The impact of EPI-based distortion correction of dynamic susceptibility contrast MRI on cerebral blood volume estimation in patients with glioblastoma. Issue 132 (November 2020)

• Record Type: Journal Article
• Authors: Hovden, Ivar T.; Geier, Oliver M.; Digernes, Ingrid; Fuster-Garcia, Elies; Løvland, Grethe; Vik-Mo, Einar; Meling, Torstein R.; Emblem, Kyrre E.
• Is Part Of: European journal of radiology. Issue 132(2020)
• Publication Date: 2020-11
• Subjects: rCBV relative Cerebral Blood Volume -- rCBF relative Cerebral Blood Flow -- DSC Dynamic Susceptibility Contrast -- EPI Echo-Planar Imaging -- GE Gradient Echo -- SE Spin Echo -- FLAIR Fluid-Attenuated Inversion Recovery -- SENSE SENSitivity Encoding -- MNI Montreal Neurological Institute -- GT Ground Truth; Echo-planar imaging -- Distortion correction -- Perfusion -- Geometric distortions -- Brain tumor -- Brain regions; Medical radiology -- Periodicals; Radiology -- Periodicals; Radiologie médicale -- Périodiques; Medical radiology; Periodicals; 616.075705
• DOI: 10.1016/j.ejrad.2020.109278
• Languages: English
• ISSNs: 0720-048X

Abstract

Graphical abstract: The impact of two echo planar imaging (EPI) geometric distortion correction methods on relative cerebral blood volume (rCBV) estimation were investigated. Distortion correction lead to significant rCBV increase in cortical regions in frontal, temporal and occipital lobes as depicted. For glioblastoma subjects, precise spatial location of rCBV voxels as well as values in relation to non-EPI anatomical MRI is essential for the diagnosis and treatment follow-up, which as shown are dependent on tumor location. Highlights: Echo-planar images (EPIs) for functional-, diffusion- and perfusion MRI possess susceptibility induced geometric distortions. Estimated relative cerebral blood volume (rCBV) maps from DSC-MRI are thus sensitive to geometric distortions. Many cortical as well as some subcortical regions had increased rCBV following distortion correction on GE and SE DSC-MRI. rCBV from uncorrected EPIs may thus be underestimated in multiple brain regions, undermining treatment Abstract: Purpose: Relative cerebral blood volume (rCBV) from dynamic susceptibility contrast (DSC)-MRI is a valuable biomarker in patients with glioblastoma for assessing treatment response and predicting overall survival. DSC-MRI based on echo planar images (EPI) may possess severe geometric distortions from magnetic field inhomogeneities up to the order of centimeters. The aim of this study is to assess how much two readily available EPI-based geometric distortion correction methods, FSL TOPUP and EPIC, affect rCBV values from DSC-MRI in patients with confirmed glioblastoma. Method: We used a combined single-shot 2D gradient-echo (T2*), spin-echo (T2) EPI sequence to estimate both T2* and T2-weighted rCBV from the same contrast agent injection. Effects of distortion correction on the positive phase-encoded T2- and T2*-images were assessed in healthy anatomical brain regions in terms of Wilcoxon signed rank tests on median rCBV change and on Dice coefficients, as well as in tumor lesions in terms of Wilcoxon signed rank tests on median rCBV change. Results: Our results show that following distortion correction, both gradient-echo and spin-echo rCBV increased in cortical areas of the frontal, temporal and occipital lobe, including the posterior orbital gyri in the frontal lobe and middle frontal gyri (p < 0.0008). Similar, improved Dice coefficients were observed for gradient-echo EPI in temporal, occipital and frontal lobe. Only spin-echo rCBV in enhancing lesion increased with correction (p = 0.0002). Conclusion: Our study sheds light on the importance of performing geometric distortion correction on EPI-based MRI data before assessing functional information such as rCBV values. Our findings may indicate that uncorrected rCBV values can be underestimated from positive phase-encoding EPI and that geometric distortion correction is warranted when comparing EPI-based data to conventional MRI.