BACKGROUND: The frequency of abnormal stress SPECT myocardial perfusion imaging (MPS) decreased over the past decades despite an increase in the prevalence of cardiovascular risk factors. These findings strengthen the need to develop more effective strategies for appropriately referring patients with suspected coronary artery disease (CAD) to cardiac imaging. The aim of this study was to develop pretest assessment models for predicting abnormal stress MPS. METHODS: We included 5,601 consecutive patients with suspected CAD, who underwent stress MPS at our academic center. Two different models were considered: a basic model including age, gender, and anginal symptoms; and a clinical model including also diabetes, hypertension, hypercholesterolemia, smoking, and family history of CAD. RESULTS: In patients with abnormal MPS, the basic model classified more than 75% of patients as intermediate risk, whereas only 23% were incorrectly classified as low risk. In patients with normal MPS, 45% were correctly classified as low risk and none as high risk. Basic and clinical models had a limited discriminating capacity (area under the receiver operating characteristic curve 0.644 for basic model and 0.647 for clinical model) between individuals with and without abnormal stress MPS. The decision curve analysis demonstrates a high net benefit across a range of threshold probabilities from ~ 15% to ~30% for both models. CONCLUSIONS: A pretest risk stratification based on traditional cardiovascular risk factors has a limited value for predicting an abnormal stress MPS in patients with suspected CAD. However, selecting a proper threshold probability enhances the appropriateness of referral to stress MPS.
BACKGROUND: The frequency of abnormal stress SPECT myocardial perfusion imaging (MPS) decreased over the past decades despite an increase in the prevalence of cardiovascular risk factors. These findings strengthen the need to develop more effective strategies for appropriately referring patients with suspected coronary artery disease (CAD) to cardiac imaging. The aim of this study was to develop pretest assessment models for predicting abnormal stress MPS. METHODS: We included 5,601 consecutive patients with suspected CAD, who underwent stress MPS at our academic center. Two different models were considered: a basic model including age, gender, and anginal symptoms; and a clinical model including also diabetes, hypertension, hypercholesterolemia, smoking, and family history of CAD. RESULTS: In patients with abnormal MPS, the basic model classified more than 75% of patients as intermediate risk, whereas only 23% were incorrectly classified as low risk. In patients with normal MPS, 45% were correctly classified as low risk and none as high risk. Basic and clinical models had a limited discriminating capacity (area under the receiver operating characteristic curve 0.644 for basic model and 0.647 for clinical model) between individuals with and without abnormal stress MPS. The decision curve analysis demonstrates a high net benefit across a range of threshold probabilities from ~ 15% to ~30% for both models. CONCLUSIONS: A pretest risk stratification based on traditional cardiovascular risk factors has a limited value for predicting an abnormal stress MPS in patients with suspected CAD. However, selecting a proper threshold probability enhances the appropriateness of referral to stress MPS.