Deep learning for calcium segmentation in intravascular ultrasound images
Knowing the shape of vascular calcifications is crucial for appropriate planning and conductance of percutaneous coronary interventions. The clinical workflow can therefore benefit from automatic segmentation of calcified plaques in intravascular ultrasound (IVUS) images. To solve segmentation probl...
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
De Gruyter
2021-08-01
|
| Series: | Current Directions in Biomedical Engineering |
| Subjects: | |
| Online Access: | https://doi.org/10.1515/cdbme-2021-1021 |
| _version_ | 1818407752577318912 |
|---|---|
| author | Bargsten Lennart Riedl Katharina A. Wissel Tobias Brunner Fabian J. Schaefers Klaus Grass Michael Blankenberg Stefan Seiffert Moritz Schlaefer Alexander |
| author_facet | Bargsten Lennart Riedl Katharina A. Wissel Tobias Brunner Fabian J. Schaefers Klaus Grass Michael Blankenberg Stefan Seiffert Moritz Schlaefer Alexander |
| author_sort | Bargsten Lennart |
| collection | DOAJ |
| description | Knowing the shape of vascular calcifications is crucial for appropriate planning and conductance of percutaneous coronary interventions. The clinical workflow can therefore benefit from automatic segmentation of calcified plaques in intravascular ultrasound (IVUS) images. To solve segmentation problems with convolutional neural networks (CNNs), large datasets are usually required. However, datasets are often rather small in the medical domain. Hence, developing and investigating methods for increasing CNN performance on small datasets can help on the way towards clinically relevant results. We compared two state-of-the-art CNN architectures for segmentation, U-Net and DeepLabV3, and investigated how incorporating auxiliary image data with vessel wall and lumen annotations improves the calcium segmentation performance by using these either for pretraining or multi-task training. DeepLabV3 outperforms U-Net with up to 6.3 % by means of the Dice coefficient and 36.5 % by means of the average Hausdorff distance. Using auxiliary data improves the segmentation performance in both cases, whereas the multi-task approach outperforms the pre-training approach. The improvements of the multi-task approach in contrast to not using auxiliary data at all is 5.7 % for the Dice coefficient and 42.9 % for the average Hausdorff distance. Automatic segmentation of calcified plaques in IVUS images is a demanding task due to their relatively small size compared to the image dimensions and due to visual ambiguities with other image structures. We showed that this problem can generally be tackled by CNNs. Furthermore, we were able to improve the performance by a multi-task learning approach with auxiliary segmentation data. |
| first_indexed | 2024-12-14T09:32:50Z |
| format | Article |
| id | doaj.art-de59b24cd3524e2994684ae7d09b53ef |
| institution | Directory Open Access Journal |
| issn | 2364-5504 |
| language | English |
| last_indexed | 2024-12-14T09:32:50Z |
| publishDate | 2021-08-01 |
| publisher | De Gruyter |
| record_format | Article |
| series | Current Directions in Biomedical Engineering |
| spelling | doaj.art-de59b24cd3524e2994684ae7d09b53ef2022-12-21T23:08:01ZengDe GruyterCurrent Directions in Biomedical Engineering2364-55042021-08-01719610010.1515/cdbme-2021-1021Deep learning for calcium segmentation in intravascular ultrasound imagesBargsten Lennart0Riedl Katharina A.1Wissel Tobias2Brunner Fabian J.3Schaefers Klaus4Grass Michael5Blankenberg Stefan6Seiffert Moritz7Schlaefer Alexander8Hamburg University of Technology, Institute of Medical Technology and Intelligent Systems,Hamburg, GermanyDepartment of Cardiology, University Heart & Vascular Center Hamburg,Hamburg, GermanyPhilips Research -Hamburg, GermanyDepartment of Cardiology, University Heart & Vascular Center Hamburg,Hamburg, GermanyPhilips Research -Eindhoven, The NetherlandsPhilips Research -Hamburg, GermanyDepartment of Cardiology, University Heart & Vascular Center Hamburg,Hamburg, GermanyDepartment of Cardiology, University Heart & Vascular Center Hamburg,Hamburg, GermanyHamburg University of Technology, Institute of Medical Technology and Intelligent Systems,Hamburg, GermanyKnowing the shape of vascular calcifications is crucial for appropriate planning and conductance of percutaneous coronary interventions. The clinical workflow can therefore benefit from automatic segmentation of calcified plaques in intravascular ultrasound (IVUS) images. To solve segmentation problems with convolutional neural networks (CNNs), large datasets are usually required. However, datasets are often rather small in the medical domain. Hence, developing and investigating methods for increasing CNN performance on small datasets can help on the way towards clinically relevant results. We compared two state-of-the-art CNN architectures for segmentation, U-Net and DeepLabV3, and investigated how incorporating auxiliary image data with vessel wall and lumen annotations improves the calcium segmentation performance by using these either for pretraining or multi-task training. DeepLabV3 outperforms U-Net with up to 6.3 % by means of the Dice coefficient and 36.5 % by means of the average Hausdorff distance. Using auxiliary data improves the segmentation performance in both cases, whereas the multi-task approach outperforms the pre-training approach. The improvements of the multi-task approach in contrast to not using auxiliary data at all is 5.7 % for the Dice coefficient and 42.9 % for the average Hausdorff distance. Automatic segmentation of calcified plaques in IVUS images is a demanding task due to their relatively small size compared to the image dimensions and due to visual ambiguities with other image structures. We showed that this problem can generally be tackled by CNNs. Furthermore, we were able to improve the performance by a multi-task learning approach with auxiliary segmentation data.https://doi.org/10.1515/cdbme-2021-1021multi-task learningsmall datasetcoronary arteryvesselconvolutional neural network |
| spellingShingle | Bargsten Lennart Riedl Katharina A. Wissel Tobias Brunner Fabian J. Schaefers Klaus Grass Michael Blankenberg Stefan Seiffert Moritz Schlaefer Alexander Deep learning for calcium segmentation in intravascular ultrasound images Current Directions in Biomedical Engineering multi-task learning small dataset coronary artery vessel convolutional neural network |
| title | Deep learning for calcium segmentation in intravascular ultrasound images |
| title_full | Deep learning for calcium segmentation in intravascular ultrasound images |
| title_fullStr | Deep learning for calcium segmentation in intravascular ultrasound images |
| title_full_unstemmed | Deep learning for calcium segmentation in intravascular ultrasound images |
| title_short | Deep learning for calcium segmentation in intravascular ultrasound images |
| title_sort | deep learning for calcium segmentation in intravascular ultrasound images |
| topic | multi-task learning small dataset coronary artery vessel convolutional neural network |
| url | https://doi.org/10.1515/cdbme-2021-1021 |
| work_keys_str_mv | AT bargstenlennart deeplearningforcalciumsegmentationinintravascularultrasoundimages AT riedlkatharinaa deeplearningforcalciumsegmentationinintravascularultrasoundimages AT wisseltobias deeplearningforcalciumsegmentationinintravascularultrasoundimages AT brunnerfabianj deeplearningforcalciumsegmentationinintravascularultrasoundimages AT schaefersklaus deeplearningforcalciumsegmentationinintravascularultrasoundimages AT grassmichael deeplearningforcalciumsegmentationinintravascularultrasoundimages AT blankenbergstefan deeplearningforcalciumsegmentationinintravascularultrasoundimages AT seiffertmoritz deeplearningforcalciumsegmentationinintravascularultrasoundimages AT schlaeferalexander deeplearningforcalciumsegmentationinintravascularultrasoundimages |