Stratification of coronary artery disease patients for revascularization procedure based on estimating adverse effects


Background: Percutaneous coronary intervention (PCI), which mostly involves the use of stents, is the preferred alternative option to coronary artery bypass grafting (CABG) for treatment of atherosclerosis. In addition to the nature of surgical procedure (i.e. minimally invasive versus invasive), the main open question is which of these treatment strategies provides the optimal solution with respect to costs and related complications such as in- stent restenosis and target vessel revascularization (TVR).

Methods: Here, we provide an algorithmic solution employing hierarchical probabilistic models to assess stent related adverse events, depending on the type of stent used (drug eluting versus bare metal), and associated costs. We collected clinical and angiographic data from 3423 patients that were treated in the Deutsches Herzzentrum (Munich, Germany) between 1999 and 2006. We compare our proposed workflow to the currently predominant treatment with drug-eluting stents.

Results: The probability of adverse effects determined by angiographic restenosis (effectiveness) and hazardous events (safety) decreases by 26.0% (95% CI, 21.4 to 29.7) and 23.8% (95% CI, 17.4 to 27.5), respectively, for hazardous events at 36 and 12 months ($P$ $< 0.001$) and reduces the corrective procedure costs per patient from \$3,641 and \$2,394 to \$2,586 and \$1,704 ($P$ $< 0.001$), respectively. Analysis of target lesion revascularization yields a reduction in adverse effects by 22.5% (95% CI, 18.3 to 25.4) and 36.7% (95% CI, 28.9 to 42.2) at 36 and 12 months, respectively.

BMC Medical Informatics and Decision Making
Sebastian Pölsterl
Post-Doctoral Researcher

My research interests include machine learning for time-to-event analysis, non-Euclidean data, and biomedical applications.