Multi-centre deep learning for placenta segmentation in obstetric ultrasound with multi-observer and cross-country generalization
Abstract The placenta is crucial to fetal well-being and it plays a significant role in the pathogenesis of hypertensive pregnancy disorders. Moreover, a timely diagnosis of placenta previa may save lives. Ultrasound is the primary imaging modality in pregnancy, but high-quality imaging depends on t...
Main Authors: | , , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Nature Portfolio
2023-02-01
|
Series: | Scientific Reports |
Online Access: | https://doi.org/10.1038/s41598-023-29105-x |
_version_ | 1811165900241371136 |
---|---|
author | Lisbeth Anita Andreasen Aasa Feragen Anders Nymark Christensen Jonathan Kistrup Thybo Morten Bo S. Svendsen Kilian Zepf Karim Lekadir Martin Grønnebæk Tolsgaard |
author_facet | Lisbeth Anita Andreasen Aasa Feragen Anders Nymark Christensen Jonathan Kistrup Thybo Morten Bo S. Svendsen Kilian Zepf Karim Lekadir Martin Grønnebæk Tolsgaard |
author_sort | Lisbeth Anita Andreasen |
collection | DOAJ |
description | Abstract The placenta is crucial to fetal well-being and it plays a significant role in the pathogenesis of hypertensive pregnancy disorders. Moreover, a timely diagnosis of placenta previa may save lives. Ultrasound is the primary imaging modality in pregnancy, but high-quality imaging depends on the access to equipment and staff, which is not possible in all settings. Convolutional neural networks may help standardize the acquisition of images for fetal diagnostics. Our aim was to develop a deep learning based model for classification and segmentation of the placenta in ultrasound images. We trained a model based on manual annotations of 7,500 ultrasound images to identify and segment the placenta. The model's performance was compared to annotations made by 25 clinicians (experts, trainees, midwives). The overall image classification accuracy was 81%. The average intersection over union score (IoU) reached 0.78. The model’s accuracy was lower than experts’ and trainees’, but it outperformed all clinicians at delineating the placenta, IoU = 0.75 vs 0.69, 0.66, 0.59. The model was cross validated on 100 2nd trimester images from Barcelona, yielding an accuracy of 76%, IoU 0.68. In conclusion, we developed a model for automatic classification and segmentation of the placenta with consistent performance across different patient populations. It may be used for automated detection of placenta previa and enable future deep learning research in placental dysfunction. |
first_indexed | 2024-04-10T15:43:46Z |
format | Article |
id | doaj.art-04b9fd21165b42d7a44bdd946462a050 |
institution | Directory Open Access Journal |
issn | 2045-2322 |
language | English |
last_indexed | 2024-04-10T15:43:46Z |
publishDate | 2023-02-01 |
publisher | Nature Portfolio |
record_format | Article |
series | Scientific Reports |
spelling | doaj.art-04b9fd21165b42d7a44bdd946462a0502023-02-12T12:12:08ZengNature PortfolioScientific Reports2045-23222023-02-011311810.1038/s41598-023-29105-xMulti-centre deep learning for placenta segmentation in obstetric ultrasound with multi-observer and cross-country generalizationLisbeth Anita Andreasen0Aasa Feragen1Anders Nymark Christensen2Jonathan Kistrup Thybo3Morten Bo S. Svendsen4Kilian Zepf5Karim Lekadir6Martin Grønnebæk Tolsgaard7Copenhagen Academy for Medical Education and Simulation (CAMES) RigshospitaletTechnical University of Denmark (DTU) ComputeTechnical University of Denmark (DTU) ComputeTechnical University of Denmark (DTU) ComputeCopenhagen Academy for Medical Education and Simulation (CAMES) RigshospitaletTechnical University of Denmark (DTU) ComputeArtificial Intelligence in Medicine Lab (BCN-AIM), Universitat de BarcelonaCopenhagen Academy for Medical Education and Simulation (CAMES) RigshospitaletAbstract The placenta is crucial to fetal well-being and it plays a significant role in the pathogenesis of hypertensive pregnancy disorders. Moreover, a timely diagnosis of placenta previa may save lives. Ultrasound is the primary imaging modality in pregnancy, but high-quality imaging depends on the access to equipment and staff, which is not possible in all settings. Convolutional neural networks may help standardize the acquisition of images for fetal diagnostics. Our aim was to develop a deep learning based model for classification and segmentation of the placenta in ultrasound images. We trained a model based on manual annotations of 7,500 ultrasound images to identify and segment the placenta. The model's performance was compared to annotations made by 25 clinicians (experts, trainees, midwives). The overall image classification accuracy was 81%. The average intersection over union score (IoU) reached 0.78. The model’s accuracy was lower than experts’ and trainees’, but it outperformed all clinicians at delineating the placenta, IoU = 0.75 vs 0.69, 0.66, 0.59. The model was cross validated on 100 2nd trimester images from Barcelona, yielding an accuracy of 76%, IoU 0.68. In conclusion, we developed a model for automatic classification and segmentation of the placenta with consistent performance across different patient populations. It may be used for automated detection of placenta previa and enable future deep learning research in placental dysfunction.https://doi.org/10.1038/s41598-023-29105-x |
spellingShingle | Lisbeth Anita Andreasen Aasa Feragen Anders Nymark Christensen Jonathan Kistrup Thybo Morten Bo S. Svendsen Kilian Zepf Karim Lekadir Martin Grønnebæk Tolsgaard Multi-centre deep learning for placenta segmentation in obstetric ultrasound with multi-observer and cross-country generalization Scientific Reports |
title | Multi-centre deep learning for placenta segmentation in obstetric ultrasound with multi-observer and cross-country generalization |
title_full | Multi-centre deep learning for placenta segmentation in obstetric ultrasound with multi-observer and cross-country generalization |
title_fullStr | Multi-centre deep learning for placenta segmentation in obstetric ultrasound with multi-observer and cross-country generalization |
title_full_unstemmed | Multi-centre deep learning for placenta segmentation in obstetric ultrasound with multi-observer and cross-country generalization |
title_short | Multi-centre deep learning for placenta segmentation in obstetric ultrasound with multi-observer and cross-country generalization |
title_sort | multi centre deep learning for placenta segmentation in obstetric ultrasound with multi observer and cross country generalization |
url | https://doi.org/10.1038/s41598-023-29105-x |
work_keys_str_mv | AT lisbethanitaandreasen multicentredeeplearningforplacentasegmentationinobstetricultrasoundwithmultiobserverandcrosscountrygeneralization AT aasaferagen multicentredeeplearningforplacentasegmentationinobstetricultrasoundwithmultiobserverandcrosscountrygeneralization AT andersnymarkchristensen multicentredeeplearningforplacentasegmentationinobstetricultrasoundwithmultiobserverandcrosscountrygeneralization AT jonathankistrupthybo multicentredeeplearningforplacentasegmentationinobstetricultrasoundwithmultiobserverandcrosscountrygeneralization AT mortenbossvendsen multicentredeeplearningforplacentasegmentationinobstetricultrasoundwithmultiobserverandcrosscountrygeneralization AT kilianzepf multicentredeeplearningforplacentasegmentationinobstetricultrasoundwithmultiobserverandcrosscountrygeneralization AT karimlekadir multicentredeeplearningforplacentasegmentationinobstetricultrasoundwithmultiobserverandcrosscountrygeneralization AT martingrønnebæktolsgaard multicentredeeplearningforplacentasegmentationinobstetricultrasoundwithmultiobserverandcrosscountrygeneralization |