OUP user menu

Effect of cardiac hybrid 15O-water PET/CT imaging on downstream referral for invasive coronary angiography and revascularization rate

Ibrahim Danad, Pieter G. Raijmakers, Hendrik J. Harms, Cornelis van Kuijk, Niels van Royen, Michaela Diamant, Adriaan A. Lammertsma, Mark Lubberink, Albert C. van Rossum, Paul Knaapen
DOI: http://dx.doi.org/10.1093/ehjci/jet125 170-179 First published online: 9 July 2013


Aims This study evaluates the impact of hybrid imaging on referral for invasive coronary angiography (ICA) and revascularization rates.

Methods and results A total of 375 patients underwent hybrid 15O-water positron emission tomography (PET)/computed tomography (CT)-based coronary angiography (CTCA) imaging for the evaluation of coronary artery disease (CAD). Downstream treatment strategy within a 60-day period after hybrid PET/CTCA imaging for ICA referral and revascularization was assessed. CTCA examinations were classified as showing no (obstructive) CAD, equivocal (borderline test result), or obstructive CAD, while the PET perfusion images were classified into normal or abnormal. On the basis of CTCA imaging, 182 (49%) patients displayed no (obstructive) CAD. Only 10 (5%) patients who showed no (obstructive) CAD on CTCA were referred for ICA, which were all negative. An equivocal CT study was observed in 80 (21%) patients, among whom 56 (70%) showed normal myocardial perfusion imaging (MPI), resulting in referral rates for ICA of 18% for normal MPI and 71% for abnormal MPI, respectively. No revascularizations were performed in the presence of normal MPI, while 59% of those with abnormal MPI were revascularized. CTCA indentified obstructive CAD in 113 (30%) patients accompanied in 59 (52%) patients with abnormal MPI. Referral rate for ICA was 57% for normal MPI and 88% for those with abnormal MPI, resulting in revascularization rates of 26% and 72%, respectively.

Conclusion Hybrid 15O-water PET/CTCA imaging impacts clinical decision-making with regard to referral for ICA and revascularization procedures. Particularly, in the presence of an equivocal or abnormal CTCA, MPI could guide in the decision to refer for ICA and revascularization.

  • hybrid PET/CT imaging
  • referral for invasive coronary angiography
  • revascularization rate


The utilization of non-invasive cardiac imaging for the evaluation of coronary artery disease (CAD) has increased rapidly over the last years.1,2 Computed tomography (CT)-based coronary angiography (CTCA) is extensively used for non-invasive diagnosis of CAD. It has demonstrated to be an excellent tool for ruling out CAD due to its high sensitivity and negative predictive value.35 However, functional assessment is needed in the presence of a CTCA-graded obstructive stenosis, since anatomical imaging is limited in the assessment of the functional relevance of a stenosis.69 It has been demonstrated that revascularization of coronary arteries with angiographically obstructive stenosis without proof of ischaemia confers no prognostic or symptomatic benefit for patients and is potentially more harmful than medical therapy.8 The fact that only one-third of lesions deemed significant at invasive coronary angiography (ICA) are indeed functionally significant10 underscores the importance of using a comprehensive diagnostic strategy by performing coronary imaging and myocardial perfusion imaging (MPI). Previous data have indicated that the integration of coronary anatomy and perfusion, obtained with hybrid devices, improves both diagnostic accuracy and clinical decision-making.5,11,12 It is, therefore, expected that cardiac hybrid imaging will result in better personalized treatment strategy of CAD, including a more judicious referral of patients to invasive cardiac procedures. However, studies examining the impact of hybrid cardiac imaging on subsequent downstream resource utilization are scarce.12,13 The aim of the current study was therefore to evaluate in daily clinical practice the impact of hybrid 15O-water positron emission tomography (PET)/CTCA on referral for ICA and revascularization rates in a large clinical cohort of patients evaluated for CAD.


Study population

This is a retrospective study of 375 consecutive patients who underwent CTCA and PET MPI on a single-session hybrid PET/CT scan (Gemini TF 64, Philips Healthcare, Best, The Netherlands) for the evaluation of suspected CAD at the VU University Medical Center, Amsterdam, The Netherlands. All patients were referred from within our institution or outpatient clinic by cardiologists with several years of experience, with hybrid PET/CT for the evaluation of CAD. Exclusion criteria were atrial fibrillation, second- or third-degree atrioventricular block, impaired renal function (glomerular filtration <45 mL/min), symptomatic asthma, pregnancy, or a documented history of CAD. A history of CAD was defined as a prior percutaneous coronary intervention (PCI), coronary artery bypass graft surgery (CABG), or a previous myocardial infarction. Electrocardiography did not show signs of a previous myocardial infarction, and echocardiography showed a normal left ventricular function without wall motion abnormalities in all patients. None of the subjects underwent previous MPI and all patients presented for the first time to the cardiologist with chest pain or a high risk profile. Post-imaging treatment strategy was left to the discretion of the referring physician and was generally based on the hybrid imaging test results, as well as the clinical history and symptoms of a patient. Downstream treatment strategy after hybrid PET/CTCA imaging was assessed for referral for ICA and revascularization procedures (PCI or CABG) within a 60-day interval after PET/CT. CAD pre-test likelihood was determined according to the Diamond and Forrester criteria, using percent cut-offs of <13.4%, >87.2%, and in between for low, high, and intermediate pre-test likelihood, respectively.14 The need for written informed consent was waived by the institutional review board (local ethics committee) because of the nature of the study, which solely had clinical data collection.

PET/CTCA imaging

The PET/CTCA imaging protocol is shown in Figure 1. The PET/CTCA imaging sequence has been described in detail previously.11,15 Parametric myocardial blood flow (MBF) images, showing MBF on the voxel level, were generated and quantitative analysis was performed using the in-house developed software, Cardiac VUer.16 In short, input functions were obtained using automatic segmentation of dynamic images, after which parametric images were obtained as described previously.16

Figure 1

Cardiac 15O-water PET/CT imaging protocol. CTCA, CT coronary angiography; LD CT, low-dose CT for attenuation correction; H215O, 15O-water.

Invasive coronary angiography

ICA was performed according to standard clinical protocols. The coronary tree was divided to a 16-segment coronary artery model modified from the American Heart Association.17 Significant CAD was defined by a visually graded stenosis of ≥50%. When fractional flow reserve (FFR) measurements were performed, visual grading was overruled by FFR, where a value of ≤0.80 was considered significant. FFR was measured at the discretion of the interventional cardiologist performing the ICA procedure. The ICA and FFR procedures have been described in detail elsewhere.11

Data interpretation

All CT scans were analysed with a three-dimensional workstation (Brilliance, Philips Medical Systems, Best, the Netherlands) by an experienced radiologist and cardiologist. The axial slices were initially evaluated for the presence of significant segmental disease, and additionally, curved multiplanar reconstructions were obtained for each coronary artery. These reconstructions were used to determine coronary stenosis severity. The coronary tree was evaluated according to a 16-segment coronary artery model modified from the American Heart Association.17 The readers classified the segments as (i) no (obstructive) CAD (no stenosis in combination with a coronary artery calcium (CAC) score of zero or luminal diameter reduction <50%); (ii) equivocal CT result (equivocal lesion or unable to grade stenosis severity due to artefacts, or (iii) obstructive CAD (luminal diameter reduction ≥50%). The parametric PET stress perfusion images were visually analysed by two experienced readers, whereby a perfusion defect of at least two adjacent segments was assigned to a vascular territory. Subsequently, this regional perfusion value was used for further analyses instead of the mean MBF or coronary flow reserve (CFR) of the predefined vascular territory. After visual assessment of the images, the readers interpreted the quantitative perfusion data (i.e. hyperaemic MBF and CFR). A CFR value of <2.0 and/or hyperaemic MBF of <2.0 mL/min/g was considered abnormal according to the previously published data.1822 Discordant quantitative PET findings were seen in 69 (18%) patients, whereby in 55 (15%) patients hyperaemic MBF and in 14 (4%) patients, the CFR value was used for classification following the visual assessment of the PET images. In case of a visual reversible perfusion defect in combination with normal quantitative perfusion values, the PET scan was classified as normal. These PET results were combined with the coronary anatomy as obtained with CTCA to obtain a hybrid interpretation and final diagnosis for each PET/CT scan. Figure 2 shows an example of a hybrid PET/CTCA fusion image of a patient with a left anterior descending (LAD) artery stenosis.

Figure 2

A 52-year-old male with atypical angina. Hybrid 15O-water PET/CTCA imaging reveals a severely reduced hyperaemic perfusion (1.25 mL/min/g) in the area supplied by the LAD artery.

Statistical analysis

Continuous variables are expressed as mean value and standard deviation, whereas categorical variables are presented as frequencies or percentages. Differences in referral rates for ICA and subsequent revascularization between the imaging groups were compared using the χ2 test. To determine the most powerful predictors of referral to catheterization within 60 days after PET/CTCA (early catheterization) and referral to revascularization within this same time span, a multiple logistic regression was performed. In the univariate analysis, all available covariates were analysed. In the multivariate analysis, covariates were included when P < 0.05 after univariate analysis. Values were expressed as odds ratios (ORs) with 95% confidence intervals (CIs). A P-value of <0.05 was considered significant. All analyses were performed using the SPSS Statistics Software (version 20, IBM, Armonk, New York, USA).


Patient characteristics

In total, 375 patients underwent hybrid 15O-water PET/CTCA imaging for the evaluation of CAD. Baseline patient characteristics are given in Table 1. The mean heart rate during CTCA was 58 ± 7 bpm. A good quality CTCA scan was obtained in 257 (69%) patients, while 87 (23%) and 31 (8%) patients, respectively, had a moderate or poor CTCA study.

View this table:
Table 1

Patient characteristics (N = 375)

Age (years)58 ± 10
Males (%)192 (51%)
Body mass index (kg/m2)27 ± 4
CAD risk factors (%)
 Diabetes mellitus type II66 (18%)
 Hypertension172 (46%)
 History of smoking158 (42%)
 Hypercholesterolaemia122 (33%)
 Family history of premature CAD195 (52%)
Medication (%)
 Statins227 (61%)
 Beta-blockers220 (59%)
 Aspirin247 (66%)
 ACE-inhibitors60 (16%)
 AT-II antagonists64 (17%)
 Calcium antagonists90 (24%)
Reason for referral (%)
 Typical angina95 (25%)
 Atypical angina131 (35%)
 Non-anginal chest pain112 (30%)
 High risk, no chest discomfort37 (10%)
Pre-test likelihood of CAD (%)
 Low78 (21%)
 Intermediate225 (60%)
 High72 (19%)
Screening exercise ECG (%)276 (74%)
 Normal140 (51%)
 Inconclusive72 (26%)
 Abnormal64 (23%)
  • CAD, coronary artery disease; ACE, angiotensin-converting enzyme; AT-II, angiotensin-II-receptor; ECG, electrocardiogram.

Hybrid PET/CTCA findings

The mean CAC score was 14 ± 46, 345 ± 808, and 484 ± 679 in those patients with no (obstructive) CAD, an equivocal CT result, and obstructive CAD on CTCA, respectively. On the basis of CTCA imaging, 182 (49%) patients were graded as no (obstructive) CAD, while an equivocal CT result was observed in 80 (21%) patients. Obstructive CAD on CTCA was seen in 113 (30%) patients. Overall, PET revealed normal perfusion in 276 (74%) patients, while an abnormal MPI study was reported in 99 (26%) patients. As shown in Figure 3, the majority of patients with the CT-based absence of (obstructive) CAD displayed no ischaemia at PET MPI (n = 166, 91%), while in patients with an equivocal CT result, 56 (70%) showed normal MPI. In case of obstructive CAD on CTCA, abnormal MPI was seen in approximately half of these patients (Figure 3).

Figure 3

Pie charts illustrating the relationship between CTCA imaging and PET MPI. CTCA, CT coronary angiography; CAD, coronary artery disease; PET, positron emission tomography; MPI, myocardial perfusion imaging.

Referral for invasive coronary angiography and downstream revascularization

The PET/CTCA findings in relation to referral for ICA and downstream revascularization are illustrated in Figure 4. A total of 120 (32%) patients were referred for ICA, among whom 58 (48%) had obstructive CAD on ICA. FFR was performed in 44 (37%) patients. During 60-day follow-up, only 10 (5%) patients with the absence of (obstructive) CAD on CTCA were further evaluated by ICA, of whom none showed obstructive CAD on ICA. In patients with an equivocal CTCA study (n = 80) and normal MPI (n = 56; 70%), 10 (18%) were further evaluated by ICA, of whom 2 had obstructive CAD on ICA. In patients with an equivocal CTCA study (n = 80) and ischaemia on PET (n = 24; 30%), 17 patients (71%) were referred for ICA, of whom 10 displayed obstructive CAD and subsequently underwent a revascularization procedure. Of the subjects with obstructive CAD on CTCA (n = 113) in combination with normal MPI (n = 54; 48%), an abnormal ICA result was seen in 10 (19%) patients, of whom 8 were referred for a revascularization. Thirty-six (61%) patients with obstructive CAD on CTCA and an abnormal MPI study (n = 59; 52%) displayed obstructive CAD on ICA, of whom 34 underwent a revascularization. Conservative treatment was chosen in all patients with the absence of symptoms or relatively unfavourable coronary characteristics for revascularization. The impact of 15O-water PET/CTCA imaging on referral rates for ICA and revascularization is shown in Figure 5. Of note, the yield of CAD per ICA was 20 and 59% in patients in whom hybrid PET/CTCA revealed an indeterminate CT result in combination with normal and abnormal perfusion, respectively, compared with 32 and 69% in the group with obstructive CAD on CTCA.

Figure 4

Relationship between hybrid PET/CTCA findings and referral for ICA and the number of revascularizations. PET, positron emission tomography; CTCA, CT coronary angiography; CAD, coronary artery disease; MPI, myocardial perfusion imaging; ICA, invasive coronary angiography; FFR, fractional flow reserve.

Figure 5

Rates of referral for ICA after hybrid PET/CTCA findings (A) and rates of downstream revascularization (B). PET, positron emission tomography; CTCA, CT coronary angiography; ICA, invasive coronary angiography; CAD, coronary artery disease; MPI, myocardial perfusion imaging.

Predictors of subsequent invasive coronary angiography and revascularizations

All the demographic, clinical, and PET/CTCA variables that were associated with referral for ICA are summarized in Table 2. Univariate analysis showed that age (OR 1.07, 95% CI 1.04–1.09, P < 0.001), male gender (OR 2.69, 95% CI 1.71–4.25, P < 0.001), diabetes mellitus (OR 2.35, 95% CI 1.37–4.05, P < 0.01), hypertension (OR 1.98, 95% CI 1.28–3.08, P < 0.01), and hypercholesterolaemia (OR 2.00, 95% CI 1.27–3.15, P < 0.01) were predictive for referral to ICA. In addition, the type of chest pain, CAC score, and CTCA results showed a significant predictive value for referral to ICA. An OR for ICA was 2.54 higher for patients with typical angina than for those without anginal complaints. Obstructive CAD on CTCA was associated with a 47.6 higher OR for referral to ICA than those with no (obstructive) CAD on CT. Hyperaemic MBF seems to be a better predictor for referral to ICA (OR 10.8) than the CFR (OR 6.36), respectively (Table 2). The model that included hyperaemic MBF revealed that only diabetes mellitus (OR 2.62, 95% CI 1.04–6.61, P = 0.04), CTCA results [obstructive CAD on CTCA was associated with a 27.7 higher OR than no (obstructive) CAD], and hyperaemic MBF (OR 5.29, 95% CI 2.48–11.3, P < 0.001) were independent predictors of referral to ICA. The multivariate regression model that included CFR as a perfusion parameter showed that only the CAC score (OR 1.07, 95% CI 1.01–1.15, P = 0.03), CTCA results [obstructive CAD had a 28.9 higher OR than no (obstructive) CAD on CTCA], and CFR (OR 4.57, 95% CI 2.25–9.28, P < 0.001) were independently associated with referral to ICA (Table 2). With regard to revascularizations, univariate regression analysis showed that age (OR 1.04, 95% CI 1.01–1.07, P < 0.01), male gender (OR 2.83, 95% CI 1.53–5.26, P < 0.01), CAC score (OR 1.09 95% CI 1.05–1.14, P < 0.001), hyperaemic MBF (OR 12.2, 95% CI 6.39–23.3, P < 0.001), and the CFR (OR 5.14, 95% CI 2.85–9.29, P < 0.001) were associated with subsequent revascularizations (Table 3). In addition, patients with typical angina (a 16.6 higher OR than those without anginal complaints), high pre-test likelihood (a 8.25 higher OR than those with a low likelihood for CAD), and obstructive CAD on CTCA (a 36.7 higher OR than those with (no) obstructive CAD on CTCA) were significantly associated with a revascularization procedure (Table 3). Multivariate analysis revealed that only CTCA results, hyperaemic MBF <2.0 mL/min/g, and the CFR were independent predictors of a revascularization procedure (Table 3).

View this table:
Table 2

Demographic, clinical, CTCA, and PET variables as predictors of referral for ICA

CharacteristicsUnivariate analysisMultivariate analysis
Model 1Model 2
OR (95% CI)P-valueOR (95% CI)P-valueOR (95% CI)P-value
Age (years)1.07 (1.04–1.09)<0.0011.00 (0.96–1.04)0.951.01 (0.97–1.05)0.69
Men2.69 (1.71–4.25)<0.0011.12 (0.52–2.38)0.781.79 (0.87–3.68)0.11
Body mass index (kg/m2)1.05 (0.99–1.10)0.08
Type of chest pain<0.010.300.14
 No anginal complaints1 (–)1 (–)1 (–)
 Aspecific chest pain0.99 (0.43–2.28)0.97 (0.17–5.56)1.42 (0.24–8.39)
 Atypical angina0.95 (0.42–2.16)0.58 (0.09–3.66)0.78 (0.12–4.95)
 Typical angina2.54 (1.11–5.81)1.88 (0.20–17.5)3.31 (0.38–29.1)
Pre-test likelihood for CAD<0.0010.470.59
 Low1 (–)1 (–)1 (–)
 Intermediate2.07 (1.09–3.93)1.77 (0.42–7.39)1.34 (0.33–5.44)
 High4.57 (2.18–9.58)0.93 (0.12–7.42)0.70 (0.09–5.19)
CAD risk factors
 Diabetes mellitus (yes/no)2.35 (1.37–4.05)<0.012.62 (1.04–6.61)0.042.31 (0.93–5.78)0.07
 Hypertension (yes/no)1.98 (1.28–3.08)<0.010.91 (0.46– 1.81)0.800.89 (0.45–1.76)0.89
 Smoking history (yes/no)1.30 (0.84–2.02)0.24
 Hypercholesterolaemia (yes/no)2.00 (1.27–3.15)<0.011.62 (0.82–3.20)0.171.55 (0.78–3.08)0.21
 Family history of CAD (yes/no)0.65 (0.42–1.01)0.051.06 (0.54–2.06)0.880.94 (0.50–1.78)0.85
Exercise ECG0.38
 Non-ischaemic1 (–)
 Ischaemic1.31 (0.71–2.43)
 Non-diagnostic0.78 (0.42–1.48)
 CAC score1.33 (1.21–1.48)<0.0011.07 (1.00–1.15)0.051.07 (1.01–1.15)0.03
CTCA result<0.001<0.001<0.001
  (Non)obstructive CAD1 (–)1 (–)1 (–)
 Equivocal result8.75 (3.98–19.3)6.25 (2.53–15.5)5.67 (2.27–14.1)
 Obstructive CAD47.6 (22.2–102)27.7 (11.2–68.3)28.9 (11.7–71.3)
Quantitative PET
 Hyperaemic MBF (<2.0 mL/min/g)10.8 (6.33–18.5)<0.0015.29 (2.48–11.3)<0.001
 CFR (<2.0)6.36 (3.84–10.6)<0.0014.57 (2.25–9.28)<0.001
  • Two multivariate analyses were performed either excluding CFR and including hyperaemic MBF (model 1), or including CFR and excluding hyperaemic MBF (model 2).

  • OR, odds ratio; CI, confidence interval; CAD, coronary artery disease; ECG, electrocardiogram; CAC, coronary artery calcium (divided by 100); CTCA, CT coronary angiography; MBF, myocardial blood flow; CFR, coronary flow reserve.

View this table:
Table 3

Demographic, clinical, CTCA, and PET variables as predictors of subsequent revascularizations

CharacteristicsUnivariate analysisMultivariate analysis
Model 1Model 2
OR (95% CI)P-valueOR (95% CI)P-valueOR (95% CI)P-value
Age (years)1.04 (1.01–1.07)<0.010.98 (0.94–1.03)0.511.00 (0.95–1.04)0.83
Men2.83 (1.53–5.26)<0.011.26 (0.50–3.14)0.622.04 (0.88–4.75)0.10
Body mass index (kg/m2)0.99 (0.93–1.06)0.75
Type of chest pain<0.0010.230.05
 No anginal complaints1 (–)1 (–)1 (–)
 Aspecific chest pain2.77 (0.34–22.9)11.6 (0.87–154)20.6 (1.56–272)
 Atypical angina5.74 (0.74–44.5)14.3 (1.07–194)28.6 (2.15–381)
 Typical angina16.6 (2.17–127)20.0 (1.11–358)57.9 (3.35–1002)
Pre-test likelihood for CAD<0.0010.120.09
 Low1 (–)1 (–)1 (–)
 Intermediate1.91 (0.71–5.15)0.29 (0.06–1.37)0.22 (0.05–0.96)
 High8.25 (2.96–23.0)0.76 (1.00–6.12)0.41 (0.06–2.98)
CAD risk factors
 Diabetes mellitus (yes/no)1.14 (0.56–2.34)0.72
 Hypertension (yes/no)1.49 (0.85–2.62)0.17
 Smoking history (yes/no)1.51 (0.86–2.66)0.15
 Hypercholesterolaemia (yes/no)1.24 (0.69–2.24)0.47
 Family history of CAD (yes/no)0.62 (0.35–1.09)0.10
Exercise ECG0.20
 Non-ischaemic1 (–)
 Ischaemic2.03 (0.93–4.42)
 Non-diagnostic1.45 (0.65–3.22)
 CAC score1.09 (1.05–1.14)<0.0011.04 (0.98–1.10)0.171.05 (1.00–1.11)0.07
CTCA result<0.001<0.001<0.001
 (Non)obstructive CAD1 (–)1 (–)1 (–)
 Equivocal result9.51 (2.58–35.1)6.27 (1.47–26.7)6.66 (1.59–27.8)
 Obstructive CAD36.7 (11.0–122)18.4 (4.76–71.4)22.9 (5.99–87.9)
Quantitative PET
 Hyperaemic MBF (<2.0 mL/min/g)12.2 (6.39–23.3)<0.0016.08 (2.66–13.9)<0.001
 CFR (<2.0)5.14 (2.85–9.29)<0.0012.74 (1.30–5.76)<0.01
  • Two multivariate analyses were performed either excluding CFR and including hyperaemic MBF (model 1), or including CFR and excluding hyperaemic MBF (model 2).

  • OR, odds ratio; CI, confidence interval; CAD, coronary artery disease; ECG, electrocardiogram; CAC, coronary artery calcium (divided by 100); CTCA, CT coronary angiography; MBF, myocardial blood flow; CFR, coronary flow reserve.


The current study demonstrates that cardiac hybrid 15O-water PET/CTCA imaging provides additional information and influences clinical decision-making, leading to an apparent more judicious treatment strategy with regard to referral for ICA and revascularization procedures. The diagnosis of CAD and decision for revascularization are often based on the findings at conventional coronary angiography. However, ICA is an expensive and invasive procedure, which is associated with a low, but nevertheless important risk of procedure-related complications, including bleeding, coronary artery dissections, embolism, cardiac arrhythmias, myocardial infarction, and even death.23 Therefore, the search for non-invasive modalities for the detection of CAD has been extensive over the last decades. CTCA has emerged as a promising alternative tool for the non-invasive evaluation of coronary artery lumen and diagnosis of CAD and has been consistently shown to have a high negative predictive value of 97–99%, rendering it an excellent tool for the exclusion of CAD.35,11 In our study, only 5% of patients with no (obstructive) CAD on CTCA were referred for ICA, of whom none showed obstructive CAD on ICA. Although it is acknowledged that CTCA is an excellent technique for the exclusion of CAD, the functional relevance of a coronary stenosis cannot be assessed with anatomical imaging only, since there is frequently a discordance between anatomical and functional stenosis severity.6,7,9,10 On the other hand, 16 (9%) patients with no (obstructive) CAD on CTCA exhibited abnormal MPI, indicating that a normal CTCA does not exclude abnormal cardiac perfusion imaging. In the present study, 15O-water was used as a perfusion tracer, which provides the ability to measure MBF over a wide range of flow velocities. The large heterogeneity of perfusion values obtained with 15O-water PET might proof difficult to obtain a optimal cut-off value of high accuracy,18,22 which may result in the observed abnormal MPI results in the absence of obstructive CAD. In addition, the use of CFR, which is dependent on both baseline and hyperaemic MBF, likely contributes to these observations as a reduction in CFR might reflect adequate hyperaemic perfusion in the presence of high baseline values. However, these discrepant findings have been previously documented by Schenker et al.,24 and likely reflect, next to false positive imaging results, the presence of myocardial ischaemia due to diffuse ‘non-significant’ atherosclerotic epicardial CAD and/or microvasular dysfunction.25,26 Indeed, myocardial perfusion is governed by both the presence of an epicardial stenosis and microvascular resistance, which is related to patient characteristics such as age, gender, and CAD risk factors.22,27 Diffuse non-obstructive CAD may also have a negative impact on MBF,28 which may explain the number of patients with both abnormal CT and PET results in the absence of an obstructive epicardial stenosis on ICA. All in all, these discrepant findings reflect two distinct disease entities, namely (subclinical) atherosclerosis and ischaemia as the result of focal CAD that both require different treatment strategies.

In daily clinical practice, physicians are often faced with several factors such as patients body size, poor contrast bolus, blooming and motion artefacts, and borderline appearing stenoses that may result in equivocal CT results and subsequently repeat visits of patients for further (functional) testing. Indeed, 21% of the current study population exhibited an equivocal CTCA study in whom 70% PET imaging revealed normal perfusion. Interestingly, none of these patients had the clinical necessity to be referred for a revascularization procedure, compared with 42% in the group with abnormal PET perfusion imaging. In addition, hybrid PET/CTCA imaging also facilitates patient management in the presence of an obstructive lesion on CTCA. In line with previous studies, roughly half of CTCA deemed obstructive lesions were haemodynamically relevant.2,6 Although patients were not routinely referred for ICA, the yield of CAD by ICA after hybrid PET/CTCA imaging was 59% in patients with an equivocal CT study and abnormal MPI, and 69% in whom hybrid PET/CTCA identified obstructive stenosis and abnormal perfusion, which is favourably compared with the relatively low yield of 38% reported in a large clinical cohort.29 Consequently, a revascularization rate per angiogram of 59% was seen in patients with an equivocal CT and abnormal MPI, compared with no revascularizations in those with normal PET perfusion imaging. In addition, the revascularization rate per invasive coronary angiogram was 65 and 26% in patients who displayed obstructive CAD on CTCA in combination with abnormal and normal MPI, respectively. When compared with the results of the SPARC trial,30 referral rates for ICA in the current study are substantially higher, indicating that abnormal hybrid imaging findings compared with stand-alone pathological imaging results encourage physicians to refer patients to the catheterization laboratory. Although this indicates the adequate discriminatory value of MPI for revascularization when faced with a positive CTCA, it also indicated that a number of patients were falsely classified as non-significant CAD based on a normal MPI who subsequently did underwent revascularization. Fortunately, in absolute numbers this comprised only a minority of the entire study population [8 of 375 (2%) patients]. Not surprisingly, PET/CTCA results were strong predictors of referral for ICA and subsequent revascularizations. Furthermore, hyperaemic MBF is a stronger predictor than the CFR for both referral for ICA and revascularizations, which could be attributed to the dependency of the CFR on baseline perfusion values hampering its predictive capacity. Interestingly, the exercise electrocardiogram (ECG) results were not predictive of referral rates for ICA and downstream revascularizations. Although the exercise ECG is a widely used baseline tool for the assessment of myocardial ischaemia, its diagnostic performance for detecting obstructive CAD is only moderate.31 All in all, these findings suggest that the hybrid evaluation of CAD facilitates clinical decision-making, particularly when compared with the revascularization rate of 38% per invasive angiogram, which has been reported in a large European registry study.32 Similar results have been demonstrated with hybrid single photon emission computed tomography (SPECT)/CTCA and cadmium–zinc–telluride/CTCA devices.12,13 However, in contrast to the current study, these previous studies did not provide a detailed comparison with equivocal CT results. Thus, the current findings extend and corroborate previous work in a larger clinical cohort of patients by the inclusion of equivocal CT test results, reflecting daily clinical practice more closely. Nevertheless, a few limitations need to be mentioned. First, the current study describes the rate of downstream utilization of ICA and revascularization after cardiac PET/CTCA in a retrospective manner, which could have biased the current findings. Obviously, the decision to perform revascularization was to an important extent influenced by the results of the imaging study. For example, some patients might unjustly have been denied revascularization based on a negative MPI and vice versa. Furthermore, there is no control group that underwent only stand-alone imaging to confirm the incremental value of hybrid imaging with regard to clinical decision-making. Another limitation of the present study is the lack of information regarding the long-term effects of the combined PET/CTCA imaging protocol on patient care. The long-term impact of hybrid imaging protocols on patient management, including referrals to a catheterization lab and the reduction of the cost incurred by the protocol, requires further prospective study. Furthermore, myocardial perfusion was assessed using a 15O-water PET/CTCA protocol that necessitated the use of an on-site cyclotron and a PET/CT camera, both of which are not as easily available as the cameras involved in SPECT/CT imaging. However, 15O-water PET imaging has specific advantages over SPECT imaging, such as a high diagnostic accuracy, a low radiation dose of 2 mSv for a rest-stress 15O-water PET imaging protocol, and an imaging time of only 45 min for a combined rest-stress 15O-water PET/CTCA protocol.5,11 Furthermore, for subgroup analyses, the studied patient population was too small to draw definite conclusions. Finally, the current study population consisted of patients with predominantly an intermediate likelihood for CAD and normal cardiac function. These results might therefore not be generalized to different patient populations with previous myocardial infarction, revascularization procedures, and/or heart failure.


The combination of MPI and assessment of coronary anatomy by hybrid 15O-water PET/CTCA imaging impacts referral for ICA and downstream revascularization. Particularly, in the presence of an equivocal or abnormal CTCA, MPI acts as an arbiter to guide a judicious referral to the catheterization laboratory and revascularization strategy.


Conflict of interest: none declared.


We thank Suzette van Balen, Amina Elouahmani, Judith van Es, Robin Hemminga, Femke Jongsma, Nghi Pham, and Nasserah Sais for performing the scans; Henri Greuter, Marissa Rongen, Robert Schuit, and Kevin Takkenkamp for producing 15O-water.


View Abstract