Cristiana Corsi

Thesis propoosal 1

Master’s Thesis Proposal with Internship at Angiodroid SpA

The thesis topic concerns ANGIOPULSE, a device that functions as a true aortic impedance adapter: a counterpulsation device—specifically, a balloon positioned in the aorta—that modulates the resistance perceived by the ventricle to reduce cardiac afterload. The goal is to provide mechanical support to the heart following myocardial infarction or in cases of heart failure, helping the ventricle function more effectively during the recovery phase. This project combines engineering, cardiovascular physiology, and the development of real medical devices, with a direct clinical impact on patients.

We are seeking a candidate for a 6-month in-office internship focused on concrete and stimulating tasks:

• Integration of new sensors into the device (bidirectional helium gas flow + ECG + arterial pressure signal), working side by side with the R&D team on hardware and firmware

• Setup and use of an aortic simulator to reproduce real-world hemodynamic conditions in the lab and test the system prior to clinical validation

• Developing calibration and synchronization tools for the acquired signals (blood pressure, ECG, flow) and optimizing the firmware’s operating cycles to ensure proper balloon inflation and adjustable pressure

• Analyzing signals collected from clinical cases, with the opportunity to develop custom data processing scripts

This is an opportunity to work closely with a small technical team on a device that is already being used to treat real patients, gaining hands-on experience with the entire process—from signal processing to clinical validation. At the end of the 6 months, we might consider a more structured collaboration.

Thesis proposal 2

Radiomics-based characterization of atrial blood flow in atrial fibrillation: comparison between 4D flow MRI and computational fluid-dynamics derived flow

Topic: Atrial fibrillation (AF) alters left atrial (LA) hemodynamics and increases thromboembolic risk. In this study we want to evaluate whether radiomics, based on the quantification of spatial texture patterns, can provide reproducible LA flow biomarkers in AF. Nowadays, two different approaches are available to characterize 3D complex flow patterns in the LA: 4D flow MRI and blood velocity fields resulting from patient specific fluid dynamics (CFD) simulations [1].

We have developed a workflow to extract hemodynamic parameter maps [2] across anatomical planes within the LA chamber. The radiomics features extracted from this analysis are further analysed through nested leave-one-out cross-validation incorporating a three-stage feature selection process: statistical testing, effect-size filtering, and stability assessment. Final classification has been performed using support vector machines, and feature importance was evaluated using SHAP analysis.

We want to test this workflow on a group of AF patients in which both 4D flow MRI and CFD are available and assess the reliability of the results obtained with the two approaches to identify LA flow biomarkers for AF risk stratification.

In collaboration with: The research is being carried out in collaboration with the Laboratoire d’Imagerie Biomédicale (LIB), Sorbonne Université, INSERM, CNRS, Paris, France

Acquired skills: The thesis offers the student the opportunity to gain in-depth knowledge of cutting-edge methodologies relating to machine learning and radiomics, simulation environments and advanced medical imaging techniques.

[1] Falanga M, et al. A digital twin approach for stroke risk assessment in Atrial Fibrillation Patients. Heliyon (2024) 10:e39527. doi: 10.1016/j.heliyon.2024.e39527;

[2] Huellebrand M, et al. Radiomics-based aortic flow profile characterization with 4D phase-contrast MRI. Front Cardiovasc Med (2023) 10:1102502. doi: 10.3389/fcvm.2023.1102502