Segmenting Cervical Arteries in Phase Contrast Magnetic Resonance Imaging Using Convolutional Encoder–Decoder Networks
Phase contrast (PC) magnetic resonance imaging (MRI) is a primary method used to quantify blood flow. Cerebral blood flow (CBF) is an important hemodynamic parameter to characterize cerebrovascular and neurological diseases. However, a critical step in CBF quantification using PC MRI is vessel segme...
| Main Authors: | , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-10-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/13/21/11820 |
| _version_ | 1797632223183110144 |
|---|---|
| author | Britney Campbell Dhruv Yadav Ramy Hussein Maria Jovin Sierrah Hoover Kim Halbert Dawn Holley Mehdi Khalighi Guido A. Davidzon Elizabeth Tong Gary K. Steinberg Michael Moseley Moss Y. Zhao Greg Zaharchuk |
| author_facet | Britney Campbell Dhruv Yadav Ramy Hussein Maria Jovin Sierrah Hoover Kim Halbert Dawn Holley Mehdi Khalighi Guido A. Davidzon Elizabeth Tong Gary K. Steinberg Michael Moseley Moss Y. Zhao Greg Zaharchuk |
| author_sort | Britney Campbell |
| collection | DOAJ |
| description | Phase contrast (PC) magnetic resonance imaging (MRI) is a primary method used to quantify blood flow. Cerebral blood flow (CBF) is an important hemodynamic parameter to characterize cerebrovascular and neurological diseases. However, a critical step in CBF quantification using PC MRI is vessel segmentation, which is largely manual, and thus time-consuming and prone to interrater variability. Here, we present encoder–decoder deep learning models to automate segmentation of neck arteries to accurately quantify CBF. The PC-MRI data were collected from 46 Moyamoya (MM) patients and 107 healthy control (HC) participants. Three segmentation U-Net models (Standard, Nested, and Attention) were compared. The PC MRI images were taken before and 15 min after vasodilation. The models were assessed based on their ability to detect the internal carotid arteries (ICAs), external carotid arteries (ECAs), and vertebral arteries (VAs), using the Dice score coefficient (DSC) of overlap between manual and predicted segmentations and receiver operator characteristic (ROC) metric. Analysis of variance, Wilcoxon rank-sum test, and paired <i>t</i>-test were used for comparisons. The Standard U-NET, Attention U-Net, and Nest U-Net models achieved results of mean DSCs of 0.81 ± 0.21, and 0.85 ± 0.14, and 0.85 ± 0.13, respectively. The ROC curves revealed high area under the curve scores for all methods (≥0.95). While the Nested and Attention U-Net architectures accomplished reliable segmentation performance for HC and MM subsets, Standard U-Net did not perform as well in the subset of MM patients. Blood flow velocities calculated by the models were statistically comparable. In conclusion, optimized deep learning architectures can successfully segment neck arteries in PC MRI images and provide precise quantification of their blood flow. |
| first_indexed | 2024-03-11T11:33:52Z |
| format | Article |
| id | doaj.art-ab8723b0e7484f8983829ef156e8e332 |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-11T11:33:52Z |
| publishDate | 2023-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-ab8723b0e7484f8983829ef156e8e3322023-11-10T14:58:51ZengMDPI AGApplied Sciences2076-34172023-10-0113211182010.3390/app132111820Segmenting Cervical Arteries in Phase Contrast Magnetic Resonance Imaging Using Convolutional Encoder–Decoder NetworksBritney Campbell0Dhruv Yadav1Ramy Hussein2Maria Jovin3Sierrah Hoover4Kim Halbert5Dawn Holley6Mehdi Khalighi7Guido A. Davidzon8Elizabeth Tong9Gary K. Steinberg10Michael Moseley11Moss Y. Zhao12Greg Zaharchuk13Department of Radiology, Stanford University, Stanford, CA 94305, USADepartment of Radiology, Stanford University, Stanford, CA 94305, USADepartment of Radiology, Stanford University, Stanford, CA 94305, USADepartment of Radiology, Stanford University, Stanford, CA 94305, USADepartment of Radiology, Stanford University, Stanford, CA 94305, USADepartment of Radiology, Stanford University, Stanford, CA 94305, USADepartment of Radiology, Stanford University, Stanford, CA 94305, USADepartment of Radiology, Stanford University, Stanford, CA 94305, USADepartment of Radiology, Stanford University, Stanford, CA 94305, USADepartment of Radiology, Stanford University, Stanford, CA 94305, USADepartment of Neurosurgery, Stanford University, Stanford, CA 94305, USADepartment of Radiology, Stanford University, Stanford, CA 94305, USADepartment of Radiology, Stanford University, Stanford, CA 94305, USADepartment of Radiology, Stanford University, Stanford, CA 94305, USAPhase contrast (PC) magnetic resonance imaging (MRI) is a primary method used to quantify blood flow. Cerebral blood flow (CBF) is an important hemodynamic parameter to characterize cerebrovascular and neurological diseases. However, a critical step in CBF quantification using PC MRI is vessel segmentation, which is largely manual, and thus time-consuming and prone to interrater variability. Here, we present encoder–decoder deep learning models to automate segmentation of neck arteries to accurately quantify CBF. The PC-MRI data were collected from 46 Moyamoya (MM) patients and 107 healthy control (HC) participants. Three segmentation U-Net models (Standard, Nested, and Attention) were compared. The PC MRI images were taken before and 15 min after vasodilation. The models were assessed based on their ability to detect the internal carotid arteries (ICAs), external carotid arteries (ECAs), and vertebral arteries (VAs), using the Dice score coefficient (DSC) of overlap between manual and predicted segmentations and receiver operator characteristic (ROC) metric. Analysis of variance, Wilcoxon rank-sum test, and paired <i>t</i>-test were used for comparisons. The Standard U-NET, Attention U-Net, and Nest U-Net models achieved results of mean DSCs of 0.81 ± 0.21, and 0.85 ± 0.14, and 0.85 ± 0.13, respectively. The ROC curves revealed high area under the curve scores for all methods (≥0.95). While the Nested and Attention U-Net architectures accomplished reliable segmentation performance for HC and MM subsets, Standard U-Net did not perform as well in the subset of MM patients. Blood flow velocities calculated by the models were statistically comparable. In conclusion, optimized deep learning architectures can successfully segment neck arteries in PC MRI images and provide precise quantification of their blood flow.https://www.mdpi.com/2076-3417/13/21/11820deep learningphase contrast MRIblood flowMoyamoya |
| spellingShingle | Britney Campbell Dhruv Yadav Ramy Hussein Maria Jovin Sierrah Hoover Kim Halbert Dawn Holley Mehdi Khalighi Guido A. Davidzon Elizabeth Tong Gary K. Steinberg Michael Moseley Moss Y. Zhao Greg Zaharchuk Segmenting Cervical Arteries in Phase Contrast Magnetic Resonance Imaging Using Convolutional Encoder–Decoder Networks Applied Sciences deep learning phase contrast MRI blood flow Moyamoya |
| title | Segmenting Cervical Arteries in Phase Contrast Magnetic Resonance Imaging Using Convolutional Encoder–Decoder Networks |
| title_full | Segmenting Cervical Arteries in Phase Contrast Magnetic Resonance Imaging Using Convolutional Encoder–Decoder Networks |
| title_fullStr | Segmenting Cervical Arteries in Phase Contrast Magnetic Resonance Imaging Using Convolutional Encoder–Decoder Networks |
| title_full_unstemmed | Segmenting Cervical Arteries in Phase Contrast Magnetic Resonance Imaging Using Convolutional Encoder–Decoder Networks |
| title_short | Segmenting Cervical Arteries in Phase Contrast Magnetic Resonance Imaging Using Convolutional Encoder–Decoder Networks |
| title_sort | segmenting cervical arteries in phase contrast magnetic resonance imaging using convolutional encoder decoder networks |
| topic | deep learning phase contrast MRI blood flow Moyamoya |
| url | https://www.mdpi.com/2076-3417/13/21/11820 |
| work_keys_str_mv | AT britneycampbell segmentingcervicalarteriesinphasecontrastmagneticresonanceimagingusingconvolutionalencoderdecodernetworks AT dhruvyadav segmentingcervicalarteriesinphasecontrastmagneticresonanceimagingusingconvolutionalencoderdecodernetworks AT ramyhussein segmentingcervicalarteriesinphasecontrastmagneticresonanceimagingusingconvolutionalencoderdecodernetworks AT mariajovin segmentingcervicalarteriesinphasecontrastmagneticresonanceimagingusingconvolutionalencoderdecodernetworks AT sierrahhoover segmentingcervicalarteriesinphasecontrastmagneticresonanceimagingusingconvolutionalencoderdecodernetworks AT kimhalbert segmentingcervicalarteriesinphasecontrastmagneticresonanceimagingusingconvolutionalencoderdecodernetworks AT dawnholley segmentingcervicalarteriesinphasecontrastmagneticresonanceimagingusingconvolutionalencoderdecodernetworks AT mehdikhalighi segmentingcervicalarteriesinphasecontrastmagneticresonanceimagingusingconvolutionalencoderdecodernetworks AT guidoadavidzon segmentingcervicalarteriesinphasecontrastmagneticresonanceimagingusingconvolutionalencoderdecodernetworks AT elizabethtong segmentingcervicalarteriesinphasecontrastmagneticresonanceimagingusingconvolutionalencoderdecodernetworks AT garyksteinberg segmentingcervicalarteriesinphasecontrastmagneticresonanceimagingusingconvolutionalencoderdecodernetworks AT michaelmoseley segmentingcervicalarteriesinphasecontrastmagneticresonanceimagingusingconvolutionalencoderdecodernetworks AT mossyzhao segmentingcervicalarteriesinphasecontrastmagneticresonanceimagingusingconvolutionalencoderdecodernetworks AT gregzaharchuk segmentingcervicalarteriesinphasecontrastmagneticresonanceimagingusingconvolutionalencoderdecodernetworks |