Support development of the medical implant screening as part of the Federated Health Nordic Innovation project using federated learning and natural language processing.

Ongoing:	Yes
Start:	2024-08-23

Before an MRI (Magnetic Resonance Imaging) examination, it is crucial to know if a patient has medical implants, such as pacemakers, which can pose significant risks during the procedure. However, identifying and verifying these implants is challenging because current medical record systems are not structured to easily extract this information. The process is often manual, time-consuming, and involves various experts reviewing entire patient records—posing a safety risk due to potential errors.

The aiMPLANT project addresses this issue by developing an AI solution that automatically identifies medical implants from patient records, reducing the time and effort required while improving accuracy. Building on previous work in implant terminology extraction, the goal is to enhance the model’s robustness in detecting a wide range of devices across diverse clinical texts. By using federated learning from multiple nordic sources, we aim to improve patient safety by ensuring critical implant information is not overlooked during MRI examinations.

The AIDA Data Hub supports the project in developing Natural Language Processing (NLP) techniques as well as running federated learning experiments together with the nordic partners.