Statistical issues: MRI to investigate injuries to the brachial plexus
Category: Imaging Studies
This meeting will bring together statisticians, clinicians and other methodologists for discussion of common statistical issues in imaging studies. On this occasion, Mr Ryckie Wade, NIHR Doctoral Research Fellow in Plastic & Reconstructive Surgery at the University of Leeds, will present his work evaluating the role of diffuse-weighted MRI to diagnose injury to the nerve roots of the brachial plexus. Small group discussion will follow to discuss the statistical issues of the study design and analysis; these issues will be widely applicable to other imaging modalities.
Summary: The brachial plexus is a network of nerves which supplies the upper limbs with movement and feeling. It is composed of five nerve roots, per side. The most common site of injury is the root. Injuries are always unilateral. To diagnose injury to the roots, surgeons perform exploratory surgery because current imaging is not sufficiently accurate. Therefore, exploratory surgery serves as the reference standard (injured vs healthy). The study is investigating how diffusion MRI measurements vary in healthy vs injured roots. The outcomes are scaled and normally distributed. Patients are scanned twice before surgery (close time points) and twice after (months apart).
The meeting will include small group discussion of statistical issues of this study design and analysis, including:
- Estimating agreement between 2 preoperative scans
- Estimating differences between injured and uninjured nerves
- Modelling change over time for injured and uninjured nerves
- Dealing with missing (predictor) data
- Addressing effect modifiers (age may modify the outcomes of interest)
A full program will be made available shortly.
Places are limited, so please register through Eventbrite at your earliest convenience. This event is free of charge.