MRI texture features as biomarkers to predict MGMT methylation status in glioblastomas

Research output: Contribution to journalArticle

50 Citations (Scopus)

Abstract

Purpose: Imaging biomarker research focuses on discovering relationships between radiological features and histological findings. In glioblastoma patients, methylation of the O6-methylguanine methyltransferase (MGMT) gene promoter is positively correlated with an increased effectiveness of current standard of care. In this paper, the authors investigate texture features as potential imaging biomarkers for capturing the MGMT methylation status of glioblastoma multiforme (GBM) tumors when combined with supervised classification schemes. Methods: A retrospective study of 155 GBM patients with known MGMT methylation status was conducted. Co-occurrence and run length texture features were calculated, and both support vector machines (SVMs) and random forest classifiers were used to predict MGMT methylation status. Results: The best classification system (an SVM-based classifier) had a maximum area under the receiver-operating characteristic (ROC) curve of 0.85 (95% CI: 0.780.91) using four texture features (correlation, energy, entropy, and local intensity) originating from the T2-weighted images, yielding at the optimal threshold of the ROC curve, a sensitivity of 0.803 and a specificity of 0.813. Conclusions: Results show that supervised machine learning of MRI texture features can predict MGMT methylation status in preoperative GBM tumors, thus providing a new noninvasive imaging biomarker.

Original languageEnglish (US)
Pages (from-to)2835-2844
Number of pages10
JournalMedical Physics
Volume43
Issue number6
DOIs
StatePublished - Jun 1 2016

Fingerprint

Methyltransferases
Glioblastoma
Methylation
Biomarkers
ROC Curve
Entropy
Standard of Care
Neoplasms
Retrospective Studies
Research
Genes
Support Vector Machine

Keywords

  • glioblastoma multiforme
  • imaging biomarkers
  • MGMT
  • MRI
  • random forest
  • support vector machines

ASJC Scopus subject areas

  • Biophysics
  • Radiology Nuclear Medicine and imaging

Cite this

@article{df49d28225b74e5da02ab46412212ae4,
title = "MRI texture features as biomarkers to predict MGMT methylation status in glioblastomas",
abstract = "Purpose: Imaging biomarker research focuses on discovering relationships between radiological features and histological findings. In glioblastoma patients, methylation of the O6-methylguanine methyltransferase (MGMT) gene promoter is positively correlated with an increased effectiveness of current standard of care. In this paper, the authors investigate texture features as potential imaging biomarkers for capturing the MGMT methylation status of glioblastoma multiforme (GBM) tumors when combined with supervised classification schemes. Methods: A retrospective study of 155 GBM patients with known MGMT methylation status was conducted. Co-occurrence and run length texture features were calculated, and both support vector machines (SVMs) and random forest classifiers were used to predict MGMT methylation status. Results: The best classification system (an SVM-based classifier) had a maximum area under the receiver-operating characteristic (ROC) curve of 0.85 (95{\%} CI: 0.780.91) using four texture features (correlation, energy, entropy, and local intensity) originating from the T2-weighted images, yielding at the optimal threshold of the ROC curve, a sensitivity of 0.803 and a specificity of 0.813. Conclusions: Results show that supervised machine learning of MRI texture features can predict MGMT methylation status in preoperative GBM tumors, thus providing a new noninvasive imaging biomarker.",
keywords = "glioblastoma multiforme, imaging biomarkers, MGMT, MRI, random forest, support vector machines",
author = "Panagiotis Korfiatis and Timothy Kline and Lucie Coufalova and Lachance, {Daniel H} and Parney, {Ian F} and Carter, {Rickey E.} and Buckner, {Jan Craig} and Erickson, {Bradley J}",
year = "2016",
month = "6",
day = "1",
doi = "10.1118/1.4948668",
language = "English (US)",
volume = "43",
pages = "2835--2844",
journal = "Medical Physics",
issn = "0094-2405",
publisher = "AAPM - American Association of Physicists in Medicine",
number = "6",

}

TY - JOUR

T1 - MRI texture features as biomarkers to predict MGMT methylation status in glioblastomas

AU - Korfiatis, Panagiotis

AU - Kline, Timothy

AU - Coufalova, Lucie

AU - Lachance, Daniel H

AU - Parney, Ian F

AU - Carter, Rickey E.

AU - Buckner, Jan Craig

AU - Erickson, Bradley J

PY - 2016/6/1

Y1 - 2016/6/1

N2 - Purpose: Imaging biomarker research focuses on discovering relationships between radiological features and histological findings. In glioblastoma patients, methylation of the O6-methylguanine methyltransferase (MGMT) gene promoter is positively correlated with an increased effectiveness of current standard of care. In this paper, the authors investigate texture features as potential imaging biomarkers for capturing the MGMT methylation status of glioblastoma multiforme (GBM) tumors when combined with supervised classification schemes. Methods: A retrospective study of 155 GBM patients with known MGMT methylation status was conducted. Co-occurrence and run length texture features were calculated, and both support vector machines (SVMs) and random forest classifiers were used to predict MGMT methylation status. Results: The best classification system (an SVM-based classifier) had a maximum area under the receiver-operating characteristic (ROC) curve of 0.85 (95% CI: 0.780.91) using four texture features (correlation, energy, entropy, and local intensity) originating from the T2-weighted images, yielding at the optimal threshold of the ROC curve, a sensitivity of 0.803 and a specificity of 0.813. Conclusions: Results show that supervised machine learning of MRI texture features can predict MGMT methylation status in preoperative GBM tumors, thus providing a new noninvasive imaging biomarker.

AB - Purpose: Imaging biomarker research focuses on discovering relationships between radiological features and histological findings. In glioblastoma patients, methylation of the O6-methylguanine methyltransferase (MGMT) gene promoter is positively correlated with an increased effectiveness of current standard of care. In this paper, the authors investigate texture features as potential imaging biomarkers for capturing the MGMT methylation status of glioblastoma multiforme (GBM) tumors when combined with supervised classification schemes. Methods: A retrospective study of 155 GBM patients with known MGMT methylation status was conducted. Co-occurrence and run length texture features were calculated, and both support vector machines (SVMs) and random forest classifiers were used to predict MGMT methylation status. Results: The best classification system (an SVM-based classifier) had a maximum area under the receiver-operating characteristic (ROC) curve of 0.85 (95% CI: 0.780.91) using four texture features (correlation, energy, entropy, and local intensity) originating from the T2-weighted images, yielding at the optimal threshold of the ROC curve, a sensitivity of 0.803 and a specificity of 0.813. Conclusions: Results show that supervised machine learning of MRI texture features can predict MGMT methylation status in preoperative GBM tumors, thus providing a new noninvasive imaging biomarker.

KW - glioblastoma multiforme

KW - imaging biomarkers

KW - MGMT

KW - MRI

KW - random forest

KW - support vector machines

UR - http://www.scopus.com/inward/record.url?scp=84969855998&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84969855998&partnerID=8YFLogxK

U2 - 10.1118/1.4948668

DO - 10.1118/1.4948668

M3 - Article

VL - 43

SP - 2835

EP - 2844

JO - Medical Physics

JF - Medical Physics

SN - 0094-2405

IS - 6

ER -