Annals of Nuclear Medicine

, Volume 27, Issue 2, pp 132–139

Quantitative evaluation improves specificity of myocardial perfusion SPECT in the assessment of functionally significant intermediate coronary artery stenoses: a comparative study with fractional flow reserve measurements


    • Department of Nuclear MedicineAnkara Oncology Research and Training Hospital
  • Umit O. Akdemir
    • Department of Nuclear MedicineGazi University School of Medicine
  • Sinan A. Kocaman
    • Department of CardiologyGazi University School of Medicine
  • Asife Sahinarslan
    • Department of CardiologyGazi University School of Medicine
  • Timur Timurkaynak
    • Department of CardiologyGazi University School of Medicine
  • Mustafa Unlu
    • Department of Nuclear MedicineGazi University School of Medicine
Original Article

DOI: 10.1007/s12149-012-0666-4

Cite this article as:
Sahiner, I., Akdemir, U.O., Kocaman, S.A. et al. Ann Nucl Med (2013) 27: 132. doi:10.1007/s12149-012-0666-4



Myocardial perfusion SPECT (MPS) is a noninvasive method commonly used for assessment of the hemodynamic significance of intermediate coronary stenoses. Fractional flow reserve (FFR) measurement is a well-validated invasive method used for the evaluation of intermediate stenoses. We aimed to determine the association between MPS and FFR findings in intermediate degree stenoses and evaluate the added value of quantification in MPS.


Fifty-eight patients who underwent intracoronary pressure measurement in the catheterization laboratory to assess the physiological significance of intermediate (40–70 %) left anterior descending (LAD) artery lesions, and who also underwent stress myocardial perfusion SPECT either for the assessment of an intermediate stenosis or for suspected coronary artery disease were analyzed retrospectively in the study. Quantitative analysis was performed using the 4DMSPECT program, with visual assessment performed by two experienced nuclear medicine physicians blinded to the angiographic findings. Summed stress scores (SSS) and summed difference scores (SDS) in the LAD artery territory according to the 20 segment model were calculated. A summed stress score of ≥3 and an SDS of ≥2 were assumed as pathologic, indicating significance of the lesion; a cutoff value of 0.75 was used to define abnormal FFR. Both visual and quantitative assessment results were compared with FFR using Chi-square (χ²) test.


The mean time interval between two studies was 13 ± 11 days. FFR was normal in 45 and abnormal in 13 patients. Considering the FFR results as the gold standard method for assessing the significance of the lesion, the sensitivity and specificity of quantitative analysis determining the abnormal flow reserve were 85 and 84 %, respectively, while visual analysis had a sensitivity of 77 % and a specificity of 51 %. There was a good agreement between the observers (κ = 0.856). Summed stress and difference scores demonstrated moderate inverse correlations with FFR values (r = −0.542, p < 0.001 and r = −0.506, p < 0.001, respectively).


Quantitative analysis of the myocardial perfusion SPECT increases the specificity in evaluating the significance of intermediate degree coronary lesions.


Intermediate degree coronary artery stenosisMyocardial perfusion SPECTFractional flow reserveFunctional significance


Coronary angiography is an invasive imaging modality which is frequently used in the diagnosis and follow-up of patients with coronary artery disease. Intermediate coronary artery stenoses with unknown physiologic significance are detected in some of these patients. Assessment of the severity and physiologic significance of the coronary lesion is of great importance for the deciding between percutaneous coronary intervention (PCI) and medical follow-up in the management of these patients [1]. Myocardial perfusion imaging and catheter-based invasive measurements of flow and pressure are among the commonly used methods for the assessment of functional significance of the coronary lesions.

Myocardial perfusion scintigraphy (MPS) is the most widely used technique in nuclear cardiology. It provides valuable information for the diagnosis of the coronary artery disease, assessment of disease prognosis as well as evaluating the efficacy of the medical treatment and guiding decisions regarding revascularization. Distribution of perfusion along the whole myocardium can be imaged in a single study. Decreased uptake of the radiopharmaceutical in the affected region is seen in the presence of significant coronary artery disease or in the case of abnormal flow reserve. Relative differences in the regional myocardial perfusion can be evaluated visually or quantitatively during the maximal vasodilatation produced by exercise or pharmacological stress. The role of MPS in the decision of PCI in intermediate degree coronary stenoses has been shown [2]. However, in the presence of multivessel or microvascular disease, endothelial dysfunction, and in the case of lesion existence distal to the target lesion; MPS may not be useful for the evaluation of the lesion severity [3].

Fractional flow reserve (FFR) is calculated as the ratio of coronary artery pressure distal to the target lesion to the aortic pressure measured during maximal vasodilatation. FFR is an invasive technique derived by pressure measurements during coronary angiography and has an important role guiding the decision to perform PCI in intermediate coronary stenoses with undetermined functional significance [4]. Deferral of percutaneous intervention in patients with intermediate stenoses with FFR values above 0.75 has been shown to be safe in different studies [4, 5].

The aim of our study was to evaluate the relation between the findings of the MPS and fractional flow reserve, and the additive effect of quantitative values obtained by MPS to the visual evaluation in the assessment of the severity of the lesion when FFR results are accepted as the gold standard.

Materials and methods

The study group was selected from among the patients who underwent coronary angiography between 1998 and 2009 in Gazi University Hospital, Ankara, Turkey. Patients who had undergone both FFR measurement and myocardial perfusion imaging for the assessment of the significance of intermediate degree stenosis (40–70 %) were evaluated. Patients with a previous revascularization history of either surgical bypass or percutaneous intervention, evidence of ongoing infection or inflammation, and or recent acute coronary syndrome either with or without ST-segment elevation (1 month before enrollment) were excluded from the study. A total of 75 patients met the criteria for eligibility for the study. Non-LAD lesions (n = 6), data of patients with MPS using Tc-99 m SESTAMIBI (n = 4) and patients with a history of prior myocardial infarction in LAD territory (n = 6) were excluded. One patient was excluded due to severe motion artifact in MPS. The remaining 58 patients were included in the study. The MPS and coronary angiography data were analyzed retrospectively. The study was approved by the local ethics committee.

Coronary angiography and fractional flow reserve

Multidirectional coronary angiography was performed using Judkins technique in all patients. FFR measurement had been performed by the following technique: all patients were anticoagulated with at least 5000 units of intravenous unfractionated heparin before the procedure. A 0.014-inch RadiPressureWire XT (Radi Medical Systems, Uppsala, Sweden) pressure-monitoring guide wire was introduced through a 6–8 French guiding catheter and calibrated. Then, the pressure wire was advanced through the coronary artery until it was positioned distal to the stenosis [6, 7]. After recording the baseline distal intracoronary pressure, an intracoronary adenosine bolus starting at a dose of 30 μg was administered (≥30 μg bolus, mean 93 ± 58 μg). An incremental dose approach was applied in patients with FFR between 0.85 and 0.75 to achieve maximum hyperemia [8]. The mean coronary pressure distal to the lesion was recorded. FFR was calculated by dividing the mean distal intracoronary pressure by the mean arterial pressure. This procedure was repeated twice and the minimum FFR calculation was used to determine the severity of the lesion. A lesion was accepted as pathologic and functionally significant if FFR < 0.75 after adenosine administration.

201Tl myocardial perfusion SPECT

All patients fasted for at least 6 h before the study. In 52 patients, dynamic exercise on a treadmill using the Bruce Protocol was carried out until the patients reached ≥85 % of the age-predicted maximum heart rate (APMHR) (maximal age-predicted heart rate = 220 − age) with the administration of 2–3 mCi of 201Tl at peak exercise. Exercise was continued for one more minute after the injection. Acquisition of the SPECT images was started 3–10 min after the injection of the radiopharmaceutical. Six patients were scheduled for pharmacological stress MPS due to inability to reach 85 % of APMHR on the treadmill (n = 2), or the presence of orthopedic problems (n = 4). Caffeine-containing beverages and medications that could interfere with the vasodilatory effects of adenosine were discontinued for at least 12 h before the study. Dipyridamole at a dose of 0.57 mg/kg over 4 min was administered to induce vasodilation in the coronary arteries. Injection of the tracer was performed 3–5 min after completion of the infusion. Single photon emission tomography was performed using a dual-headed gamma camera (General Electric, Optima NX) equipped with low energy all purpose collimators. Acquisition was performed using a 180° circular orbit, a 64 × 64 matrix size and a acquisition time of 32 frames of 40 s. Redistribution images were gathered using the same parameters 3–4 h after the stress imaging [9].

Evaluation of MPS data

Standard filtered back projection using a Butterworth filter was performed without applying attenuation correction. Stress and redistribution tomographic images were displayed side by side in the short axis, horizontal long-axis and vertical long-axis for visual assessment. Two experienced nuclear medicine physicians (M.U. and O.A.), who had no knowledge of the angiographic data evaluated the scintigraphic data for the presence of ischemia. The final decision was made by consensus in the case of discordant results. Perfusion SPECT images were scored automatically based on the 20-segment model using the 4DMSpect 2.2 program (Xeleris 1.0, General Electric) [10]. Summed stress scores (SSS) and summed difference scores (SDS) were calculated for the segments 1, 2, 7, 8, 13, 14, 15, 19 and 20 representing LAD territory (Fig. 1). An SSS of ≥3 and SDS of ≥2 were accepted as pathologic showing the presence of ischemia.
Fig. 1

Schematic representation of 20-segment model of left ventricular myocardium

Statistical analysis

Continuous variables were given as mean ± SD and categorical variables as percentages. Continuous variables were compared using the independent t test, and the χ² test was used for the categorical variables. All tests of significance were two tailed. Pearson correlation was calculated to investigate the relationship between FFR and MPS findings. Sensitivity and specificity were calculated to predict the ability of MPS evaluation (both visual and scoring) to identify functional significance of a lesion in comparison with FFR. The SPSS statistical software (SPSS 20.0 for Windows Inc., Chicago, IL, USA) was used for all statistical calculations.


Fifty-eight patients (19 females, 39 males) were included in the study. Mean time interval between FFR measurement and MPS was 13 ± 11 days. FFR values were pathologic (<0.75) in 13 patients and normal in the rest. There were no significant differences between patients with normal and pathologic FFR results considering coronary artery disease risk factors except for male gender and high-density lipoprotein levels. Patient characteristics, angiography and SPECT findings are shown in Table 1.
Table 1

Demographic characteristics and clinical background of the patients with normal and pathologic FFR results


FFR < 0.75 (n = 13)

FFR ≥ 0.75 (n = 45)

p value


Age (years)

60 ± 12

56 ± 8


Gender (male)

12 (92)

27 (60)


Total cholesterol (mg/dL)

190 ± 53

196 ± 47


LDL (mg/dL)

122 ± 46

118 ± 42


Triglycerides (mg/dL)

166 ± 89

181 ± 79


HDL (mg/dL)

38 ± 8

43 ± 7


Creatinine (mg/dL)

1.3 ± 1.7

1.0 ± 0.1


Hemoglobin (mg/dl)

14.0 ± 2.2

14.6 ± 1.5



5 (39)

27 (60)



6 (46)

18 (40)


Smoking history

6 (46)

23 (51)


Family history

5 (39)

15 (33)


Previous MI in non-LAD territory

1 (8)

5 (11)


Ejection fraction (%)

58 ± 10

60 ± 4




95 %

96 %



50 %

60 %


Beta blockers

86 %

88 %


Calcium channel blockers

26 %

30 %



70 %

84 %


Angiography findings

QCA (%)

48.5 ± 8.5

47.0 ± 8.0


Visual lesion severity (%)

56 ± 11

49 ± 11


Adenosine (μg)

94 ± 59

91 ± 58


NDVs (≥70 %)

1.0 ± 0.8

0.7 ± 0.8


Multivessel disease (%)

8 (62)

17 (38)


SPECT measurements





Exercise time (min)





4.7 ± 2.4

0.9 ± 1.7



3.2 ± 2.2

0.6 ± 1.2


Data presented as mean ± SD, numbers in parentheses indicate percentages

LDL low-density lipoprotein, HDL high-density lipoprotein, QCA quantitative coronary angiography, NDVs number of diseased vessels, SDS summed difference score, SSS summed stress score, EF ejection fraction, ACEI/ARB angiotensin inhibitors/angiotensin receptor blockers, NS not significant, APMHR age-predicted maximum heart rate

Dynamic exercise on a treadmill was the choice of stress in 52 (90 %) of the patients and pharmacologic stress in 6 (10 %), who were not able to perform exercise due to various problems. Mean percentage of APMHR was 95.94 ± 10.0 % (range 85–135 %) and mean duration of exercise was 8.37 ± 2.53 min (range 2.47–14.30 min) for patients who performed exercise on a treadmill. Exercise was terminated in 3 patients with sudden onset tachycardia and it was symptom limited in 15 patients after 85 % threshold of APMHR (fatigue in 14, angina in 1 patient). The remaining patients performed exercise until ≥90 % of APMHR was reached.

The degree of stenoses by visual assessment in coronary angiography in patients with FFR values below 0.75 was found to be significantly higher than that in the normal FFR group (p < 0.001) while no significant differences were found between the groups when QCA was used for stenosis percentage (Table 1). Mean percent stenosis diameters in QCA were not significantly different between the two FFR groups and between groups according to MPS data as indicated by either visual or quantitative evaluation (Table 2). When the significance of the lesion in the coronary artery supplying the related region was assessed using FFR values; the sensitivity and specificity of the visual evaluation of myocardial perfusion SPECT images for the presence of ischemia were 77 and 51 %, respectively (Table 3). The agreement between two physicians was 90 % for visual interpretation. Summed stress and difference scores demonstrated moderate inverse correlations with FFR values (r = −0.542, p < 0.001 and r = −0.506, p < 0.001, respectively). (Fig. 2a, b). χ² analysis showed that FFR and SSS values were related (χ² = 26974; p < 0.001). Sensitivity and specificity of pathologic SSS were 85 and 84 %, and of pathologic SSS and SDS were 61 and 89 % respectively, relying on FFR values for the assessment of lesion significance (Tables 4, 5).
Table 2

Comparison of stenosis diameter in QCA and MPS

Quantitative coronary angiography (%)


p value


Visual ischemia (n = 21)

48.3 ± 10.1

Visual non-ischemia (n = 37)

48.3 ± 10.4



SSS ≥3 (n = 19)

51.7 ± 8.6

SSS <3 (n = 39)

46.6 ± 10.6



SSS ≥3 and SDS ≥2 (n = 14)

52.1 ± 7

SSS <3 or SDS <2 (n = 44)

46.9 ± 10.8


Data presented as mean ± SD

Table 3

Comparison of FFR with visual assessment of MPS

Presence of Ischemia



<0.75 (n)

≥0.75 (n)

Ischemic (n)











Fig. 2

a Correlation of FFR with Summed Stress Score. b Correlation of FFR with Summed Difference Score

Table 4

Comparison of FFR and SSS




<0.75 (n)

≥0.75 (n)

≥3 (n)




<3 (n)








Table 5

Comparison of FFR and SSS together with SDS




<0.75 (n)

≥0.75 (n)

SSS ≥ 3 and SDS ≥ 2 (n)




SSS < 3 or SDS < 2 (n)









Appropriate patient and lesion selection for percutaneous coronary intervention creates a great challenge in terms of cost efficiency and reduction of possible complications [11]. Assessment of the functional significance of intermediate degree coronary stenoses is of great importance for avoidance of unnecessary interventions. Quantitative coronary angiography has been shown to be inaccurate in defining the functional significance of coronary artery stenoses [12, 13]. Noninvasive techniques such as myocardial perfusion SPECT, myocardial perfusion PET, dobutamine stress echocardiography (DSE), contrast echocardiography and quantitative cardiac magnetic resonance imaging as well as invasive procedures like intravascular ultrasonography, measurement of fractional flow reserve and coronary flow reserve can be used for the assessment of the significance of intermediate degree coronary artery stenoses [1416]. In this study, it was shown that there is a moderate correlation between fractional flow reserve values measured for the intermediate degree stenoses and summed stress scores calculated by myocardial perfusion SPECT.

In the present study, SSS and SDS showed moderate inverse correlations with FFR values whereas, no correlation was found between FFR and visual analysis on MPS, indicating the added value of quantification in MPS. A good, but far from perfect concordance has been shown in a meta-analysis of 21 studies comparing the results of noninvasive techniques (MPS and DSE) and fractional flow reserve for the evaluation of myocardial ischemia [13]. These data were largely driven by the results of the nuclear medicine tests. It has also been suggested in the same analysis that quantitative coronary angiography fails to predict the functional significance of coronary lesions, with FFR providing important independent data against QCA. There was an extensive lack of concordance between QCA and FFR especially for intermediate degree (40–70 %) lesions. It has also been shown that the decision regarding revascularization for intermediate degree stenoses can be safely based on FFR values, and the long-term prognostic value of FFR is higher compared to MPS in multivessel coronary artery disease [4, 5, 1720]. Furthermore, a large multicenter prospective study showed that PCI guided by FFR significantly reduces clinical events in multivessel coronary artery disease [21]. As the current literature lacks evidence indicating that the presence of ischemia on MPS is associated with a worse outcome considering an intermediate degree stenosis, FFR should be regarded as the gold standard for the evaluation of the physiologic significance of stenosis with indeterminate significance.

Various quantitative software packages have provided high diagnostic performances for the detection of coronary artery disease [22, 23]. To the best of our knowledge, the diagnostic accuracy of quantitative analysis programs has never been tested for assessment of intermediate degree stenoses. In our study, quantitative evaluation of myocardial perfusion images for myocardial ischemia using 4DMSpect 2.2 software increased sensitivity and specificity compared to visual-only interpretation when FFR values were accepted as the reference. Since revascularization of a particular stenosis has drawbacks of its own, trying to overcome the problem of the slightest sign of ischemia may bias the decision regarding intervention. The presence of mild and moderate ischemia was reported as well as severe reversible perfusion defects in the present study. Using a threshold both for SDS and SSS might have eliminated the cases which were assessed as slightly ischemic by visual-only interpretation. Specificity of visual interpretation of myocardial perfusion SPECT in our study was far below the accepted specificity values for detection of coronary artery disease. However, values considering the diagnostic accuracy of MPS were mostly decided by studies on arteries with a luminal stenosis of ≥70 %, where perfusion changes are expected to be more pronounced. For this reason, quantitative evaluation of the myocardial perfusion images seems to be superior to visual evaluation for assessment of the functional significance of the intermediate degree coronary stenoses. In our opinion, the results of quantitative analysis should be taken into account strictly both by the nuclear medicine physician and interventional cardiologist for assessment of the functional severity of intermediate degree stenoses.

Ischemic threshold of FFR < 0.75 was used in our study for the evaluation of lesion significance. Previous studies demonstrated that deferral of revascularization of intermediate degree stenoses with FFR ≥ 0.75 was safe [4]. Moreover, routine measurement of FFR in patients with multivessel coronary artery disease was proposed to decrease adverse events related to unnecessary PCI of stenoses that are not inducing ischemia [21]. Based on the results of large multicenter studies evaluating the role of FFR measurement on patient outcome, stenting is always justified for a stenosis with FFR ≤ 0.75. However, the decision for intervention may be questioned in 0.76–0.80 interval and should also be based on information regarding clinical characteristics of the patient as well as noninvasive imaging studies [24]. Using the FFR threshold of ≤80 for the presence of ischemia did not alter our results, with quantification still improving the evaluation ischemia (data not presented).

Both stress and difference scores in LAD territory demonstrated moderate inverse correlations with FFR values in our study. Fractional flow reserve and MPS evaluate the presence of myocardial ischemia relying on completely different principles. Myocardial perfusion imaging is based on the relative flow reserve of different territories of the myocardium, and normal perfusion in at least one region is necessary for the detection of the perfusion abnormality. On the other hand, fractional flow reserve is lesion specific and derived from the ratio of the coronary pressure distal to the lesion and the aortic pressure. Considering these fundamental differences between the two techniques, absolute concordance of the results should not be expected.

Fractional flow reserve measurement technique has some limitations as well. It has been suggested that a cut-off value of 0.75 may be low for patients having left ventricular hypertrophy, left ventricular dysfunction and conditions resulting in microvascular disorders such as diabetes mellitus and hyperlipidemia [25, 26]. Moreover, non-ischemic FFR results would be expected in the case of exercise-induced spasm despite the presence of apparently tight stenosis [24]. Patients with risk factors which may cause microvascular dysfunction were included in our study and this makes the FFR cut-off value of 0.75 questionable for the assessment of lesion significance.

Sensitivity and specificity of the MPS scores to detect FFR values below 0.75 for the target vessel have been shown to be lower in patients with prior myocardial infarction [27]. Six patients (10 %) in our study group had myocardial infarction in non-LAD territories and this may have influenced the concordance between the two techniques.

Non-homogeneity of the study group is another limitation of the present study. Owing to the multifactorial etiopathogenesis of coronary artery disease it included patients with various risk factors such as diabetes, hypertension, hyperlipidemia and cigarette smoking which may cause endothelial and microvascular dysfunction. This may have caused false-positive results on the myocardial perfusion imaging. Our study group was too small to further subanalyze the effects of these different risk factors. Furthermore, 25 (43 %) of the patients had a significant lesion in at least one coronary artery other than the left anterior descending coronary artery. Poor concordance between FFR and MPS has been previously demonstrated in patients with multivessel disease, probably owing to the lower sensitivity and specificity of the MPS scores in such patients [3, 17, 28]. Induction of ischemia in MPS was either with exercise (90 %) or with dipyridamole (10 %) in the present study. However, considering that the reported sensitivity and specificity of the two techniques are similar, it is unlikely that the different protocols of inducing ischemia affected our results [29]. There was a lack of constant adenosine dosage for pharmacologic vasodilatation for FFR measurement in the patient population. Although it can be speculated that adenosine dosage may affect the degree of vasodilatation, no significant difference was found in FFR after different dosages of adenosine in other studies [30, 31]. In addition to that, it is believed that an adenosine dose greater than 40 μg is not needed for maximal hyperemia and does not produce additional dilatation. In our study, different doses were preferred by the operator; however, no significant difference was observed at different adenosine doses, and no statistically significant differences were found for adenosine doses in the FFR groups.

Our study suggests that quantitative evaluation of the myocardial perfusion images increases the specificity and sensitivity of MPS in detecting the functional significance of indeterminate coronary lesions, as well as having better concordance with FFR values. Myocardial perfusion SPECT, being one of the most extensively evaluated methods showing inducible ischemia, is widely available and has great value in guiding the decision for intervention of an intermediate degree stenosis. Thus, scores of underperfusion demonstrated either on stress or difference data are of great value in the assessment of lesions with unknown functional significance.

Conflict of interest

None declared.

Copyright information

© The Japanese Society of Nuclear Medicine 2012