Machine learning-enhanced screening funnel for clinical trials in Alzheimer's disease

Area of Science:

• Neuroscience
• Medical Imaging
• Machine Learning

Background:

• Alzheimer's disease (AD) clinical trials are lengthy and costly due to slow, variable disease progression.
• Recruiting suitable participants for early-phase AD studies is challenging.

Purpose of the Study:

• To present a novel screening paradigm integrating disease progression models to enhance AD clinical trial efficiency.
• To identify appropriate candidates for early-phase Alzheimer's disease clinical studies.

Main Methods:

• Enhancing traditional screening funnels with machine learning models, including 3D convolutional neural networks and ensemble models.
• Integrating neuroimaging, demographic, genetic, and clinical data for predictive modeling.

Main Results:

• The approach predicts clinical progression with an area under the curve of 0.836.
• Optimized screening could reduce subject numbers by 55%, recruitment time by 13 months, and PET scans by 72%.

Conclusions:

• This enhanced screening funnel accelerates Alzheimer's disease therapy development by reducing patient burden and trial timelines.
• The customizable funnel improves efficiency for early-phase clinical studies.