Deep learning model predicts Alzheimer’s disease more accurately than conventional methods

Takeaway

  • Novel deep learning prognostic model for disease progression from mild cognitive impairment (MCI) to Alzheimer’s disease (AD) achieves better performance than state-of-the-art alternative imaging feature extraction methods.

Why this matters

  • Existing methods for predicting progression of MCI to AD rely on relatively simple imaging measures (tissue density, cortical thickness) and binary classifications (progressive or stable MCI) that do not accommodate the heterogeneous nature of those with the disease.

  • Using deep learning may help accurately and economically predict which patients with MCI will progress to AD and within what timeframe, potentially improving clinical trial enrollment and treatment strategies.