Pi-PE: A Pipeline for Pulmonary Embolism Detection using Sparsely Annotated 3D CT Images

Deepta Rajan, David Beymer, Shafiqul Abedin, Ehsan Dehghan
Proceedings of the Machine Learning for Health NeurIPS Workshop, PMLR 116:220-232, 2020.

Abstract

Pulmonary embolisms (PE) are known to be one of the leading causes for cardiac-related mortality. Due to inherent variabilities in how PE manifests and the cumbersome nature of manual diagnosis, there is growing interest in leveraging AI tools for detecting PE. In this paper, we build a two-stage detection pipeline that is accurate, computationally efficient, robust to variations in PE types and kernels used for CT reconstruction, and most importantly, does not require dense annotations. Given the challenges in acquiring expert annotations in large-scale datasets, our approach produces state-of-the-art results with very sparse emboli contours (at 10mm slice spacing), while using models with significantly lower number of parameters. We achieve AUC scores of 0.94 on the validation set and 0.85 on the test set of highly severe PEs. Using a large, real-world dataset characterized by complex PE types and patients from multiple hospitals, we present an elaborate empirical study and provide guidelines for designing highly generalizable pipelines.

Cite this Paper

BibTeX
@InProceedings{pmlr-v116-rajan20a, title = {{Pi-PE: A Pipeline for Pulmonary Embolism Detection using Sparsely Annotated 3D CT Images}}, author = {Rajan, Deepta and Beymer, David and Abedin, Shafiqul and Dehghan, Ehsan}, booktitle = {Proceedings of the Machine Learning for Health NeurIPS Workshop}, pages = {220--232}, year = {2020}, editor = {Dalca, Adrian V. and McDermott, Matthew B.A. and Alsentzer, Emily and Finlayson, Samuel G. and Oberst, Michael and Falck, Fabian and Beaulieu-Jones, Brett}, volume = {116}, series = {Proceedings of Machine Learning Research}, month = {13 Dec}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v116/rajan20a/rajan20a.pdf}, url = {https://proceedings.mlr.press/v116/rajan20a.html}, abstract = {Pulmonary embolisms (PE) are known to be one of the leading causes for cardiac-related mortality. Due to inherent variabilities in how PE manifests and the cumbersome nature of manual diagnosis, there is growing interest in leveraging AI tools for detecting PE. In this paper, we build a two-stage detection pipeline that is accurate, computationally efficient, robust to variations in PE types and kernels used for CT reconstruction, and most importantly, does not require dense annotations. Given the challenges in acquiring expert annotations in large-scale datasets, our approach produces state-of-the-art results with very sparse emboli contours (at 10mm slice spacing), while using models with significantly lower number of parameters. We achieve AUC scores of 0.94 on the validation set and 0.85 on the test set of highly severe PEs. Using a large, real-world dataset characterized by complex PE types and patients from multiple hospitals, we present an elaborate empirical study and provide guidelines for designing highly generalizable pipelines.} }
Endnote
%0 Conference Paper %T Pi-PE: A Pipeline for Pulmonary Embolism Detection using Sparsely Annotated 3D CT Images %A Deepta Rajan %A David Beymer %A Shafiqul Abedin %A Ehsan Dehghan %B Proceedings of the Machine Learning for Health NeurIPS Workshop %C Proceedings of Machine Learning Research %D 2020 %E Adrian V. Dalca %E Matthew B.A. McDermott %E Emily Alsentzer %E Samuel G. Finlayson %E Michael Oberst %E Fabian Falck %E Brett Beaulieu-Jones %F pmlr-v116-rajan20a %I PMLR %P 220--232 %U https://proceedings.mlr.press/v116/rajan20a.html %V 116 %X Pulmonary embolisms (PE) are known to be one of the leading causes for cardiac-related mortality. Due to inherent variabilities in how PE manifests and the cumbersome nature of manual diagnosis, there is growing interest in leveraging AI tools for detecting PE. In this paper, we build a two-stage detection pipeline that is accurate, computationally efficient, robust to variations in PE types and kernels used for CT reconstruction, and most importantly, does not require dense annotations. Given the challenges in acquiring expert annotations in large-scale datasets, our approach produces state-of-the-art results with very sparse emboli contours (at 10mm slice spacing), while using models with significantly lower number of parameters. We achieve AUC scores of 0.94 on the validation set and 0.85 on the test set of highly severe PEs. Using a large, real-world dataset characterized by complex PE types and patients from multiple hospitals, we present an elaborate empirical study and provide guidelines for designing highly generalizable pipelines.
APA
Rajan, D., Beymer, D., Abedin, S. & Dehghan, E.. (2020). Pi-PE: A Pipeline for Pulmonary Embolism Detection using Sparsely Annotated 3D CT Images. Proceedings of the Machine Learning for Health NeurIPS Workshop, in Proceedings of Machine Learning Research 116:220-232 Available from https://proceedings.mlr.press/v116/rajan20a.html.

Related Material