Brachial plexus nerve segmentation

Background

Regional nerve block is a common anaesthesia technique used for surgery on the extremities. A successful block requires excellent anaesthesia experience including the ability to identify the appropriate nerves and surrounding tissues on ultrasound and good skills with a needle.

Previous studies have primarily focussed on the usage of ultrasound which has shown that ultrasound increases the success rate of regional nerve blocks. Some studies, however, have found that even with ultrasound assistance, a relatively high failure rate persists. This failure rate has largely been attributed to operators with limited experience and insufficient ultrasound skills.

A failed nerve block not only results in a bad experience for the patient, it might even lead to damage to the patients’ health and in some cases complications could even threaten life.

It’s essential to recognise ultrasound anatomy when performing nerve blocks, however this may sometimes be hampered by patients’ habitus.

Aim

In this research line, we aim to develop, validate and deploy algorithms that assist in the interpretation of ultrasound imaging for brachial plexus blockade. This research line is led by Rob Tolboom.

Click on the cards below to learn more about the various projects.

Projects

Dataset creation with manual segmentation

The goal of this project is to gather 500 ultrasound images of the brachial plexus in 250 healthy adults and manually annotate the brachial plexus.

Read more →
Training U-net model

The goal of this project is to use the dataset with manual segmentation masks to train an U-net model.

Read more →

People

Rob Tolboom

Rob Tolboom

Radboudumc, Nijmegen

Erik Dortangs

Erik Dortangs

ETZ, Tilburg

Nicola Pezzotti

Nicola Pezzotti

Assistant Professor

TU Eindhoven / Philips Research

Ruud van Sloun

Ruud van Sloun

Assistant Professor

TU Eindhoven / Philips Research