Deep learning-based detection and segmentation-assisted management of brain metastases.
MRI
brain metastases
deep learning
fully convolution network
stereotactic radiotherapy
Journal
Neuro-oncology
ISSN: 1523-5866
Titre abrégé: Neuro Oncol
Pays: England
ID NLM: 100887420
Informations de publication
Date de publication:
15 04 2020
15 04 2020
Historique:
pubmed:
24
12
2019
medline:
28
4
2021
entrez:
24
12
2019
Statut:
ppublish
Résumé
Three-dimensional T1 magnetization prepared rapid acquisition gradient echo (3D-T1-MPRAGE) is preferred in detecting brain metastases (BM) among MRI. We developed an automatic deep learning-based detection and segmentation method for BM (named BMDS net) on 3D-T1-MPRAGE images and evaluated its performance. The BMDS net is a cascaded 3D fully convolution network (FCN) to automatically detect and segment BM. In total, 1652 patients with 3D-T1-MPRAGE images from 3 hospitals (n = 1201, 231, and 220, respectively) were retrospectively included. Manual segmentations were obtained by a neuroradiologist and a radiation oncologist in a consensus reading in 3D-T1-MPRAGE images. Sensitivity, specificity, and dice ratio of the segmentation were evaluated. Specificity and sensitivity measure the fractions of relevant segmented voxels. Dice ratio was used to quantitatively measure the overlap between automatic and manual segmentation results. Paired samples t-tests and analysis of variance were employed for statistical analysis. The BMDS net can detect all BM, providing a detection result with an accuracy of 100%. Automatic segmentations correlated strongly with manual segmentations through 4-fold cross-validation of the dataset with 1201 patients: the sensitivity was 0.96 ± 0.03 (range, 0.84-0.99), the specificity was 0.99 ± 0.0002 (range, 0.99-1.00), and the dice ratio was 0.85 ± 0.08 (range, 0.62-0.95) for total tumor volume. Similar performances on the other 2 datasets also demonstrate the robustness of BMDS net in correctly detecting and segmenting BM in various settings. The BMDS net yields accurate detection and segmentation of BM automatically and could assist stereotactic radiotherapy management for diagnosis, therapy planning, and follow-up.
Sections du résumé
BACKGROUND
Three-dimensional T1 magnetization prepared rapid acquisition gradient echo (3D-T1-MPRAGE) is preferred in detecting brain metastases (BM) among MRI. We developed an automatic deep learning-based detection and segmentation method for BM (named BMDS net) on 3D-T1-MPRAGE images and evaluated its performance.
METHODS
The BMDS net is a cascaded 3D fully convolution network (FCN) to automatically detect and segment BM. In total, 1652 patients with 3D-T1-MPRAGE images from 3 hospitals (n = 1201, 231, and 220, respectively) were retrospectively included. Manual segmentations were obtained by a neuroradiologist and a radiation oncologist in a consensus reading in 3D-T1-MPRAGE images. Sensitivity, specificity, and dice ratio of the segmentation were evaluated. Specificity and sensitivity measure the fractions of relevant segmented voxels. Dice ratio was used to quantitatively measure the overlap between automatic and manual segmentation results. Paired samples t-tests and analysis of variance were employed for statistical analysis.
RESULTS
The BMDS net can detect all BM, providing a detection result with an accuracy of 100%. Automatic segmentations correlated strongly with manual segmentations through 4-fold cross-validation of the dataset with 1201 patients: the sensitivity was 0.96 ± 0.03 (range, 0.84-0.99), the specificity was 0.99 ± 0.0002 (range, 0.99-1.00), and the dice ratio was 0.85 ± 0.08 (range, 0.62-0.95) for total tumor volume. Similar performances on the other 2 datasets also demonstrate the robustness of BMDS net in correctly detecting and segmenting BM in various settings.
CONCLUSIONS
The BMDS net yields accurate detection and segmentation of BM automatically and could assist stereotactic radiotherapy management for diagnosis, therapy planning, and follow-up.
Identifiants
pubmed: 31867599
pii: 5684915
doi: 10.1093/neuonc/noz234
pmc: PMC7158643
doi:
Types de publication
Journal Article
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
505-514Informations de copyright
© The Author(s) 2019. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
Références
Hematol Oncol Clin North Am. 2012 Aug;26(4):933-47
pubmed: 22794291
J Magn Reson Imaging. 2010 Jan;31(1):85-93
pubmed: 20027576
Clin Oncol (R Coll Radiol). 2015 May;27(5):316-23
pubmed: 25662094
Front Oncol. 2017 Sep 25;7:221
pubmed: 28993798
IEEE Trans Cybern. 2019 Dec 18;:
pubmed: 31869812
IEEE Trans Med Imaging. 2016 May;35(5):1240-1251
pubmed: 26960222
Neuroimaging Clin N Am. 2016 Nov;26(4):601-620
pubmed: 27712796
Med Image Anal. 2017 Jan;35:18-31
pubmed: 27310171
Brain Tumor Pathol. 2016 Jan;33(1):1-12
pubmed: 26496727
Neuro Oncol. 2018 Mar 27;20(4):567-575
pubmed: 29016814
Clin Oncol (R Coll Radiol). 2020 Jan;32(1):13-25
pubmed: 31301960
Nat Rev Dis Primers. 2019 Jan 17;5(1):5
pubmed: 30655533
Radiology. 2019 Mar;290(3):607-618
pubmed: 30667332
IEEE Trans Med Imaging. 2018 Jul;37(7):1562-1573
pubmed: 29969407
IEEE Trans Med Imaging. 2019 Dec;38(12):2914-2925
pubmed: 31135354
JAMA. 2006 Jun 7;295(21):2483-91
pubmed: 16757720
Cancer Manag Res. 2018 Jul 05;10:1889-1899
pubmed: 30013395
IEEE Trans Med Imaging. 2006 Sep;25(9):1145-57
pubmed: 16967800
Magn Reson Imaging. 2012 Jul;30(6):824-36
pubmed: 22521993
IEEE Trans Cybern. 2019 Mar;49(3):1123-1136
pubmed: 29994385
Biochim Biophys Acta. 2006 Jul;1758(7):883-91
pubmed: 16787638
Annu Rev Biomed Eng. 2017 Jun 21;19:221-248
pubmed: 28301734
Lancet Oncol. 2015 Jun;16(6):e270-8
pubmed: 26065612
J Neuroimaging. 2016 Mar-Apr;26(2):240-6
pubmed: 26235208
J Magn Reson Imaging. 2016 Sep;44(3):642-52
pubmed: 26934581
Lung Cancer. 2019 Jan;127:12-18
pubmed: 30642539
IEEE Trans Pattern Anal Mach Intell. 2019 Jul;41(7):1559-1572
pubmed: 29993532