Week of 2026-04-27 — 5 new papers
Cardiovascular CFD, hemodynamics, and AI modelling (PINN · neural operators · surrogates · digital twins). Auto-curated from OpenAlex + arXiv, classified with Claude.
At a glance
| Tier | Topic | Count |
|---|---|---|
| A | Boundary conditions (inlet/outlet, Windkessel) | 1 |
| B | Turbulence modelling (RANS, LES, DNS) | 0 |
| C | V&V and uncertainty quantification | 0 |
| D | Physiology & scaling laws | 0 |
| E | Imaging & WSS measurement | 2 |
| F | AI, ML & digital-twin pipelines | 2 |
Highlights this week
- Patient-specific computational fluid dynamics analysis of portal vein hemodynamics before and after balloon angioplasty following living donor liver transplantation: a proof-of-concept study — F Yang et al. · Frontiers in Bioengineering and Biotechn
- Bayesian Physics-Informed Neural Network for Multiscale Sepsis Modelling: Reference Implementation, Synthetic Cohort Generator, and External Validation Harness — Gustav Olaf Yunus Laitinen-Fredriksson Lundström-Imanov · Zenodo (CERN European Organization for N
- Validation of Aortic Blood Flow Simulations During Extracorporeal Circulation Using Phase Contrast Magnetic Resonance Imaging — Anna Kathrin Assmann et al. · Artificial Organs
Tier A — Boundary conditions (inlet/outlet, Windkessel)
Patient-specific computational fluid dynamics analysis of portal vein hemodynamics before and after balloon angioplasty following living donor liver transplantation: a proof-of-concept study
F Yang et al. · Frontiers in Bioengineering and Biotechnology · 2026 · DOI
Patient-specific CFD framework reconstructs portal venous geometry from imaging and uses Doppler-derived inlet conditions with three-element Windkessel outlet boundary conditions to evaluate hemodynamics before and after balloon angioplasty for portal vein stenosis post-LDLT. Analysis shows angioplasty reduces trans-stenotic pressure gradients, normalizes WSS patterns, and remodels intrahepatic flow distribution over 6 months.
Why it matters: Demonstrates practical 0D–3D coupling (Windkessel BC) and longitudinal patient-specific CFD for surgical surveillance, directly relevant to clinically motivated boundary condition implementation and hemodynamic outcome prediction.
Related from the AortaCFD corpus:
- Kwon et al. (2014) · Computational fluid dynamics analysis of the blood flow in patient-specific arteries
- Ali et al. (2024) · Computational fluid dynamics modeling of coronary artery blood flow using OpenFOAM: Validation with the food and drug administration benchmark nozzle model (p. 9) — DOI
Tier E — Imaging & WSS measurement
Validation of Aortic Blood Flow Simulations During Extracorporeal Circulation Using Phase Contrast Magnetic Resonance Imaging
Anna Kathrin Assmann et al. · Artificial Organs · 2026 · DOI
Phase-contrast MRI was used to validate CFD simulations of aortic blood flow under extracorporeal circulation (ECC) in vivo, comparing retrograde and antegrade cannulation strategies. The study demonstrates that retrograde ECC preserves cerebral perfusion and suggests personalized simulations could optimize cannulation to reduce complications.
Why it matters: V&V of cardiovascular CFD against 4D flow MRI in a clinically relevant, complex scenario (ECC) is directly valuable; however, the primary contribution is imaging-based validation rather than methodological CFD advancement.
Related from the AortaCFD corpus:
- Rispoli et al. (2015) · Computational fluid dynamics simulations of blood flow regularized by 3D phase contrast MRI (p. 2) — DOI
- Dyverfeldt et al. (2006) · Quantification of intravoxel velocity standard deviation and turbulence intensity by generalizing phase-contrast MRI — DOI
The emerging role of 4D flow MRI in abdominal interventional radiology
Ryota Hyodo et al. · Abdominal Radiology · 2026 · DOI
A clinical review of 4D flow MRI applications in abdominal interventional radiology, covering aortic interventions, portal venous interventions, and hepatic outflow assessment. The paper highlights endoleak characterization and portal hemodynamic quantification as promising use cases while acknowledging technical limitations and the need for outcome-driven validation.
Why it matters: CFD practitioners use 4D flow MRI as ground-truth data for model validation and inlet/outlet condition specification; this review clarifies clinical capabilities and artifacts relevant to patient-specific modeling workflows.
Related from the AortaCFD corpus:
- Markl et al. (2012) · 4D flow MRI (p. 10) — DOI
- Steinman & Migliavacca (2018) · Editorial: Special Issue on Verification, Validation, and Uncertainty Quantification of Cardiovascular Models: Towards Effective VVUQ for Translating Cardiovascular Modelling to Clinical Utility (p. 3) — DOI
Tier F — AI, ML & digital-twin pipelines
Bayesian Physics-Informed Neural Network for Multiscale Sepsis Modelling: Reference Implementation, Synthetic Cohort Generator, and External Validation Harness
Gustav Olaf Yunus Laitinen-Fredriksson Lundström-Imanov · Zenodo (CERN European Organization for Nuclear Research) · 2026 · DOI
A Bayesian PINN framework couples multiscale sepsis pathophysiology—including pathogen-immune dynamics, cytokine kinetics, Windkessel hemodynamics, and microcirculation Navier-Stokes—with uncertainty quantification via ensemble posterior inference. External validation on MIMIC-IV, eICU-CRD, and HiRID compares nine baselines (LSTM, Neural ODE, vanilla PINN, XGBoost, etc.) for sepsis mortality prediction.
Why it matters: Demonstrates PINN application to coupled 0D-3D hemodynamics with physiological constraints and rigorous uncertainty quantification; relevant for practitioners interested in physics-informed surrogates and digital-twin workflows, though primary focus is systems biology rather than pure CFD.
Related from the AortaCFD corpus:
- Kissas et al. (2020) · Machine learning in cardiovascular flows modeling: Predicting arterial blood pressure from non-invasive 4D flow MRI data using physics-informed neural networks (p. 23) — DOI
- Mukherjee & others (2025) · Survey and perspective on verification, validation, and uncertainty quantification of digital twins for precision medicine (p. 4) — DOI
Bayesian Physics-Informed Neural Network for Multiscale Sepsis Modelling: Reference Implementation, Synthetic Cohort Generator, and External Validation Harness
Gustav Olaf Yunus Laitinen-Fredriksson Lundström-Imanov · Open MIND · 2026 · DOI
A Bayesian PINN framework couples multiscale sepsis pathophysiology—from pathogen-immune dynamics and cytokine cascades through Windkessel macrocirculation and axisymmetric Navier-Stokes microcirculation with nitric oxide vasoplegia—using composite physics and data loss terms with Stein variational gradient descent. External validation against MIMIC-IV, eICU-CRD, and HiRID benchmarks nine baselines including Neural ODE and vanilla PINN variants.
Why it matters: Demonstrates PINN architecture and UQ methodology (Bayesian ensemble) for a physiologically rich hemodynamic-systemic model with open-source implementation; relevant for practitioners building patient-specific digital twins and learning operators from intensive care data.
Related from the AortaCFD corpus:
- Kissas et al. (2020) · Machine learning in cardiovascular flows modeling: Predicting arterial blood pressure from non-invasive 4D flow MRI data using physics-informed neural networks (p. 23) — DOI
- Mukherjee & others (2025) · Survey and perspective on verification, validation, and uncertainty quantification of digital twins for precision medicine (p. 4) — DOI
Methodology, tier definitions and scope caveats: see the Paper Digest landing page. Classifier threshold this run: 0.5.
Found an issue or have a suggestion for this page?
Open a GitHub issue