Table1_Comparison of the RF-CL and CACS-CL models to estimate the pretest probability of obstructive coronary artery disease and predict prognosis in patients with stable chest pain and diabetes mellitus.docx

BackgroundThe most appropriate tool for estimating the pretest probability (PTP) of obstructive coronary artery disease (CAD) in patients with diabetes mellitus (DM) and stable chest pain (SCP) remains unknown. Therefore, we aimed to validate and compare two recent models, namely, the risk factor-weighted clinical likelihood (RF-CL) model and coronary artery calcium score (CACS)-weighted clinical likelihood (CACS-CL) model, in these patient populations.
MethodsA total of 1,245 symptomatic patients with DM, who underwent CACS and coronary computed tomographic angiography (CCTA) scan, were identified and followed up. PTP of obstructive CAD for each patient was estimated using the RF-CL model and CACS-CL model, respectively. Area under the receiver operating characteristic curve (AUC), net reclassification improvement (NRI), and integrated discrimination improvement (IDI) were used to assess the performance of models. The associations of major adverse cardiovascular events (MACE) with risk groups were evaluated using Cox proportional hazards regression.
ResultsCompared with the RF-CL model, the CACS-CL model revealed a larger AUC (0.856 vs. 0.782, p = 0.0016), positive IDI (12%, p < 0.0001) and NRI (34%, p < 0.0001), stronger association to MACE (hazard ratio: 0.26 vs. 0.38) and less discrepancy between observed and predicted probabilities, resulting in a more effective risk assessment to optimize downstream clinical management.
ConclusionAmong patients with DM and SCP, the incorporation of CACS into the CACS-CL model resulted in a more accurate estimation for PTP and prediction of MACE. Utilizing the CACS-CL model, instead of the RF-CL model, might have greater potential to avoid unnecessary and omissive cardiovascular imaging testing with minimal cost.