Machine learning tool for intra-operative neurophysiological monitoring

This technology is a tool that uses machine learning to autonomously detect intra-operative neuromonitoring anomalies during spine surgery.

Unmet Need: Human variation in intra-operative neurophysiological monitoring

Intra-operative neurophysiological monitoring (IONM) during spine surgery depends on the interpretation of complex data by highly trained technicians, neurologists, or neurophysiologists. The effectiveness of IONM is highly biased by human variation in monitoring and can affect the neurological safety of patients undergoing surgery.

The Technology: Machine learning model for standardizing intra-operative neurophysiological monitoring

This technology uses a machine learning model to identify anomalous patterns of motor-evoked action potentials (MEPs) in patients undergoing spine surgery. It is trained on a patient’s baseline myoelectric signal prior to instrumentation to generate a patient-specific signature which is then used to identify anomalies during the rest of the case. Notable anomalies or deviations in myoelectric signals are highly predictive of adverse changes in a patient’s neurologic condition. This technology can improve patient safety during spine surgery and serve as an additional warning system for surgical personnel in real time.

This technology has been validated in a proof-of-concept retrospective analysis of 84 patients.

Applications:

• Detection of anomalies in intra-operative neurophysiological monitoring (IONM) during spine surgery
• Detection of IONM anomalies during other surgeries, including but not limited to vascular, neurology, otolaryngologic surgery
• Standardization of IONM interpretation across surgeries

Advantages:

• Automated and unbiased intra-operative neurophysiological monitoring (IONM)
• Cost-effective compared to the training of qualified technicians
• Identification of subtle signal changes that can point to anomalies
• Early detection of adverse events in real time (demonstrated detection prior to detection by trained human IONM team)

Lead Inventor:

Varun Arvind, M.D.

Related Publications:

Tech Ventures Reference:

• IR CU25236

• Licensing Contact: Jerry Kokoshka