In the ENTRAPME project we aim to devise, validate and implement a machine-learning based clinical decision support system for diagnosing nerve entrapments.

The project is funded by the South-Eastern Norway Regional Health Authority.

Background

Entrapment neuropathies such as carpal tunnel syndrome and ulnar entrapment affect around 6% of the population and is one of the musculoskeletal diseases resulting in the most lost workdays. With a correct and timely diagnosis, the prognosis is good. The diagnostics of entrapment neuropathies rely heavily on nerve conduction studies (NCS), which are interpreted by physician specialists in clinical neurophysiology. However, this is both costly and inefficient, there are few such specialists in Norway with little prospect of improvement, and many hospitals thusly remain underserved. An artificial intelligence (AI) based clinical decision support with pattern recognition can increase diagnostic efficiency and reduce the burden on neurophysiological expertise. We conservatively aim to reduce personnel cost for diagnosing entrapments by 20-30 %, constituting a large economic impact worldwide. In Norway alone, the project would impact at least 15-20 000 subjects each year.  

Aims

The long-term aim is improved diagnostics for patients suffering from entrapment disorders. 

The short-term aims are: 

• Finalize database and prepare infrastructure for clinical validation 
• Develop a novel machine-learning based model to recognize specific disease patterns in NCS
• Validation of the model in clinical practice with a feasibility study
• Implement software in commercially available format in collaboration with industrial partner.