Single-photon emission computed tomography appropriateness: does it matter for patient outcomes?

I n this issue of Circulation, Doukky et al 1 report findings from a cohort of 1511 patients from 11 outpatient community based practices (20 primary care physicians and 2 cardiologists) in the Chicago metropolitan area. The patients underwent single-photon emission computed tomographic (SPECT) myocardial perfusion imaging and were then followed up for 27±10 months for major adverse cardiac events: death, death or myocardial infarction, and cardiac death or myocardial infarction. The SPECT studies were categorized on the basis of the 2009 appropriate use criteria (AUC) as appropriate, uncertain, or inappropriate. The investigators report that 823 patients (54.5%) underwent SPECT scans that were classified as appropriate or uncertain and 688 patients (45.5%) underwent SPECT scans that were classified as inappropriate. In those patients whose SPECT scans were appropriate or uncertain, abnormal scans were of significant value in predicting major adverse cardiac events with hazard ratios of 3.1 to 3.7 compared with normal scans. However, in those patients undergoing SPECT classified as inappropriate, abnormal SPECT scans did not achieve statistical significance in predicting major adverse cardiac events, although the hazard ratios ranged from 2.3 to 11.8. Regardless of the appropriate-ness of SPECT, the presence of ischemia on SPECT, reflected in the summed difference score (SDS), predicted subsequent coronary angiography and revascularization. As the investigators indicate, this is the first large study validating the prog-nostic implications of SPECT AUC, further supporting its clinical utility. In this editorial, we examine both the internal and external validity of this study to place it in context for evidence-based clinicians. Is the study internally valid? That is, do the data justify the conclusions? This study has a number of strengths, beginning with the size of the cohort and the several-year duration of follow up. Although follow-up was >99% complete, a number of patients were excluded (n=182) or lost to follow-up (n=14). To their credit, the authors systematically compare these patients with those patients with follow-up data. The differences were modest, suggesting that the exclusion of these patients did not bias the results. The SPECT studies were all acquired on a single dedicated cardiac SPECT camera and interpreted by a single expert nuclear cardiologist, eliminating the variability inherent in different equipment and different interpreters. The SPECT results were categorized into published groupings of the summed stress score and SDS, permitting comparison with extensive previous literature using the summed stress score and SDS. There were some potential methodological …