Smith WH, Segalla D, Beller GA, Watson DD,. Quantitative gated single photon emission computed tomography imaging: accounts-based method for display and measurement of regional and global ventricular systolic function. J Nucl Cardiol 1997; 4:451– 463. 13. Aarnoudse WH, Botman KJ, Pijls NH. False-negative myocardial scintigraphy in balanced three-vessel disease, revealed by coronary pressure measurement. Int J Cardiovasc Intervent 2003;5:67–71. 14. Casella G, Leibig M, Schiele TM, Schrepf R, Seelig V, Stempfle HU, Erdin P, Rieber J, Konig A, Siebert U, Klauss V. Are high doses of intracoronary adenosine an alternative to standard intravenous adenosine for the assessment of fractional flow reserve? Am Heart J 2004; 148:590 –595.
Myocardial perfusion imaging (MPI) techniques are based on the concept of relative flow reserve (i.e., hyperemic flow in a stenotic artery vs hyperemic flow in a nonstenotic artery) and require the presence of ⱖ1 normal vascular bed to demonstrate ischemia. These techniques are often applied to determine the significance of disease in an individual coronary artery. Although this is likely of use in patients with single-vessel coronary disease, the ability of MPI to determine the significance of disease in each artery in the setting of multivessel disease (MVD) has not been demonstrated and is in question.1,2 In a recent study, despite angiographically severe 3-vessel disease, myocardial perfusion showed no defect or a single-vessel disease pattern in 54% of patients.3 Explanations for this discrepancy include a lack of spatial resolution to discriminate individual vascular beds and/or the presence of “balanced” ischemia, with patients with MVD lacking a normal reference bed. Fractional flow reserve (FFR) is an invasive pressure-based methodology that determines the physiologic significance of disease within a specific coronary artery with FFR ⬍0.75, which is considered “ischemic.”4 –9 FFR may be better than Cardiovascular Division, Department of Medicine, University of Virginia Health System, Charlottesville, Virginia. Manuscript received August 23, 2006; revised manuscript received and accepted November 7, 2006. *Corresponding author: Tel: 434-924-2420; fax: 434-982-0901. E-mail address: [email protected] (M. Ragosta). 0002-9149/07/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2006.11.035
MPI at determining the significance of disease in individual arteries. Accordingly, we hypothesized that, in patients with MVD, MPI would fail to identify ischemia in all vascular zones subtended by arteries with total occlusion or a FFR ⬍0.75. Methods The study group was selected from 366 consecutive and prospectively identified patients who underwent stressgated technetium-99m sestamibi single-photon emission computed tomographic (SPECT) imaging followed by coronary arteriography. To incorporate all potentially flowlimiting lesions, we considered all arteries with ⱖ50% stenosis by angiography. MVD, defined as ⬎50% stenosis by visual analysis in the proximal to mid segment of ⱖ2 of the 3 major epicardial arteries or their major branches, was present in 56 patients. The remaining 310 patients had normal arteries, minimal atherosclerosis, or significant single-vessel disease. Of those with MVD, 20 patients were excluded because the operator declined (n ⫽ 5), technical difficulties in obtaining reliable FFR measurements (n ⫽ 2), or the patient required coronary bypass surgery to revascularize, and the operator felt it unsafe to instrument the artery (n ⫽ 13). Thus, the study cohort consisted of 36 patients with 88 diseased vessels. The study was approved by the institutional committee on human research. www.AJConline.org
Coronary Artery Disease/FFR Versus Nuclear Scintigraphy in MVD
897
Figure 1. Fourteen-segment model used to define vascular zones by perfusion imaging. Tracer uptake for each segment is determined relative to the segment with the greatest uptake and expressed as percent uptake. The upper number for each segment represents stress uptake and the lower number represents uptake at rest. Upper apical segments shown are averaged to represent segment 1 and lower apical segments are averaged to represent segment 2. A significant defect is defined as ⬍75% uptake and reversibility is present if stress uptake is ⬍75% and increases on imaging at rest.
Study protocol overview: For each patient, FFR was assessed in each artery with ⬎50% stenosis determined by visual estimation of the angiogram by the operator and later measured by offline quantitative angiography. Because it is not possible to measure FFR in a totally occluded artery, vessels with total occlusion were a priori considered “significant.” Quantitative SPECT perfusion images were analyzed using a 14-segment model. For each patient, the SPECT perfusion segments correlating with the artery with ⬎50% narrowing were determined based on the angiogram, taking into consideration coronary dominance and unique vascular territories supplied by each vessel. Concordance between angiographic occlusion, FFR and SPECT perfusion imaging was determined for each coronary artery with ⬎50% narrowing. Clinical characteristics were extracted from a prospectively collected, clinical database (Clinical Automated Office Solutions, Intelligent Business Solutions, Inc., Winston-Salem, North Carolina). Coronary angiography: Patients underwent coronary angiography within a mean of 7 ⫾ 8 days after SPECT imaging with no clinical event, change in status, or revascularization procedure occurring between examinations. Coronary angiography was performed in multiple oblique projections using standard techniques. The decision to perform FFR was based on the operator’s visual assessment of coronary artery narrowing ⬎50%. Off-line quantitative coronary angiography was performed to further characterize stenoses by a single observer blinded to the clinical, SPECT, and FFR data. For patients with total occlusion,
collaterals to the occluded artery were graded 0 to 3, where 0 represents absent collaterals, 1 represents minimal collaterals to small branches, 2 represents near complete filling of the occluded artery, and 3 represents complete filling of the occluded artery. Coronary angiographic images were analyzed with a computer-assisted program (DICOM, Heartlab, Inc., Westerly, Rhode Island). Single frames demonstrating the most severe lumen narrowing without foreshortening were selected for analysis. Angiographic catheter diameter was used as a calibration standard. Luminal diameter proximal or just distal to the moderate stenosis (reference lumen) and minimal lumen diameter at the site of the moderate lesion were determined for all lesions. Percent diameter stenosis was calculated as the ratio of minimal lumen diameter to reference lumen diameter. Determination of FFR: Using a 6Fr guide catheter without side holes and after administration of intravenous heparin (50 U/kg), a 0.014-inch sensor-tipped high-fidelity Pressure Wire (RADI Medical, Uppsala, Sweden) was set at 0, calibrated and aligned, and advanced through the guiding catheter. The pressure sensor was positioned beyond the stenosis in the distal portion of the artery. Phasic and mean aortic pressures and phasic and mean coronary pressures distal to the stenosis were measured under maximum coronary hyperemia induced by administration of intracoronary adenosine (30- to 40-g bolus in the right coronary artery or 80- to 100-g bolus in the left coronary artery) with measurements recorded continuously for 15 seconds after adenosine bolus. Hemodynamic data were digitally stored on a
898
The American Journal of Cardiology (www.AJConline.org)
computer system (WITT Biomedical, Melbourne, Florida), allowing offline analysis. During off-line analysis, FFR was calculated as the ratio of mean distal intracoronary pressure to mean aortic pressure at the time of peak hyperemia at the point of maximum trans-stenotic pressure gradient. FFR values ⬍0.75 and total artery occlusions were classified as significant lesions. SPECT perfusion imaging: SPECT perfusion imaging during stress and at rest was performed as previously described.10 –12 Briefly, a weight-adjusted dose of 240 to 300 MBq of technetium-99m sestamibi was administered at rest and 750 to 900 MBq was given at peak stress ⱖ3 hours after the first injection. Stress testing was performed by exercise using a standard Bruce protocol in 20 patients (with all 20 patients achieving adequate exercise, i.e., ⱖ85% maximum predicted heart rate) and by dipyridamole infusion in 16 patients. Images were obtained 60 minutes after tracer injection. Left ventricular ejection fraction and regional function were determined from gated images acquired on a Picker Prism 3000 3-headed gamma camera using a lowenergy high-resolution parallel-hole collimator (Picker, Cleveland, Ohio) and University of Virginia quantitatively gated SPECT software as previously described.11,12 For myocardial perfusion, quantitative tracer uptake in each segment was determined and relative tracer activity in each segment compared with a gender-specific normal database. Scintigraphic interpretation: First, the angiographer determined the SPECT scintigraphic segments corresponding to each artery in which FFR was assessed based on review of the coronary angiograms. Two reviewers blinded to the results of the angiograms and FFR measurements then interpreted the scintigraphic images. Figure 1 shows an example of the 14-segment scintigraphic model used. In general, the apical segments (segments 1 and 2) were assigned to the left anterior descending artery unless the coronary angiogram showed that these segments were supplied from the posterior descending artery of the right coronary artery or a dominant circumflex artery. Segments 3, 4, 9, and 10 were assigned to the left anterior descending artery. Dominant right coronary arteries were assigned segments 5, 6, 11, and 12, and the circumflex artery was assigned segments 7, 8, 13, and 14. Adjustments in these assignments were made depending on the presence of unique anatomic circumstances. Each vascular zone was classified as showing “no perfusion defect” if all segments in the zone had normal quantitative uptake. A vascular zone was classified as showing a “perfusion defect” if quantitative tracer uptake in any segment within the vascular zone on the stress image was ⬍75%. Each zone with a perfusion defect was further classified as “reversible” if any of the segments with a stress defect quantitatively improved tracer uptake on the image at rest or “fixed” if any of the segments with a stress defect showed no change or a further decrease in tracer uptake on the image at rest. Statistical analysis: Data were analyzed with RS/1 6.0.1 (Domain Manufacturing Corp., Burlington, Massachusetts). All normally distributed data were expressed as mean ⫾ 1 SD; data not normally distributed were expressed as median
Table 1 Patient characteristics (n ⫽ 36) No. of vascular beds Left anterior descending coronary artery Left circumflex coronary artery Right coronary artery Ramus/diagonal Total occlusion Age (yrs) Men Diabetes mellitus Hypertension Left ventricular hypertrophy Previous Q-wave myocardial infarction Previous coronary bypass surgery Ejection fraction (%) 2-Vessel coronary disease 3-Vessel coronary disease ⱖ1 Occluded coronary artery Exercise stress Dipyridamole stress
88 27 (31%) 27 (31%) 31 (35%) 3 (3%) 20 (23%) 62 ⫾ 11 27 (75%) 16 (44%) 28 (78%) 5 (14%) 7 (19%) 1 (3%) 51 ⫾ 10 21 (58%) 15 (42%) 19 (53%) 20 (56%) 16 (44%)
Figure 2. Relation between FFR and SPECT perfusion imaging in nonoccluded coronary arteries. The lower limit cut-off value (dashed line) represents ischemia of 0.75.
(25th, 75th percentile). Comparisons between groups were performed, and group differences of continuous factors were compared using Wilcoxon signed-rank tests. Group differences of categorical variables were compared using chi-square tests or, in cases of small cells, Fishers’ exact test. All p values are from 2-sided tests, and p values ⬍0.05 were considered statistically significant. Results Clinical characteristics: Clinical characteristics are listed in Table 1. Stress testing and subsequent catheterization were performed for evaluation of typical stable angina in 26 patients (72%), dyspnea without chest pain in 4 patients (11%), atypical chest pain in 3 patients (8%), and as a preoperative risk assessment in 3 asymptomatic patients (8%). Twenty-one patients (58%) had 2-vessel disease, and the remainder had 3-vessel disease (42%). Coronary angiographic and hemodynamic data: All 88 study arteries were narrowed ⱖ50% by quantitative
Coronary Artery Disease/FFR Versus Nuclear Scintigraphy in MVD
Figure 3. Chart shows number of vascular zones and relation between FFR and SPECT imaging. Concordant vascular zones are those with a SPECT defect and a FFR ⬍0.75 or total occlusion and those with no SPECT defect and a FFR ⬎0.75. Discordant vascular zones are those with no SPECT defect and a FFR ⬎0.75 or total occlusion and those with a SPECT defect but a FFR ⬎0.75.
Table 2 Comparison between concordant and discordant vascular territories Variable
Catheterization data Diameter stenosis (%) Left anterior descending coronary artery Right coronary artery Circumflex/ramus/diagonal Total coronary occlusion FFR ⬍0.75 Perfusion imaging data Exercise stress No perfusion defect Fixed defect Reversible defect Uptake at rest Stress uptake Delta uptake Wall motion abnormality
Concordant Zones (n ⫽ 61)
Discordant Zones (n ⫽ 27)
p Value
76 ⫾ 17 19 (31%)
74 ⫾ 15 8 (30%)
0.55 1.0
22 (36%) 20 (34%) 16 (26%) 16 (26%)
9 (33%) 10 (37%) 4 (15%) 18 (67%)
1.0 0.81 0.28 0.004
33 (54%) 29 (48%) 4 (7%) 28 (46%) 79 ⫾ 13% 74 ⫾ 14% 5 ⫾ 6% 32 (52%)
16 (59%) 22 (81%) 3 (11%) 2 (7%) 82 ⫾ 10% 81 ⫾ 9% 1 ⫾ 5% 12 (44%)
0.82 0.004 0.67 ⬍0.005 0.41 0.04 0.005 0.64
angiographic analysis (mean 76 ⫾ 16%). Total occlusion was observed in 20 of 88 lesions (23%) in 19 of 36 patients (53%). In the 68 arteries without total occlusion, the FFR was ⬍0.75 in 34 (50%) with a mean FFR of 0.75 ⫾ 0.14. Thus, FFR was ⬍0.75 or the artery was totally occluded in 54 of 88 arteries (61%). SPECT perfusion data and relation across angiography, FFR, and SPECT perfusion: No perfusion abnormality was seen in 51 of 88 vascular zones (58%). Of the 37 zones with a perfusion abnormality, reversibility was seen in 30 and fixed defects in 7. The relation between FFR and SPECT perfusion imaging for the 68 nonoccluded vascular zones is shown in Figure 2. In 47 zones with no perfusion abnormality, 18 (38%) had a FFR ⬍0.75. In 15 zones with reversible defects, FFR was ⬍0.75 in 13 (87%); in 6 zones
899
with fixed defects, FFR was ⬍0.75 in 3 (50%). Thus, combining reversible and fixed defects, concordance between FFR and SPECT perfusion imaging was observed in only 45 of 68 nonoccluded vascular zones (66%), and discordance was observed in 23 of 68 (34%). In 18 of these 23 discordant zones, discordance was due to a FFR ⬍0.75 despite no perfusion abnormality. The remaining 5 discordant vascular zones showed perfusion defects but a FFR ⬎0.75 (2 of these showing reversible and 3 showing fixed defects). For the 20 totally occluded vessels in which FFR could not be directly measured, 15 of 20 vascular zones (75%) showed a reversible perfusion defect, 1 of 20 (5%) showed a fixed defect, and only 4 of 20 (20%) had no perfusion defect. Comparison of the quantitative SPECT perfusion data between vascular territories subtended by an occluded artery and vascular territories subtended by a nonoccluded artery showed significantly worse quantitative perfusion during imaging under stress (64 ⫾ 11% vs 80 ⫾ 11%, p ⬍0.001) and at rest (72 ⫾ 12% vs 82 ⫾ 11%) than nonoccluded zones. There was no relation between extent of angiographic collateral and presence of a perfusion defect. In the 16 zones with a perfusion defect, 12 had good collaterals (grade 2 to 3) and 4 had poor collaterals (grade 0 to 1). For the 4 zones with no perfusion defect, 3 had good collaterals and 1 had no collateral. When all 88 vascular zones were included, concordance across angiographic findings, FFR, and SPECT perfusion imaging was present in 61 of 88 vascular zones (69%), and discordance was observed in 27 of 88 (31%; Figure 3). In 22 of the 27 vascular zones with discordance, discordance was due to a FFR ⬍0.75 or total coronary occlusion despite no perfusion abnormality. Analysis of concordant and discordant vascular zones: Vascular zones demonstrating concordance between FFR and MPI were not statistically significantly different from discordant vascular zones in terms of stenosis severity, presence or absence of total coronary occlusion, location of artery, or method of stress induction (exercise or pharmacologic; Table 2). The main discrepancy between concordant and discordant vascular zones was due to a larger proportion of discordant zones with a FFR ⬍0.75 despite no perfusion abnormality. There was no difference between concordant and discordant zones in terms of presence of wall motion abnormality within the vascular zone. Although quantitative uptake of sestamibi at rest was similar, stress uptake was significantly lower in concordant zones with a greater change in uptake from stress to uptake at rest consistent with the larger proportion of reversible defects in concordant zones. Comparison between patients with concordant and discordant vascular zones: Complete concordance among nuclear perfusion imaging, coronary angiographic, and FFR findings was observed in only 14 patients (39%). Thus, 22 of 38 patients (61%) with MVD had ⱖ1 discordant vascular zone. Patients with concordant and discordant zones were similar in terms of clinical characteristics, presence of typical angina, method of stress induction, left ventricular func-
900
The American Journal of Cardiology (www.AJConline.org)
Figure 4. Examples of nuclear perfusion imaging, coronary angiographic, and FFR findings in a 63-year-old man with increasing angina pectoris and 3-vessel coronary disease. (A) Perfusion images show a reversible defect only in the distribution of the left anterior descending coronary artery (LAD). (B) Coronary angiography shows a severe stenosis in the LAD with a FFR of 0.45 and thus concordant with perfusion images. However, there is also a moderate lesion in the left circumflex artery (LCX) with a FFR of 0.64, demonstrating discordance between nuclear imaging and FFR findings. (C) The right coronary artery has a moderate lesion in the proximal portion with a FFR of 0.78, concordant with nuclear imaging. This is an example of how perfusion images identify only the most severe stenosis in the setting of MVD.
tion, and presence of ⱖ1 totally occluded vessel. Importantly, a significantly larger proportion of patients with discordant zones had 3-vessel coronary disease (59% vs
21%, p ⫽ 0.04) and normal SPECT perfusion scans (41% vs 7%, p ⬍0.05) compared with patients with concordant zones.
Coronary Artery Disease/FFR Versus Nuclear Scintigraphy in MVD
Reasons for discordance between FFR and SPECT perfusion in MVD: The various reasons responsible for a discrepancy among the angiograms, FFR result, and MPI in the 22 patients with ⱖ1 discordant zone were explored. The most common reason for discordance, seen in 11 of 22 patients (50%), occurred when SPECT perfusion identified 1 stenosis and failed to identify a second or third significant stenosis (a FFR ⬍0.75 or total occlusion). In 8 of 11 of these cases, a perfusion defect was associated just with the “worst” stenosis, i.e., the artery with the lowest FFR; normal perfusion was present in other zones despite significant stenoses. In 7 of 22 patients (32%), SPECT perfusion imaging was completely normal in all zones despite a low FFR in ⱖ1 zone. In the remaining 4 patients, a perfusion defect was present despite a FFR ⬎0.75; in 3 of these patients the perfusion defect was fixed. Figure 4 shows an example of a patient with angina pectoris and discordance in whom nuclear perfusion imaging identified the zone supplied by the most severe stenosis and failed to identify a significant stenosis in another zone.
Discussion There are several important findings of the present study. First, in patients with MVD, 36% of vascular zones with no perfusion abnormalities on MPI were ischemic as defined by the presence of total coronary occlusion or a FFR ⬍0.75. Totally occluded arteries were more likely to be concordant and associated with perfusion defects than patent arteries (80% vs 31%), so most of the disparity was observed in patent arteries of which 18 of 47 (38%) had a FFR ⬍0.75 despite normal perfusion. This is important because it implies that significant stenoses would be overlooked if a clinician relied on perfusion scintigraphy to determine the significance of an ambiguous stenosis. Second, routine clinical variables could not discriminate zones more likely to show discordance between FFR and nuclear scintigraphy. Stenosis severity, vessel involved, and type of stress protocol used were similar in concordant and discordant zones. Patients with ⱖ1 discordant zone were similar to patients with complete concordance with the exception that patients with discordant zones were more likely to have 3-vessel disease. This is an important observation, because this is a particularly high-risk group in whom it is important to choose optimal treatment. Third, the reason for discordance supported our study hypothesis because most cases of discordance were due primarily to perfusion imaging correctly identifying the most severe stenosis but not identifying other zones subtended by significant lesions. The phenomenon of balanced ischemia in which perfusion appears normal because of severe flow limitation in all vascular zones was not observed in this study but has been demonstrated by other investigators using FFR methodology.13 There are several limitations to this study. Selection bias may have influenced the results because the study enrolled consecutive patients undergoing stress testing followed by angiography with no attempt made to stan-
901
dardize criteria for catheterization after stress testing, and an unknown number of patients underwent catheterization without a stress test. Ideally, FFR should have been measured in every patent vessel, not just those with moderate degrees of stenosis. Although unlikely, it is possible that a “normal”-appearing vessel had a FFR ⬍0.75. However, including these angiographic normal arteries could only have strengthened the data because no patient with 2-vessel disease had perfusion defects in the territory of the normal-appearing artery. Another limitation relates to our analysis combining fixed and reversible defects. Some fixed defects may represent severe ischemia, so we considered any defect as potentially ischemic. However, most defects were reversible, and the major results would not have changed if the few fixed defects were excluded. We chose high-dose intracoronary adenosine to induce hyperemia because this is the method most commonly used in the clinical setting. Other investigators use intravenous infusion (140 g/kg/min). Although comparison between these 2 routes has shown greater hyperemia with the intravenous route, the difference in FFR calculation was small, with only 10% of patients with a FFR ⬎0.75 by 60 g of intracoronary adenosine changing to a FFR ⬍0.75 with intravenous infusion.14 Similar results applied to our study would not have changed the conclusions and intravenous adenosine could only have strengthened our data and led to a further increase in disparity between nuclear scintigraphic and FFR assessments.
1. Daou D, Delahaye N, Vilain D, Lebtahi R, Faraggi M, Le Guludec D. Identification of extensive coronary artery disease: incremental value of exercise Tl-201 SPECT to clinical and stress test variables. J Nucl Cardiol 2002;9:161–168. 2. Zaacks SM, Ali A, Parrillo JE, Barron JT. How well does radionuclide dipyridamole stress testing detect three-vessel coronary artery disease and ischemia in the region supplied by the most stenotic vessel? Clin Nucl Med 1999;24:35– 41 3. Lima RSL, Watson DD, Goode AR, Siadaty MS, Ragosta M, Beller GA, Samady H. Incremental value of combined perfusion and function over perfusion alone by gated SPECT myocardial perfusion imaging for detection of severe three vessel coronary artery disease. J Am Coll Cardiol 2003;42:64 –70. 4. Pijls NHJ, van Son JAM, Kirkeeide RL, De Bruyne B, Gould KL. Experimental basis of determining maximum coronary, myocardial, and collateral blood flow by pressure measurements for assessing functional stenosis severity before and after percutaneous transluminal coronary angioplasty. Circulation 1993;87:1354 –1367. 5. De Bruyne B, Baudhuin T, Melin JA, Pijls NH, Sys SU, Bol A, Paulus WJ, Heyndrickx GR, Wijns W. Coronary flow reserve calculated from pressure measurements in humans: validation with positron emission tomography. Circulation 1994;89:1013–1022. 6. Pijls NHJ, Van Gelder B, Van der Voort P, Peels K, Bracke FALE, Bonnier HJRM, El Gamal MIH. Fractional flow reserve: a useful index to evaluate the influence of an epicardial coronary stenosis on myocardial blood flow. Circulation 1995;92:3183–3193. 7. De Bruyne B, Bartunek, J, Sys SU, Heyndrickx GR. Relation between myocardial fractional flow reserve calculated from coronary pressure measurements and exercise-induced myocardial ischemia. Circulation 1995;92:39 – 46. 8. Pijls NHJ, De Bruyne B, Peels K, van der Voort PH, Bonnier HJRM, Bartunek J, Koolen JJ. Measurement of fractional flow reserve to assess the functional severity of coronary-artery stenoses. N Engl J Med 1996;334:1703–1708.
902
The American Journal of Cardiology (www.AJConline.org)
9. Bishop AH, Samady H. Fractional flow reserve: critical review of an important physiologic adjunct to angiography. Am Heart J 2004;147: 792– 802. 10. Taillefer R, DePuey EG, Udelson JE, Beller GA, Latour Y, Reeves F. Comparative diagnostic accuracy of Tl-201 and Tc-99m sestamibi SPECT imaging (perfusion and ECG-gated SPECT) in detecting coronary artery disease in women. J Am Coll Cardiol 1997;29: 69 –77. 11. Calnon DA, Kastner RJ, Smith WH, Segalla D, Beller GA, Watson DD. Validation of a new counts-based gated single photon emission computed tomography method for quantifying left ventricular systolic function: comparison with equilibrium radionuclide angiography. J Nucl Cardiol 1997;4:464 – 471.
12. Calnon DA, Kastner RJ, Smith WH, Segalla D, Beller GA, Watson DD,. Quantitative gated single photon emission computed tomography imaging: accounts-based method for display and measurement of regional and global ventricular systolic function. J Nucl Cardiol 1997; 4:451– 463. 13. Aarnoudse WH, Botman KJ, Pijls NH. False-negative myocardial scintigraphy in balanced three-vessel disease, revealed by coronary pressure measurement. Int J Cardiovasc Intervent 2003;5:67–71. 14. Casella G, Leibig M, Schiele TM, Schrepf R, Seelig V, Stempfle HU, Erdin P, Rieber J, Konig A, Siebert U, Klauss V. Are high doses of intracoronary adenosine an alternative to standard intravenous adenosine for the assessment of fractional flow reserve? Am Heart J 2004; 148:590 –595.