(Closed) Postdoc position on Personalized modeling of lung poromechanics in COVID-19

Jobs
Published

September 22, 2020

Context & Objectives The COVID-19 infection manifests itself in its severe forms as acute alveolar-interstitial and vascular disease of varying severity, the long-term course of which is unknown. A development towards fibrosing pneumopathy could affect 10 to 30% of survivors of severe forms, which would make it a public health problem due to the secondary handicap. A better understanding of the regional physiological mechanisms in the evolution of COVID-19 pneumonia in their severe forms would allow to anticipate the development or not of sequelae, in particular fibrosing.

In the past few years, the MΞDISIM team at École Polytechnique and Inria, in close collaboration with clinicians (pulmonologists and radiologists) from Avicenne AP-HP Hospital/Sorbonne Paris Nord University/INSERM, as well as image processing experts from Telecom SudParis, has developed a poromechanical model of the human lung, together with an estimation pipeline allowing to personalize the model to a patient based on clinical images [1]. Until now the focus of the group was on idiopathic pulmonary fibrosis; however, the consortium has just received a grant from the ANR to adapt the image-based patient-specific modeling pipeline to COVID-19 patients.

More precisely, the project concerns the processing of CT images of patients followed at the Avicenne AP-HP hospital for an assessment of vascular remodeling and regional lung mechanics/compliance. These parameters obtained at 2-3 months of the COVID-19 infection will be correlated with clinical, CT, and functional data at 6 months and 1 year of follow-up. To do so, developments are required at the modeling level (to adapt the model to the new disease), on the numerical method level (to automatize the few remaining manual steps) as well as at the implementation level (to streamline the image à model pipeline).

Keywords Pulmonary Biomechanics; Image-based Modeling; Finite Element Method; Data assimilation

Candidate profile We are looking for a postdoc. The funding is for one year, but could be further extended if we receive other grants. The candidate will have a fair understanding of continuum mechanics, with if possible knowledge of finite strains, biomechanics, and numerical methods. He/She will be at ease with computational pipelines, and also have an interest in the application in pulmonology, especially for interacting with clinical collaborators.

Work environment The work will take place within the MΞDISIM team (joint between École Polytechnique & Inria) and within the Solid Mechanics Laboratory, on the École Polytechnique campus. It will be co-supervised by Martin Genet & Dominique Chapelle, and in direct contact with clinicians and image processing scientists. It should start as soon as possible.

Bibliograpy
[1]  C. Patte, M. Genet, C. Fetita, P.-Y. Brillet and D. Chapelle, “Mécanique pulmonaire personnalisée: modélisation et estimation—Application à la fibrose pulmonaire”, 14ème colloque national en calcul des structures (CSMA), Giens, France, 2019.

Contacts martin.genet@polytechnique.edu, dominique.chapelle@inria.fr