|
| HOME | SUBSCRIPTIONS | CURRENT ISSUE | PAST ISSUES | CARDIOSOURCE | SEARCH | HELP | FEEDBACK |
|
|
|
|||||||||
|
|||||||||||
|
J Am Coll Cardiol, 2006; 48:1029-1039, doi:10.1016/j.jacc.2006.06.025
(Published online 15 August 2006). © 2006 by the American College of Cardiology Foundation |
Division of Cardiology, University of Ottawa Heart Institute, Ottawa, Ontario, Canada.
Manuscript received October 3, 2005; revised manuscript received January 26, 2006, accepted March 7, 2006.
* Reprint requests and correspondence: Dr. Rob S. B. Beanlands, University of Ottawa Heart Institute, 40 Ruskin Street, Ottawa, Ontario, Canada K1Y 4W7. (Email: rbeanlands{at}ottawaheart.ca).
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
BACKGROUND: 82Rb PET MPI accurately diagnoses coronary artery disease (CAD). However, there are limited data evaluating its prognostic value.
METHODS: Follow-up (3.1 ± 0.9 years) was obtained in 367 patients who underwent dipyridamole 82Rb PET MPI. Patients were divided into groups based on their summed stress score (SSS): group I, normal (<4); group II, mild (4 to 7); and group III, moderate (8 to 11) to severe (
12).
RESULTS: There were significant differences among patients in the 3 SSS groups for hard events (cardiac death and myocardial infarction [MI]) (p < 0.001) and total cardiac events (hard events, revascularization and hospitalization) (p < 0.001). The annual hard events rates were 0.4%, 2.3%, and 7.0% in the normal, mild, and moderate-severe groups, respectively. In adjusted survival models, 82Rb PET SSS was the strongest predictor of total cardiac events and a significant predictor of hard events. Among patients referred for PET after 99mTc single-photon emission computed tomography, the annual total event rate was higher with abnormal versus normal SSS on PET (15.2% vs. 1.3%, p < 0.001). In patients with obesity, the annual total event rate was 11.1% with an abnormal scan and 1.5% with a normal scan (p < 0.001).
CONCLUSIONS: This study shows that 82Rb PET MPI has significant prognostic value for predicting cardiac events, including death and MI. It also seems to have prognostic value in patients whose diagnosis remains uncertain after single-photon emission computed tomography MPI and in obese patients. The prognostic value of PET MPI may improve the management of cardiac patients.
| ||||||||||||||
Positron emission tomography (PET) MPI has been widely used as an accurate means to diagnose CAD (8–14). Rubidium-82 (82Rb) is a PET perfusion tracer produced from a strontium-82 (82Sr)/82Rb generator and is widely used in centers without immediate access to a cyclotron (9,11,13–16). Rubidium-82 PET MPI is highly specific and sensitive for the diagnosis of CAD (11,13,14,17–19). However, data supporting the prognostic value of 82Rb PET MPI are limited (20). The primary goal of the current study was to evaluate the prognostic value of 82Rb PET imaging for the prediction of cardiac events.
Although SPECT MPI is a highly useful noninvasive diagnostic modality for the detection of CAD, clinicians are sometimes faced with equivocal or nondiagnostic results. Patients with equivocal scans have increased risk compared with a normal population (21). Further risk stratification of these patients is often required. Rubidium-82 PET MPI may be clinically useful for this indication (22). However, data to support this approach are in fact quite limited, and no studies have evaluated the prognostic value of PET MPI in such patients. A secondary objective of this study was to evaluate the prognostic value of 82Rb PET MPI among a subset of patients referred for 82Rb PET imaging after SPECT perfusion imaging.
The prevalence of obesity (body mass index [BMI] 30 kg/m2) has dramatically increased since 1990 with >20% of the U.S. adult population being reported as obese (23,24). Obesity has a significant adverse impact on cardiovascular morbidity (25,26). Obese patients require accurate diagnostic tools for the evaluation of their cardiac risk. Current modalities often have diagnostic difficulties in the obese population (27–30). Although PET MPI is often applied in patients with obesity, its prognostic value in this population has not been well studied. An additional secondary objective of the current study was to evaluate the prognostic value of 82Rb PET MPI in a subset of patients with obesity.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Because abnormal scans can result in early revascularization and potentially alter outcome, 11 (2.9%) patients with early revascularization (in the first 3 months after 82Rb PET MPI) were excluded from the prognostic portion of the analysis. Similar approaches have been applied in previous studies (2,3,31,32). Thus, the prognostic data presented in the current study are based on a population of 367 patients.
The study was approved by the University of Ottawa Heart Institute Human Research Ethics Board. Informed consent was obtained from all subjects. In cases in which patients died before follow-up, consent was obtained from the next of kin.
PET acquisition protocol. Patients were positioned in a whole-body PET scanner (ECAT ART, Siemens/CTI Knoxville, Tennessee), a 4-min transmission scan was performed for attenuation correction (33). Immediately after the transmission scan, 400 to 2000 MBq (8 MBq/kg) of 82Rb was administered intravenously over 1 min. A 7.5-min acquisition was initiated 2.5 min after tracer administration (14,33–36).
Dipyridamole myocardial perfusion 82Rb PET imaging.
Patients were instructed to fast for 6 h and to abstain from caffeine-containing products for >12 h before pharmacological stress. After the rest 82Rb PET imaging, a pharmacological stress test was performed with the infusion of dipyridamole (0.14 mg/kg/min over 5 min). Stress imaging was acquired using the identical protocol as the rest imaging. The 82Rb (400 to 2,000 MBq [8 MBq/kg]) was infused 3 min after completion of the dipyridamole infusion (33,35,36). During dipyridamole stress, symptoms, heart rate, blood pressure, and the electrocardiogram were monitored continuously.
PET image interpretation.
Regional myocardial perfusion was assessed using the 17-segment model recommended by the Cardiac Imaging Committee of the American Heart Association and the American Society of Nuclear Cardiology (36,37). Each segment was scored by two expert observers blinded to clinical data using a five-point scoring system (0 = normal, 1 = mildly reduced, 2 = moderately reduced, 3 = severely reduced, 4 = absent uptake) (36,38,39). Differences between observers were settled by a third observer. The sum of segment scores at stress (summed stress score [SSS]), scores at rest (summed rest score [SRS]), and differences between stress and rest score (summed difference score [SDS]) were calculated (4). Patients were divided into groups based on their SSS (3,4,14,21). Because we used a 17-segment model according to American Society of Nuclear Cardiology recommendation (36), we adjusted the validated SPECT 20-segment model scoring to be equivalent percentages of the total possible score in the myocardium. Thus, a scan was considered normal if the SSS was <4, mildly abnormal if SSS was between 4 and 7, moderately abnormal if SSS was between 8 and 11, and severely abnormal if SSS was 12. This normal/abnormal cutoff represents a similar percentage of the myocardium used as a cutoff score in previous PET studies (14). The moderately abnormal and severely abnormal groups were combined because of the small number of patients in the moderately abnormal group. The SSS was the primary focus of our analysis.
Patients were also divided into groups based on their SDS combined with their SSS: patients with a SDS 2 were considered to have myocardial ischemia; patients with an SDS <2 with normal SSS were considered normal; patients with SDS <2 but abnormal SSS were considered to have scar without ischemia.
Image quality was evaluated by two experts blinded to clinical or other imaging data and scored on a three-point scale: suboptimal = 0 (artifact such that a definitive conclusion could not be made), fair = 1 (an artifact but interpretation was made), good = 2 (conclusive interpretation).
Patient follow-up. Patient follow-up was performed using a scripted telephone interview by health care professionals blinded to the patient’s MPI data. The mean follow-up period was 3.1 ± 0.9 years. Cardiac events were confirmed using the best information that was available. Hard cardiac events were defined as either cardiac death (confirmed by review of death certificate, hospital chart, or physician’s records) or nonfatal MI (considering the appropriate combination of elevated cardiac enzymes, electrocardiogram changes, and ischemic symptoms, from hospital chart or physician’s records). Late (>3 months) revascularization (coronary artery bypass grafting [CABG] or percutaneous coronary intervention [PCI]) or the need for other hospitalization for cardiac causes (e.g., unstable angina, chest pain, heart failure) other than MI were defined as other cardiac events. Hence total cardiac events were a composite of cardiac death, nonfatal MI, late revascularization, and cardiac hospitalization.
If the patients had hard cardiac events, patients in the study only had one hard cardiac event, death or nonfatal MI. For total cardiac events, some patients had two or more events (e.g., nonfatal MI and revascularization). In such patients, the first event date was considered for the survival analysis.
Subgroup populations—patients referred for PET after SPECT MPI. To further evaluate the prognostic value of 82Rb PET MPI, we assessed the subgroup of patients with recent 99mTc SPECT MPI. The goal of this ancillary subgroup analysis was to evaluate the prognostic value of 82Rb PET MPI in cases of patients referred after SPECT, in which uncertainty about the diagnosis remained either based on the SPECT report or in which SPECT findings did not support the clinical impression. Clinical SPECT MPI reports were reviewed by two observers blinded to 82Rb PET imaging data to categorize SPECT image findings and image quality.
A subgroup of 102 patients was identified who had had SPECT MPI within 6 months before 82Rb PET MPI studies. Among these, one patient who had unstable angina between SPECT and 82Rb PET MPI was excluded. Two patients had early revascularization. These two patients were included in image quality analysis but excluded from the secondary outcome analysis. Five patients with thallium-201 (201Tl) SPECT MPI were also excluded. In addition, four patients had clear-cut moderate-severe abnormalities on SPECT that had similar location, extent, and severity of defect as their 82Rb PET MPI studies. These patients were included in the 99mTc SPECT versus 82Rb PET image quality analysis but excluded from the secondary outcome analysis, so as not to bias the prognostic value data in favor of PET. Therefore, the subgroup was composed of 96 patients for image quality and 90 patients for secondary outcome analysis, who were referred for PET after 99mTc SPECT MPI.
Among these 96 patients, 50 patients had exercise; 45 had dipyridamole stress 99mTc SPECT MPI, and 1 had rest SPECT imaging but did not get dipyridamole because of very poor image quality at rest, rendering the images equivocal.
99mTc SPECT imaging. The SPECT MPI was performed at either the University of Ottawa Heart Institute (n = 92) or other referring institutions (n = 10). For 99mTc SPECT MPI studies, a same-day, rest-stress 99mTc tetrofosmin or 99mTc sestamibi injection protocol was used. For the rest studies, tomographic imaging was initiated 30 to 45 min after 296 to 370 MBq of 99mTc tetrofosmin or 99mTc sestamibi tracer injection. The 99mTc tomographic images were obtained using a standard 30-min 180° SPECT acquisition protocol (40).
The stress study was performed approximately 120 min after the rest imaging. For patients who underwent symptom-limited treadmill exercise using the Bruce protocol, a dose of 925 to 1,110 MBq of 99mTc tetrofosmin or 99mTc sestamibi was injected at peak exercise and the patients continued exercise for an additional 1 to 2 min. For those who underwent dipyridamole stress testing, 3 min after the completion of dipyridamole (0.14 mg/kg/min over 5 min), tracer was injected intravenously. Tomographic imaging was initiated 30 to 45 min after the tracer injection. The same acquisition protocol for rest was used for stress (40). Based on clinical reports, a similar protocol was followed for the patients referred from other institutes.
Image quality and diagnostic certainty with 99mTc SPECT and 82Rb PET MPI. All clinical SPECT reports were reviewed. Image quality was classified as described for 82Rb PET MPI; suboptimal = 0, fair = 1, good = 2.
Regarding diagnostic certainty of the images, reports that were definitively indicated as normal or abnormal were considered conclusive, whereas diagnostic uncertainty was defined as studies in which reports were not conclusive and were considered to be probably normal, equivocal, or probably abnormal (41). Diagnostic certainty was scored as conclusive = 1 or uncertainty = 0. A total of 96 patients were evaluated for image quality and diagnostic certainty analysis.
PET MPI in patients with obesity.
There are limited data on the value of MPI in patients with obesity (defined as BMI 30 kg/m2) (42); thus, we further evaluated the prognostic value of PET MPI in this subgroup.
The mean BMI of our study population was 30.0 ± 14.3 kg/m2. There were 134 (36.5%) patients with BMI 30 kg/m2 (obese), 136 (37.1%) patients with BMI 25 to 29.9 kg/m2 (overweight), and 97 (26.4%) patients with a BMI <25 kg/m2 (normal). The subgroup population for this ancillary analysis was the obese population (n = 134).
Statistical analysis. Continuous measures are presented as means ± SD. For the univariable analysis of differences, Wilcoxon rank sum tests were used for continuous measures. Categorical measures are presented as frequencies with percentages. Frequencies were compared using Fisher exact tests for the univariable analysis.
The prognostic value of SSS for unadjusted as well as adjusted hard (cardiac death and MI) and total cardiac event rates were assessed for the full population. All unadjusted comparisons of event rates were based on survival analysis log-rank tests. Annual event rates were calculated by dividing the 3-year Kaplan-Meier event rates by 3.
For the risk-adjusted analysis, multivariable Cox proportional hazard models were used to assess the independent prognostic value of the SSS controlling for baseline patient characteristics and creating adjusted survival curves. To prevent overfitting of the model, the number of variables considered for modeling were reduced by combining the coronary risk factors (hyperlipidemia, hypertension, diabetes, smoking, family history of CAD) into one variable (>2 coronary risk factors). Also to limit the number of variables, the baseline variables that showed significance in the univariable analysis were considered individually with SSS in the model. From this testing, for the hard cardiac events, age >65 years, history of MI, and known CAD met the threshold for entry of p < 0.05. These variables were considered together in the stepwise selection of the final multivariable model for hard cardiac events (cardiac death and MI), which included only three covariates (SSS, age >65 years, and history of MI). For total cardiac events, men, history of MI and CABG, known CAD, angina symptoms, and patients with two or more coronary risk factors met the threshold for entry of p < 0.05. These variables were considered together in the stepwise selection of the final multivariable model for total cardiac events, which included four covariates (SSS, gender [male], history of MI, and two or more coronary risk factors).
The multivariable models with and without SSS were produced to determine the incremental value of SSS. A statistically significant increase in the global chi-square value of the models with SSS defined an incremental prognostic value of SSS. The percent of contribution of each covariate to the global chi-square value was also reported to show the relative importance of the variables in the hard (cardiac death and MI) and total cardiac events models.
For the secondary analyses of the subgroups populations, the small number of events combined with the small sample size in the subgroups limited the ability to perform the multivariable analysis. Thus, only unadjusted survival analysis was performed for these subpopulations.
The concordance between 82Rb PET and 99mTc SPECT MPI for classifying the image quality and diagnostic certainty was determined using kappa statistics. The mean image quality score was compared using the Wilcoxon signed rank test. The comparison of diagnostic certainty of the images (conclusive = 1, uncertainty = 0) was performed using the McNemar test.
A p value < 0.05 was considered statistically significant. Statistical calculations were carried out using SAS software (version 9.1.2, SAS Institute Inc., Cary, North Carolina).
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
|
Univariable analysis of baseline demographics for cardiac events. The results of univariable analysis of baseline demographics (not including PET parameters) are summarized in Table 2. Patients with hard cardiac events (cardiac death and nonfatal MI) were older, were more often male, and had a greater history of MI or CABG, known CAD, angina symptoms, and diabetes mellitus. Patients with any cardiac event were older, were more often male, and had a greater history of MI, CABG, PCI, known CAD, angina symptoms, coronary risk factors (2 or more), diabetes mellitus, and smoking.
|
|
|
Adjusted survival curves illustrate that SSS groups were significantly different for hard cardiac events (cardiac death and nonfatal MI) (Fig. 1A) and total cardiac events (Fig. 1B).
|
2); group III, scar without ischemia (SDS <2; abnormal SSS). The unadjusted survival free of hard cardiac events (cardiac death and nonfatal MI) was significantly different among patients with scar without ischemia (group III), those with ischemia (group II), and those with normal SSS and no ischemia (group I) (p < 0.001 for hard cardiac events, p < 0.001 for cardiac death). Ischemia-related events (nonfatal MI and late revascularization) were significantly different among patients with ischemia (group II), those with scar without ischemia (group III), and those with normal SSS and no ischemia (group I) (p = 0.034 for MI, p < 0.001 for late revascularization).
|
Compared with the main study cohort, the patients who were referred for PET after 99mTc SPECT were younger, had less history of revascularization, had less known CAD, and had less severe SSS compared with the total cohort (total cohorts vs. SPECT patients: age, 59.4 ± 10.9 vs. 56.6 ± 9.8 years [p = 0.023]; history of CABG, 12.3% vs. 3.3% [p = 0.012]; history of PCI, 20.2% vs. 5.6% [p < 0.001]; known CAD, 40.3% vs. 23.3% [p = 0.003]; SSS distribution [p = 0.002], respectively). There was a trend for less history of MI (p = 0.051).
These patients had high BMI similar to the total cohort. There was no significant difference in gender between the two cohorts (gender [male], 45.8% vs. 48.8%, p = 0.638). There was also no significant difference in other parameters, including follow-up time, angina symptoms, chronic heart failure class, and coronary risk factors (p = NS).
Image quality comparison between 99mTc SPECT AND 82Rb PET MPI
In this specific study population, the overall concordance of image quality between 99mTc SPECT and 82Rb PET MPI was 22% (21 of 96, kappa = 0.02) (Table 6). The image quality was superior in PET compared with SPECT. The mean quality score of PET was 1.94 ± 0.24, and that of SPECT was 0.71 ± 0.78 (p < 0.001).
|
Cardiac events in patients referred for PET after SPECT MPI
Among the 90 patients included in the outcome analysis, 11 (12.2%) patients had an abnormal 82Rb PET SSS (8 mildly abnormal and 3 severely abnormal) and 79 patients had a normal SSS.
During follow-up, there were a total of 11 cardiac events: 1 nonfatal MI, 7 late revascularizations, and 3 cardiac hospitalizations. One patient had both a nonfatal MI and late revascularization. Thus, 10 patients (11.1%) suffered at least one cardiac event.
Among 11 patients with an abnormal 82Rb PET SSS, 7 patients had a cardiac event with a high annual total cardiac event rate (15.2%), compared with 3 patients suffering a cardiac event in those with normal 82Rb PET MPI, for an annual total cardiac event rate of 1.3%. Unadjusted survival free of any (total) cardiac events was significantly different between normal and abnormal SSS groups (p < 0.001) (Fig. 2).
|
During follow-up, there were a total of 22 cardiac events: 5 cardiac deaths, 2 nonfatal MIs, 8 late revascularizations, and 7 cardiac hospitalizations occurred in 20 patients. In the subgroup of 134 obese patients, 45 patients with an abnormal SSS had 7 hard cardiac events, with a high annual hard cardiac event rate (6.0%) compared with 0% in those with normal 82Rb PET MPI (Table 7). Unadjusted survival free from hard cardiac events was significantly different between normal and abnormal SSS groups (p < 0.001).
|
|
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
In subgroup analyses the data suggests that in patients with recent SPECT MPI and in those with obesity, 82Rb PET MPI has prognostic value and is able to identify higher-risk patients.
Prognostic value of PET MPI. The PET MPI has high spatial and temporal resolution compared with other nuclear imaging techniques (11,16). It yields high sensitivity and specificity for the detection of CAD (11,18,19,43). However, data supporting the prognostic value of PET MPI are limited.
One previous study by Marwick et al. (20) evaluated the prognostic value of 82Rb PET MPI in a population composed mostly of patients with known CAD (70%). They showed that 82Rb PET defect severity was associated with cardiac death and total events rates. Mortality rates for normal and abnormal defects were 0.9% per year and 4.3% per year. Our current study supports the notion that cardiac events increase with the severity of 82Rb PET MPI SSS. However, the two studies differ in patient population and event rates. In the study by Marwick et al. (20), most of the patients (70%) had CAD, a large number (22%) of patients had early revascularization, and normal perfusion on PET MPI was present in only 24%. In contrast, the current study, 71% of patients had normal stress perfusion and 2.9% patients had early revascularization. Thus, Marwick et al. (20) studied a higher-risk population than did our current study. This higher-risk population may not reflect the patients for whom PET has been suggested to be most useful and cost-effective, namely those with a low and intermediate pretest likelihood of CAD (44).
Our patient population more closely resembles recent SPECT study populations in which 20% to 50% of patients had known CAD (3,6,21,32,45,46) and early revascularization rates of <7% (3). In the current study, 40% of patients were known to have CAD, which may explain the lower event rates and lower early revascularization rate compared with the study by Marwick et al. (20). Our data indicate that the extent of 82Rb PET MPI SSS has prognostic value and applicability in a lower-risk patient population. Also of note, the current study showed that in the adjusted survival models, 82Rb PET MPI SSS was the strongest predictor of total cardiac events and was also a significant predictor of hard cardiac events.
Marwick et al. (20) reported that ischemia was not related to cardiac death and total cardiac events, which they attributed to their high early revascularization rate (22%) among patients with ischemia. This high revascularization rate made it more difficult to evaluate the prognostic value of ischemia on 82Rb PET MPI. In the present study, the cardiac death rate was significantly higher in patients with fixed defects (scar) compared with patients with ischemia or normal scans. However, ischemic events (nonfatal MI and late revascularization) were greater among the ischemic group. The current data suggest that the extent of ischemia as identified by 82Rb PET MPI may also provide useful prognostic information.
We have recently reported a very low hard cardiac event rate (0.2% over 27 months of follow-up; 0.09% per year) in a group of patients with 82Rb PET MPI clinically reported as normal (47). This very low event rate differs from the current study and is attributed to: 1) a more strict definition of normal (based on clinical reports); 2) shorter follow-up; and 3) exclusion of patients with left bundle branch block or ventricular pacing, in the previous study. In addition to the longer follow-up, broader inclusion criteria, and use of SSS, the current study also evaluated the prognostic value of abnormal SSS on PET MPI.
The current study also differs from previous work and is unique because it evaluates the prognostic value in important subgroups in which PET may have additional value, namely in patients whose diagnosis remains uncertain after SPECT MPI and in patients with obesity.
Prognostic value of SPECT MPI. The prognostic value of exercise and pharmacological SPECT MPI has been well established (5–7). Previous SPECT studies showed annual hard event rates of 0.2% to 0.8% for a normal SPECT MPI study (2,3,21). The demographics and SSS in the current study are similar to those of previous SPECT studies (2,3,21). Also, the increasing event rate with worsening defect size along with the excellent event free survival in patients with normal PET MPI are consistent with those of previous SPECT studies (2–4,21,45).
Previous SPECT MPI studies showed that cardiac death is better predicted by extent of myocardial scar, whereas myocardial ischemia seems to predict ischemic events (6,32,48). Similar findings were observed in the current study for PET MPI.
Prognostic value of PET in patients referred after SPECT MPI. An equivocal SPECT MPI study is an American College of Cardiology/American Heart Association/American Society of Nuclear Cardiology guidelines Class 1 indication for the use of PET MPI (22). This recommendation was based on consensus given the advantages of PET imaging. However, studies evaluating this approach have been limited. MacIntyre et al. (49) reported that patients who had false-negative 201Tl SPECT but true positive 82Rb PET had a significant change in their treatment strategies. However, to our knowledge, there has been no previous study evaluating the prognostic value of PET in the clinical setting when the diagnosis remains uncertain after SPECT MPI. The importance of this group of patients has been evaluated by Berman et al. (21), who reported that an equivocal SPECT study was associated with increased cardiac events and thus these patients may require additional risk stratification. In the current study, a subgroup of 90 patients whose diagnosis remained uncertain after 99mTc SPECT MPI was evaluated and had been referred for 82Rb PET MPI. In this subgroup, the annual total cardiac event rate was higher in patients with abnormal 82Rb PET MPI SSS than in those with normal PET MPI (15.2% vs. 1.3%). Unadjusted survival from total cardiac events was significantly different between normal and abnormal PET MPI SSS groups (p < 0.001). This study provides data to support the consensus opinion in the American College of Cardiology/American Heart Association/American Society of Nuclear Cardiology guidelines for PET MPI. Further outcome studies in this population are required.
In this specific subpopulation of the current study, the image quality concordance between 82Rb PET MPI and 99mTc SPECT MPI was poor. Image quality was superior with 82Rb PET compared with 99mTc SPECT MPI. In these selected patients, the added value of PET likely reflects its accuracy and its ability to correct for attenuation. Attenuation-corrected SPECT MPI is currently applied in limited centers (50). New methods of attenuation correction are being developed for both SPECT and PET. These may further improve diagnostic accuracy and prognostic value for both modalities. Further studies are needed to evaluate these evolving approaches.
Prognostic value of PET in obesity. Accurate diagnostic tools are required to evaluate the cardiac risk of obese patients. However, standard myocardial imaging with SPECT MPI and echocardiography may have technical difficulties in some patients with obesity (27–30). In the present study, 82Rb PET MPI distinguished cardiac event risk in obese patients with known or suspected CAD. The annual hard cardiac event rates were 0% in those with a normal SSS and 6.0% in those with an abnormal SSS, indicating low and high risk, respectively (5,7,51).
The high spatial resolution and attenuation correction of PET help explain the current findings in the obese population (11). The current study supports the clinical utility of 82Rb PET MPI in patients with obesity.
Study limitations. In the present study, the total hard event rate was low, at 4.6% in the overall study population. However, this hard event rate is similar to previous SPECT studies, which range between 2.6% and 5.5% in patients with suspected or known CAD (3,21,45,46,52). Our sample size and number of hard cardiac events were smaller. However, these were sufficient to detect significant differences in hard and total cardiac event rates among 82Rb PET MPI SSS groups. Post-hoc sample size calculations showed a power >80% with the univariable unadjusted survival analysis log-rank test for these comparisons.
The multivariable Cox proportional hazard models are limited by the number of events. Accordingly, to prevent overfitting of the multivariable Cox proportional hazard models, we reduced the number of variables entered into the model as described in the statistical methods section. With the adjusted model for cardiac death and nonfatal MI only containing three variables (SSS, age, and history of MI), the benefits of 10 events per variable are almost achieved. With the adjusted model for total cardiac events containing four variables and more than 10 events per variable, the results are more powerful. Post-hoc sample size calculations for the adjusted models for total events and hard events showed the power to be 83% and 64%, respectively. Statistically significant differences in the event-free survival (for both total events and hard events) were observed in the current study. However, an adjusted model approach for hard events in future studies may require a larger sample size.
In the current study, 8.3% of patients were lost to follow-up. This is somewhat higher than some but is similar to other previous reports, with rates ranging from 3.8% to 7.0% (3,21,32,45). This may in part be related to the timing of follow-up, which was longer than in most previous SPECT studies, and to the fact that some of our patients are referred from other centers and regions, which can sometimes make follow-up more difficult. Although it is not possible to determine whether events occurred or not, the similarity with the main population suggests that there was no bias of demographic or SSS among the patients who were lost to follow-up.
Recently, electrocardiogram gated acquisition has been applied to PET MPI; this as well as left ventricular mass and flow measurements may add value to PET imaging data. However, measurement of left ventricular mass, flow quantification, and gated acquisition were not routinely performed when these patients underwent their PET imaging in the year 2000. Thus data on these parameters are not available.
The SPECT subgroup was a selective population; therefore, the data must be interpreted with caution. The sample size was small but still large enough to show a significant difference in total cardiac event rates for normal compared with abnormal 82Rb PET MPI. Post-hoc sample size calculations showed a power >80% with the univariable unadjusted survival analysis log rank test for this comparison. The small number of events in this subgroup limited the ability to perform the multivariable analysis. Events in this subset were predominantly revascularization or hospitalization. Berman et al. (21) reported that increased total cardiac events among patients with equivocal SPECT were also predominantly revascularizations. Further study with a larger sample size is required in this population.
The sample size for the obese subgroup was also small. However, post-hoc sample size calculations also showed a power >80% with the univariable unadjusted survival analysis. Although, the present data show significant differences in hard and total cardiac event rates for normal compared with abnormal 82Rb PET MPI, further studies are required with a larger cohorts to confirm our results.
Conclusions. The 82Rb PET MPI has significant prognostic value for predicting cardiac events. Normal-perfusion PET carries an excellent prognosis. With the adjusted survival model, 82Rb PET SSS was the strongest predictor of total cardiac events and was also a significant predictor of hard cardiac events. The 82Rb PET MPI also seems to have prognostic value in patients whose diagnosis remains uncertain after SPECT MPI and in patients with obesity. The prognostic value of PET MPI may improve the management of cardiac patients.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  K. L. Gould, T. Pan, C. Loghin, N. P. Johnson, and S. Sdringola Reducing Radiation Dose in Rest-Stress Cardiac PET/CT by Single Poststress Cine CT for Attenuation Correction: Quantitative Validation J. Nucl. Med., May 1, 2008; 49(5): 738 - 745. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. R. Vesely and V. Dilsizian Nuclear Cardiac Stress Testing in the Era of Molecular Medicine J. Nucl. Med., March 1, 2008; 49(3): 399 - 413. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J Knuuti and F M Bengel Positron emission tomography and molecular imaging Heart, March 1, 2008; 94(3): 360 - 367. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. M. Bateman Business Aspects of Cardiovascular Computed Tomography: Tackling the Challenges J. Am. Coll. Cardiol. Img., January 1, 2008; 1(1): 111 - 118. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. L. Gould Not All Randomized Trials Are Equal J. Am. Coll. Cardiol., November 13, 2007; 50(20): 2013 - 2015. [Full Text] [PDF]
|
|
|
|
|
|
R. J. Gibbons, P. A. Araoz, and E. E. Williamson The Year in Cardiac Imaging J. Am. Coll. Cardiol., September 4, 2007; 50(10): 988 - 1003. [Full Text] [PDF]
|
|
|
|
|
|
K. L. Gould, T. Pan, C. Loghin, N. P. Johnson, A. Guha, and S. Sdringola Frequent Diagnostic Errors in Cardiac PET/CT Due to Misregistration of CT Attenuation and Emission PET Images: A Definitive Analysis of Causes, Consequences, and Corrections J. Nucl. Med., July 1, 2007; 48(7): 1112 - 1121. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. F. Di Carli, S. Dorbala, J. Meserve, G. El Fakhri, A. Sitek, and S. C. Moore Clinical Myocardial Perfusion PET/CT J. Nucl. Med., May 1, 2007; 48(5): 783 - 793. [Abstract] [Full Text] [PDF]
|
|
