All publications

Journal article / 2026

Geometric deep learning for local growth prediction on abdominal aortic aneurysm surfaces

Dieuwertje Alblas, Patryk Rygiel, Julian Suk, Kaj O. Kappe, Marieke Hofman, Christoph Brune, Kak Khee Yeung, Jelmer M. Wolterink

Artificial Intelligence in Medicine · Published

A vascular surface and local physical features enter LaB-GATr, which predicts deformation vectors and the future aneurysm surface.
Local growth prediction combines vascular geometry, multiphysical features and the follow-up interval to estimate changes to the aneurysm surface. Figure 1 · Alblas et al. (2025 preprint) · Alblas et al. · CC BY 4.0. Original figure from arXiv:2506.08729v2.

Predicting local changes in vessel shape

This study models aneurysm progression on the vascular surface, retaining the spatial structure that a single diameter measurement cannot capture. A transformer combines geometry with local physical features to estimate growth between follow-up scans.

The method is evaluated on longitudinal CT angiography and an external hospital dataset. The work explores the potential of directional growth prediction for more individualized surveillance.

Cite this work

@article{aaagrowth2026,
  title = {Geometric deep learning for local growth prediction on abdominal aortic aneurysm surfaces},
  author = { Dieuwertje Alblas and Patryk Rygiel and Julian Suk and Kaj O. Kappe and Marieke Hofman and Christoph Brune and Kak Khee Yeung and Jelmer M. Wolterink },
  year = {2026},
  journal = {Artificial Intelligence in Medicine},
  volume = {172},
  eid = {103323},
  doi = {10.1016/j.artmed.2025.103323},
  eprint = {2506.08729},
  archivePrefix = {arXiv},
  url = { https://doi.org/10.1016/j.artmed.2025.103323 }
}