49 The use of virtual stent planner in patient selection for invasive coronary angiography

Introduction Appropriate patient selection for percutaneous coronary intervention (PCI) is of utmost importance. Computed tomography coronary angiography (CTCA) with fractional flow reserve (FFRCT) has been shown to alter patient management. However post PCI FFR is a crucial determinant of outcomes with higher target vessel failure, myocardial infraction and revascularisation rates associated with a post PCI FFR <0.8. CTCA based virtual stent planner (HeartFlow planner®) allows users to undertake virtual PCI and model FFRCT outcomes prior to invasive angiography. Given most PCI is undertaken for morbidity benefit a modelled PCI that does not achieve an adequate post procedural FFRCT may influence patient selection for PCI. Methods Prospective patients (30/01/23–30/2/24) listed for a clinically indicated invasive coronary angiogram (ICA) with a preceding CTCA and stent planner were included. Three interventional cardiologists, blinded to stent planner, were asked to review CTCA(FFRCT) data and select a treatment strategy of guideline directed medical therapy (GDMT), PCI or multi-disciplinary meeting including possible coronary artery bypass grafting (MDT+/-CABG) on a per-patient and per-vessel basis. Stent planner was then unblinded and theoretical management decisions re-recorded. A real world comparison of the theoretical stent number, length and predicted duration of procedure was made in a sub-set that underwent PCI within the study period. Results 25 cases, 100 vessels were included. 64% male, 63+/-12 years. Initial strategy was PCI in 92% (23/25) and MDT+/-CABG in 8% (2/25) of cases, [Fleiss' free-marginal kappa 0.72 (0.54–0.9)]. On a per-patient basis stent planner reclassified strategy to 76% (19/25) PCI, 16% (4/25) MDT+/-CABG and 8% (2/25) GDMT, [Fleiss' free-marginal kappa 0.6 (0.4–0.8)]. On a per-vessel basis PCI cases were reduced by 12%, 4% and 8% in the left anterior descending, left circumflex and right coronary arteries respectively. A subset (n=12) underwent PCI. Predicted procedure duration (minutes), stent number and length (mm) was 67+/-17 vs 71.4+/-23 (p=.2), 1.3+/-0.2 vs 1.3+/-0.5 (p=.4) and 30.8+/-5.9 vs 42.2+/-18.6 (p=.04) respectively for stent planner prediction versus real world findings. Conclusions The integration of a virtual stent planner may alter management decisions, decreasing the frequency with which PCI is chosen. Stent planner may facilitate prediction of a PCI strategy and subsequent procedural length but more work in these areas is required. Conflict of Interest None