Robust deep learning-based segmentation of glioblastoma on routine clinical mri scans using sparsified training
Authors
Eijgelaar, R. S.Visser, M.
Müller, D. M. J.
Barkhof, F.
Vrenken, H.
van Herk, Marcel
Bello, L.
Conti Nibali, M.
Rossi, M.
Sciortino, T.
Berger, M. S.
Hervey-Jumper, S.
Kiesel, B.
Widhalm, G.
Furtner, J.
Robe, P.
Mandonnet, E.
De Witt Hamer, P. C.
de Munck, J. C.
Witte, M. G.
Affiliation
Department of Radiation Oncology, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, the NetherlandsIssue Date
2020
Metadata
Show full item recordAbstract
Purpose: To improve the robustness of deep learning-based glioblastoma segmentation in a clinical setting with sparsified datasets. Materials and methods: In this retrospective study, preoperative T1-weighted, T2-weighted, T2-weighted fluid-attenuated inversion recovery, and postcontrast T1-weighted MRI from 117 patients (median age, 64 years; interquartile range [IQR], 55-73 years; 76 men) included within the Multimodal Brain Tumor Image Segmentation (BraTS) dataset plus a clinical dataset (2012-2013) with similar imaging modalities of 634 patients (median age, 59 years; IQR, 49-69 years; 382 men) with glioblastoma from six hospitals were used. Expert tumor delineations on the postcontrast images were available, but for various clinical datasets, one or more sequences were missing. The convolutional neural network, DeepMedic, was trained on combinations of complete and incomplete data with and without site-specific data. Sparsified training was introduced, which randomly simulated missing sequences during training. The effects of sparsified training and center-specific training were tested using Wilcoxon signed rank tests for paired measurements. Results: A model trained exclusively on BraTS data reached a median Dice score of 0.81 for segmentation on BraTS test data but only 0.49 on the clinical data. Sparsified training improved performance (adjusted P < .05), even when excluding test data with missing sequences, to median Dice score of 0.67. Inclusion of site-specific data during sparsified training led to higher model performance Dice scores greater than 0.8, on par with a model based on all complete and incomplete data. For the model using BraTS and clinical training data, inclusion of site-specific data or sparsified training was of no consequence. Conclusion: Accurate and automatic segmentation of glioblastoma on clinical scans is feasible using a model based on large, heterogeneous, and partially incomplete datasets. Sparsified training may boost the performance of a smaller model based on public and site-specific data.Supplemental material is available for this article.Published under a CC BY 4.0 license.Citation
Eijgelaar RS, Visser M, Müller DMJ, Barkhof F, Vrenken H, van Herk M, et al. Robust Deep Learning–based Segmentation of Glioblastoma on Routine Clinical MRI Scans Using Sparsified Training. Radiology: Artificial Intelligence. 2020 Sep 1;2(5):e190103.Journal
Radiology: Artificial IntelligenceDOI
10.1148/ryai.2020190103PubMed ID
33937837Additional Links
https://dx.doi.org/10.1148/ryai.2020190103Type
ArticleLanguage
enae974a485f413a2113503eed53cd6c53
10.1148/ryai.2020190103
Scopus Count
Collections
Related articles
- Generative Adversarial Networks to Synthesize Missing T1 and FLAIR MRI Sequences for Use in a Multisequence Brain Tumor Segmentation Model.
- Authors: Conte GM, Weston AD, Vogelsang DC, Philbrick KA, Cai JC, Barbera M, Sanvito F, Lachance DH, Jenkins RB, Tobin WO, Eckel-Passow JE, Erickson BJ
- Issue date: 2021 May
- Radiologist-Level Performance by Using Deep Learning for Segmentation of Breast Cancers on MRI Scans.
- Authors: Hirsch L, Huang Y, Luo S, Rossi Saccarelli C, Lo Gullo R, Daimiel Naranjo I, Bitencourt AGV, Onishi N, Ko ES, Leithner D, Avendano D, Eskreis-Winkler S, Hughes M, Martinez DF, Pinker K, Juluru K, El-Rowmeim AE, Elnajjar P, Morris EA, Makse HA, Parra LC, Sutton EJ
- Issue date: 2022 Jan
- A Deep Learning Segmentation Pipeline for Cardiac T1 Mapping Using MRI Relaxation-based Synthetic Contrast Augmentation.
- Authors: Bhatt N, Ramanan V, Orbach A, Biswas L, Ng M, Guo F, Qi X, Guo L, Jimenez-Juan L, Roifman I, Wright GA, Ghugre NR
- Issue date: 2022 Nov
- Application of deep learning for automatic segmentation of brain tumors on magnetic resonance imaging: a heuristic approach in the clinical scenario.
- Authors: Di Ieva A, Russo C, Liu S, Jian A, Bai MY, Qian Y, Magnussen JS
- Issue date: 2021 Aug
- Clinical Evaluation of a Multiparametric Deep Learning Model for Glioblastoma Segmentation Using Heterogeneous Magnetic Resonance Imaging Data From Clinical Routine.
- Authors: Perkuhn M, Stavrinou P, Thiele F, Shakirin G, Mohan M, Garmpis D, Kabbasch C, Borggrefe J
- Issue date: 2018 Nov