Unsupervised deep clustering and reinforcement learning can accurately segment MRI brain tumors with very small training sets Conference Paper


Authors: Stember, J. N.; Shalu, H.
Title: Unsupervised deep clustering and reinforcement learning can accurately segment MRI brain tumors with very small training sets
Conference Title: International Symposium on Intelligent Informatics (ISI 2022)
Abstract: Lesion segmentation in medical imaging is key to evaluating treatment response. We have recently shown that reinforcement learning can be applied to radiological images for lesion localization. Furthermore, we demonstrated that reinforcement learning addresses important limitations of supervised deep learning, namely, it can eliminate the requirement for large amounts of annotated training data and can provide valuable intuition lacking in supervised approaches. However, we did not address the fundamental task of lesion/structure-of-interest segmentation. Here we introduce a method combining unsupervised deep learning clustering with reinforcement learning to segment brain lesions on MRI. We initially clustered images using unsupervised deep learning clustering to generate candidate lesion masks for each MRI image. The user then selected the best mask for each of 10 training images. We then trained a reinforcement learning algorithm to select the masks. We tested the corresponding trained deep Q network on a separate testing set of 10 images. For comparison, we also trained and tested a U-net supervised deep learning network on the same set of training/testing images. Whereas the supervised approach quickly overfits the training data and predictably performed poorly on the testing set (16% average Dice score), the unsupervised deep clustering and reinforcement learning achieved an average Dice score of 83%. We have demonstrated a proof-of-principle application of unsupervised deep clustering and reinforcement learning to segment brain tumors. The approach represents human-allied AI that requires minimal input from the radiologist without the need for hand-traced annotation. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
Keywords: treatment response; magnetic resonance imaging; brain; medical imaging; tumors; brain tumors; learning systems; learning algorithms; deep learning; radiological images; reinforcement learning; reinforcement learnings; training sets; small training; testing sets; clustering learning; learning clustering; lesion segmentations
Journal Title Smart Innovation, Systems and Technologies
Volume: 333
Conference Dates: 2022 Aug 31-Sep 2
Conference Location: Trivandrum, India
ISBN: 2190-3018
Publisher: Springer  
Location: Singapore
Date Published: 2023-04-01
Start Page: 255
End Page: 270
Language: English
DOI: 10.1007/978-981-19-8094-7_19
PROVIDER: scopus
DOI/URL:
Notes: This conference paper was published in a book titled "International Symposium on Intelligent Informatics: Proceedings of ISI 2022" (ISBN: 978-981-19-8093-0) -- The event was also held virtually --Export Date: 1 May 2023 -- Source: Scopus
Altmetric
Citation Impact
BMJ Impact Analytics
MSK Authors
  1. Joseph Nathaniel Stember
    19 Stember