Summary of Post The Cardiovascular Research Institute (CVRI) Dublin is seeking a Postdoctoral Researcher to lead the development of non-invasive AI models for assessing Right Ventricular-Pulmonary Arterial (RV-PA) coupling.
This role is a core component of the DUO MAX project's mission to provide a holistic "Precision Risk Stratification" for patients with tricuspid regurgitation.
The successful candidate will develop machine learning models to estimate mean Pulmonary Artery Pressure (mPAP) and derive prognostic coupling indices from echocardiographic data (TAPSE, FAC, and flow dynamics).
By moving beyond simple anatomical measures, your work will provide clinicians with a functional "physiologic signature" of the patient, significantly improving the prediction of outcomes following tricuspid valve intervention.
You will work closely with both the Computer Vision team and the Clinical Imaging Lead to integrate these hemodynamic insights into the final Clinical Decision Support System (CDSS).
Specifically, the duties of the post are: The applicant will work in Prof. Soliman's lab at both RCSI and Mater Private Hospital.
Key Responsibilities The Senior Computational Cardiovascular Modeller will provide overall scientific and technical expertise for the Right Ventricle-Pulmonary Artery coupling.
Specifically, the duties include: Develop patient-specific pulmonary artery CFD models using CT or MRI data Perform geometry reconstruction, cleaning, simplification, and mesh generation Implement physiological inlet and outlet boundary conditions, including lumped-parameter (e.g., Windkessel) models Calibrate CFD models using available hemodynamic reference data (e.g., right heart catheterization, implantable pressure sensors) Validate CFD-predicted pressures through quantitative agreement analyses (e.g., bias, error metrics, Bland–Altman) Conduct sensitivity analyses to assess the impact of modelling Document modelling workflows, assumptions, and limitations in a reproducible and transparent manner Hemodynamic Modeling: Develop and validate machine learning models (e.g., XGBoost, Random Forests, or Neural Networks) for the non-invasive estimation of mPAP and RV-PA coupling.
Feature Engineering: Extract and refine physiological features from echocardiographic measures including TAPSE, Fractional Area Change (FAC), and Doppler-derived indices.
Prognostic Analytics: Correlate AI-derived RV-PA indices with longitudinal patient outcomes to establish clinical utility and risk thresholds.
Data Integration: Collaborate with the NLP