Clinical applications​

Quantitative cardiac MRI (T1, T2 mapping, cardiac function assessment etc…); bowel motility assessment by dynamic MRI; breast cancer detection/characterization by MRI; study of cardiac remodelling in post-infarct patients; catheter ablation of ventricular tachycardia; prediction of the risk of ventricular tachycardia from MRI.

ELECTRA: Electrical property imaging by MRI: Application to the MR-safety of medical devices

French ANR programme - Enterprise-academia collaboration (2021-25)

Involvement : Principal investigator

Academic collaborators : ICUBE (Strasbourg), LMR (Reims), CIC-IT (Nancy)

Enterprise collaborator : Healtis (Nancy)

Description :

Knowledge of the electrical properties (EP, conductivity and permittivity) of human tissues is required to estimate, by numerical simulation, the heating of tissues during an MRI examination in the presence of implanted or accessory medical devices. Currently, simulations assume that EP values in a given tissue are the same for each patient (values obtained from ex-vivo studies on animals), regardless of age or sex.

The aim of ELECTRA is to develop new MRI tools for imaging EP in-vivo. Our original approach relies on new reconstruction methods which are robust to transitions of EP at interfaces between tissues. By creating a database of 120 subjects with marked differences in age, we will be able to validate those measures, to establish reference values in humans as a function of age, and to interpret the observed variations across the population. Finally those results will be applied to MRI safety investigations for medical devices (Healtis company).

The consortium comprises three academic teams who have already worked together to develop MRI of EP, with complementary expertise in MRI reconstruction (IADI), acquisition and MRI signal modelling (ICUBE) and solving inverse problems associated with Maxwell equations (LMR). The Healtis company will bring its expertise in MR-compatibility (simulation and testing) and will transfer the results to new services. CIC-IT will provide support for constituting the volunteer database and for its dissemination.

The economic impact of ELECTRA affects the market of implanted medical devices, in particular active devices (cardiac and neurologic stimulators, cochlear implants etc…), which is growing fast.  The issue of MR-compatibility of these devices is now essential. International standards regarding the assessment of heating risks in the case of passives devices (ASTM F2182) and active devices ((ISO/TS 10974) are evolving rapidly, and Healtis is actively participating in the workgroups of these standards.  Electromagnetic simulation is increasingly present. The main challenge is to standardize the methods for validating patient models (the precise EP values for tissues, and their variability across the population, are still an open question), device models, and simulation results in terms of specific absorption rate (SAR) or temperature elevation in the patient.

The scientific ambition of ELECTRA is to make MR imaging of EP a new potential biomarker for diagnostic imaging. The new expected results in EP reconstruction (mathematics and applied methods) will be key to improve the technique, and to validate it more and more realistic phantoms. The volunteer study will allow reference in-vivo PE values to be established for a wide range of tissues (head and trunc). This study is intended to be the first one to show in-vivo, in humans, age-related changes of conductivity in white matter, associated with structural changes.

The developments in ELECTRA will have many implications for diagnostic imaging (EP measurements as a new marker of fibrosis, inflammation, lesion severity etc…) and for the implementation of personalized SAR models. The latter will be useful for developing high-field MRI, but also for optimizing clinical MRI in general. Indeed, all MRI systems are currently restrained, based on SAR estimation in a worst-case scenario from generic patient models. A personalized SAR model, based on a patient-specific EP map, would allow these constraints to be relaxed. This would result in a gain in image quality and/or acquisition duration and ultimately an improvement in cost-efficiency of MRI.

BraCoil: Smart bra-shaped MRI breast coil

French ANR programme – International collaboration, France-Austria (2017-21)

Involvement : Leader of WP4 (Motion-compensated reconstruction)

Main collaborators : University of Vienna (Austria), IR4M (Orsay)

Description :

BraCoil aims at developping new hardware and software tools to improve MR mammography. Bra-shaped flexible MRI coils are being developed (Vienna and IR4M partners) to improve the signal-to-noise ratio of breast MR images, and motion correction techniques are developed to allow the examination to be performed in the supine position (as opposed to the prone position in today’s clinical practice), with the patient breathing freely during the whole examination.

Vectorized Levenberg-Marquardt optimization for T1 mapping MRI (Matlab and GPU code)

The code below is the source code used in the paper:


Liu S; Bustin A; Ferry P; Codreanu A; Burschka D; Menini A; Odille F

A vectorized Levenberg-Marquardt model fitting algorithm for efficient post-processing of cardiac T1 mapping MRI.

Computers in biology and medicine, 96 , pp. 106-115, 2018. doi:10.1016/j.compbiomed.2018.03.009