AI and Dynamic MRI to Assess Abdominal Hernias

Niamh Belton | (opens in a new window)niamhbelton@gmail.com

Together, we’re focused on our goals of;

1. Automating the analysis of abdominal MRIs.
Manually segmenting muscles and calculating functional metrics can take hours per patient. We're developing an end-to-end AI tool that completes these tasks in seconds, capturing essential indicators like muscle displacement and inter-muscle distance, a known risk factor for recurrence.

2. Using generative AI to model patient outcomes.
By comparing how abdominal walls function before and after surgery, our models aim to pinpoint the features most predictive of surgical success and failure.

Our vision is to build AI tools that empowers clinicians, supports patients, and reduces recurrence through smarter, more personalised care.

• Assistant professor Dr. Kathleen Curran affiliated with the School of Medicine, University College Dublin, Science Foundation Ireland Centre for Research Training in Machine Learning and Insight Centre for Data Analytics, Dublin, Ireland. She is also the director of Machine Learning in Medical Imaging Research Group. Dr. Kathleen Curran has a PhD in Computer Science (University College London) and >20 years experience leading innovative methods research applied to real-world medical imaging and diagnostics applications.

• David Bendahan from Aix-Marseille University is an expert in musculoskeletal magnetic resonance. David is the director of the National Centre for Research Science (CNRS) in Aix-Marseille university, France and is a member of the Center for Magnetic Resonance in Biology and Medicine in France.

• Assistant professor Dr. Aonghus Lawlor from UCD and the Insight Centre for Data Analytics.

• Niamh Belton, former PhD student in UCD and a member of SFI funded ML-Labs with a background in Data Science and Machine Learning. Niamh’s PhD focused on the specific task of leveraging Machine Learning and AI for medical imaging problems.

• Thierry Bege has an MD and practices as a surgeon.

• Catherine Masson is a senior researcher at the Laboratory of Biomechanics and Applications (LBA), Aix Marseille University.

• Victoria Joppin, former PhD student from Aix-Marseille university with a background in mechanical engineering. Victoria’s PhD focused on investigating the dynamics of the abdominal muscles in healthy vs. patients with hernias and predicting the occurrence/recurrence of abdominal hernias.

Publications

Joppin, V., Belton, N., Hostin, M.A., Bellemare, M.E., Lawlor, A., Curran, K., Bège, T., Masson,
C. and David, B., 2024, July. Automatic muscle segmentation on healthy abdominal MRI
using nnUNet. In Medical Imaging with Deep Learning (MIDL2024).

Belton, N., Joppin, V., Lawlor, A., Curran, K.M., Masson, C., Bege, T. and Bendahan, D., 2024.
DyABD: A Dataset and Technique for Synthetically Generating Dynamic Abdominal MRIs
with Dual Class and Anatomically Conditioned Diffusion Models. In Medical Imaging with
Deep Learning.