Differentiation of solitary brain metastasis from glioblastoma multiforme: a predictive multiparametric approach using combined MR diffusion and perfusion

Adam Herman Bauer, William Erly, Franklin G. Moser, Marcel Maya, Kambiz Nael

Research output: Contribution to journalArticle

31 Citations (Scopus)

Abstract

Introduction: Solitary brain metastasis (MET) and glioblastoma multiforme (GBM) can appear similar on conventional MRI. The purpose of this study was to identify magnetic resonance (MR) perfusion and diffusion-weighted biomarkers that can differentiate MET from GBM. Methods: In this retrospective study, patients were included if they met the following criteria: underwent resection of a solitary enhancing brain tumor and had preoperative 3.0 T MRI encompassing diffusion tensor imaging (DTI), dynamic contrast-enhanced (DCE), and dynamic susceptibility contrast (DSC) perfusion. Using co-registered images, voxel-based fractional anisotropy (FA), mean diffusivity (MD), K<sup>trans</sup>, and relative cerebral blood volume (rCBV) values were obtained in the enhancing tumor and non-enhancing peritumoral T2 hyperintense region (NET2). Data were analyzed by logistic regression and analysis of variance. Receiver operating characteristic (ROC) analysis was performed to determine the optimal parameter/s and threshold for predicting of GBM vs. MET. Results: Twenty-three patients (14 M, age 32–78 years old) met our inclusion criteria. Pathology revealed 13 GBMs and 10 METs. In the enhancing tumor, rCBV, K<sup>trans</sup>, and FA were higher in GBM, whereas MD was lower, neither without statistical significance. In the NET2, rCBV was significantly higher (p = 0.05) in GBM, but MD was significantly lower (p < 0.01) in GBM. FA and K<sup>trans</sup> were higher in GBM, though not reaching significance. The best discriminative power was obtained in NET2 from a combination of rCBV, FA, and MD, resulting in an area under the curve (AUC) of 0.98. Conclusion: The combination of MR diffusion and perfusion matrices in NET2 can help differentiate GBM over solitary MET with diagnostic accuracy of 98 %.

Original languageEnglish (US)
Pages (from-to)697-703
Number of pages7
JournalNeuroradiology
Volume57
Issue number7
DOIs
StatePublished - Apr 7 2015

Fingerprint

Glioblastoma
Magnetic Resonance Spectroscopy
Perfusion
Neoplasm Metastasis
Brain
Anisotropy
Diffusion Tensor Imaging
ROC Curve
Brain Neoplasms
Area Under Curve
Neoplasms
Analysis of Variance
Retrospective Studies
Biomarkers
Logistic Models
Regression Analysis
Pathology
Cerebral Blood Volume

Keywords

  • Diffusion tensor imaging
  • Dynamic contrast enhancement
  • Dynamic susceptibility contrast
  • Glioblastoma multiforme
  • Intracranial metastasis

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging
  • Clinical Neurology
  • Cardiology and Cardiovascular Medicine

Cite this

Differentiation of solitary brain metastasis from glioblastoma multiforme : a predictive multiparametric approach using combined MR diffusion and perfusion. / Bauer, Adam Herman; Erly, William; Moser, Franklin G.; Maya, Marcel; Nael, Kambiz.

In: Neuroradiology, Vol. 57, No. 7, 07.04.2015, p. 697-703.

Research output: Contribution to journalArticle

@article{c77a4e2d4a934beeab5adb466d575d2c,
title = "Differentiation of solitary brain metastasis from glioblastoma multiforme: a predictive multiparametric approach using combined MR diffusion and perfusion",
abstract = "Introduction: Solitary brain metastasis (MET) and glioblastoma multiforme (GBM) can appear similar on conventional MRI. The purpose of this study was to identify magnetic resonance (MR) perfusion and diffusion-weighted biomarkers that can differentiate MET from GBM. Methods: In this retrospective study, patients were included if they met the following criteria: underwent resection of a solitary enhancing brain tumor and had preoperative 3.0 T MRI encompassing diffusion tensor imaging (DTI), dynamic contrast-enhanced (DCE), and dynamic susceptibility contrast (DSC) perfusion. Using co-registered images, voxel-based fractional anisotropy (FA), mean diffusivity (MD), Ktrans, and relative cerebral blood volume (rCBV) values were obtained in the enhancing tumor and non-enhancing peritumoral T2 hyperintense region (NET2). Data were analyzed by logistic regression and analysis of variance. Receiver operating characteristic (ROC) analysis was performed to determine the optimal parameter/s and threshold for predicting of GBM vs. MET. Results: Twenty-three patients (14 M, age 32–78 years old) met our inclusion criteria. Pathology revealed 13 GBMs and 10 METs. In the enhancing tumor, rCBV, Ktrans, and FA were higher in GBM, whereas MD was lower, neither without statistical significance. In the NET2, rCBV was significantly higher (p = 0.05) in GBM, but MD was significantly lower (p < 0.01) in GBM. FA and Ktrans were higher in GBM, though not reaching significance. The best discriminative power was obtained in NET2 from a combination of rCBV, FA, and MD, resulting in an area under the curve (AUC) of 0.98. Conclusion: The combination of MR diffusion and perfusion matrices in NET2 can help differentiate GBM over solitary MET with diagnostic accuracy of 98 {\%}.",
keywords = "Diffusion tensor imaging, Dynamic contrast enhancement, Dynamic susceptibility contrast, Glioblastoma multiforme, Intracranial metastasis",
author = "Bauer, {Adam Herman} and William Erly and Moser, {Franklin G.} and Marcel Maya and Kambiz Nael",
year = "2015",
month = "4",
day = "7",
doi = "10.1007/s00234-015-1524-6",
language = "English (US)",
volume = "57",
pages = "697--703",
journal = "Neuroradiology",
issn = "0028-3940",
publisher = "Springer Verlag",
number = "7",

}

TY - JOUR

T1 - Differentiation of solitary brain metastasis from glioblastoma multiforme

T2 - a predictive multiparametric approach using combined MR diffusion and perfusion

AU - Bauer, Adam Herman

AU - Erly, William

AU - Moser, Franklin G.

AU - Maya, Marcel

AU - Nael, Kambiz

PY - 2015/4/7

Y1 - 2015/4/7

N2 - Introduction: Solitary brain metastasis (MET) and glioblastoma multiforme (GBM) can appear similar on conventional MRI. The purpose of this study was to identify magnetic resonance (MR) perfusion and diffusion-weighted biomarkers that can differentiate MET from GBM. Methods: In this retrospective study, patients were included if they met the following criteria: underwent resection of a solitary enhancing brain tumor and had preoperative 3.0 T MRI encompassing diffusion tensor imaging (DTI), dynamic contrast-enhanced (DCE), and dynamic susceptibility contrast (DSC) perfusion. Using co-registered images, voxel-based fractional anisotropy (FA), mean diffusivity (MD), Ktrans, and relative cerebral blood volume (rCBV) values were obtained in the enhancing tumor and non-enhancing peritumoral T2 hyperintense region (NET2). Data were analyzed by logistic regression and analysis of variance. Receiver operating characteristic (ROC) analysis was performed to determine the optimal parameter/s and threshold for predicting of GBM vs. MET. Results: Twenty-three patients (14 M, age 32–78 years old) met our inclusion criteria. Pathology revealed 13 GBMs and 10 METs. In the enhancing tumor, rCBV, Ktrans, and FA were higher in GBM, whereas MD was lower, neither without statistical significance. In the NET2, rCBV was significantly higher (p = 0.05) in GBM, but MD was significantly lower (p < 0.01) in GBM. FA and Ktrans were higher in GBM, though not reaching significance. The best discriminative power was obtained in NET2 from a combination of rCBV, FA, and MD, resulting in an area under the curve (AUC) of 0.98. Conclusion: The combination of MR diffusion and perfusion matrices in NET2 can help differentiate GBM over solitary MET with diagnostic accuracy of 98 %.

AB - Introduction: Solitary brain metastasis (MET) and glioblastoma multiforme (GBM) can appear similar on conventional MRI. The purpose of this study was to identify magnetic resonance (MR) perfusion and diffusion-weighted biomarkers that can differentiate MET from GBM. Methods: In this retrospective study, patients were included if they met the following criteria: underwent resection of a solitary enhancing brain tumor and had preoperative 3.0 T MRI encompassing diffusion tensor imaging (DTI), dynamic contrast-enhanced (DCE), and dynamic susceptibility contrast (DSC) perfusion. Using co-registered images, voxel-based fractional anisotropy (FA), mean diffusivity (MD), Ktrans, and relative cerebral blood volume (rCBV) values were obtained in the enhancing tumor and non-enhancing peritumoral T2 hyperintense region (NET2). Data were analyzed by logistic regression and analysis of variance. Receiver operating characteristic (ROC) analysis was performed to determine the optimal parameter/s and threshold for predicting of GBM vs. MET. Results: Twenty-three patients (14 M, age 32–78 years old) met our inclusion criteria. Pathology revealed 13 GBMs and 10 METs. In the enhancing tumor, rCBV, Ktrans, and FA were higher in GBM, whereas MD was lower, neither without statistical significance. In the NET2, rCBV was significantly higher (p = 0.05) in GBM, but MD was significantly lower (p < 0.01) in GBM. FA and Ktrans were higher in GBM, though not reaching significance. The best discriminative power was obtained in NET2 from a combination of rCBV, FA, and MD, resulting in an area under the curve (AUC) of 0.98. Conclusion: The combination of MR diffusion and perfusion matrices in NET2 can help differentiate GBM over solitary MET with diagnostic accuracy of 98 %.

KW - Diffusion tensor imaging

KW - Dynamic contrast enhancement

KW - Dynamic susceptibility contrast

KW - Glioblastoma multiforme

KW - Intracranial metastasis

UR - http://www.scopus.com/inward/record.url?scp=84937518680&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84937518680&partnerID=8YFLogxK

U2 - 10.1007/s00234-015-1524-6

DO - 10.1007/s00234-015-1524-6

M3 - Article

C2 - 25845813

AN - SCOPUS:84937518680

VL - 57

SP - 697

EP - 703

JO - Neuroradiology

JF - Neuroradiology

SN - 0028-3940

IS - 7

ER -