. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Added value of hybrid myocardial perfusionSPECT and CT coronary angiography in thediagnosis of coronary artery diseaseJeroen Schaap1,2*, Joris A.H. de Groot3, Koen Nieman4,5, W. Bob Meijboom4,5,S. Matthijs Boekholdt6, Robert M. Kauling1, Martijn C. Post1, Jan A. Van der Heyden1,Thom L. de Kroon7, Benno J.W.M. Rensing1, Karel G.M. Moons3,and J. Fred Verzijlbergen2,8

1Department of Cardiology, Amphia Hospital, PO Box 90158, 4800 RK, Breda, The Netherlands; 2Department of Nuclear Medicine, St. Antonius Hospital, Nieuwegein, The Netherlands;3Julius Center for Health Sciences and Primary Care, UMC Utrecht, Utrecht, The Netherlands; 4Department of Cardiology, Erasmus Medical Center, Rotterdam, The Netherlands;5DepartmentofRadiology, ErasmusMedicalCenter,Rotterdam,The Netherlands; 6DepartmentofCardiology,AcademicMedicalCenter,Amsterdam,The Netherlands; 7Department ofCardio-Thoracic Surgery, St. Antonius Hospital, Nieuwegein, The Netherlands; and 8Department of Nuclear Medicine, Erasmus Medical Center, Rotterdam, The Netherlands

Received 10 March 2014; accepted after revision 17 June 2014; online publish-ahead-of-print 29 July 2014

Aims Hybrid single photon emission computed tomography (SPECT)/coronary computed tomography angiography (CCTA)has only been evaluated for its diagnostic accuracy as a single test in patients suspected of significant coronary arterydisease (CAD). Added value of hybrid SPECT/CCTA beyond usual clinical work-up, or use of each of these tests separ-ately, remains unclear. We evaluated the added value of hybrid myocardial perfusion SPECT (SPECT) and CCTA, beyondpre-test likelihood and exercise stress ECG (X-ECG), in the diagnosis of CAD.

Methodsand results

Two hundred and five patients with stable angina pectoris and intermediate-to-high pre-test likelihood were prospect-ively included. All patients underwent clinical history and examination, X-ECG, stress and rest SPECT, coronary calciumscoring (CCS) and CCTA. Fractional flow reserve measurement ,0.80 or a lesion .50% on coronary angiography (CA)served as reference standard for significant CAD. Multiple imputation was used to correct for missing test results(17–20%). Added value of hybrid SPECT/CCTA to the basic model of pre-test likelihood plus X-ECG was quantifiedusing logistic regression analysis. Model differences were then assessed using differences in C-index and in net reclassifi-cation improvement (NRI). The basic model had a C-index of 0.73 (95%CI 0.66–0.80). This significantly increased to 0.85(95%CI 0.80–0.91) by addition of only SPECT, to 0.90 (95%CI 0.85–0.94) when adding only CCTA, and to 0.96 (95%CI0.92–0.99) when adding hybrid SPECT/CCTA. The accompanying NRIs were 0.82 (95%CI 0.62–1.02), 0.86 (95%CI0.66–1.06) and 1.57 (95%CI 1.11–1.59) respectively.

Conclusion Current analysis resembles clinical routine of layered testing and shows that hybrid SPECT/CCTA imaging has a substan-tially higher yield than standalone SPECT or CCTA in the diagnostic workup of patients suspected of significant CAD.

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Keywords Added value † Hybrid imaging † Myocardial perfusion SPECT † Coronary computed tomography angiography

IntroductionPatients with a high pre-test likelihood of significant coronary arterydisease (CAD) have a class 1A indication for invasive coronary angi-ography (CA).1 However, only 36% of patients referred for CA proveto suffer from significant CAD.2 A reliable non-invasive imaging

technique that accurately discriminates between the presence orabsence of significant CAD in these patients would be of great value.

The work-up of patients with anginal complaints in many out-patient clinics routinely starts with exercise stress electrocardio-graphy (X-ECG).3 Several reports describe the use of coronarycalcium scoring (CCS) for diagnostic purposes in symptomatic

* Corresponding author. Tel: +31 76 5955593; fax: +31 765954745. E-mail: jeroenschaap01@gmail.com

Published on behalf of the European Society of Cardiology. All rights reserved. & The Author 2014. For permissions please email: journals.permissions@oup.com.

European Heart Journal – Cardiovascular Imaging (2014) 15, 1281–1288doi:10.1093/ehjci/jeu135

by guest on October 30, 2014

Dow

nloaded from

patients.4,5 However, current guidelines do not recommend usingthese tests in patients with intermediate-to-high pre-test likelihood(based on gender, age, clinical history and examination).1,6 –8 Assuch, many patients are referred for further non-invasive imaging todefer from CA or to exclude high risk CAD.1,3,9 From all non-invasiveimaging modalities, including myocardial perfusion single photonemission computed tomography (SPECT), positron emission tomog-raphy (PET), CT coronary angiography (CCTA), stress echocardiog-raphy and cardiac MRI, hybrid SPECT or PET/CCTA probably havethe highest diagnostic yield. The non-invasive evidence of absenceor presence of ischaemia on SPECT or PET as a result of a .50% cor-onary luminal loss on CCTA resulted in sensitivity and specificityvalues of 94–96% and 92–100% for the diagnosis of significantCAD.10–15

However, hybrid SPECT/CCTA has only been evaluated for itsdiagnostic accuracy as a single test. But in the routine work-up ofpatients with angina pectoris not only the result of the last imagingprocedure is taken into account, also clinical history, examinationand the results of other tests are always considered to decidewhether or not to perform CA.16 As such, the single test accuracyestimates do not indicate whether pre-test likelihood and X-ECGeither in combination with standalone SPECT, with standaloneCCTA, or with hybrid SPECT/CCTA is the best combination ofdiagnostic procedures in these patients.

We therefore performed a study to evaluate the added value ofhybrid SPECT/CCTA imaging beyond pre-test likelihood, X-ECG,and standalone SPECT or CCTA in patients with an intermediate-to-high pre-test likelihood of significant CAD.

Methods

Patient populationWe prospectively conducted a single centre cross-sectional study, inNieuwegein, the Netherlands. Patients with chest pain symptoms andan intermediate to high pre-test likelihood of CAD were consecutivelyrecruited for this study from September 2010 until December 2011.History was taken by physicians and chest pain was categorized accordingto traditional criteria.3,17–19 Pre-test likelihood was calculated according

to Diamond and Forrester criteria.17 The boundaries for low, intermedi-ate and high pre-test likelihood were set at ,20, 20–80, and .80%respectively.1,17 Exclusion criteria were history of surgical revasculariza-tion (CABG), stent implantation (PCI), unstable cardiac condition,cardiac rhythm other than sinus rhythm, known severe heart failure orvalvular heart disease, (possible) pregnancy, contraindication to receiveiodinated contrast agents andpulmonary conditionprohibiting successfulSPECTorCCTA imaging. The study conformed to theprinciples outlinedin the declaration of Helsinki and was approved by the local ethics com-mittee. Written informed consent was obtained from all patients. Fur-thermore, the study was conducted and reported in accord withSTARD guidelines.20

Diagnostic work-upAt day one of the diagnostic work-up all patients underwent X-ECG orpharmacological stress preceding hybrid stress SPECT and CCTA.Within 14 days, on the second day of the diagnostic work-up, restSPECT acquisition and CA with fractional flow reserve (FFR) measure-ments were scheduled (Figure 1). CA and FFR measurements served asthe reference test for significant CAD (see below). The interpretationof all diagnostic tests occurred blinded for all other modalities. Physicianswere, however, aware of gender, age, and clinical history during each testinterpretation.21

Prior to the first dayof the diagnostic work-up all patients stopped ratelimiting or anti-anginal medications. Stress SPECT, CCS, and CCTA wereacquired on a Hybrid SPECT-CT system, CardioMD gamma camera andBrilliance 64-slice CT scanner (Philips Medical Systems, Best, the Nether-lands). The components share a common table that is aligned to allowsequential image acquisition. Either X-ECG or pharmacological stress(adenosine at standard rate of 0.14 mg/kg/min over 6 min) were per-formed prior to stress SPECT acquisition. Inability to perform bicyclestress testing or failure to reach 85% of the predicted maximum heartrate was reason to switch to pharmacological stress. Directly followingstress SPECTacquisitionCCS andCCTA imagingwasperformed.All par-ticipants presenting with baseline heart rates over 60 bpm received100 mg oral metoprolol tartrate at the start of the stress test precedingSPECT imaging. Patients with an initial heart rate below 60 bpm received50 mg metoprolol tartrate. Pharmacokinetics of oral metoprolol do notallow for an immediate effect that would interfere with X-ECG acquisi-tion. Preceding CCTA additional intravenous metoprolol was adminis-tered to achieve a heart rate below 62 bpm. In patients with a normal

Figure 1: Sequence of diagnostic work-up. X-ECG, exercise stress electrocardiography. SPECT, myocardial perfusion SPECT. CCS, coronarycalcium score. CCTA, CT coronary angiography. CA, invasive coronary angiography.

J. Schaap et al.1282

by guest on October 30, 2014

Dow

nloaded from

stress SPECT (normal perfusion and gated data) no rest SPECT wasacquired. Inability to perform acquisition of prospectively triggeredCCTA (heart rate .62–64 bpm) was reason to exclude patients(N ¼ 31, 15.1%) from the study. For radiation dose saving purposes thestudy protocol did not allow for retrospectively gated CCTA. Mean radi-ation dose for SPECT was 6.3+2.5 mSv and mean radiation dose forCCTA was 4.1+2.2 mSv. CA and FFR measurements were performedin all patients with a completed non-invasive imaging sequence, includingthose with normal non-invasive test results to refrain from any selectionbias. During FFR measurement, distal coronary pressure was measuredwith a pressure guidewire (Certus Wire, Radi Medical Systems,Uppsala, Sweden) during maximal hyperaemia induced by intravenousadenosine. Mean radiation dose for CA was 10.5+4.8 mSv. No compli-cations occurred during non-invasive imaging, CA was complicated byone (0.6%) anaphylactic reaction requiring the administration of clemas-tine and hydrocortisone.

Index testsExercise stress electrocardiographyBicycle stress electrocardiography was performed according to a previ-ously described stepwise protocol.22 All X-ECG results were reviewedin consensus by two experienced physicians blinded for all other diagnos-tic tests. An abnormal X-ECG was defined as ≥0.1 mV horizontal shift ofthe ST segment at 80 ms after the J-point in three consecutive beats.22,23

Non-conclusive X-ECG was defined as typical angina during stress,.30 mmHg decrease in systolic blood pressure, uninterpretable ECGas a result of an unstable baseline or inability to reach at least 85% ofthe predicted heart rate without electrocardiographic evidence ofischaemia. Normal X-ECG wasdefined as the absence of anyof the afore-mentioned criteria. The predicted heart rate was corrected for length,age, and sex, making use of a standardized table.22

Myocardial perfusion SPECTSPECT imaging was performed as previously described using a weight-adjusteddose of400–600 MBqof 99mTc-sestamibi.6 Attenuationcorrec-tion was performed for all images using a non-enhanced CT scan. Stressand rest SPECT images were interpreted in consensus by two experi-enced nuclear medicine physicians blinded for all other diagnostic testsas previously described.6,24,25 In brief, summed stress score andsummed rest score were calculated after scoring for perfusion defectsaccording to a five-point scoring system (0 ¼ normal to 4 ¼ absenceof tracer uptake) and a 17-segment model. Gated data and semi quanti-tative perfusion data were assessed using the QGS/QPS softwarepackage (Cedars-Sinai Medical Centre, Los Angeles, CA, USA). For thedefinition of normal, non-conclusive and abnormal myocardial perfusionon SPECTaccording to segmental scoresweused the definition providedby Abidov et al.24 Only the definition these authors provided for anequivocal SPECT was used to define a SPECT non-conclusive. SPECTwith a non-conclusive perfusion score was regarded abnormal.

Coronary computed tomography angiographyCCTA was acquired with a prospectively ECG-triggered scan mode aspreviously described and according to the guidelines provided by thesociety of cardiovascular computed angiography.26 Prior to CCTA, anon-enhanced scan to calculate the CCS and determine scan lengthwas performed. CCTAs were read in consensus by two experiencedcardiologists blinded for all other diagnostic tests. They categorizedobstructive severity according to a 16-segment model and five categoriesof luminal loss: no CAD (0% area stenosis), mild lesion severity (0–50%area stenosis), intermediate lesion severity (50–70% area stenosis),severe lesions (70–99% area stenosis), and total occlusion. All unevalu-able segments for the category of luminal loss (e.g. due to motion artefact

or severe calcifications) were categorized as non-conclusive. Thenumber of CCTAs non-conclusive due to calcifications was 26 (16.0%).Subsequently a vessel-based analysis was performed from which a diag-nosis on a per-patient level was determined. Vessels were categorizedas abnormal if at least one segment had a percent diameter stenosis.50%. All non-conclusive segments were defined abnormal.

Hybrid SPECT/CCTA imaging analysisAccording to our pre-defined protocol all patients were included inhybrid imaging evaluation. Interpretation was performed in consensusby two independent experienced physicians blinded for X-ECG, CA,and FFR measurements. Vessel-based analysis was performed fromwhich a diagnosis on a per-patient level was determined. The interpret-ation hybrid SPECT/CCTA images led to the categorization of allvessels and patients as normal or abnormal, as we previously describedin detail.13 In brief, concordant normal or abnormal findings by SPECTand CCTA led to a similar interpretation of the hybrid imaging result.In patients with a normal CCTA and a discordant abnormal or non-conclusive SPECT, the hybrid imaging result was defined normal. Inpatients with a non-conclusive CCTA, the normal or abnormal SPECTresultwasdecisive for the hybrid imaging result. In patients with an abnor-malCCTAandadiscordantnormal SPECT,CCTAprevailedonly in thosepatients with a definitely abnormal CCTA. In patients with an abnormalCCTA and a discordant non-conclusive SPECT, the hybrid imagingresult was defined abnormal.27

Reference test: invasive coronary angiographyA FFR measurement ,0.80 was defined abnormal and served as the ref-erence test for significant CAD. FFR measurements were acquired in 104patients (61%) of patients that underwent CA. All other angiograms wereinterpreted in consensus by two experienced cardiologists for obstruct-ive severityaccording toa16-segmentmodel similar to the interpretationof CCTA. Interpretation occurred blinded for all other diagnostic modal-ities. They were aware of age, gender, and clinical history.21 Lesions with.50% obstructive disease were considered abnormal (significant CAD).Vessel-based analysis was performed from which diagnostic accuracy ona per-patient level was determined.

CAwas acquired on Allura catheterization equipment (Philips MedicalSystems, Best, the Netherlands) via femoral or radial artery access.Biplane views were acquired from all major coronary arteries. FFR mea-surements were performed as previously described.28

Statistical analysisFirst, in accordance with recent methodological recommendations, allmissing test results in our study were multiply (10 times) imputed usingregression techniques.29 –31 Then four different logistic regressionmodels were fitted in each of the 10 individual imputed datasets and sub-sequently combined across imputations using Rubin’s Rule,32,33 resultingin four ‘combined’ models. These ‘combined’ models were then used tocalculate the added value measures [i.e. Area under the receiver operat-ing characteristic curve and net reclassification improvement (NRI)] foreach model. To resemble routine clinical work-up34,35 we first fittedthe basic logistic regression model (Model 1) including the pre-test like-lihood of CAD plus X-ECG as independent diagnostic predictors (indextests) and CADpresence (yes/no)based on CAas outcome. The model’sC-index or area under the receiver operating characteristics curve (Rocarea) was computed. The added value of additional imaging tests (i.e.SPECT, CCTA, and hybrid SPECT/CCTA) beyond the basic model wasquantified by consecutively extending the basic model with these testsand assessing the increase in C-index. The models of interest areshown in Table 1. We included CCS in the model evaluating CCTA and

Added value of hybrid myocardial perfusion SPECT and CT coronary angiography 1283

by guest on October 30, 2014

Dow

nloaded from

Hybrid SPECT/CCTA, since acquisition of CCS is commonly performedprior to CCTA and Hybrid SPECT/CCTA. Furthermore, the number ofpatients correctly reclassified after adding the various imaging modalitiesto the basic model, was expressed using the NRI. This measure, summar-izing the reclassification table, is calculated by subtracting the proportionof subjects classified worse by the new model from the proportion ofpatients that is better classified by the new model. As cut-off values forthe reclassification table and related measures, we used three commonCAD probability categories; low (probability ,20%), intermediate(20–80%), and high (.80%) for CAD.1,16,36 Data were analysed usingR (Version 2.13.1).37

Results

Study populationTwo hundred and five adult patients were included in the study.Patient characteristics are displayed in Table 2.

Diagnostic value of pre-test likelihoodand exercise stress electrocardiography(Model 1)In the basic multivariable model for detecting CAD the independentvariables were pre-test likelihood and X-ECG. Pre-test likelihoodalone, used as a continuous variable, had a C-index of 0.67 (0.59–0.74). The combination of pre-test likelihood and X-ECG (Model 1;Table 1) resulted in a C-index of 0.73 (95% CI; 0.66–0.80) (Figure 2).

Added diagnostic value of SPECT, CCTA,and Hybrid SPECT/CCTA (Models 2–4)Adding further imaging techniques, i.e. SPECT, CCTA, and HybridSPECT/CCTA (Models 2–4; Table 1) to the basic model (Model 1;Table 1) resulted in a significant increase in C-index from 0.73(95%CI; 0.66–0.80) to 0.85 (0.80–0.91), 0.90 (0.85–0.94), and0.96 (0.92–0.99), respectively (Figure 2).

Reclassification by Hybrid SPECT/CCTAUsing the basic model (Model 1), the estimated probability of CADwas calculated for each patient. Of the patients with a low estimatedprobability of CAD (,20%, n ¼ 16), 13 did not have significant CAD

(negative predictive value; NPV 81%) (Table 3). In those with a highpredicted probability (.80%, n ¼ 13), eight had significant CAD(positive predictive value; PPV 62%), Table 3. Adding HybridSPECT/CCTA (Model 4) increased the number of patients with esti-mated low probability from 16 to 107, of whom 101 did not have sig-nificant CAD (NPV 94%). The number of high probability patientsincreased from 13 to 97, of whom 91 indeed had significant CAD(PPV 94%). Addition of Hybrid SPECT/CCTA did not change the esti-mated probability class (low, intermediate, or high) in 22 patients(11%). Hybrid SPECT/CCTA correctly increased the estimated diag-nostic probability in 83 patients with CAD, and 89 patients withoutCAD were correctly reclassified to lower probability groups. Fivepatients with CAD were incorrectly reclassified to lower probabilitygroups and six patients without CAD were incorrectly reclassified tohigher probability groups (Table 3). The imaging characteristics ofthese patients who were incorrectly reclassified, were a very challen-ging CCTA (CCS . 800 or motion artifacts) in combination witheither a false positive or false negative SPECT (n ¼ 5), discordantSPECT and CCTA findings (n ¼ 3), or an unavailable reference test(n ¼ 3). The NRI for the addition of hybrid SPECT/CCTA wastherefore 1.57 (95%CI; 1.31–1.83) (Table 3). The NRI for theaddition of SPECT or CCS + CCTA was 0.82 (0.62–1.02) and 0.86(0.66–1.06), respectively.

DiscussionThis study shows that in the clinical work-up of patients with anintermediate-to-high pre-test likelihood, addition of hybrid SPECT/CCTA imaging analysis yields the most accurate diagnosis of the pres-ence or absence of significant CAD (C-index 0.96). Further, additionof hybrid SPECT/CCTA to pre-test information and X-ECG betterreclassifies patients when compared with addition of only SPECTor CCTA (NRI of 1.57 vs. 0.82 and 0.86 respectively).

Non-invasive tests beyond pre-test likelihoods are needed to cor-rectly reclassify patients from high probability to low probability andthus correctly defer them from CA, since the clinical reality is thatonly 36% of patients referred for CA prove to suffer from significantCAD.1,2 In the subset of patients with an intermediate-to-highpre-test likelihood of CAD, the routinely performed X-ECG orCCS alone are not sufficient to provide adequate discrimination

Table 1 Models of interest

Model 1: Probability of coronary artery disease (basic modela) ¼

1/(1 + exp-(22.496 + 0.026 × pre-test likelihood + 1.014 × X-ECG))

Model 2: Probability of coronary artery disease after adding SPECT to Model 1 ¼

1/(1 + exp-(22.186 + 0.008 × pre-test likelihood + 0.800 × X-ECG + 2.2628 × SPECT))

Model 3: Probability of coronary artery disease after CCS and CCTA to Model 1 5

1/(1 + exp-(23.989 + 0.012 × pre-test likelihood + 1.098 × X-ECG + 0.001 × CCS + 3.3160 × CCTA))

Model 4: Probability of coronary artery disease after adding CCS and hybrid SPECT/CCTA to Model 1 ¼

1/(1 + exp-(22.450 – 0.010 × pre-test likelihood + 0.555 × X-ECG + 0.001 × CCS + 5.393 × hybrid SPECT/CCTA))

This table reports the ‘combined’ logistic regression Models 1–4, which were first fitted in each of the 10 individual imputed datasets and subsequently combined across imputationsusing Rubin’s Rule. Pre-test likelihood was used as a continuous variable in all models.X-ECG, exercise stress electrocardiography; SPECT, myocardial perfusion SPECT; CCS, coronary calcium score; CCTA, CT coronary angiography.aBasic model ¼ pre-test likelihood + X-ECG.

J. Schaap et al.1284

by guest on October 30, 2014

Dow

nloaded from

between patients with and without CAD.6,7,38 This is also demon-strated in the current study (C-index of 0.73 and 0.75, respectively,vs. pre-test likelihood alone with a C-index of 0.67. Performance ofstandalone SPECT or CCTA in addition to the aforementionedtests seems much more accurate (C-index of 0.85 and 0.90, respect-ively). However, available literature describes a limited diagnostic ac-curacy of SPECT or CCTA in high pre-test likelihood patients.39 –41

The high negative predictive value of CCTA could be challengeddue to increased amount of coronary calcifications in high pre-test

Figure 2: Areas under the receiver operator characteristicscurves for all diagnostic models. Model 1 (basic model) ¼ Pre-test likelihood + X-ECG, Model 2 ¼ basic model + SPECT,Model 3 ¼ basic model + CCTA, Model 4 ¼ basic model +hybrid SPECT/CCTA. X-ECG, exercise stress electrocardio-graphy. SPECT, myocardial perfusion SPECT. CCTA, CT coronaryangiography.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 2 Characteristics of the patients included in theanalysis, N 5 205

Patient characteristics N (%) Missings(%)

Age (years)a 62.7+9.7 0 (0)

Gender, men/women 127 (62)/78 (38) 0 (0)

Risk factors 0 (0)

Smoking

Current 44 (21)

Past 35 (17)

Hypertension 134 (65)

Diabetes mellitus 35 (17)

Dyslipidaemia 130 (63)

Family history of prematureatherosclerosis

120 (59)

Risk factors ≥3 105 (51)

Medicationb 0 (0)

Acetylsalicylic acid 151 (74)

Betablocker 120 (59)

Calcium channel blocker 43 (21)

Nitrate 36 (18)

Statin 137 (67)

ACE-i/ARB 85 (42)

Diuretic 58 (28)

Angina pectoris 0 (0)

Non-anginal 1 (1)

Atypical 67 (33)

Typical 137 (67)

CaCS Angina class 0 (0)

Class I 14 (7)

Class II 172 (84)

Class III 19 (9)

Class IV 0 (0)

Pre-test likelihoodc 0.87 (0.22–0.95) 0 (0)

Low 0 (0)

Intermediate 90 (44)

High 103 (50)

Unknownd 12 (6)

Exercise stress electrocardiography(N ¼ 158)c

37 (18)d

Normal 119 (58)

Abnormal 43 (21)

Non-conclusive 6 (3)

Coronary calcium score (N ¼ 204)a 536+800 1 (1)

Myocardial perfusion SPECT (N ¼ 204) 1 (1)

Normal 115 (56)

Abnormal 89 (43)

CT coronary angiography (N ¼ 163) 41 (20)

Normal 69 (34)

Abnormal 95 (46)

Hybrid SPECT/CCTA (N ¼ 163) 41 (20)

Normal 91 (44)

Abnormal 73 (36)

Continued

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table2 Continued

Patient characteristics N (%) Missings(%)

Invasive coronary angiography (N¼ 171)f 34 (17)

Normal 85 (41)

Abnormal 86 (42)

Values are shown as number and percentage unless otherwise noted.ACE-i, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker;CaCS, Canadian Cardiovascular Society.aValues expressed as mean+ standard deviation.bMedication use at time of inclusion until conclusion of the complete diagnosticwork-up.cValues expressed as median (range).dPatients with atypical angina pectoris and .70 years old.eAdenosine stress was performed in 37 (18%) patients prior to stress SPECTacquisition.fOf 7 patients without a completed non-invasive imaging sequence invasive coronaryangiography was acquired with a clinical indication.

Added value of hybrid myocardial perfusion SPECT and CT coronary angiography 1285

by guest on October 30, 2014

Dow

nloaded from

likelihood patients.42,43 Diagnostic performance of SPECT in thesepatients is known to be limited, due to a given amount of non-conclusive SPECT results due to small perfusion defects, inadequateheart rate response or discrepant clinical data (anginal complaintsduring stress or positive X-ECG result).44 The advantage of hybridimaging lies in the compensation for the weak properties of CCTAand SPECT, and the dissolution of non-conclusive results of both asstandalone imaging procedures. As such, to overcome the diagnosticuncertainty in high pre-test likelihood patients and in patients withnon-conclusive imaging results, hybrid SPECT/CCTA has beenproposed.10,27,45

The yet available studies show that when used in isolation this testhas a very promising diagnostic value with a sensitivity 89–99% andspecificity of 67–95%.10 –12 The current study not only confirmsthis finding, but more importantly proofs that hybrid SPECT/CCTAhas added value over standalone SPECT or CCTA in a highpre-test likelihood population. As such hybrid imaging is specificallyrelevant for the subset of patients with continued diagnostic uncer-tainty despite conclusive imaging results or patients with non-conclusive imaging results per se. And it has the potential to guidedecisions on the necessity of revascularization.46

Several limitations to our study should be acknowledged. Recentreports have shown that the model we used to calculate pre-test like-lihood CAD (Diamond and Forrester) heavily overestimates thepresence of CAD.47,48 The presence of a non-sinusal heart rhythmwas an exclusion criterion for our study; in clinical practice this

could constitute a considerable number of patients. Hybrid SPECT/CCTA imaging was scheduled in all patients. However for reasonsof cost-effectiveness and radiation dose saving purposes hybridimaging should not be the primary imaging strategy for all patientspresenting with chest pain symptoms. Also, additional data areneeded to appreciate the benefit of hybrid imaging weighed againstits cost. We had missing test results and are well aware of the factthat missing data rarely occur completely at random and may there-fore hamper the results. However, the number of missing data waslimited and we used multiple imputations to alleviate the problemof missing test results. Three (1.5%) patients had non-conclusiveSPECT results and no significant area stenosis in epicardial coronaryarteries on CCTA, but did prove to suffer from significant CAD.These patients could suffer from microvascular disease whichcannot be completely appreciated by SPECT since it is unable toquantify absolute myocardial perfusion. We chose a FFR cut-offlevel of ,0.80, as the reference standard for significant CAD.However, FFR is not a direct measure of low-flow ischaemia anddoes not reflect absolute perfusion or coronary flow reservewhich are determinants of ischaemia.49 FFR measurements wereavailable in 61% of patients that underwent CA, whereas for the re-mainder we used an anatomical reference test (.50% obstructivedisease as diagnostic for significant CAD), which is known to be im-perfect. According to available literature it is safe to assume thatlesions with ,50% stenosis have no hemodynamic significance.50

FFR measurements were available from 78% of patients with angio-graphically 50–70% and 68% of patients with 70–95% obstructivedisease. We chose a boundary of 20–80% for an intermediate prob-ability of significant CAD; this is supported by available literature.1

However, boundaries of 10–90% or 15–85% are also commonlyused. According to our analysis a boundary of 10–90% would leave24 patients (12%) and a boundary of 15–85% would leave 1 patient(0.5%) with and without CAD in the intermediate group afterhybrid SPECT/CCTA imaging analysis vs. one patient (0.5%) if aboundary of 20–80% was used for intermediate probability. TheNRI of 1.57 for the addition of hybrid SPECT/CCTA is indeed veryhigh. In part, this is explained by the fact that these models includeimaging modalities that evaluate the same characteristics of coronaryartery disease. Also, this high value occurs as a result of the chosenprobability thresholds of 20 and 80%, 176 of 205 patients (86%)are estimated to have an intermediate probability using the basicmodel (pre-test likelihood and X-ECG). The extended model is farbetter in predicting low probability (,20%) and high probability(.80%) patients, thus leading to a high NRI. Healthcare policiesadvocate a stepwise imaging protocol in the diagnostic work-up ofpatients, for radiation dose saving purposes and a cost effective useof diagnostic modalities. As such, a hybrid imaging approach shouldnot be the primary imaging strategy for all patients presenting withangina pectoris.

ConclusionsOur study is the first to demonstrate the added diagnostic value of thecomplementary nature of hybrid SPECT/CCTA imaging. With an ana-lysis resembling clinical routine of layered testing, the impressive diag-nostic accuracyofhybrid SPECT/CCTA beyondX-ECG and SPECTorCCTA is demonstrated in intermediate-to-high pre-test likelihood

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 3 Reclassification table according to low (<20%),intermediate (20–80%), and high probability (>80%)groups: basic model (Model 1) vs. hybrid SPECT/CCTA(Model 4)

Model 4: basicmodela 1 hybridSPECT/CCTA

Model 1: basic modela

Probability group for CAD <20% 20–80% >80% Total

Patients with CAD

,20% 1 0 2 3

20–80% 5 0 81 86

.80% 0 0 8 8

Total 6 0 91 97

Patients without CAD

,20% 13 0 0 13

20–80% 84 0 6 90

.80% 4 1 0 5

Total 101 1 6 108

Reclassification improvement is 0.80 among patients with CAD (83–5 of 97) and0.77 among patients without CAD (89–6 of 108), resulting in a net reclassificationimprovement of 0.80 + 0.77 ¼ 1.57 (95% CI; 1.11–1.59).Italicized numbers are patients classified in agreement according to the model withand without hybrid SPECT-CT.X-ECG, exercise stress electrocardiography; SPECT, myocardial perfusion SPECT;CCS, coronary calcium score; CCTA, CT coronary angiography; CAD, coronaryartery disease.aBasic model ¼ pre-test likelihood + X-ECG.

J. Schaap et al.1286

by guest on October 30, 2014

Dow

nloaded from

patients. We show that these patients can be accurately reclassifiedinto low- or high probability groups of CAD, as such deferring themfrom CA and probably indicate revascularization respectively.

Conflict of interest: None declared.

References1. Wijns W, Kolh P, Danchin N, Di MC, Falk V, Folliguet T et al. Guidelines on myocar-

dial revascularization: The Task Force on Myocardial Revascularization of theEuropean Society of Cardiology (ESC) and the European Association for Cardio-Thoracic Surgery (EACTS). Eur Heart J 2010;31:2501–55.

2. Patel MR, Peterson ED,Dai D, Brennan JM, RedbergRF, Anderson HVet al. Low diag-nostic yield of elective coronary angiography. N Engl J Med 2010;362:886–95.

3. FoxK, GarciaMA, Ardissino D, Buszman P, Camici PG, Crea F et al. Guidelines on themanagement of stable angina pectoris: executive summary: The Task Force onthe Management of Stable Angina Pectoris of the European Society of Cardiology.Eur Heart J 2006;27:1341–81.

4. Oudkerk M, Stillman AE, Halliburton SS, Kalender WA, Mohlenkamp S,McCollough CH et al. Coronary artery calcium screening: current status and recom-mendations from the European Society of Cardiac Radiology and North AmericanSociety for Cardiovascular Imaging. Int J Cardiovasc Imaging 2008;24:645–71.

5. Budoff MJ, Achenbach S, Blumenthal RS, Carr JJ, Goldin JG, Greenland P et al. Assess-ment of coronary artery disease by cardiac computed tomography: a scientific state-ment from the American Heart Association Committee on Cardiovascular Imagingand Intervention, Council on Cardiovascular Radiology and Intervention, and Com-mittee on Cardiac Imaging, Council on Clinical Cardiology. Circulation 2006;114:1761–91.

6. Schaap J, Kauling RM, Boekholdt SM, Post MC, Van der Heyden JA, de Kroon TL et al.Usefulness of coronary calcium scoring to myocardial perfusion SPECT in the diag-nosis of coronary artery disease in a predominantly high risk population. Int J Cardi-ovasc Imaging 2013;29:677–84.

7. Gottlieb I, Miller JM, Arbab-Zadeh A, Dewey M, Clouse ME, Sara L et al. The absenceof coronary calcification does not exclude obstructive coronary artery disease orthe need for revascularization in patients referred for conventional coronary angiog-raphy. J Am Coll Cardiol 2010;55:627–34.

8. SchroederS, AchenbachS, Bengel F, BurgstahlerC,Cademartiri F, de FPet al. Cardiaccomputed tomography: indications, applications, limitations, and training require-ments: report of a Writing Group deployed by the Working Group Nuclear Cardi-ology and Cardiac CT of the European Society of Cardiology and the EuropeanCouncil of Nuclear Cardiology. Eur Heart J 2008;29:531–56.

9. Gibbons RJ, Abrams J, Chatterjee K, Daley J, Deedwania PC, Douglas JS et al. ACC/AHA 2002 guideline update for the management of patients with chronic stableangina—summary article: a report of the American College of Cardiology/AmericanHeart Association Task Force on practice guidelines (Committee on the Manage-ment of Patients With Chronic Stable Angina). J Am Coll Cardiol 2003;41:159–68.

10. Rispler S, Keidar Z, Ghersin E, Roguin A, Soil A, Dragu R et al. Integrated single-photon emission computed tomography and computed tomography coronary angi-ography for the assessment of hemodynamically significant coronary artery lesions.J Am Coll Cardiol 2007;49:1059–67.

11. Gaemperli O, Schepis T, Valenta I, Husmann L, Scheffel H, Duerst V et al. Cardiacimage fusion from stand-alone SPECT and CT: clinical experience. J Nucl Med2007;48:696–703.

12. Slomka PJ, Cheng VY, Dey D, Woo J, Ramesh A, Van KS et al. Quantitative analysis ofmyocardial perfusion SPECT anatomically guided by coregistered 64-slice coronaryCT angiography. J Nucl Med 2009;50:1621–30.

13. Schaap J, Kauling RM, Boekholdt SM, Nieman K, Meijboom WB, Post MC et al. Incre-mental diagnostic accuracy of hybrid SPECT/CT coronary angiography in a popula-tion with an intermediate to high pre-test likelihood of coronary artery disease.Eur Heart J Cardiovasc Imaging 2013;14:642–9.

14. Kajander S, Joutsiniemi E, Saraste M, Pietila M, Ukkonen H, Saraste A et al. Cardiacpositron emission tomography/computed tomography imaging accurately detectsanatomically and functionally significant coronary artery disease. Circulation 2010;122:603–13.

15. Gaemperli O, Husmann L, Schepis T, Koepfli P, Valenta I, Jenni W et al. Coron-ary CT angiography and myocardial perfusion imaging to detect flow-limitingstenoses: a potential gatekeeper for coronary revascularization? Eur Heart J2009;30:2921–9.

16. Moons KG, de Groot JA, Linnet K, Reitsma JB, Bossuyt PM. Quantifying the addedvalue of a diagnostic test or marker. Clin Chem 2012;58:1408–17.

17. Diamond GA, Forrester JS. Analysis of probability as an aid in the clinical diagnosis ofcoronary-artery disease. N Engl J Med 1979;300:1350–8.

18. Fraker TD Jr, Fihn SD, Gibbons RJ, Abrams J, Chatterjee K, Daley J et al. 2007 chronicangina focused update of the ACC/AHA 2002 guidelines for the management ofpatients with chronic stable angina: a report of the American College of Cardi-ology/American Heart Association Task Force on Practice Guidelines WritingGroup to develop the focused update of the 2002 Guidelines for the managementof patients with chronic stable angina. Circulation 2007;116:2762–72.

19. Diamond GA. Right answer, wrong question: on the clinical relevance of the cardio-vascular history. Circulation 2011;124:2377–9.

20. Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis CA, Glasziou PP, Irwig LM et al. Towardscomplete and accurate reporting of studies of diagnostic accuracy: the STARD Initia-tive. Ann Intern Med 2003;138:40–4.

21. Moons KG, Grobbee DE. When should we remain blind and when should our eyesremain open in diagnostic studies? J Clin Epidemiol 2002;55:633–6.

22. Ascoop CA, van Zeijl LG, Pool J, Simoons M. Cardiac exercise testing-I indications,staff, equipment, conduct and procedures. Guidelines for cardiac exercise testing.Neth J Cardiol 1994;2:1–11.

23. Gibbons RJ, Balady GJ, Bricker JT, Chaitman BR, Fletcher GF, Froelicher VF et al.ACC/AHA 2002 guideline update for exercise testing: summary article. A reportof the American College of Cardiology/American Heart Association Task Forceon Practice Guidelines (Committee to Update the 1997 Exercise Testing Guide-lines). J Am Coll Cardiol 2002;40:1531–40.

24. Abidov A, Hachamovitch R, Hayes SW, Friedman JD, Cohen I, Kang X et al. Areshades of gray prognostically useful in reporting myocardial perfusion single-photonemission computed tomography? Circ Cardiovasc Imaging 2009;2:290–8.

25. BermanDS, AbidovA, Kang X, Hayes SW, Friedman JD, Sciammarella MG et al. Prog-nostic validation of a 17-segment score derived from a 20-segment score for myo-cardial perfusion SPECT interpretation. J Nucl Cardiol 2004;11:414–23.

26. Abbara S, Arbab-Zadeh A, Callister TQ, Desai MY, Mamuya W, Thomson L et al.SCCT guidelines for performance of coronary computed tomographic angiography:a report of the Society of Cardiovascular Computed Tomography Guidelines Com-mittee. J Cardiovasc Comput Tomogr 2009;3:190–204.

27. Schaap J, Kauling RM, Boekholdt SM, Post MC, Van der Heyden JA, de Kroon TL et al.Zero coronary calcium in the presence of severe isolated left main stenosis detectedby CT coronaryangiography in apatientwith typical angina andequivocalmyocardialperfusion SPECT. J Nucl Cardiol 2012;19:165–8.

28. Pijls NH, De Bruyne B, Peels K, Van Der Voort PH, Bonnier HJ, Bartunek JKJ et al.Measurement of fractional flow reserve to assess the functional severity ofcoronary-artery stenoses. N Engl J Med 1996;334:1703–8.

29. Rubin DB. Inferences and missing data. Biometrika 1976;63:581–90.30. Donders AR, van der Heijden GJ, Stijnen T, Moons KG. Review: a gentle introduction

to imputation of missing values. J Clin Epidemiol 2006;59:1087–91.31. de Groot JA, Janssen KJ, Zwinderman AH, Moons KG, Reitsma JB. Multiple

imputation to correct for partial verification bias revisited. Stat Med 2008;27:5880–9.

32. Vergouwe Y, Royston P, Moons KG, Altman DG. Development and validation of aprediction model with missing predictor data: a practical approach. J Clin Epidemiol2010;63:205–14.

33. Rubin DB. Multiple Imputation for non Respons in Surveys. New york: Wiley; 1987.34. Moons KG, Biesheuvel CJ, Grobbee DE. Test research versus diagnostic research.

Clin Chem 2004;50:473–6.35. Steyerberg EW. Clinical Prediction Models: A Practical Approach to Development, Valid-

ation and Updating. New York: Springer; 2008.36. Pencina MJ, D’Agostino RB Sr, D’Agostino RB Jr, Vasan RS. Evaluating the added pre-

dictive abilityof anewmarker: fromareaunder the ROCcurve to reclassificationandbeyond. Stat Med 2008;27:157–72.

37. R: A language and environment for statistical computing [computer program]. RFoundation for Statistical Computing; 2004.

38. Dedic A, Rossi A, Ten Kate GJ, Neefjes LA, Galema TW, Moelker A et al. First-lineevaluation of coronary artery disease with coronary calcium scanning or exerciseelectrocardiography. Int J Cardiol 2013;163:190–5.

39. Underwood SR, Anagnostopoulos C, Cerqueira M, Ell PJ, Flint EJ, Harbinson M et al.Myocardial perfusion scintigraphy: the evidence. Eur J Nucl Med Mol Imaging 2004;31:261–91.

40. Groutars RG, Verzijlbergen JF, Tiel-van Buul MM, Zwinderman AH, Ascoop CA, vanHemel NM et al. The accuracy of 1-day dual-isotope myocardial SPECT in a popula-tion with high prevalence of coronary artery disease. Int J Cardiovasc Imaging 2003;19:229–38.

41. Hamirani YS, Isma’eel H, Larijani V, Drury P, Lim W, Bevinal M et al. The diagnosticaccuracy of 64-detector cardiac computed tomography compared with stressnuclear imaging in patients undergoing invasive cardiac catheterization. J ComputAssist Tomogr 2010;34:645–51.

42. Budoff MJ, Dowe D, Jollis JG, Gitter M, Sutherland J, Halamert E et al. Diagnostic per-formance of 64-multidetector row coronary computed tomographic angiographyfor evaluation of coronary artery stenosis in individuals without known coronaryartery disease: results from the prospective multicenter ACCURACY (Assessment

Added value of hybrid myocardial perfusion SPECT and CT coronary angiography 1287

by guest on October 30, 2014

Dow

nloaded from

by Coronary Computed Tomographic Angiography of Individuals Undergoing Inva-sive Coronary Angiography) trial. J Am Coll Cardiol 2008;52:1724–32.

43. Meijboom WB, Meijs MF, Schuijf JD, Cramer MJ, Mollet NR, van Mieghem CAet al. Diagnostic accuracy of 64-slice computed tomography coronary angiog-raphy: a prospective, multicenter, multivendor study. J Am Coll Cardiol 2008;52:2135–44.

44. Abidov A, RaffGL. Value of coronary CTA in patients with knownor suspected CADand non-diagnostic initial myocardial perfusion testing: current evidence and clinicalconsiderations. J Nucl Cardiol 2010;17:1101–6.

45. Flotats A, Knuuti J, Gutberlet M, Marcassa C, Bengel FM, Kaufmann PA et al. Hybridcardiac imaging: SPECT/CT and PET/CT. A joint position statement by the EuropeanAssociation of Nuclear Medicine (EANM), the European Society of Cardiac Radi-ology (ESCR) and the European Council of Nuclear Cardiology (ECNC). Eur JNucl Med Mol Imaging 2011;38:201–12.

46. Schaap J, de Groot JA, Nieman K, Meijboom WB, Boekholdt SM, Post MC et al.Hybrid myocardial perfusion SPECT/CT coronary angiography and invasive

coronary angiography in patients with stable angina pectoris lead to similar treat-ment decisions. Heart 2013;99:188–94.

47. Genders TS, Steyerberg EW, Alkadhi H, Leschka S, Desbiolles L, Nieman K et al. Aclinical prediction rule for the diagnosis of coronary artery disease: validation, updat-ing, and extension. Eur Heart J 2011;32:1316–30.

48. Genders TS, Steyerberg EW, Hunink MG, Nieman K, Galema TW, Mollet NR et al.Prediction model to estimate presence of coronary artery disease: retrospectivepooled analysis of existing cohorts. BMJ 2012;344:e3485.

49. Gould KL, Johnson NP, Bateman TM, Beanlands RS, Bengel FM, Bober R et al. Ana-tomic versus physiologic assessment of coronary artery disease. Role of coronaryflow reserve, fractional flow reserve, and positron emission tomography imagingin revascularization decision-making. J Am Coll Cardiol 2013;62:1639–53.

50. Tonino PA, Fearon WF, De BB, Oldroyd KG, Leesar MA, Ver Lee PN et al. Angio-graphic versus functional severity of coronary artery stenoses in the FAME studyfractional flow reserve versus angiography in multivessel evaluation. J Am CollCardiol 2010;55:2816–21.

IMAGE FOCUS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

doi:10.1093/ehjci/jeu124Online publish-ahead-of-print 27 June 2014

A notch in time: midventricular hypertrophic cardiomyopathyKevin C. Ong*, Jeffrey B. Geske, Kyle W. Klarich, and Rick A. Nishimura

Division of Cardiovascular Diseases, Department of Internal Medicine, Mayo Clinic, Gonda 6, 200 First Street SW, Rochester, MN 55905, USA

* Corresponding author. Tel: +1 507 284 3545; Fax: +1 507 266 7929, Email: ong.kevin@mayo.edu

Imaging of a 62-year-old woman with hypertroph-ic cardiomyopathy revealed mid-septal hyper-trophy (22 mm), an apical aneurysm (Panel A),and midventricular obstruction (MVO) with anunusual continuous-wave Doppler (CWD)profile (Panel B). Initially, CWD showed an in-crease in midventricular systolic velocities to2.7 m/s (Panels B1 and C1). In midsystole, MVObecame sufficiently severe to cause attenuationof distal flow (Panels B2, double headed arrowand C2) in the plane of CWD interrogation coin-ciding with appearance of a notch (Panel B2, aster-isk). However, proximal flow was still present(Panel C2), indicating that MVO was incompleteas opposed to complete where all flow wouldbe absent. Flow just downstream of obstructionconverges to the vena contracta, where velocityis highest allowing calculation of peak gradient.Since there was dropout of the jet containingvena contracta velocity, the notch actually repre-sented flow velocities in the neck of obstructionand did not represent midventricular gradientswhich continued to increase. Dynamic MVO ledto aprogressive reduction in orifice size, increasedafterload, and a corresponding decrease in proximal midsystolic flow velocities (Panel B2, asterisk) analogous to the lobster claw abnor-mality described in left ventricular outflow tract obstruction.1 Partial relief of MVO in late systole permitted the return of distal flow cor-responding to the second CWD peak velocity (Panels B3 and C3) and continuing into early diastole (Panels B4 and C4), during which time peakflow velocities now represented midventricular gradients. Simultaneous appearance of apical forward flow and mitral inflow is created by anapex–base gradient that persists into early diastole as a result of high apical pressures and relaxation of the left ventricular basal segments.

Conflict of interest: none declared.

Reference1. Sherrid MV, Gunsburg DZ, Pearle G. Mid-systolic drop in left ventricular ejection velocity in obstructive hypertrophic cardiomyopathy—the lobster claw abnormality.

J Am Soc Echocardiogr 1997;10:707–12.

Published on behalf of the European Society of Cardiology. All rights reserved. & The Author 2014. For permissions please email: journals.permissions@oup.com.

J. Schaap et al.1288

by guest on October 30, 2014

Dow

nloaded from