Resting Radionuclide Myocardial Perfusion Imaging in a Chest Pain Center Including an Overnight Delayed Image Acquisition Protocol

MATERIALS AND METHODS

RESULTS

Four-hundred eighty-four patients were enrolled. Five were lost to follow-up, resulting in 479 patients in the final analysis. The study population demographics are shown in Table 1. The age ranged from 20 to 80 y with a mean age of approximately 45 y. The ratio of males to females was approximately equal. African–Americans made up the largest racial demographic, followed by Whites. Cardiovascular risk factor prevalence was typical for a low-risk population with the exception of smoking and cocaine abuse, both of which were high. Cardiovascular medication use was low in the study population. No individual class of medication was used in >10% of the study population.

Subjects underwent a protocol-driven evaluation in the CP center and received gated SPECT MPI. Four-hundred thirty-four patients had normal resting MPI and 45 had abnormal MPI consistent with ischemia. In the normal MPI group, 3 (0.7%) patients had cardiac events. In the abnormal MPI group, 10 (22.2%) patients had cardiac events (Table 2). The difference in cardiac events between the 2 groups was highly significant statistically (P < 0.001). The test characteristics of gated SPECT MPI in predicting 30-d events were as follows: sensitivity, 76.9%; specificity, 92.4%; positive predictive value, 22.2%; and negative predictive value, 99.3%.

Among the 3 patients in the normal MPI group with cardiac events, 2 had abnormal coronary angiographic studies and PCI and the third had a MI on the same day as the scan. This last patient was later found to have a history of CAD and presented with CP in the setting of significant anemia and several comorbidities. Cardiac biomarkers were mildly elevated, and the patient was treated conservatively. The details of these 3 “false-negative” scans are shown in Table 3. CP had resolved for >2 h before tracer injection in each of the 3 patients. Patients with cardiac events in the abnormal MPI group included 4 cases of MI undergoing PCI and 1 case of MI managed conservatively. The remaining 5 patients had significant coronary stenosis at catheterization and underwent PCI.

Forty-four (9.1%) patients presented in the early morning hours and, therefore, underwent a delayed-image acquisition protocol. Details of this patient group are shown in Table 4. Most patients received tracer injection during or shortly after the cessation of CP. The average time from injection to image acquisition was 3 h 20 min, with the longest delay remaining under 6 h. Forty-one of these patients had normal scans and were free of cardiac events. Three had abnormal perfusion scans without cardiac events. All 3 had normal coronary angiography or normal follow-up stress MPI. None of the patients undergoing delayed image acquisition had a normal perfusion scan and later suffered a cardiac event.

DISCUSSION

Gated SPECT MPI has been shown to add significant diagnostic information to the standard ED CP evaluation in previous studies (3,5,6). One of the impediments to implementation of gated-SPECT MPI in EDs has been the logistic challenge of providing prompt image acquisition and interpretation 24 h per day, 7 d per week. The present study illustrates that—through the use of prompt tracer injection combined with flexible nuclear medicine technologist scheduling and image interpretation by local teleradiology—accurate results can be obtained. Previous reports of resting MPI in the ED excluded patients presenting in the late evening, in the early morning, or on weekends when nuclear cardiology staff are unavailable (9,10). A portion of our study population underwent a delayed imaging protocol. This protocol involved radionuclide injection by trained personnel in the early morning hours with image acquisition after 6 am when nuclear medicine technologists arrived at the hospital. ^99mTc perfusion imaging reflects relative myocardial blood flow at the time of tracer injection. There is no significant redistribution; therefore, imaging can occur some time after injection. There is a risk of image degradation with a prolonged injection-imaging interval as the agent decays. The half-life of ^99mTc is approximately 6 h. Most of the imaging in our study, including the overnight delayed acquisition, occurred within one half-life. Of the 44 patients undergoing delayed image acquisition, 3 had abnormal scans and none of these patients had events. More importantly, none of the patients with a normal perfusion scan had cardiac events. Our results indicate that prompt tracer injection followed by delayed image acquisition did not diminish the high negative predictive value of gated-SPECT MPI.

The overall results of the current study are consistent with previous reports. Kontos, et al. conducted a study in which low-to-moderate risk patients underwent rest MPI in the ED and were followed for 5 d (3). For the combined endpoint of MI or revascularization, the sensitivity and negative predictive value of rest MPI were 81% and 99%, respectively—very similar to our findings. That study had a more stringent definition of positive scan findings, requiring both a perfusion defect and regional wall motion abnormality. Equivocal scans were classified as normal. The primary utility of resting cardiac perfusion imaging in a population of patients presenting to the ED with CP is in identification of patients at low risk for cardiac events. These low-risk patients may be discharged from the ED with appropriate outpatient follow-up. The current study adds to the literature a large, single-center experience showing the utility of gated SPECT MPI in identifying patients at low risk for cardiac events at 30-d follow-up.

The present study also provides details of the 3 patients who had perfusion scans interpreted as normal in the ED and later had cardiac events. Although the number of these patients is small, they are of great concern to practitioners in the ED who make decisions to release patients to home with chest discomfort. Some similarities were noted in this group. All 3 had radionuclide tracer injection after CP had resolved, in some cases several hours later. In 2 patients the cardiac event was abnormal angiography requiring PCI. One patient with a normal perfusion scan had a MI. This patient had tracer injection 5 h after resolution of CP and was admitted despite the normal MPI due to her overall clinical presentation. Furthermore, this patient was found to have a history of CAD, an exclusion criterion for this study. The appropriateness of resting MPI in this individual is questionable. Similar to our current study, previous publications have reported administration of ^99mTc tracer for several hours after the resolution of CP (2). The detailed description of the patients with normal scans having cardiac events in our study suggests that the accuracy of resting MPI could be improved further if tracer injection occurred during or shortly after CP.

This study is a single-center, case-control analysis of MPI of patients with a complaint of CP in an urban ED. Our institution handles a high volume of cardiac nuclear imaging from both the ED and the nuclear cardiology laboratory. Generalization of our findings to another practice setting may be difficult. Also, the lack of a control group and nonrandomized patient assignment fails to control for potential patient selection bias.

CONCLUSION

A normal resting gated SPECT radionuclide MPI scan predicts a very low incidence of 30-d cardiac events. Abnormal scans identify patients at higher risk for 30-d events who should receive further diagnostic testing for atherosclerotic CAD. A delayed image acquisition protocol did not diminish the accuracy of resting MPI for prediction of cardiac events in the present study. Lastly, radionuclide injection after resolution of CP may be associated with a higher rate of false-negative MPI results.