Sensitivity and predictors of false-negative SPECT myocardial perfusion imaging in a high-burden coronary artery disease population: a retrospective analysis using revascularization as the reference standard.

backgroundCoronary artery disease (CAD) places a disproportionate burden on rural Appalachian communities; Scioto County, Ohio has among the highest CAD-related hospitalization rates in the state. Single-photon emission computed tomography (SPECT) myocardial perfusion imaging is widely used, yet its performance in high-burden populations with substantial comorbidity is poorly characterized. We evaluated the sensitivity of SPECT for predicting subsequent revascularization and identified predictors of false-negative results.

methodsRetrospective analysis of prospectively collected American College of Cardiology CathPCI registry data at a single community hospital, including all cardiac catheterizations from Q3 2013 to Q1 2018. Sensitivity was calculated using subsequent revascularization (percutaneous coronary intervention or coronary artery bypass grafting) as the primary reference standard, with angiographic ≥ 70% and ≥ 50% stenosis as secondary standards. Logistic regression with forward stepwise selection identified predictors of false-negative SPECT, with sex forced into the model as a sensitivity analysis. Reporting followed STROBE.

resultsOf 7,312 catheterizations, 3,116 (42.6%) led to revascularization. Among 3,189 patients with SPECT data, sensitivity for predicting revascularization was 82.0% (95% CI 79.8-84.0%); 84.6% (82.5-86.6%) excluding indeterminate and unavailable results; and 82.9% (81.1-84.6%) using anatomic ≥ 70% stenosis, very similar to the revascularization-based estimate. Territory-specific sensitivity ranged from 82.2% (mid-distal LAD, n = 1,026) to 91.2% (ramus, n = 91). Among negative-SPECT patients who proceeded to catheterization, 32.8% (226/688) required intervention, a proportion specific to this selected subgroup and not a population-level false-negative rate. Family history of premature CAD was independently associated with false-negative results (OR 1.73, 95% CI 1.25-2.40), unchanged when sex was forced into the model (OR 1.73) and attenuated but still significant under an anatomic outcome definition (OR 1.39, 95% CI 1.05-1.84).

conclusionsSPECT sensitivity for predicting revascularization was 82.0%, consistent with contemporary benchmarks and unchanged under an anatomic reference standard. Family history of premature CAD was independently associated with false-negative SPECT, a finding that, to our knowledge, has not been previously reported in the SPECT literature, though residual referral bias cannot be excluded. Where SPECT remains the primary cardiac imaging modality, these findings support careful clinical interpretation of negative SPECT results in patients with a strong family history of CAD.