Surgical Planning Laboratory - Brigham & Women's Hospital - Boston, Massachusetts USA - a teaching affiliate of Harvard Medical School

Surgical Planning Laboratory

The Publication Database hosted by SPL

All Publications | Upload | Advanced Search | Gallery View | Download Statistics | Help | Import | Log in

Transfer Learning for Domain Adaptation in MRI: Application in Brain Lesion Segmentation

Institution:
1Diagnostic Image Analysis Group, Radboud University Medical Center, Nijmegen, The Netherlands.
2Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA
3Institute for Computing and Information Sciences, Radboud University, Nijmegen, The Netherlands.
4University of British Columbia, Vancouver, BC, Canada.
5Department of Neurology, Donders Institute for Brain, Cognition and Behaviour, Diagnostic Image Analysis Group, Radboud University Medical Center, Nijmegen, The Netherlands.
Publisher:
Int Conf Med Image Comput Comput Assist Interv. MICCAI 2017
Publication Date:
Sep-2017
Volume Number:
20
Issue Number:
Pt. 3
Pages:
516-24
Citation:
Int Conf Med Image Comput Comput Assist Interv. 2017 Sep;20(Pt3):516-24.
Appears in Collections:
NCIGT, Prostate Group, SLICER, SPL
Sponsors:
P41 EB015898/EB/NIBIB NIH HHS/United States
Generated Citation:
Ghafoorian M., Mehrtash A., Kapur T., Karssemeijer N., Marchiori E., Pesteie M., Guttmann C.R.G., de Leeuw F-E., Tempany C.M., van Ginneken B., Fedorov A., Abolmaesumi P., Platel B., Wells III W.M. Transfer Learning for Domain Adaptation in MRI: Application in Brain Lesion Segmentation. Int Conf Med Image Comput Comput Assist Interv. 2017 Sep;20(Pt3):516-24.
Downloaded: 32 times. [view map]
Paper: Download, View online
Export citation:
Google Scholar: link

Magnetic Resonance Imaging (MRI) is widely used in routine clinical diagnosis and treatment. However, variations in MRI acquisition protocols result in different appearances of normal and diseased tissue in the images. Convolutional neural networks (CNNs), which have shown to be successful in many medical image analysis tasks, are typically sensitive to the variations in imaging protocols. Therefore, in many cases, networks trained on data acquired with one MRI protocol, do not perform satisfactorily on data acquired with different protocols. This limits the use of models trained with large annotated legacy datasets on a new dataset with a different domain which is often a recurring situation in clinical settings. In this study, we aim to answer the following central questions regarding domain adaptation in medical image analysis: Given a fitted legacy model, (1) How much data from the new domain is required for a decent adaptation of the original network?; and, (2) What portion of the pre-trained model parameters should be retrained given a certain number of the new domain training samples? To address these questions, we conducted extensive experiments in white matter hyperintensity segmentation task. We trained a CNN on legacy MR images of brain and evaluated the performance of the domain-adapted network on the same task with images from a different domain. We then compared the performance of the model to the surrogate scenarios where either the same trained network is used or a new network is trained from scratch on the new dataset. The domain-adapted network tuned only by two training examples achieved a Dice score of 0.63 substantially outperforming a similar network trained on the same set of examples from scratch.

Additional Material
1 File (101.059kB)
Ghafoorian-MICCAI2017-fig3.jpg (101.059kB)