Assessment of the repeatability and stability of NODDI diffusion modelling using phantoms and in-vivo acquisitions

MRI is a commonly used clinical tool for diagnosing, evaluating, and monitoring neurodegenerative diseases. One of the key contrast mechanisms in MRI is diffusion-weighted imaging (DWI), which measures the signal attenuation caused by water molecule displacement within brain tissues. This technique allows for the analysis of the brain tissue structure, specifically the white matter tracts where water diffusion aligns with the direction of the fibres. A new model called neurite orientation dispersion and density imaging (NODDI) has been proposed for characterizing brain tissue at a microscopic level. However, there is no gold standard for validating diffusion measures due to factors such as scanner type, scanning protocols, software methods, and observers. The aim of this work is to evaluate the repeatability and stability of NODDI diffusion modelling using phantoms and in vivo acquisitions. Phantoms and healthy volunteers are scanned multiple times on different days, and the acquired data is fitted to the NODDI model. The coefficient of variation is computed for phantoms to assess the consistency of the model over time, and the p-value of the paired t-test is calculated. The Bland-Altman analysis is performed for in vivo acquisitions to assess the stability and repeatability of NODDI over time. Results showed a great consistency of the NODDI metrics over time for both phantoms and in vivo acquisitions. The results also indicated stability even with magnetic field gradient coil heating, which can affect the acquired images and cause the model to not fit well. Additionally, the analysis showed that the hydration state of a healthy participant does not influence the outcome of the analysis. In conclusion, the proven stability of the NODDI model over time and under different acquisition conditions opens the way to several applications for clinical patients to study the effect and the progression of neurodegenerative diseases at a microscopic level.