Performance of Traditional Pretest Probability Estimates in Stable Patients Undergoing Myocardial Perfusion Imaging

Omar Batal, Saurabh Malhotra, Matthew Harinstein, Jeremy Markowitz, Gavin Hickey, Sunil Agarwal, Pamela Douglas, Prem Soman

Research output: Contribution to journalArticlepeer-review

7 Scopus citations


Background: The yield of myocardial perfusion imaging is low in contemporary patients with suspected coronary artery disease (CAD) selected based on American College of Cardiology Foundation/American Heart Association pretest probability estimate. We compared traditional pretest estimates of CAD probability with the prevalence of abnormal myocardial perfusion single-photon emission computed tomography (MPS). Methods: This was a cohort study from a single academic center. Consecutive stable patients without known CAD referred for stress MPS for suspected CAD between 2004 and 2011 were identified (n=15 777). Angina typicality was determined using standard criteria. Abnormal MPS perfusion was defined as a summed stress score ≥4, ischemia as summed stress score ≥4 and summed difference score ≥2, and extensive ischemia as summed difference score ≥8 using a standard, 17-segment model of the left ventricle. The pretest probability of CAD was determined using the American College of Cardiology Foundation/American Heart Association criteria. Results: Overall, 14% (n=2177) of patients had abnormal MPS of whom 11% (n=1698) had ischemia and 4% (n=684) extensive ischemia. In patients with chest pain who underwent treadmill MPS (n=4764), only 27% reported angina on the treadmill. Typical angina was associated with the highest prevalence for positive MPS (33% in men and 14% in women), ischemia (30% in men and 12% in women), and extensive ischemia (22% in men and 4% in women) when compared with other symptom categories. Prevalence of MPS abnormality was substantially lower than expected based on pretest probability estimates across most sex and age groups. In multivariable analysis, the pretest probability estimate was not an independent predictor of abnormal MPS. Conclusions: Traditional estimates of pretest probability of CAD are not predictive of MPS perfusion abnormality and overestimate its prevalence in stable patients.

Original languageEnglish (US)
Article numbere008473
JournalCirculation: Cardiovascular Imaging
Issue number10
StatePublished - Oct 1 2019

Bibliographical note

Funding Information:
Dr Soman reports grant funding to the University of Pittsburgh from Astallas Pharma for protocol related to stress testing. Dr Malhotra participates on the Speaker Bureau of Pfizer. The other authors report no conflicts.

Publisher Copyright:
© 2019 American Heart Association, Inc.


  • cohort studies
  • coronary disease
  • humans
  • male
  • radionuclide imaging


Dive into the research topics of 'Performance of Traditional Pretest Probability Estimates in Stable Patients Undergoing Myocardial Perfusion Imaging'. Together they form a unique fingerprint.

Cite this