Project

AI-Powered Precision in Cervical Spine Imaging

Short Summary

Back pain is a major public health problem in our current society. For individuals, it reduces quality of life drastically. For the society, it induces significant economic burden as it can temporarily remove individuals from the workforce and the treatment can be expensive. A part of back pain complaints stems from the cervical spine. For diagnosis and treatment planning Magnetic Resonance Imaging (MRI) is potentially the most important tool in our arsenal. In this project, we aim to improve the use of MRI for visualizing cervical spine so that during routine clinical practice, very fine level details can be visualized and correct treatment strategy can be identified.

Goals

The goal of this work is to enable visualization of fine level details for cervical spine imaging. To this end, we are tackling the problem from two different fronts. One, we are developing state-of-the-art image reconstruction algorithms that can potentially halve the acquisition times. Second, we are developing algorithms that can personalize complex anatomical models based on noisy and even corrupted images to create highly detailed digital twins of the patient from which diagnostic information can be drawn and potentially used for treatment planning. The developed algorithms are planned to be deployed in the research setting at the Balgrist university hospital.

Significance

Upon successful completion of the project, we will have the technologies that allows detailed visualization and imaging of the cervical spine. This will facilitate diagnosis and treatment planning. Concretely, we believe this will help clinical researchers understand better why certain treatments work in some cases and not others. Through this, patient stratification will be possible and, potentially, the optimal treatment for each patient will be selected based on richer information coming from the MRI.

Background

The current MRI technology for imaging the cervical spine is limited in the details of structures that can be accurately visualized. Nerve roots and nerve rootlets, structures that are responsible for the perceived pain in a large portion of the patients, cannot be visualized with the current system. New imaging sequences provide the means to visualize these structures at the cost of very long acquisition times and noise. This setting cannot be applied to patients suffering from back-pain.

Precision Imaging Project

Co-Investigators

Prof. Jean-Philippe Thiran (EPFL)
Prof. Reto Sutter (Balgrist University Hospital)

Consortium

Balgrist University Hospital

Status

In Progress

7th Call, Precision Imaging