Impact of Diffusion–Perfusion Mismatch on Predicting Final Infarction Lesion Using Deep Learning
We report a study that validates the impact of diffusion-perfusion mismatch in a deep learning (DL) model predicting the final infarction lesion from baseline magnetic resonance imaging (MRI). From 472 consecutive patients with acute ischemic stroke, we gathered baseline and follow-up MRI having int...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2022-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/9875295/ |
_version_ | 1798031480213995520 |
---|---|
author | Sungdong Lee Leonard Sunwoo Youngwon Choi Jae Hyup Jung Seung Chai Jung Joong-Ho Won |
author_facet | Sungdong Lee Leonard Sunwoo Youngwon Choi Jae Hyup Jung Seung Chai Jung Joong-Ho Won |
author_sort | Sungdong Lee |
collection | DOAJ |
description | We report a study that validates the impact of diffusion-perfusion mismatch in a deep learning (DL) model predicting the final infarction lesion from baseline magnetic resonance imaging (MRI). From 472 consecutive patients with acute ischemic stroke, we gathered baseline and follow-up MRI having intervals of 3–7 days, and initial and final infarction lesions were segmented. Four U-Net-based DL models from baseline MRI with different combinations of diffusion-weighted imaging (DWI), perfusion-weighted imaging (PWI) maps, and initial diffusion-restricted lesion prediction map (<inline-formula> <tex-math notation="LaTeX">$\text {Pred}_{\text {init}}$ </tex-math></inline-formula> map) were trained to predict the final infarction lesion. Five-fold cross-validation was used for training and testing. As an external test set, 55 patients from another institution were analyzed. Dice similarity coefficient (DSC) was compared between the models and subgroups according to the presence of lesion growth and/or diffusion-perfusion mismatch. The model using the PWI maps and <inline-formula> <tex-math notation="LaTeX">$\text {Pred}_{\text {init}}$ </tex-math></inline-formula> map showed the best mean DSC (0.422 and 0.486 for internal and external test set, respectively). This model showed better performance in predicting rapid lesion growth compared with the baseline model (mean DSC difference, 0.040; 95% confidence interval: 0.018–0.062). Using the PWI map with initial diffusion-restricted lesion prediction improved the performance of DL model in predicting the final infarction lesion from baseline MRI. |
first_indexed | 2024-04-11T19:58:13Z |
format | Article |
id | doaj.art-88a1bb4911a648c7b34b2754aff76845 |
institution | Directory Open Access Journal |
issn | 2169-3536 |
language | English |
last_indexed | 2024-04-11T19:58:13Z |
publishDate | 2022-01-01 |
publisher | IEEE |
record_format | Article |
series | IEEE Access |
spelling | doaj.art-88a1bb4911a648c7b34b2754aff768452022-12-22T04:05:53ZengIEEEIEEE Access2169-35362022-01-0110978799788710.1109/ACCESS.2022.32040489875295Impact of Diffusion–Perfusion Mismatch on Predicting Final Infarction Lesion Using Deep LearningSungdong Lee0Leonard Sunwoo1https://orcid.org/0000-0003-0374-8658Youngwon Choi2Jae Hyup Jung3Seung Chai Jung4Joong-Ho Won5Department of Statistics, Seoul National University, Seoul, Republic of KoreaDepartment of Radiology, Seoul National University Bundang Hospital, Seongnam-si, Republic of KoreaDepartment of Radiological Sciences, University of California at Los Angeles, Los Angeles, CA, USADepartment of Radiology, Seoul National University Bundang Hospital, Seongnam-si, Republic of KoreaDepartment of Radiology, Asan Medical Center, Seoul, Republic of KoreaDepartment of Statistics, Seoul National University, Seoul, Republic of KoreaWe report a study that validates the impact of diffusion-perfusion mismatch in a deep learning (DL) model predicting the final infarction lesion from baseline magnetic resonance imaging (MRI). From 472 consecutive patients with acute ischemic stroke, we gathered baseline and follow-up MRI having intervals of 3–7 days, and initial and final infarction lesions were segmented. Four U-Net-based DL models from baseline MRI with different combinations of diffusion-weighted imaging (DWI), perfusion-weighted imaging (PWI) maps, and initial diffusion-restricted lesion prediction map (<inline-formula> <tex-math notation="LaTeX">$\text {Pred}_{\text {init}}$ </tex-math></inline-formula> map) were trained to predict the final infarction lesion. Five-fold cross-validation was used for training and testing. As an external test set, 55 patients from another institution were analyzed. Dice similarity coefficient (DSC) was compared between the models and subgroups according to the presence of lesion growth and/or diffusion-perfusion mismatch. The model using the PWI maps and <inline-formula> <tex-math notation="LaTeX">$\text {Pred}_{\text {init}}$ </tex-math></inline-formula> map showed the best mean DSC (0.422 and 0.486 for internal and external test set, respectively). This model showed better performance in predicting rapid lesion growth compared with the baseline model (mean DSC difference, 0.040; 95% confidence interval: 0.018–0.062). Using the PWI map with initial diffusion-restricted lesion prediction improved the performance of DL model in predicting the final infarction lesion from baseline MRI.https://ieeexplore.ieee.org/document/9875295/Magnetic resonance imagingdisgnostic imagingdiffusion-perfusion mismatchfinal infarction lesionimage segmentationU-Net |
spellingShingle | Sungdong Lee Leonard Sunwoo Youngwon Choi Jae Hyup Jung Seung Chai Jung Joong-Ho Won Impact of Diffusion–Perfusion Mismatch on Predicting Final Infarction Lesion Using Deep Learning IEEE Access Magnetic resonance imaging disgnostic imaging diffusion-perfusion mismatch final infarction lesion image segmentation U-Net |
title | Impact of Diffusion–Perfusion Mismatch on Predicting Final Infarction Lesion Using Deep Learning |
title_full | Impact of Diffusion–Perfusion Mismatch on Predicting Final Infarction Lesion Using Deep Learning |
title_fullStr | Impact of Diffusion–Perfusion Mismatch on Predicting Final Infarction Lesion Using Deep Learning |
title_full_unstemmed | Impact of Diffusion–Perfusion Mismatch on Predicting Final Infarction Lesion Using Deep Learning |
title_short | Impact of Diffusion–Perfusion Mismatch on Predicting Final Infarction Lesion Using Deep Learning |
title_sort | impact of diffusion x2013 perfusion mismatch on predicting final infarction lesion using deep learning |
topic | Magnetic resonance imaging disgnostic imaging diffusion-perfusion mismatch final infarction lesion image segmentation U-Net |
url | https://ieeexplore.ieee.org/document/9875295/ |
work_keys_str_mv | AT sungdonglee impactofdiffusionx2013perfusionmismatchonpredictingfinalinfarctionlesionusingdeeplearning AT leonardsunwoo impactofdiffusionx2013perfusionmismatchonpredictingfinalinfarctionlesionusingdeeplearning AT youngwonchoi impactofdiffusionx2013perfusionmismatchonpredictingfinalinfarctionlesionusingdeeplearning AT jaehyupjung impactofdiffusionx2013perfusionmismatchonpredictingfinalinfarctionlesionusingdeeplearning AT seungchaijung impactofdiffusionx2013perfusionmismatchonpredictingfinalinfarctionlesionusingdeeplearning AT joonghowon impactofdiffusionx2013perfusionmismatchonpredictingfinalinfarctionlesionusingdeeplearning |