More than a decade of peer-reviewed evidence.

Despite extensive guideline coverage and broad access to cardiovascular diagnostics, Germany continues to observe substantial rates of missed and late-diagnosed coronary artery disease — the so-called German paradox. This article reviews the structural drivers of that gap and introduces Cardio Explorer® as a clinically deployable AI decision-support tool that combines clinical, laboratory, and demographic variables into a single probabilistic estimate of obstructive CAD. The paper details the Memetic Pattern Algorithm architecture, summarises validation across the Basel, LURIC and Maastricht cohorts, and discusses integration into ambulatory cardiology pathways.

Consecutive patients (n = 2,417) referred for Rubidium-82 positron emission tomography were evaluated. Pre-test probability (PTP) was calculated using the ESC 2013/2019 and ACC 2012/2021 guidelines, and a memetic pattern-based algorithm (MPA) was applied incorporating symptoms, vitals, ECG and biomarkers; ischemia was defined as a summed difference score ≥ 2. Ischemia was present in 37.1%. The MPA model was the most accurate predictor of ischemia (AUC 0.758 vs. 0.585–0.673 for the guideline models, p < 0.001). At the < 5% threshold, its sensitivity and negative predictive value to rule out ischemia were 99.1% and 96.4%. The model allocated patients more evenly across PTP categories, reduced the proportion in the intermediate (15–85%) range by 29–51%, and was the only tool to correctly predict ischemia prevalence in the very low PTP category.

Objective: To evaluate the diagnostic performance of an artificial intelligence (AI) model — Cardio Explorer® — to assess the probability of obstructive coronary artery disease (CAD) in a low- to intermediate-risk outpatient population, using only clinical and laboratory variables already available at the point of care. Design: Retrospective diagnostic accuracy study using prospectively collected charts from an outpatient cardiology clinic at Maastricht University Medical Centre (MUMC+). Reference standard combined invasive coronary angiography, coronary CT angiography, and ≥3 years event-free follow-up. Results: The algorithm achieved an AUC of 0.87 (95% CI 0.84–0.91) with high sensitivity and a strong negative predictive value, supporting its use as a non-invasive rule-out tool in the intended-use population.

This paper situates Cardio Explorer® within the framework of Predictive, Preventive and Personalized Medicine (PPPM). It describes how a Memetic Pattern Algorithm trained and validated across independent European cohorts can support individualised CAD risk estimation without exposing patients to additional radiation, contrast, or stress testing. The article details the inputs, outputs, intended clinical use and the regulatory positioning of the tool, and outlines how AI-based pre-test probability can be integrated into shared decision-making for downstream imaging or invasive evaluation.

This foundational paper introduces a memetic pattern based algorithm (MPA) for the non-invasive diagnosis and exclusion of coronary artery disease. Using clinical, demographic, and laboratory inputs, the algorithm reconstructs patient-specific risk patterns and benchmarks them against a reference library of phenotypes with known coronary status. Initial evaluation in a Basel high-prevalence cohort referred for invasive coronary angiography demonstrated the feasibility of using machine-learned pattern recognition as a triage step prior to invasive testing, establishing the methodological foundation for the Cardio Explorer® platform.

This ESC AI Congress abstract compares the third-generation Memetic Pattern Algorithm (MPA v3) — the engine behind Cardio Explorer® — to the ESC 2024 RF-CL (Risk Factor weighted Clinical Likelihood) algorithm for the estimation of obstructive coronary artery disease. Using a contemporary outpatient cohort, the analysis evaluates discrimination, calibration, and net reclassification across pre-test probability strata, with a particular focus on the low-to-intermediate risk band where guideline pathways most often produce ambiguous recommendations. MPA v3 demonstrates improved discrimination and a more favourable rule-out profile than RF-CL while preserving sensitivity, supporting its role as a non-invasive triage tool prior to advanced imaging.

Registered clinical study describing the prospective implementation of the Cardio Explorer® AI algorithm in routine outpatient cardiology practice. The protocol assesses real-world diagnostic performance, downstream resource utilisation, and impact on clinical decision-making versus standard guideline-based pathways.