Prospective prediction of cardiovascular outcomes for HORIZON-Lp(a) trial using a mechanistic QSP model of ASCVD

Background: Long, expensive outcomes trials are typically required to quantify cardiovascular (CV) benefit under lipid-lowering therapy (LLT). Mechanistic in-silico trials built on Quantitative Systems Pharmacology (QSP) offer a complementary pathway to anticipate efficacy and inform trial design. Objective: Using a mechanistic Quantitative Systems Pharmacology (QSP) framework, this study aims to prospectively predict the Pelacarsen efficacy on Major Advanced Coronary Event (MACE) outcomes of the HORIZON-Lp(a) study. Methods: We improved a previously published and validated QSP model with new revascularization endpoints, calibrated composite MACE outcomes on FOURIER trial data, validated the model and in particular the Lp(a) - MACE relationship on ODYSSEY-OUTCOMES trial and CGPS cohort data. We then conducted a in silico simulation of the HORIZON-Lp(a) study to prospectively predict MACE outcomes and Pelacarsen efficacy and performed a sensitivity analysis to determine the impact of sources of uncertainty on predictions.Results: We predicted a palette of predictions for 3P-MACE, 4P-MACE and their individual components for different values of median follow-up time and contextual effect in Placebo arm of HORIZON-Lp(a) trial which are currently unknown, as well as for the Lp(a) >= 90 mg/dL subgroup of interest.Conclusions: Within a credibility framework inspired by recent work on ASCVD mechanistic modeling, we delivered pre-readout prospective predictions of the upcoming HORIZON-Lp(a) trial, suggesting a consistent clinical benefit in lowering Lp(a). Validation of results with upcoming trial results will challenge the credibility of the ASCVD model and the robustness of predictive QSP approaches.