Training Robust T1-Weighted Magnetic Resonance Imaging Liver Segmentation Models Using Ensembles of Datasets With Different Contrast Protocols and Liver Disease Etiologies

Authors

Nihil Patel
Adrian Celaya
Mohamed Eltaher
Rachel Glenn
Kari Brewer Savannah
Kristy K Brock
Jessica I Sanchez
Tiffany L Calderone
Darrel Cleere
Ahmed Elsaiey
Matthew Cagley
Nakul Gupta
David Victor
Laura Beretta
Eugene J Koay
Tucker J Netherton
David T Fuentes

Publication Date

9-9-2024

Journal

Scientific Reports

Abstract

Image segmentation of the liver is an important step in treatment planning for liver cancer. However, manual segmentation at a large scale is not practical, leading to increasing reliance on deep learning models to automatically segment the liver. This manuscript develops a generalizable deep learning model to segment the liver on T1-weighted MR images. In particular, three distinct deep learning architectures (nnUNet, PocketNet, Swin UNETR) were considered using data gathered from six geographically different institutions. A total of 819 T1-weighted MR images were gathered from both public and internal sources. Our experiments compared each architecture's testing performance when trained both intra-institutionally and inter-institutionally. Models trained using nnUNet and its PocketNet variant achieved mean Dice-Sorensen similarity coefficients>0.9 on both intra- and inter-institutional test set data. The performance of these models suggests that nnUNet and PocketNet liver segmentation models trained on a large and diverse collection of T1-weighted MR images would on average achieve good intra-institutional segmentation performance.

Keywords

Humans, Magnetic Resonance Imaging, Deep Learning, Liver, Liver Diseases, Contrast Media, Image Processing, Computer-Assisted, Liver Neoplasms

DOI

10.1038/s41598-024-71674-y

PMID

39251664

PMCID

PMC11385384

PubMedCentral® Posted Date

9-9-2024

PubMedCentral® Full Text Version

Post-print

Published Open-Access

yes