Pseudo-domains in imaging data improve prediction of future disease status in multi-center studies

Matthias Perkonigg; Peter Mesenbrink; Alexander Goehler; Miljen Martic; Ahmed Ba-Ssalamah; Georg Langs

Pseudo-domains in imaging data improve prediction of future disease status in multi-center studies

Image and Video Processing 2021-11-16 v1 Computer Vision and Pattern Recognition Machine Learning

Authors: Matthias Perkonigg , Peter Mesenbrink , Alexander Goehler , Miljen Martic , Ahmed Ba-Ssalamah , Georg Langs

Abstract

In multi-center randomized clinical trials imaging data can be diverse due to acquisition technology or scanning protocols. Models predicting future outcome of patients are impaired by this data heterogeneity. Here, we propose a prediction method that can cope with a high number of different scanning sites and a low number of samples per site. We cluster sites into pseudo-domains based on visual appearance of scans, and train pseudo-domain specific models. Results show that they improve the prediction accuracy for steatosis after 48 weeks from imaging data acquired at an initial visit and 12-weeks follow-up in liver disease

Keywords

multi-center studies imaging data pseudo-domains site clustering disease prediction data heterogeneity liver steatosis

Cite

@article{arxiv.2111.07634,
  title  = {Pseudo-domains in imaging data improve prediction of future disease status in multi-center studies},
  author = {Matthias Perkonigg and Peter Mesenbrink and Alexander Goehler and Miljen Martic and Ahmed Ba-Ssalamah and Georg Langs},
  journal= {arXiv preprint arXiv:2111.07634},
  year   = {2021}
}

Comments

Accepted at Medical Imaging Meets NeurIPS 2021

Pseudo-domains in imaging data improve prediction of future disease status in multi-center studies

Abstract

Keywords

Cite

Comments

Related papers