GLISTRboost: Combining multimodal MRI segmentation, registration, and biophysical tumor growth modeling with gradient boosting machines for glioma segmentation

Spyridon Bakas, Ke Zeng, Aristeidis Sotiras, Saima Rathore, Hamed Akbari, Bilwaj Gaonkar, Martin Rozycki, Sarthak Pati, Christos Davatzikos

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

We present an approach for segmenting low- and high-grade gliomas in multimodal magnetic resonance imaging volumes. The proposed approach is based on a hybrid generative-discriminative model. Firstly, a generative approach based on an Expectation-Maximization framework that incorporates a glioma growth model is used to segment the brain scans into tumor, as well as healthy tissue labels. Secondly, a gradient boosting multi-class classification scheme is used to refine tumor labels based on information from multiple patients. Lastly, a probabilistic Bayesian strategy is employed to further refine and finalize the tumor segmentation based on patient-specific intensity statistics from the multiple modalities. We evaluated our approach in 186 cases during the training phase of the BRAin Tumor Segmentation (BRATS) 2015 challenge and report promising results. During the testing phase, the algorithm was additionally evaluated in 53 unseen cases, achieving the best performance among the competing methods.

Original languageEnglish (US)
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages144-155
Number of pages12
Volume9556
ISBN (Print)9783319308579
DOIs
StatePublished - 2016
Externally publishedYes
Event1st International Workshop on Brainlesion, Brainles 2015 Held in Conjunction with International Conference on Medical Image Computing for Computer-Assisted Intervention, MICCAI 2015 - Munich, Germany
Duration: Oct 5 2015Oct 5 2015

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9556
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other1st International Workshop on Brainlesion, Brainles 2015 Held in Conjunction with International Conference on Medical Image Computing for Computer-Assisted Intervention, MICCAI 2015
CountryGermany
CityMunich
Period10/5/1510/5/15

Fingerprint

Tumor Growth
Boosting
Magnetic resonance imaging
Registration
Tumors
Tumor
Segmentation
Gradient
Modeling
Labels
Multi-class Classification
Expectation Maximization
Magnetic Resonance Imaging
Magnetic resonance
Growth Model
Modality
Brain
Statistics
Tissue
Imaging techniques

Keywords

  • Brain tumor
  • Brain tumor growth model
  • BRATS challenge
  • Expectation maximization
  • Glioma
  • Gradient boosting
  • Multimodal MRI
  • Probabilistic model
  • Segmentation

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Bakas, S., Zeng, K., Sotiras, A., Rathore, S., Akbari, H., Gaonkar, B., ... Davatzikos, C. (2016). GLISTRboost: Combining multimodal MRI segmentation, registration, and biophysical tumor growth modeling with gradient boosting machines for glioma segmentation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9556, pp. 144-155). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9556). Springer Verlag. https://doi.org/10.1007/978-3-319-30858-6_13

GLISTRboost : Combining multimodal MRI segmentation, registration, and biophysical tumor growth modeling with gradient boosting machines for glioma segmentation. / Bakas, Spyridon; Zeng, Ke; Sotiras, Aristeidis; Rathore, Saima; Akbari, Hamed; Gaonkar, Bilwaj; Rozycki, Martin; Pati, Sarthak; Davatzikos, Christos.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9556 Springer Verlag, 2016. p. 144-155 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9556).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Bakas, S, Zeng, K, Sotiras, A, Rathore, S, Akbari, H, Gaonkar, B, Rozycki, M, Pati, S & Davatzikos, C 2016, GLISTRboost: Combining multimodal MRI segmentation, registration, and biophysical tumor growth modeling with gradient boosting machines for glioma segmentation. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 9556, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9556, Springer Verlag, pp. 144-155, 1st International Workshop on Brainlesion, Brainles 2015 Held in Conjunction with International Conference on Medical Image Computing for Computer-Assisted Intervention, MICCAI 2015, Munich, Germany, 10/5/15. https://doi.org/10.1007/978-3-319-30858-6_13
Bakas S, Zeng K, Sotiras A, Rathore S, Akbari H, Gaonkar B et al. GLISTRboost: Combining multimodal MRI segmentation, registration, and biophysical tumor growth modeling with gradient boosting machines for glioma segmentation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9556. Springer Verlag. 2016. p. 144-155. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-30858-6_13
Bakas, Spyridon ; Zeng, Ke ; Sotiras, Aristeidis ; Rathore, Saima ; Akbari, Hamed ; Gaonkar, Bilwaj ; Rozycki, Martin ; Pati, Sarthak ; Davatzikos, Christos. / GLISTRboost : Combining multimodal MRI segmentation, registration, and biophysical tumor growth modeling with gradient boosting machines for glioma segmentation. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9556 Springer Verlag, 2016. pp. 144-155 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{2fdeb22014394d19b74aadabb7e300de,
title = "GLISTRboost: Combining multimodal MRI segmentation, registration, and biophysical tumor growth modeling with gradient boosting machines for glioma segmentation",
abstract = "We present an approach for segmenting low- and high-grade gliomas in multimodal magnetic resonance imaging volumes. The proposed approach is based on a hybrid generative-discriminative model. Firstly, a generative approach based on an Expectation-Maximization framework that incorporates a glioma growth model is used to segment the brain scans into tumor, as well as healthy tissue labels. Secondly, a gradient boosting multi-class classification scheme is used to refine tumor labels based on information from multiple patients. Lastly, a probabilistic Bayesian strategy is employed to further refine and finalize the tumor segmentation based on patient-specific intensity statistics from the multiple modalities. We evaluated our approach in 186 cases during the training phase of the BRAin Tumor Segmentation (BRATS) 2015 challenge and report promising results. During the testing phase, the algorithm was additionally evaluated in 53 unseen cases, achieving the best performance among the competing methods.",
keywords = "Brain tumor, Brain tumor growth model, BRATS challenge, Expectation maximization, Glioma, Gradient boosting, Multimodal MRI, Probabilistic model, Segmentation",
author = "Spyridon Bakas and Ke Zeng and Aristeidis Sotiras and Saima Rathore and Hamed Akbari and Bilwaj Gaonkar and Martin Rozycki and Sarthak Pati and Christos Davatzikos",
year = "2016",
doi = "10.1007/978-3-319-30858-6_13",
language = "English (US)",
isbn = "9783319308579",
volume = "9556",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "144--155",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - GLISTRboost

T2 - Combining multimodal MRI segmentation, registration, and biophysical tumor growth modeling with gradient boosting machines for glioma segmentation

AU - Bakas, Spyridon

AU - Zeng, Ke

AU - Sotiras, Aristeidis

AU - Rathore, Saima

AU - Akbari, Hamed

AU - Gaonkar, Bilwaj

AU - Rozycki, Martin

AU - Pati, Sarthak

AU - Davatzikos, Christos

PY - 2016

Y1 - 2016

N2 - We present an approach for segmenting low- and high-grade gliomas in multimodal magnetic resonance imaging volumes. The proposed approach is based on a hybrid generative-discriminative model. Firstly, a generative approach based on an Expectation-Maximization framework that incorporates a glioma growth model is used to segment the brain scans into tumor, as well as healthy tissue labels. Secondly, a gradient boosting multi-class classification scheme is used to refine tumor labels based on information from multiple patients. Lastly, a probabilistic Bayesian strategy is employed to further refine and finalize the tumor segmentation based on patient-specific intensity statistics from the multiple modalities. We evaluated our approach in 186 cases during the training phase of the BRAin Tumor Segmentation (BRATS) 2015 challenge and report promising results. During the testing phase, the algorithm was additionally evaluated in 53 unseen cases, achieving the best performance among the competing methods.

AB - We present an approach for segmenting low- and high-grade gliomas in multimodal magnetic resonance imaging volumes. The proposed approach is based on a hybrid generative-discriminative model. Firstly, a generative approach based on an Expectation-Maximization framework that incorporates a glioma growth model is used to segment the brain scans into tumor, as well as healthy tissue labels. Secondly, a gradient boosting multi-class classification scheme is used to refine tumor labels based on information from multiple patients. Lastly, a probabilistic Bayesian strategy is employed to further refine and finalize the tumor segmentation based on patient-specific intensity statistics from the multiple modalities. We evaluated our approach in 186 cases during the training phase of the BRAin Tumor Segmentation (BRATS) 2015 challenge and report promising results. During the testing phase, the algorithm was additionally evaluated in 53 unseen cases, achieving the best performance among the competing methods.

KW - Brain tumor

KW - Brain tumor growth model

KW - BRATS challenge

KW - Expectation maximization

KW - Glioma

KW - Gradient boosting

KW - Multimodal MRI

KW - Probabilistic model

KW - Segmentation

UR - http://www.scopus.com/inward/record.url?scp=84961647384&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84961647384&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-30858-6_13

DO - 10.1007/978-3-319-30858-6_13

M3 - Conference contribution

AN - SCOPUS:84961647384

SN - 9783319308579

VL - 9556

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 144

EP - 155

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

PB - Springer Verlag

ER -