2022-2025

Hand Fracture Surgical Video Analysis for AI-Based Training and Surgical Safety

The goal of this project is to develop AI based tools for the identification and segmentation of surgical stages and surgical errors dedicated to minimally invasive surgical procedures, particularly for distal radial fracture osteosynthesis. Such developments will promote surgical confidence by supervising tasks, providing warnings, and offer an automated analysis of the procedure for training purposes. The system will therefore help to prevent adverse events and train junior surgeons.

  • PhD student: C. Graëff
  • Co-supervisors: P. Liverneaux & T. Lampert, ICube & HUS

Multiscale mechanical model for skeletal muscle using homogenization approach coupling imaging, modeling and genetic expression

Through the alliance of four French and American partner laboratories, this project is at the crossroads of mechanics, imaging and genetics. Using data from mechanical tests and medical imaging, a multi-scale biomechanical model of mouse skeletal muscle is proposed to study the influence of the Klf10 gene on muscle functionality through its different scales.

This project aims to provide an original understanding of the impact of microscopic constituents of healthy, or pathological skeletal muscle, on its overall mechanical behavior. Through an innovative coupling of imaging, biomechanics and genetics, it aims to identify experimentally and numerically the influence of the Klf10 gene on muscle properties via the development of a multiscale homogenized mechanical model of mouse skeletal muscle.
 

  • PhD student: A. Loumeaud
  • Co-supervisors: D. George, S. Bensamoun & S. Chatelin, ICube & Biomechanics and Bioengineering Laboratory (UTC)

Deep learning based medical image outpainting for adaptive radiotherapy

Left: axial kVCT of a head and neck patient with a 50 cm FOV. Right: axial MVCT of the same patient, with a 39 cm FOV. Property of ICANS

This interdisciplinary project is at the crossroads of radiotherapy, image processing and medical physics. Using deep learning methods, its objective is the prediction of patient anatomy outside the image boundaries ("outpainting") in the context of adaptive radiotherapy treatments. The project will be carried out in collaboration between a clinical radiotherapy department and a laboratory specialized in medical image processing.
 

  • PhD student: C. Boily
  • Co-supervisors: P. Meyer, D. Antoni & A. Lallement, ICube & ICANS

Robust intra-cell semantic segmentation for quantifying metastasis stages

The automatic analysis of high-resolution images in structural biology represents an important challenge due to the volume of data, the complexity of the content and the high inter-acquisition variability. The objective of this project is to develop robust methods for the segmentation of cellular objects in order to analyze their interactions. This work will aim at characterizing the stages of the metastatic cascade in the context of cancer research.
 

  • PhD student: A. Stenger
  • Co-supervisors: B. Naegel & P. Schultz, ICube & IGBMC

Patient-specific study of the internal cemented screws for the pelvis metastasis

Percutaneous cementoplasty with screws is a minimally invasive procedure that involves filling metastasized bone with orthopedic cement (PMMA) around internal screws under fluoroscopic control. Mechanically strengthening the bone and limiting the pain caused by lesions are the two main goals of the procedure. This doctoral project examines the assistance to percutaneous cementoplasty with a patient-specific approach, in order to improve the mechanical stability of damaged bones and thus improve medical care.

  • PhD student: C. Sieffert
  • Co-supervisors: J. Garnon, B. Bayle & L. Meylheuc, ICube & HUS

University hospital institutes (IHU): a new paradigm for innovation in the health sector?

Fabric’innov - Credit: Flora

The main objective of this research project is to provide us with a better understanding of what makes an IHU management model successful. On the one hand, from a theoretical perspective, the project would fill some gaps in the management science literature, which, despite considerable empirical evidence (Mériade et al., 2019), still neglects the issue of tension in the medical domain. On the other hand, from an empirical perspective, the project would benefit the whole IHU French network and, more broadly, all the health institutions facing similar tensions.

  • PhD student: L. Jeanneau
  • Supervisor: S. Bianchini, S. Bollinger & E. Ruiz, BETA