1 3

ORIGINAL ARTICLE

Ann Nucl Med (2008) 22:899–909DOI 10.1007/s12149-008-0191-7

P. Georgoulias (*) · C. Tzavara · S. Giannakou · V. Valotassiou · I. FezoulidisDepartment of Nuclear Medicine, University Hospital of Larissa, Mezourlo, Larissa 41110, Greecee-mail: pgeorgoul@med.uth.gr

N. DemakopoulosDepartment of Nuclear Medicine, NIMTS Hospital, Athens, Greece

I. TsougosDepartment of Medical Physics, University Hospital of Larissa, Larissa, Greece

P. XaplanterisDepartment of Cardiology, NIMTS Hospital, Athens, Greece

Incremental prognostic value of 99mTc-tetrofosmin myocardial SPECT after percutaneous coronary intervention

Panagiotis Georgoulias · Chara Tzavara Nikolaos Demakopoulos · Stavroula Giannakou Varvara Valotassiou · Ioannis Tsougos Petros Xaplanteris · Ioannis Fezoulidis

events. Receiver-operating characteristic (ROC) analysis was used to test the prognostic ability of SSS and SDS for cardiac events. Cox proportional hazards models were used to evaluate the incremental value of SPECT variables.Results Cardiac death occurred in 12 (4.9%) patients and non-fatal myocardial infarction in 20 (8.1%) patients. In addition, 60 (24.4%) patients underwent a late revas-cularization procedure. Using ROC analysis the optimal cut-offs of SSS (AUC = 0.94; 95% CI 0.92–0.97) and SDS (AUC = 0. 76; 95% CI 0.70–0.82) for the prediction of cardiac events were 10 and 1.7, respectively. Multiple Cox regression analyses revealed that SSS > 10 (HR = 24.2; 95% CI 7.44–78.79) and SDS > 1.7 (HR = 2.72; 95% CI 1.23–6.00) provided incremental prognostic value over clinical and exercise test data for the composite end points of any cardiac event.Conclusions 99mTc-tetrofosmin myocardial gated-SPECT, performed 6 months post-percutaneous coro-nary intervention (PCI), provides incremental prognostic information for the prediction of cardiac events in asymptomatic patients after PCI.

Keywords Incremental prognostic value · PCI · Myocardial SPECT · 99MTC-tetrofosmin

Introduction

Percutaneous transluminal coronary angioplasty (PTCA) is a routine indispensable therapeutic method in selected patients with coronary artery disease (CAD). This method has been widely applied worldwide, despite the high restenosis rate of 20%–65% in patients without stenting [1–3]. More recently, the introduction of

Received: 3 April 2008 / Accepted: 9 July 2008© The Japanese Society of Nuclear Medicine 2008

AbstractObjective Percutaneous transluminal coronary angio-plasty is a well-established therapeutic method in selected patients with coronary artery disease. The aim of this study was to assess the incremental prognostic value of technetium-99m (99mTc)-tetrofosmin myocardial gated-single-photon emission computed tomography (SPECT) in asymptomatic patients after coronary artery stenting.Methods A total of 246 consecutive patients (aged 55.5 ± 8.2 years, 182 men) participated in the study with a median follow-up of 9.5 years (interquartile range 5.8–10.5 years). All patients underwent exercise gated-SPECT myocardial imaging within 5–7 months. Myocardial scintigrams were performed using 99mTc-tetrofosmin, and were evaluated calculating the summed stress score (SSS), summed rest score (SRS), and summed difference score (SDS) indexes. Cardiovascular death and non-fatal myocardial infarction were considered as hard cardiac events, and late revascularization (>3 months after myocardial SPECT) procedures as soft

900 Ann Nucl Med (2008) 22:899–909

1 3

intracoronary stenting has resulted in a dramatic decline in acute complications and also a signifi cant reduction of the restenosis rate to an average of 20%–25% [4–8]. As a result, the use of and indications for percutaneous coronary interventions (PCIs) have greatly expanded [9].

Even though the restenosis rate has in fact been reduced by intracoronary stenting, it still remains the major drawback of PCI and also an important clinical problem. Many studies have documented the lack of reliability of chest pain as a criterion of restenosis, even as the diagnostic accuracy of exercise electrocardiogra-phy testing is also limited [10–12]. On the other hand, coronary angiography is considered as the “gold stan-dard” for detecting restenosis, but its invasive character is a major drawback for its application to follow-up patient post-PCI.

Myocardial perfusion single-photon emission com-puted tomography (SPECT) has been useful for the assessment of CAD and the identifi cation of critical ste-noses before angioplasty, as well as the follow-up evalu-ation of patients after the intervention and the early detection of restenosis [10–17]. However, on the prog-nostic value of myocardial SPECT, limited data have been reported with the usual tracers (201Tl, 99mTc-MIBI) in patients after PCI, although there are currently scarce data to suggest the prognostic value of 99mTc-tetrofosmin myocardial SPECT in patients after PCI [18–26]. There-fore, the aim of this prospective study was to evaluate the incremental prognostic value of 99mTc-tetrofosmin myocardial gated-SPECT in patients without cardiac symptoms after coronary artery stenting, over clinical and exercise testing variables.

Materials and methods

Population

The study population consisted of 246 consecutive patients, who underwent a successful PCI combined with a stent implementation (routine stenting), between January 1996 and December 1999. This group consisted of 182 men and 64 women, ranging in age from 32 years to 77 years (mean age 55.5 ± 8.2 years). All patients reported angina before the intervention, but remained asymptomatic afterward. Angioplasty and stenting were performed according to standard guidelines [5, 8]. Coro-nary arteries were submitted to revascularization only in the presence of initial narrowing ≥70%. Angioplasty success was defi ned as the dilation of at least one stenosis ≥70% of the luminal diameter resulting in a residual ste-nosis of <30% in diameter (a gain of ≥40% with respect

to pre-angioplasty values) and without major complica-tions (death, myocardial infarction, or coronary artery bypass grafting). Vessel luminal stenoses ≥50%, demon-strated by coronary angiography were considered as hemodynamically signifi cant. Patients were considered to have undergone incomplete revascularization if they had stenosis of ≥50% in any coronary artery not sub-jected to revascularization.

Eighty-four patients (34.1%) had a previous myocar-dial infarction. A total of 132 patients (53.7%) had single-vessel disease: 90 of the left anterior descending artery, 14 of the left circumfl ex artery, and 28 of the right coronary artery, and they had successful PCI of the ste-nosed vessel. Ninety-six patients (39%) had two-vessel disease: 61 underwent a PCI of both stenosed vessels and 35 had a successful PCI of one vessel. Eighteen patients (7.3%) had three-vessel disease: four of them had suc-cessful one-vessel revascularization, 10 had two-vessel intervention and the other four had three-vessel angio-plasty (Table 1).

All patients underwent symptom-limited exercise testing combined with single-photon emission computed tomography (SPECT) myocardial perfusion imaging, using 99mTc-tetrofosmin, about 6 months (5–7 months) after PCI. The routine follow-up using myocardial scin-tigraphy formed the basis of the study.

Pregnant women were excluded, as well as patients whose myocardial perfusion imaging might have been affected by factors other than myocardial ischemia. We therefore excluded patients with left bundle branch block, cardiomyopathy, severe valvular disease, and also those with an implanted pacemaker. We also excluded patients who had undergone coronary artery bypass grafting (CABG), suffered a myocardial infarction within a month before PCI, with contraindication to or inability to perform treadmill testing or to achieve a satisfactory exercise level because of a noncardiac condi-tion (peripheral vascular disease, sciatica, neuropathy, disability, etc.) and patients taking digoxin (owing to its prolonged effect). Moreover, medications that could possibly infl uence patient performance on exercise testing and the related variables, were temporarily withdrawn (for about fi ve half-lives). β-blockers were discontinued gradually (within a week—depending on the medication and the dose), with complete discontinuation at least 48 h before and during the study. Calcium channel antagonists and nitrates were discontinued 48 h and 24 h before and during the study, respectively. Additionally, any patient whose medication had not been discontinued as described above was excluded. Finally, patients with either angina, myocardial infarction or coronary bypass between PCI and SPECT imaging, those who were lost to follow-up or died of a non-cardiac cause during

Ann Nucl Med (2008) 22:899–909 901

1 3

follow-up and those who underwent early revasculariza-tion (≤3 months after myocardial SPECT: 6 CABG and 19 PCI), were excluded from the prognostic analysis.

A systolic blood pressure of 140 mmHg or greater at rest and/or a diastolic blood pressure of 90 mmHg or greater at rest, or treatment with antihypertensive medi-cines, was considered as hypertension. Diagnoses of dia-betes mellitus and lipid disorders were derived from the interviews with the patients and the use of corresponding medications. Obesity was considered as a body mass index (BMI-calculated as weight in kilograms divided by height in meters squared) value of 30.0 or greater.

In accordance with the Hospital Ethics Committee guidelines, all patients gave informed consent for their

complete participation before testing, and a brief struc-tured interview from which we obtained data on symptoms, medications, previous cardiac events, coro-nary risk factors and cardiac or non-cardiac diagnoses. Finally, prior to the study, patients were given written directions on radioprotection.

Follow-up

Follow-up data were obtained by phone contact with the patients, their relatives and patients’ general practitioner or cardiologist, during visits to the clinic and/or review of the patients’ hospital records. Cardiovascular death and non-fatal myocardial infarction were considered as

Table 1 Clinical characteristics of the study group and trends across the order categories: no events, soft events, and hard cardiac events

No event Soft event Hard event Soft/hard eventMean (SD) Mean (SD) Mean (SD) Mean (SD)

Clinical dataMen, n (%) 107 (69.5) 51 (85.0) 24 (75.0) 75 (81.5)Age (years) 57.3 (7.2) 56.2 (9.6) 55.0 (6.9) 55.7 (8.9)Obese, n (%) 23 (14.9) 16 (26.7) 5 (16.1) 21 (23.1)Hypertension, n (%) 45 (29.2) 21 (35) 12 (37.5) 33 (35.9)Lipid disorder 97 (63.0) 45 (75.0) 20 (62.5) 65 (70.6)Smoking, n (%) 100 (64.9) 46 (76.7) 22 (68.8) 68 (73.9)Diabetes, n (%) 43 (27.9) 11 (18.3) 6 (18.8) 17 (18.5)Family history, n (%) 112 (72.7) 34 (56.7) 22 (68.8) 56 (60.9)Nitrates, n (%)** 137 (89) 59 (98.3) 32 (100) 91 (98.9)Calcium channel antagonists, n (%) 132 (85.7) 50 (83.3) 28 (87.5) 78 (84.8)Beta-blockers, n (%) 67 (43.5) 35 (58.3) 6 (18.8) 41 (44.6)Previous myocardial infarction, n (%)*** 31 (20.1) 25 (41.7) 28 (87.5) 53 (57.6)

No of diseased vessels, n (%)* 1 91 (59.1) 27 (45.0) 14 (43.8) 41 (44.6) 2–3 63 (40.9) 33 (55.0) 18 (56.3) 51 (55.4)Incomplete revascularization, n (%) 55 (35.7) 27 (45.0) 17 (53.1) 44 (47.8)

Exercise variablesAngina during exercise, n (%) 6 (3.9) 14 (23.3) 4 (12.5) 18 (19.6)Resting heart rate 114.5 (15.1) 109.9 (16.8) 112 (13.6) 110.6 (15.7)Max heart rate 149.7 (17.0) 147.6 (20.1) 148.2 (19.4) 147.8 (19.8)Exercise duration 12.1 (2.7) 12.3 (2.9) 10.8 (2.9) 11.8 (3)Abnormal ST-segment response, n (%) 46 (32.4) 24 (41.4) 7 (21.9) 26 (32.5)Double product (×100) 26.4 (5.2) 27 (5.3) 26.3 (5.6) 26.8 (5.4)Metabolic equivalents 9.9 (1.7) 9.9 (1.8) 9.8 (1.7) 9.9 (1.8)Chronotropic response 89.1 (15.0) 92.6 (13.7) 92.1 (16.9) 0.9 (0.1)HRR 35.2 (16.6) 37.7 (18.9) 36.2 (17.1) 37.2 (18.3)Duke score 8.6 (6.3) 5.6 (10.8) 8.3 (4.5) 6.5 (9.2)

Scintigraphic variablesTID index 0.97 (0.14) 0.93 (0.14) 0.92 (0.17) 1.0 (0.1)LVEF*** 62.2 (7.3) 54.7 (13.8) 42.9 (14.3) 50.6 (15.1)LVD, n (%)*** 0 (0) 8 (13.3) 10 (31.3) 18 (19.6)SRS*** 2.3 (2.6) 10.2 (8.1) 19.9 (8.7) 13.6 (9.5)SSS*** 4.9 (3.6) 16.2 (7.2) 25.4 (7.7) 19.4 (8.5)SDS*** 2.6 (4.1) 6 (3.4) 5.5 (4.0) 5.8 (4)SSS > 10*** 22 (14.3) 57 (95.0) 32 (100.0) 89 (96.7)SDS > 1.7*** 74 (48.1) 55 (91.7) 29 (90.6) 84 (91.3)

Mean (standard deviation)HRR heart-rate recovery, bpm beats per minute, SSS summed stress score, SRS summed rest score, SDS summed difference score, TID transient ischemic dilation, LVEF left ventricular ejection fraction, LVD left ventricular dilation* P < 0.05; ** P < 0.01; *** P < 0.001 for trend according to the three groups: no event, soft event, hard event

902 Ann Nucl Med (2008) 22:899–909

1 3

hard cardiac events, and revascularization (≥3 months after myocardial SPECT, including PCI and CABG) as soft events. Death was considered cardiovascular if it was caused by “Diseases of the Circulatory System” according to the Tenth International Classifi cation of Diseases (numbers I00–I99) [27]. In the presence of mul-tiple causes, a hierarchical preference was adopted with cancer in advanced stages and cardiovascular death in that order [28]. A reviewer, who was blinded to the clini-cal, exercise, myocardial SPECT data, and the hypothe-sis of the study, assessed the cause of death by examination of death certifi cates. All patients included in the study were followed up for at least 8 months. The follow-up period was completed in patients suffering from a hard or soft event. The mean follow-up period was 8.3 years (SD = 2.9) with median equal to 9.5 years (interquartile range from 5.8 to 10.5).

Exercise testing

After discontinuing cardioactive medication, 6–12 h fasting and avoiding smoking or engaging in heavy phys-ical activity for at least 3 h before the examination, patients underwent symptom-limited treadmill exercise testing (Bruce Protocol). Data on symptoms and esti-mated workload in metabolic equivalents (METs) were obtained; functional capacity in METs was estimated using standard tables [29]. As a criterion of ischemic ST-segment response, we considered greater than 1 mm horizontal or down-sloping ST-segment depression, 80 ms after the J point or more than 1 mm of additional ST-segment rise in leads without pathologic Q waves. Chronotropic response was defi ned as the percentage of maximal age-predicted heart rate achieved at peak exer-cise (peak heart rate/220-age); a value of 85% or greater was considered to be normal [30]. The reduction in heart-rate from its value at peak exercise to the rate 1 min later was determined as the heart-rate recovery (HRR) [31].

The Duke treadmill score equals {exercise time (in min) − [5 times the ST-segment deviation, during or after exercise (in millimeters)] − 4 times the angina index (which has a value of 0 if there is no angina, 1 if angina occurs, and 2 if angina is the reason for stopping the test)} [32].

SPECT myocardial perfusion imaging

Myocardial perfusion studies were performed using technetium-99M (99MTC) tetrofosmin (Myoview, Amersham) [33, 34].

Scintigraphies were obtained after the intravenous injection of the radiotracer (dose, 185–296 MBq—based on weight), 1–2 min before the cessation of exercise.

Twenty minutes after radiopharmaceutical administra-tion, 250 ml of milk (3.5% fat) was ingested for optimum excretion of the tracer by the gallbladder followed by two glasses of water 10–15 min later and acquisition started 35–45 min after the administration; 4 h after the fi rst injection, resting 99mTc-tetrofosmin SPECT was obtained, using a 555–888 MBq dose (based on weight). Images were acquired from the 45° right anterior oblique to 45° left posterior oblique position in step-and-shoot mode (64 projections, matrix 64 × 64, pixel size 6.8 mm, 25 s per projection for stress and 40 s per projection for rest scintigrams). The high-count rest scans were acquired as gated-SPECT studies (8 frames per cardiac cycle) and the left ventricular ejection fraction (LVEF) and end diastolic volume were calculated using Sopha Medical Vision software. Eleven patients with severe arrhythmic problems underwent MPI without using electrocardio-graphic gating [35]. The acquisitions were obtained with a two-headed camera (Sopha Medical Vision, model DST-XL) equipped with low-energy, parallel-hole, high-resolution collimators. The energy window was set at ±10% symmetrically along the 140 keV photopeak and data processing was accomplished with a Butterworth fi lter (cut-off 0.35, order 4.0), without attenuation or scatter correction. We performed polar and three-dimen-sional (3D) mapping (transient ischemic dilation index-TID was also calculated) in all studies and the presence of left ventricular dilation (LVD) was estimated [34].

According to earlier reports [34, 36], for SPECT interpretation, the myocardium of the left ventricle was divided into 17 segments. Two independent experienced observers blindly evaluated the reconstructed images, the polar maps and the 3D images of both stress and rest studies; this was done by scoring the uptake in each of the 17 regions, using a fi ve-point scoring system (0 normal uptake, 1 mildly reduced uptake, 2 moderately reduced uptake, 3 severely reduced uptake, and 4 no uptake) [34, 36]. If counts were reduced in a segment and this was judged to be the result of attenuation artifact, the score was 0 [34].

Ischemia was considered in every region with an uptake higher than 0 at stress imaging and a reduction of the score by at least one unit at rest, whereas an irre-versible defect was defi ned as a region with a score at stress >1 without an improvement at rest.

Finally, the “summed stress score” (SSS) and “summed rest score” (SRS) were obtained by adding the scores of the segments at stress and rest studies. In the case of dis-cordance between the two observers, the view of a third observer was requested and the disagreement was resolved by consensus. Subsequently, a “summed difference score” (SDS) was estimated by subtracting the “SRS” from the “SSS” to assess defect reversibility [34, 36].

Ann Nucl Med (2008) 22:899–909 903

1 3

Statistical analysis

Quantitative variables were expressed as mean values (SD) or as median values (interquartile range). Qualita-tive variables are expressed as absolute and relative fre-quencies. The scintigraphic variables SSS and SDS were tested for their ability to predict the composite outcome (soft and hard cardiac events) using receiver-operating characteristic (ROC) curves. The overall performance of the ROC analysis was quantifi ed by computing area under the curve (AUC). An area of 1 indicated perfect performance, whereas 0.5 indicated a performance that was not different than chance. Using ROC analysis we determined the optimal sensitivity and specifi city of using various cut-off values for the prediction of cardiac events. For the comparisons of proportions chi-square tests were used. Life table analyses were used to calculate cumulative survival rate (standard errors) for specifi c time intervals. Incidence rates were presented as the number of events/1000 person-years and provided by the SSS and SDS cut-offs. Event rates were calculated divid-ing the number of events occurring during the study period by the number of person-years of observation. For all clinical, exercise, and scintigraphic variables, P values for trend in rates for severity of disease by the ordered groups, no events, soft events, and hard events, were also estimated. The prognostic value of each vari-able for cardiac events was fi rst assessed by univariate Cox regression analysis [37]. Variables that showed a signifi cant association with the outcome were included in the multivariate Cox proportional-hazard model in a stepwise method. First the best prediction model using clinical and exercise data was obtained and later the incremental value of scintigraphic variables was evalu-ated by their addition to the previous model. The incre-mental prognostic value of the scintigraphic variables was evaluated by a statistically signifi cant increase in the global chi-square of the model. The assumption of pro-portional hazards was evaluated by testing for interac-tion with a continuous time variable. Kaplan–Meier survival estimates for cardiac events were graphed over the follow-up period. All reported P values are two-tailed. Statistical signifi cance was set at P < 0.05, and analyses were conducted using STATA statistical soft-ware (version 6.0).

Results

A total of 246 patients, 182 men and 64 women, with a mean age of 55.5 years (SD = 8.2 years), participated in the study. The mean follow-up period was 8.3 years (SD = 2.9) with median equal to 9.5 years (interquartile range

from 5.8 years to 10.5 years). During the follow-up period, hard cardiac events occurred in 32 (13%) patients. Specifi cally, cardiac death occurred in 12 (4.9%) patients and non-fatal myocardial infarction in 20 (8.1%) patients. Soft cardiac events (late revascularization) occurred in 60 (24.4%) of the patients (14 CABG, 46 repeat PCIs). The mean time interval between patients’ examination and cardiac events was 6.0 years (SD = 3.0 years) for hard events and 7.6 years (SD = 2.7 years) for soft events. Sixty-six (26.8%) patients had a normal myocardial SPECT about 6 months after the PCI. A total of 112 (45.5%) patients had only reversible defects, 61 (24.8%) patients had both reversible, and irreversible defects whereas 7 (2.8%) patients had only irreversible defects in their scintigram. The clinical, exercise, and scinti-graphic data for those without events, with soft cardiac events, with hard cardiac events or the composite end-point of both soft and hard cardiac events are listed in Table 1. From variables concerning clinical and exercise data an increased trend in rates according to the three groups, no event, soft events, and hard events was found for the use of nitrates, the existence of prior myocardial infarction, the number of diseased vessels and incom-plete revascularization. Additionally, an increased trend for the earlier ordered categories was found for all scin-tigraphic variables except for TID index (Table 1).

Receiver-operating characteristic curve analysis (Fig. 1) showed that the optimal cut-off of SSS for the predic-tion of any cardiac event was 10 with sensitivity equal to 96.7% and specifi city equal to 85.7. Similarly, the SDS value of 1.7 represented the optimal cut-off for the pre-diction of cardiac events (sensitivity 91.3% and specifi c-ity 52%) (Fig. 2). The AUC was 0.94 (95% CI 0.92–0.97) and 0.76 (95% CI 0.70–0.82), for SSS and SDS, respec-

Fig. 1 Receiver-operating characteristic (ROC) curve for the pre-diction of cardiac events from summed stress score (SSS)

904 Ann Nucl Med (2008) 22:899–909

1 3

tively, which signifi cantly differs from 0.5 (P < 0.001). Concerning only hard cardiac events, ROC analysis demonstrated a cut-off equal to 11.7 for SSS (sensitivity 96.9% and specifi city 91.6%) and 1.7 for SDS (sensitivity 90.6% and specifi city 52%). The AUCs were 0.98 (95% CI 0.96–1.00) and 0.72 (95% CI 0.63–0.81), for SSS and SDS, respectively, (P < 0.001); 14.3% and 48.1% of patients without events had an SSS score over 10 and SDS over 1.7, whereas the aforementioned propor-tions for patients with composite cardiac events were 96.7% and 91.3% (P < 0.001 for all comparisons). Kaplan–Meier estimates for free-cardiac event rates for SSS score over 10 and SDS over 1.7 are presented in Figs. 3 and 4.

For patients with SSS score more than 10, the cumu-lative event-free rate for the fi rst year was 98.2% (SE = 1.3%), for 5 years 80.9% (SE = 3.7%) and for 10 years 36.4% (SE = 4.7%) whereas for patients with SSS score

less than 10 the cumulative hard event-free rates for 1 year was 100% (SE = 0%) for 5 years 99.2% (SE = 0.8%) and for 10 years 96.9% (SE = 1.8%). The cumulative event-free rates for 1 years, 5 years, and 10 years were 98.7% (SE = 0.9%), 86.5% (SE = 2.7%), and 50.5% (SE = 4.3%) for patients with SDS more than 1.7% and 100% (SE = 0%), 98.8% (SE = 1.2%) and 75.1% (SE = 8.3%) for patients with SDS less than 1.7, respectively. The composite cardiac event rate/1000 patient-years that have been found for SSS score more than 10 was 110 (95% CI 86.05–130) and for SSS score less than 10 was 2.5 (95% CI 0.81–7.74). The composite cardiac event rate/1000 patient-years for SDS more than 1.7 was 65.51 (95% CI 52.89–81.12) and for SDS less than 1.7 was 10.54 (95% CI 5.27–21.07).

As for clinical and exercise testing data, univariate analysis identifi ed male sex, hypertension, the use of nitrates, prior myocardial infarction, number of diseased vessels, incomplete revascularization, angina during exercise and chronotropic response as signifi cant predic-tors for the composite endpoint (Table 2). Multiple Cox proportional hazard analysis selected sex, angina during stress testing, prior myocardial infarction and number of diseased vessels as independent predictors. The global chi-square for this model was 44.42 (P < 0.001). The gated-SPECT variables, namely, LVEF, left ventricle dilation, SSS, and SDS were signifi cant univariate pre-dictors for cardiac events (Table 2). In the multivariate model, LVEF, SSS, and SDS were shown to signifi cantly enhance the predictive ability of the clinical/exercise model (Table 2). Prior myocardial infarction and the number of diseased vessels were no longer signifi cant following the addition of scintigraphic variables (Table 2). The global chi-square for this model was 92.16 (P < 0.001) and signifi cantly increased compared with the one of the earlier models (P < 0.001), declaring the signifi cant

Fig. 2 ROC curve for the prediction of cardiac events from summed difference score (SDS)

Fig. 3 Kaplan–Meier estimates for cardiac events according to SSS levels

Fig. 4 Kaplan–Meier estimates for cardiac events according to SDS levels

Ann Nucl Med (2008) 22:899–909 905

1 3

incremental value of scintigraphic variables (Table 2). The hazard ratios for LVEF, SSS, and SDS were 0.98, 1.05, and 1.06, respectively (Table 2). The results were similar when SSS and SDS were included in the multiple analysis as binary variables using the cut-offs provided by the ROC analyses. The adjusted hazard ratios for SSS more than 10 and SDS more than 1.7 were 24.2 and 2.72, respectively (Table 2).

Discussion

Myocardial SPECT imaging is considered to be a reli-able method for a realistic approach to myocardial per-fusion, suitable for the evaluation of patients who have

undergone PCI and the detection of restenosis [9–17]. Although the prognostic value of myocardial SPECT in patients with known or suspected CAD is well estab-lished, routine testing in patients after PCI is not gener-ally recommended unless symptoms recur [9, 10, 21, 22, 24]. On the other hand, silent ischemia has been reported as a strong predictor of cardiac death and events in patients with CAD, whereas several studies have shown that asymptomatic restenosis is quite frequent in patients following PCI [18, 22, 24]. In addition, the long-term prognostic value of myocardial scanning in asymptom-atic patients after PCI has not been well established [18, 22–25].

99mTc-tetrofosmin comprises an interesting alternative to myocardial perfusion imaging, beyond the methods

Univariate Multivariate

HR (95% CI) HR (95% CI)

Clinical dataMen, n (%) 2.15 (1.26–3.65)** 1.78 (1.03–3.09)*Age (years) 0.97 (0.93–1.02)Obese, n (%) 1.35 (0.81–2.24)Hypertension, n (%) 1.69 (1.08–2.64)*Lipid disorder 1.08 (0.69–1.71)Smoking, n (%) 1.3 (0.82–2.08)Diabetes, n (%) 0.62 (0.36–1.06)Family history, n (%) 1.05 (0.68–1.61)Nitrates, n (%) 9.04 (1.26–65.01)**Calcium channel antagonists, n (%) 1.28 (0.72–2.28)Beta-blockers, n (%) 0.88 (0.58–1.33)Previous myocardial infarction, n (%) 2.35 (1.54–3.57)***

No. of diseased vessels, n (%) 1, reference 2–3 1.92 (1.25–2.95)**Incomplete revascularization, n (%) 1.82 (1.19–2.17)**

Exercise variablesAngina during exercise, n (%) 3.53 (2.07–6.01)*** 2.02 (1.17–3.47)**Resting heart rate 0.99 (0.98–1.01)Max heart rate 1.00 (0.99–1.01)Exercise duration 0.93 (0.86–1.00)Abnormal ST-segment response, n (%) 0.76 (0.48–1.22)Double product (×100) 1.00 (0.96–1.04)Metabolic equivalents 1.02 (0.9–1.14)Chronotropic response 4.5 (1.12–18.14)*HRR 1.00 (0.99–1.01)Duke score 0.98 (0.96–1.01)

Scintigraphic variablesTID index 1.23 (0.24–6.26)LVEF 0.95 (0.93–0.96)*** 0.98 (0.96–0.99)**LVD, n (%) 5.71 (3.34–9.78)***SRS 1.07 (1.05–1.09)***SSS 1.08 (1.06–1.1)*** 1.05 (1.01–1.08)**SDS 1.10 (1.06–1.14)*** 1.06 (1.01–1.10)*SSS > 10 42.43 (13.41–134.18)*** 24.2 (7.44–78.79)***SDS > 1.7 6.4 (3.11–13.28)*** 2.72 (1.23–6.00)*

Table 2 Univariate and multivariate Cox proportional hazards models for the prediction of any cardiac events

Hazard ratio (95% confi dence interval)* P < 0.05; ** P < 0.01; *** P < 0.001HRR heart-rate recovery, bpm beats per minute, SSS summed stress score, SRS summed rest score, SDS summed difference score, TID transient ischemic dilation, LVEF left ventricular ejection fraction, LVD left ventricular dilation

906 Ann Nucl Med (2008) 22:899–909

1 3

using 201Tl and 99mTc-MIBI. This agent combines the fi ne physical properties of Tc-99m with easy and fast prepa-ration [33, 35, 38–40]. The tracer’s reliability in the assessment of CAD and also in detecting restenosis has been reported earlier, whereas the prognostic value of myocardial SPECT using this agent has been well dem-onstrated [33, 35, 38–40].

This study provides a long-term follow-up (8.3 ± 2.9 years) of asymptomatic patients after 99mTc-tetrofosmin myocardial SPECT, for the evaluation of its incremental prognostic value following PCI. The incremental prog-nostic value of SPECT over clinical and exercise testing parameters has been reported in previous studies [18, 23, 26, 41]. On the other hand, on reviewing the literature, we have not found studies evaluating the incremental value of SPECT variables using specifi c cut-offs of SSS and SDS values.

In the Cox multiple analysis model gated-SPECT parameters were independently associated with an increased risk for cardiac events. Furthermore, we deter-mined that gated-SPECT variables contributed a statis-tically signifi cant increment to prognostic information over that provided by the clinical and the exercise testing data.

The prognostic validity of gated-SPECT data is further confi rmed by their proportional increased values in concordance with the ordered categories of patients: without events, soft events, hard events; providing further information that patients with severe myocardial ischemia, indicated via SPECT parameters, are at increased risk for subsequent cardiac events. This is the fi rst study evaluating the optimal cut-off values for SSS and SDS applying ROC analysis for the prediction of the composite end points of any cardiac event in asymp-tomatic patients post-PCI, with a median follow-up of 9.5 years. SSS > 10 and SDS > 1.7 were found to be the optimal cut-offs for the prognosis of any cardiac event.

Patients with SSS < 10 or SDS < 1.7 exhibited a very low rate of cardiac events per 1000 patient-years (2.5 and 10.5, respectively), whereas the corresponding values for patients with SSS > 10 or SDS > 1.7 were 110 and 65.5, respectively. However, the specifi city of SDS is lower than that of SSS (52% vs. 85.7%) for the prediction of any cardiac event. As the SDS value is calculated by subtracting the SSS value from the SRS value, it is dependent on the tracer’s uptake both at stress and rest. Thus, SDS is mostly an index of defect reversibility, whereas SSS is an index of defect extension and intensity better refl ecting the severity of ischemia (in viable myo-cardial regions).

In multiple Cox analysis, SSS > 10 and SDS > 1.7 were associated with approximately 24 and 2.7 times greater risk for the composite cardiac events, respectively. Other

independent predictors were angina during exercise testing, male sex, and LVEF.

Other studies have already reported the prognostic importance of MPI, using either 201Tl or 99mTc-MIBI as radiotracers, to predict long-term outcome following PCI, despite the fact that they included both symptom-atic and asymptomatic patients [18–26].

Cottin et al. [19] studied 152 patients 5 ± 2 months after stenting with stress-rest thallium SPECT and reported a signifi cantly worse prognosis of ischemic patients compared with non-ischemic ones (58% had major cardiac events or revascularization vs. 11%, P < 0.001). In addition, Ho et al. [20] studying the late outcome of a series of 211 patients who underwent a SPECT 201Tl MPI 1–3 years after PTCA, reported that the SSS index was signifi cantly associated with the end point of cardiac death or myocardial infarction. Similarly, Acampa et al. [22], in a study population of 206 patients who underwent a 99mTc-MIBI myocardial SPECT 12–18 months post-PCI, found that SSS and SDS were signifi cant predictors of cardiac events, whereas the occurrence of cardiac events was higher in the presence of ischemia at SPECT in symptomatic and symptom-free patients (both P < 0.001). In a larger study, Zellweger et al. [18] evaluated 356 patients (both symptomatic and asymptomatic) with 99mTc-MIBI myo-cardial SPECT 6 months after stenting and the follow-up was for an average of about 4 years. A higher degree of ischemia measured by SDS was associated with a higher rate of death, myocardial infarction and repeat revascularization, whereas SPECT imaging added incre-mental information for the prediction of critical events. Moreover, similar results were reported by Solodky et al. [24] recently. Comparably, Zhang et al. [25] studied 318 patients who underwent a 99mTc-MIBI SPECT 10 ± 12 months after PCI reported that SSS was the best independent predictor for hard cardiac events (cardiac death or myocardial infarction), whereas SDS was the best independent predictor for late revascularization (both P < 0.001). Additionally, Elhendy et al. [41], study-ing 381 patients 4.5 ± 3.2 years after myocardial revas-cularization (CABG in 201 patients, PCI in 180 patients) with 99mTc-tetrofosmin, found that an abnormal MPI indicative of myocardial ischemia was independently associated with the composite endpoints of hard cardiac events and late revascularization, adding signifi cant incremental prognostic value to clinical data in the pre-diction of hard cardiac events (P < 0.01). However, in these studies, there is a general agreement that chest pain symptoms are of limited value in predicting future cardiac events, whereas Zellweger et al. reported a worse outcome of patients with symptomatic ischemia com-pared to patients with silent ischemia, although this was

Ann Nucl Med (2008) 22:899–909 907

1 3

closely related to the extent of ischemia, as it is evaluated with MPI [18, 22, 24, 25, 41].

Recently, Galassi et al. [23] studied 322 consecutive patients with 99mTc-tetrofosmin after incomplete revas-cularization and found that nuclear data provided signifi cant incremental prognostic value for composite cardiac events compared with the clinical, exercise testing, and angiographic fi ndings (P < 0.01).

Nevertheless, our study has some potential limita-tions. First, as we did not include any not included symptomatic patients in our study we did not evaluate the signifi cance of chest pain symptoms on long-term prognosis. Nevertheless, there are suffi cient published data demonstrating the prognosis of symptomatic patient post-PCI, reporting that cardiac symptoms are of limited clinical signifi cance, and this was beyond the scope of our study [22, 24, 25]. Besides, these patients are usually guided by the referring physician to an emergent MPI examination.

Second, no systematic follow-up angiography was performed and angiographic results are not included in this study. However, we have previously reported our results comparing myocardial perfusion SPECT imaging with 99mTc-tetrofosmin and coronary angiography to detect restenosis 6 months after PCI, and this was not the aim of the present study.

Third, using 8-frame/cardiac cycle acquisition, mild underestimation of LVEF could occur [35, 42]. Ideally, it would be desirable for the number of frames to be as high as possible to better follow changes in myocardial position and volumes; however, this consideration is bal-anced by the need for each projection image to contain adequate counts without markedly increasing the acqui-sition time or radiotracer dose. Moreover, LVEF under-estimation has been reported to be uniform over a wide range of ejection fractions, and thus we do not expect a signifi cant infl uence on data analysis [42]. On the other hand, it is generally agreed that 8-frame gated-SPECT imaging does not allow for meaningful measurement of LV diastolic function, and so we have not included dia-stolic function parameters in the analysis [42].

Finally, comparable with other investigators, we did not include deaths of noncardiac origin in hard events (all-cause mortality), to avoid blurred results [19, 21, 22, 24, 25]. However, we consider that our results are biased to null as the assessment of the cause of death was blinded to clinical, exercise, myocardial SPECT data and the hypothesis of the study.

Hence, our results suggest that the presence of risk factors is unable to predict cardiac events in asymptom-atic patients after stenting. These results are consistent with several previous studies although diabetes has been reported to be an independent predictor of critical events

[18, 22, 24, 25, 41]. A possible explanation could be that diabetic patients are closely followed and generally maintain well controlled glucose plasma levels, post-PCI. In addition, similar to other reports, the angio-graphic results pre-PCI and also the number of dilated vessels (complete or incomplete revascularization) did not prove to have of independent prognostic values [19, 22].

Exercise electrocardiogram has been repeatedly dem-onstrated to be inaccurate in detecting restenosis follow-ing PCI [10–17]. In our study, between exercise testing variables only angina during exercise testing was found to have a signifi cant independent predictive value for cardiac events. These results are in general concordance with the data of previous studies, despite the fact that they used other radiotracers for myocardial perfusion imaging [18, 19, 22, 23–25].

Conclusions: clinical implications

In conclusion, our results indicate that gated-SPECT using 99mTc-tetrofosmin as the radiotracer provides sig-nifi cant information on the risk of hard and soft events. The cut-off of 10 for SSS and 1.7 for SDS had signifi cant predictive ability for any cardiac event and could be useful in clinical practice. Furthermore, SPECT results yielded incremental prognostic information, in addition to that provided by clinical and exercise testing data for cardiac events.

References

1. Serruys PW, Luijten HE, Beatt KJ, Geuskens R, de Feyter PJ, van den Brand M, et al. Incidence of restenosis after successful coronary angioplasty: a time-related phenomenon. A quanti-tative angiographic study in 342 consecutive patients at 1, 2, 3, and 4 months. Circulation 1988;77:361–71.

2. Califf RM, Fortin DF, Frid DJ, Harlan WR, Ohman EM, Bengtson JR, et al. Restenosis after coronary angioplasty: an overview. J Am Coll Cardiol 1991;17:2B–13B.

3. Gruentzig AR, King SB, Schlumpf M, Siegenthaler W. Long-term follow-up after percutaneous transluminal coronary angioplasty: the early Zurich experience. N Engl J Med 1987;316;1127–32.

4. Goy JJ, Eeckhout E. Intracoronary stenting. Lancet 1998;351:1943–9.

5. Serruys P, de Jaegere P, Kiemeneij F, Macaya C, Rutsch W, Heyndrickx G, et al. A comparison of balloon-expandable-stent implantation with balloon angioplasty in patients with coronary artery disease. Benestent Study Group. N Engl J Med 1994;331:489–95.

6. Kastrati A, Schömig A, Elezi S, Dirschinger J, Mehilli J, Schühlen H, et al. Prognostic value of the modifi ed American College of Cardiology/American Heart Association stenosis morphology classifi cation for long-term angiographic and

908 Ann Nucl Med (2008) 22:899–909

1 3

clinical outcome after coronary stent placement. Circulation 1999;100:1285–90.

7. De Jaegere PP, Eefting FD, Popma JJ, Serruys PW. Clinical trials on intracoronary stenting. Semin Interv Cardiol 1996;1:233–45.

8. Fischman DL, Leon MB, Baim DS, Schatz RA, Savage MP, Penn I, et al. A randomized comparison of coronary-stent placement and balloon angioplasty in the treatment of coro-nary artery disease. Stent Restenosis Study Investigators. N Engl J Med 1994;331:496–501.

9. Giedd KN, Bergmann SR. Myocardial perfusion imaging fol-lowing percutaneous coronary intervention. The importance of restenosis, disease progression, and directed reintervention. J Am Coll Cardiol 2004;43:328–36.

10. Adams GL, Ambati SR, Adams JM, Borges-Neto S. Role of nuclear imaging after coronary revascularization. J Nucl Cardiol 2006;13:163–9.

11. Hecht HS, Shaw RE, Bruce TR, Ryan C, Stertzer SH, Myler RK. Usefulness of tomographic thallium-201 imaging for detection of restenosis after percutaneous transluminal coro-nary angioplasty. Am J Cardiol 1990;66:1314–8.

12. Hecht HS, Shaw RE, Chin HL, Ryan C, Stertzer SH, Myler RK. Silent ischemia after coronary angioplasty: evaluation of restenosis and extent of ischemia in asymptomatic patients by tomographic thallium-201 exercise imaging and comparison with symptomatic patients. J Am Coll Cardiol 1991;17:670–7.

13. Marie PY, Danchin N, Karcher G, Grentzinger A, Juilliere Y, Olivier P, et al. Usefulness of exercise SPECT-thallium to detect asymptomatic restenosis in patients who had angina before coronary angioplasty. Am Heart J 1993;126:571–7.

14. Milavetz JJ, Miller TD, Hodge DO, Holmes DR, Gibbons RJ. Accuracy of single-photon emission tomography myocardial perfusion imaging in patients with stents in native coronary arteries. Am J Cardiol 1998;82:857–61.

15. Kosa I, Blasini R, Schneider-Eicke J, Neumann FJ, Matsunari I, Neverve J, et al. Myocardial perfusion scintigraphy to evalu-ate patients after coronary stent implantation. J Nucl Med 1998;39:1307–11.

16. Milan E, Zoccarato O, Terzi A, Ettori F, Leonzi O, Nicoli L, et al. Technetium-99m-sestamibi SPECT to detect restenosis after successful percutaneous coronary angioplasty. J Nucl Med 1996;37:1300–5.

17. Georgoulias P, Demakopoulos N, Kontos A, Xaplanteris P, Thomadakis K, Mortzos G, et al. Tc-99m tetrofosmin myo-cardial perfusion imaging before and six months after percu-taneous transluminal coronary angioplasty. Clin Nucl Med 1998;23:678–82.

18. Zellweger MJ, Weinbacher M, Zutter AW, Jeger RV, Mueller-Brand J, Kaiser C, et al. Long-term outcome of patients with silent versus symptomatic ischemia six months after percuta-neous coronary intervention and stenting. J Am Coll Cardiol 2003;42:33–40.

19. Cottin Y, Rezaizadeh K, Touzery C, Barillot I, Zeller M, Prevot S, et al. Long-term prognostic value of 201Tl single-photon emission computed tomographic myocardial perfu-sion imaging after coronary stenting. Am Heart J 2001;141:999–1006.

20. Ho KT, Miller TD, Holmes DR, Hodge DO, Gibbons RJ. Long-term prognostic value of Duke treadmill score and exercise thallium-201 imaging performed one to three years after percutaneous transluminal coronary angioplasty. Am J Cardiol 1999;84:1323–7.

21. Kaminek M, Myslivecek M, Skvarilova M, Husak V, Koranda P, Lang O. Prognostic value of myocardial perfusion tomo-

graphic imaging in patients after percutaneous transluminal coronary angioplasty. Clin Nucl Med 2000;25:775–8.

22. Acampa W, Petretta M, Florimonte L, Mattera A, Cuocolo A. Prognostic value of exercise cardiac tomography performed late after percutaneous coronary intervention in symptomatic and symptom-free patients. Am J Cardiol 2003;91:259–63.

23. Galassi AR, Grasso C, Azzarelli S, Ussia G, Moshiri S, Tamburino C. Usefulness of exercise myocardial scintigraphy in multivessel coronary disease after incomplete revasculariza-tion with coronary stenting. Am J Cardiol 2006;97:207–15.

24. Solodky A, Assali AR, Mats I, Ben-Gal T, Kornowski R, Battler A, et al. Prognostic value of myocardial perfusion imaging in symptomatic and asymptomatic patients after percutaneous coronary intervention. Cardiology 2007;107:38–43.

25. Zhang X, Liu X, He ZX, Shi R, Yang M, Gao R, et al. Long-term prognostic value of exercise 99mTc-MIBI SPET myocardial perfusion imaging in patients after percutaneous coronary intervention. Eur J Nucl Med Mol Imaging 2004;31:655–62.

26. Rajagopal V, Gurm HS, Brunken RC, Pothier CE, Bhatt DL, Lauer MS. Prediction of death or myocardial infarction by exercise single photon emission computed tomography perfu-sion scintigraphy in patients who have had recent coronary artery stenting. Am Heart J 2005;149:534–40.

27. Coster CD, Quan H, Finlayson A, Gao M, Halfon P, Humphries K, et al. Identifying priorities in methodological research using ICD-9-CM and ICD-10 administrative data: report from an international consortium. BMC Health Ser-vices Research 2006;6:77–82.

28. Pitsavos CH, Chrysohoou C, Panagiotakos DB, Kokkinos P, Skoumas J, Papaioannou I, et al. Exercise capacity and heart rate recovery as predictors of coronary heart disease events, in patients with heterozygous familial hypercholesterolemia. Atherosclerosis 2004;173:347–52.

29. Fletcher GF, Balady G, Froelicher VF, Hartley LH, Haskell WL, Pollock ML. Exercise standards: a statement for health-care professionals from the American Heart Association Writing Group. Circulation 1995;91:580–615.

30. Lauer MS, Francis GS, Okin PM, Pashkow FJ, Snader CE, Marwick TH. Impaired chronotropic response to exercise stress testing as a predictor of mortality. JAMA 1999;281:524–9.

31. Georgoulias P, Orfanakis A, Demakopoulos N, Xaplanteris P, Mortzos G, Vardas P, et al. Abnormal heart rate recovery immediately after treadmill testing: correlation with clinical, exercise testing, and myocardial perfusion parameters. J Nucl Cardiol 2003;10:498–505.

32. Rahimi K, Thomas A, Adam M, Hayerizadeh BF, Schuler G, Secknus MA. Implications of exercise test modality on modern prognostic markers in patients with known or suspected coro-nary artery disease: Treadmill versus bicycle. Eur J Cardiovasc Prev Rehabil 2006;13:45–50.

33. Beller GA, Zaret BL. Contributions of nuclear cardiology to diagnosis and prognosis of patients with coronary artery disease. Circulation 2000;101:1465–78.

34. Kapur A, Latus KA, Davies G, Dhawan RT, Eastick S, Jarritt PH, et al. A comparison of three radionuclide myocardial perfusion tracers in clinical practice: the ROBUST study. Eur J Nucl Med 2002;29:1608–16.

35. Nakajima K, Kusuoka H, Nishimura S, Yamashina A, Nishimura T. Normal limits of ejection fraction and volumes determined by gated SPECT in clinically normal patients without cardiac events: a study based on the J-ACCESS data-base. Eur J Nucl Med Mol Imaging 2007;34:1088–96.

Ann Nucl Med (2008) 22:899–909 909

1 3

36. Cerqueira MD, Weissman NJ, Dilsizian V, Jacobs AK, Kaul S, Laskey WK, et al. Standardized myocardial segmentation and nomenclature for tomographic imaging of the heart: a statement for healthcare professionals from the Cardiac Imaging Committee of the Council on Clinical Cardiology of the American Heart Association. J Nucl Med 2002;9:240–5.

37. Cox DR. Regression models and life tables. J R Stat Soc B 1972;34:187–202.

38. Schinkel AFL, Elhendy A, van Domburg RT, Bax JJ, Vourvouni EC, Bountioukos M. Incremental value of exercise technetium-99m tetrofosmin myocardial perfusion single-photon emission computed tomography for the prediction of cardiac events. Am J Cardiol 2003;91:408–11.

39. Sridhara BS, Braat S, Rigo P, Itti R, Cload P, Lahiri A. Com-parison of myocardial perfusion imaging with technetium-

99m tetrofosmin versus thallium-201 in coronary artery disease. Am J Cardiol 1993;72:1015–9.

40. Georgoulias P, Demakopoulos N, Kontos A, Xaplanteris P, Tomadakis K, Mortzos G, et al. 99mTc-tetrofosmin myocar-dial perfusion imaging: a comparison with coronary angiog-raphy. Nuklearmedizin 1996;35:153–5.

41. Elhendy A, Schinkel AF, van Domburg RT, Bax JJ, Valkema R, Poldermans D. Risk stratifi cation of patients after myocar-dial revascularization by stress Tc-99m tetrofosmin myocar-dial perfusion tomography. J Nucl Cardiol 2003;10:615–22.

42. Germano G, Berman DS. Regional and global ventricular function and volumes from single-photon emission computed tomography perfusion imaging. In: Zarret BL, Beller GA III, editors. Clinical nuclear cardiology. state of the art and future directions. Philadelphia: Elsevier Mosby; 2005. p. 189–202.