Artificial Intelligence for individual profiling and prediction: probing the FractaL dimension of brain MRI in FriEdreich ataXia and SCAs using the ENIGMA-Ataxia international meta-dataset (FLEX-AI)

Friedreich (FRDA) and (SCAs) spinocerebellar ataxias are rare, slowly progressing and highly debilitating diseases. Understanding and predicting inter-individual variability, and treatment monitoring indices for use in pending drug and gene therapy trials are urgently required. Identification of imaging indices of disease progression could allow reduction of the sample sizes needed in trials evaluating new therapeutic approaches. The fractal dimension (FD) is a promising quantitative index of structural brain complexity that can be derived from conventional magnetic resonance images (MRI), with the potential to provide further insights into the changes underlying abnormal brain development and aging in these diseases. In the project, we will evaluate the FD as a novel imaging indices. Using artificial intelligence techniques, we will evaluate the added predictive value of the FD feature to predict, at a single-patient level, clinical deterioration in FRDA and SCAs in the ENIGMA ataxia meta-dataset. The ENIGMA-Ataxia international consortium collects brain MRI, genetic and clinical data of patients with inherited cerebellar disorders from 21 sites world-wide. The ENIGMA ataxia meta-dataset is the largest available today including >800 patients and about 800 healthy subjects. We will use the ENIGMA-Ataxia platform to quantify FD abnormalities associated with abnormal development and/or neurodegeneration in large samples of FRDA, SCA1, SCA2, SCA3, SCA6 and SCA7. This work will provide leading innovation and cutting-edge computational approaches, leveraging off of an established international collaboration, to better understand disease pathophysiology and identify strategies to improve clinical trial design.