Semi-supervised Learning with Contrastive and Topology Losses for Catheter Segmentation and Misplacement Prediction

Tianyu Hwang, Chih-Hung Wang, Holger R Roth, Dong Yang, Can Zhao, Chien-Hua Huang, Weichung Wang
Medical Imaging with Deep Learning, PMLR 227:1239-1253, 2024.

Abstract

Chest X-ray images are often used to determine the proper placement of catheters, as incorrect placement can lead to severe complications. With the advent of deep learning, computer-aided detection methods have been developed to assist radiologists in identifying catheter misplacement by detecting and highlighting the catheter’s path. However, obtaining large, pixel-wise labeled datasets can be challenging due to the labor-intensive nature of annotation. To address this issue, we proposed a novel semi-supervised learning method that combines contrastive loss and topology loss. This method takes advantage of the known topological properties of catheters and does not require extensive labeling. We collected 7,378 chest X-ray images from the *****, which were labeled for misplacement of nasogastric and endotracheal tube catheters, and included pixel-wise annotation. Moreover, the CLiP dataset was used as an unlabeled dataset for semi-supervised learning. We used a hybrid U-Net architecture with an added classification head to perform simultaneous segmentation of the catheter and misplacement classification. Our model achieved an average AUC of 0.977 for classification and a average Dice score of 0.598 for segmentation.

Cite this Paper

BibTeX
@InProceedings{pmlr-v227-hwang24a, title = {Semi-supervised Learning with Contrastive and Topology Losses for Catheter Segmentation and Misplacement Prediction}, author = {Hwang, Tianyu and Wang, Chih-Hung and Roth, Holger R and Yang, Dong and Zhao, Can and Huang, Chien-Hua and Wang, Weichung}, booktitle = {Medical Imaging with Deep Learning}, pages = {1239--1253}, year = {2024}, editor = {Oguz, Ipek and Noble, Jack and Li, Xiaoxiao and Styner, Martin and Baumgartner, Christian and Rusu, Mirabela and Heinmann, Tobias and Kontos, Despina and Landman, Bennett and Dawant, Benoit}, volume = {227}, series = {Proceedings of Machine Learning Research}, month = {10--12 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v227/hwang24a/hwang24a.pdf}, url = {https://proceedings.mlr.press/v227/hwang24a.html}, abstract = {Chest X-ray images are often used to determine the proper placement of catheters, as incorrect placement can lead to severe complications. With the advent of deep learning, computer-aided detection methods have been developed to assist radiologists in identifying catheter misplacement by detecting and highlighting the catheter’s path. However, obtaining large, pixel-wise labeled datasets can be challenging due to the labor-intensive nature of annotation. To address this issue, we proposed a novel semi-supervised learning method that combines contrastive loss and topology loss. This method takes advantage of the known topological properties of catheters and does not require extensive labeling. We collected 7,378 chest X-ray images from the *****, which were labeled for misplacement of nasogastric and endotracheal tube catheters, and included pixel-wise annotation. Moreover, the CLiP dataset was used as an unlabeled dataset for semi-supervised learning. We used a hybrid U-Net architecture with an added classification head to perform simultaneous segmentation of the catheter and misplacement classification. Our model achieved an average AUC of 0.977 for classification and a average Dice score of 0.598 for segmentation.} }
Endnote
%0 Conference Paper %T Semi-supervised Learning with Contrastive and Topology Losses for Catheter Segmentation and Misplacement Prediction %A Tianyu Hwang %A Chih-Hung Wang %A Holger R Roth %A Dong Yang %A Can Zhao %A Chien-Hua Huang %A Weichung Wang %B Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2024 %E Ipek Oguz %E Jack Noble %E Xiaoxiao Li %E Martin Styner %E Christian Baumgartner %E Mirabela Rusu %E Tobias Heinmann %E Despina Kontos %E Bennett Landman %E Benoit Dawant %F pmlr-v227-hwang24a %I PMLR %P 1239--1253 %U https://proceedings.mlr.press/v227/hwang24a.html %V 227 %X Chest X-ray images are often used to determine the proper placement of catheters, as incorrect placement can lead to severe complications. With the advent of deep learning, computer-aided detection methods have been developed to assist radiologists in identifying catheter misplacement by detecting and highlighting the catheter’s path. However, obtaining large, pixel-wise labeled datasets can be challenging due to the labor-intensive nature of annotation. To address this issue, we proposed a novel semi-supervised learning method that combines contrastive loss and topology loss. This method takes advantage of the known topological properties of catheters and does not require extensive labeling. We collected 7,378 chest X-ray images from the *****, which were labeled for misplacement of nasogastric and endotracheal tube catheters, and included pixel-wise annotation. Moreover, the CLiP dataset was used as an unlabeled dataset for semi-supervised learning. We used a hybrid U-Net architecture with an added classification head to perform simultaneous segmentation of the catheter and misplacement classification. Our model achieved an average AUC of 0.977 for classification and a average Dice score of 0.598 for segmentation.
APA
Hwang, T., Wang, C., Roth, H.R., Yang, D., Zhao, C., Huang, C. & Wang, W.. (2024). Semi-supervised Learning with Contrastive and Topology Losses for Catheter Segmentation and Misplacement Prediction. Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 227:1239-1253 Available from https://proceedings.mlr.press/v227/hwang24a.html.

Related Material