TY - JOUR
T1 - Automatic segmentation of cerebral infarcts in follow-up computed tomography images with convolutional neural networks
AU - Sales Barros, Renan
AU - Tolhuisen, Manon L.
AU - Boers, Anna M. M.
AU - Jansen, Ivo
AU - Ponomareva, Elena
AU - Dippel, Diederik W. J.
AU - van der Lugt, Aad
AU - van Oostenbrugge, Robert J.
AU - van Zwam, Wim H.
AU - Berkhemer, Olvert A.
AU - Goyal, Mayank
AU - Demchuk, Andrew M.
AU - Menon, Bijoy K.
AU - Mitchell, Peter
AU - Hill, Michael D.
AU - Jovin, Tudor G.
AU - Davalos, Antoni
AU - Campbell, Bruce C. V.
AU - Saver, Jeffrey L.
AU - Roos, Yvo B. W. E. M.
AU - Muir, Keith W.
AU - White, Phil
AU - Bracard, Serge
AU - Guillemin, Francis
AU - Olabarriaga, Silvia Delgado
AU - Majoie, Charles B. L. M.
AU - Marquering, Henk A.
PY - 2020/9/1
Y1 - 2020/9/1
N2 - Background and purpose: Infarct volume is a valuable outcome measure in treatment trials of acute ischemic stroke and is strongly associated with functional outcome. Its manual volumetric assessment is, however, too demanding to be implemented in clinical practice. Objective: To assess the value of convolutional neural networks (CNNs) in the automatic segmentation of infarct volume in follow-up CT images in a large population of patients with acute ischemic stroke. Materials and methods: We included CT images of 1026 patients from a large pooling of patients with acute ischemic stroke. A reference standard for the infarct segmentation was generated by manual delineation. We introduce three CNN models for the segmentation of subtle, intermediate, and severe hypodense lesions. The fully automated infarct segmentation was defined as the combination of the results of these three CNNs. The results of the three-CNNs approach were compared with the results from a single CNN approach and with the reference standard segmentations. Results: The median infarct volume was 48 mL (IQR 15-125 mL). Comparison between the volumes of the three-CNNs approach and manually delineated infarct volumes showed excellent agreement, with an intraclass correlation coefficient (ICC) of 0.88. Even better agreement was found for severe and intermediate hypodense infarcts, with ICCs of 0.98 and 0.93, respectively. Although the number of patients used for training in the single CNN approach was much larger, the accuracy of the three-CNNs approach strongly outperformed the single CNN approach, which had an ICC of 0.34. Conclusion: Convolutional neural networks are valuable and accurate in the quantitative assessment of infarct volumes, for both subtle and severe hypodense infarcts in follow-up CT images. Our proposed three-CNNs approach strongly outperforms a more straightforward single CNN approach.
AB - Background and purpose: Infarct volume is a valuable outcome measure in treatment trials of acute ischemic stroke and is strongly associated with functional outcome. Its manual volumetric assessment is, however, too demanding to be implemented in clinical practice. Objective: To assess the value of convolutional neural networks (CNNs) in the automatic segmentation of infarct volume in follow-up CT images in a large population of patients with acute ischemic stroke. Materials and methods: We included CT images of 1026 patients from a large pooling of patients with acute ischemic stroke. A reference standard for the infarct segmentation was generated by manual delineation. We introduce three CNN models for the segmentation of subtle, intermediate, and severe hypodense lesions. The fully automated infarct segmentation was defined as the combination of the results of these three CNNs. The results of the three-CNNs approach were compared with the results from a single CNN approach and with the reference standard segmentations. Results: The median infarct volume was 48 mL (IQR 15-125 mL). Comparison between the volumes of the three-CNNs approach and manually delineated infarct volumes showed excellent agreement, with an intraclass correlation coefficient (ICC) of 0.88. Even better agreement was found for severe and intermediate hypodense infarcts, with ICCs of 0.98 and 0.93, respectively. Although the number of patients used for training in the single CNN approach was much larger, the accuracy of the three-CNNs approach strongly outperformed the single CNN approach, which had an ICC of 0.34. Conclusion: Convolutional neural networks are valuable and accurate in the quantitative assessment of infarct volumes, for both subtle and severe hypodense infarcts in follow-up CT images. Our proposed three-CNNs approach strongly outperforms a more straightforward single CNN approach.
KW - CT
KW - stroke
KW - technique
KW - thrombectomy
UR - http://www.scopus.com/inward/record.url?scp=85077045654&partnerID=8YFLogxK
U2 - https://doi.org/10.1136/neurintsurg-2019-015471
DO - https://doi.org/10.1136/neurintsurg-2019-015471
M3 - Article
C2 - 31871069
SN - 1759-8478
VL - 12
SP - 848
EP - 852
JO - Journal of NeuroInterventional Surgery
JF - Journal of NeuroInterventional Surgery
IS - 9
M1 - 015471
ER -