Variability of Cerebral MRI Diffusion Parameters by Using a Patented Normalization Process vs no Use

Overview

BrainTale has developed a standardization approach based on averaging measurements in predefined brain regions of interest and use of reference data acquired from healthy volunteers under conditions (MRI machine, acquisition protocol) identical to those used for the examination of the patient. The present study is intended to support the normalization step on healthy volunteers in clinical centres that will be equipped with the BrainTale medical device software dedicated to clinical routine practice and to assess, through a multicentre study, the impact of this normalization step on the variability of the MRI diffusion parameters. The data collected will support characterization and modelisation of the variability to explore further biais corrections methods. The study will be conducted by neuroradiologists of the neuroradiology departments of five clinical centres in France.

Advances in medical imaging have made it possible, from the 1990s, to be able to visualize structural microlesions of the brain thanks to post-processing of the Magnetic Resonance Imaging (MRI) data acquired by the sequence using the diffusion tensor. Mathematical models such as the diffusion tensor have made it possible to quantify in each of the volumetric units of the brain (voxels) measurements correlated to the microstructure of neuronal axons such as the Fractional Anisotropy (FA), the Mean Diffusivity (MD), the Radial Diffusivity (RD) or the Axial Diffusivity (AD). These measurements thus make it possible to quantify the microstructural alterations. However, the use of Diffusion Tensor Imaging (DTI) technology has limitations and currently, there is no "gold-standard" to validate diffusion measurements, which are currently dependent on acquisition protocols, post-processing software and observers. To make these diffusion parameters usable in a clinical context, BrainTale has developed a standardization approach based on averaging measurements in predefined brain regions of interest and use of reference data acquired from healthy volunteers under conditions (MRI machine, acquisition protocol) identical to those used for the examination of the patient. This process requires the acquisition of 10 healthy volunteers to calibrate a new MRI protocol, which greatly limits access to the technology in clinical routine and limits the potential technological evolutions of the acquisition. The study aims to assess the impact of a patented normalization process on the interindividual variability of Fractional Anisotropy (FA) measurements derived from cerebral Diffusion Tensor Imaging (DTI) acquisitions. A total of 60 healthy volunteers will be included. One or two acquisition will be performed during a single on-site visit in order to collect diffusion parameters outcome data for further analyses (with or without normalization process). After each acquisition and before inclusion of the following healthy volunteer, anonymised subject's data will be transferred to BrainTale on a secured web platform. BrainTale will be in charge of the Quality Control (QC) of the acquired data and will attribute a QC-passed or QC-failed.