Characterizing Peritumoral Tissue Using DTI-Based Free Water Elimination

AbstractTumor types are classically distinguished based on biopsies of the tumor itself, as well as a radiological interpretation using diverse MRI modalities. In the current study, the overarching goal is to demonstrate that primary (glioblastomas) and secondary (brain metastases) malignancies can be differentiated based on the microstructure of the peritumoral region. This is achieved by exploiting the extracellular water differences between vasogenic edema and infiltrative tissue and training a convolutional neural network (CNN) on the Diffusion Tensor Imaging (DTI)-derived free water volume fraction. We obtained 85% accuracy in discriminating extracellular water differences between local patches in the peritumoral area of 66 glioblastomas and 40 metastatic patients in a cross-validation setting. On an independent test cohort consisting of 20 glioblastomas and 10 metastases, we got 93% accuracy in discriminating metastases from glioblastomas using majority voting on patches. This level of accuracy surpasses CNNs trained on other conventional DTI-based measures such as fractional anisotropy (FA) and mean diffusivity (MD), that have been used in other studies. Additionally, the CNN captures the peritumoral heterogeneity better than conventional texture features, including Gabor and radiomic features. Our results demonstrate that the extracellular water content of the peritumoral tissue, as captured by the free water volume fraction, is best able to characterize the differences between infiltrative and vasogenic peritumoral regions, paving the way for its use in classifying and benchmarking peritumoral tissue with varying degrees of infiltration.

Download Full-text

Opposing white matter microstructure abnormalities in 22q11.2 deletion and duplication carriers

Translational Psychiatry ◽

10.1038/s41398-021-01703-1 ◽

2021 ◽

Vol 11 (1) ◽

Author(s):

Johanna Seitz-Holland ◽

Monica Lyons ◽

Leila Kushan ◽

Amy Lin ◽

Julio E. Villalon-Reina ◽

...

Keyword(s):

White Matter ◽

Psychiatric Morbidity ◽

Free Water ◽

Water Model ◽

22Q11.2 Deletion ◽

Tissue Properties ◽

Diffusion Weighted Images ◽

Diffusion Weighted ◽

Water Elimination ◽

Number Variation

AbstractDeletions and duplications at the 22q11.2 locus are associated with significant neurodevelopmental and psychiatric morbidity. Previous diffusion-weighted magnetic resonance imaging (MRI) studies in 22q11.2 deletion carriers (22q-del) found nonspecific white matter (WM) abnormalities, characterized by higher fractional anisotropy. Here, utilizing novel imaging and processing methods that allow separation of signal contribution from different tissue properties, we investigate whether higher anisotropy is driven by (1) extracellular changes, (2) selective degeneration of secondary fibers, or (3) volumetric differences. We further, for the first time, investigate WM microstructure in 22q11.2 duplication carriers (22q-dup). Multi-shell diffusion-weighted images were acquired from 26 22q-del, 19 22q-dup, and 18 healthy individuals (HC). Images were fitted with the free-water model to estimate anisotropy following extracellular free-water elimination and with the novel BedpostX model to estimate fractional volumes of primary and secondary fiber populations. Outcome measures were compared between groups, with and without correction for WM and cerebrospinal fluid (CSF) volumes. In 22q-del, anisotropy following free-water elimination remained significantly higher compared with controls. BedpostX did not identify selective secondary fiber degeneration. Higher anisotropy diminished when correcting for the higher CSF and lower WM volumes. In contrast, 22q-dup had lower anisotropy and greater extracellular space than HC, not influenced by macrostructural volumes. Our findings demonstrate opposing effects of reciprocal 22q11.2 copy-number variation on WM, which may arise from distinct pathologies. In 22q-del, microstructural abnormalities may be secondary to enlarged CSF space and more densely packed WM. In 22q-dup, we see evidence for demyelination similar to what is commonly observed in neuropsychiatric disorders.

Download Full-text

Free water elimination and mapping from diffusion MRI

Magnetic Resonance in Medicine ◽

10.1002/mrm.22055 ◽

2009 ◽

Vol 62 (3) ◽

pp. 717-730 ◽

Cited By ~ 389

Author(s):

Ofer Pasternak ◽

Nir Sochen ◽

Yaniv Gur ◽

Nathan Intrator ◽

Yaniv Assaf

Keyword(s):

Diffusion Mri ◽

Free Water ◽

Water Elimination

Download Full-text

Deep Learning with Synthetic Diffusion MRI Data for Free-Water Elimination in Glioblastoma Cases

Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - Lecture Notes in Computer Science ◽

10.1007/978-3-030-00931-1_12 ◽

2018 ◽

pp. 98-106 ◽

Cited By ~ 2

Author(s):

Miguel Molina-Romero ◽

Benedikt Wiestler ◽

Pedro A. Gómez ◽

Marion I. Menzel ◽

Bjoern H. Menze

Keyword(s):

Deep Learning ◽

Diffusion Mri ◽

Free Water ◽

Water Elimination

Download Full-text

Distinct Tumor Signatures using Deep Learning-based Characterization of the Peritumoral Microenvironment in Glioblastomas and Brain Metastases

10.21203/rs.3.rs-374535/v1 ◽

2021 ◽

Author(s):

Zahra Riahi Samani ◽

Drew Parker ◽

Ronald Wolf ◽

Wes Hodges ◽

Steven Brem ◽

...

Keyword(s):

Brain Metastases ◽

Volume Fraction ◽

Diffusion Tensor ◽

Vasogenic Edema ◽

Mean Diffusivity ◽

Free Water ◽

Water Volume ◽

Extracellular Water ◽

Peritumoral Tissue ◽

Water Volume Fraction

Abstract Tumor types are classically distinguished based on biopsies of the tumor itself, as well as a radiological interpretation using diverse MRI modalities. In the current study, the overarching goal is to demonstrate that primary (glioblastomas) and secondary (brain metastases) malignancies can be differentiated based on the microstructure of the peritumoral region. This is achieved by exploiting the extracellular water differences between vasogenic edema and infiltrative tissue and training a convolutional neural network (CNN) on the Diffusion Tensor Imaging (DTI)-derived free water volume fraction map of the peritumoral regions of 54 metastases and 89 glioblastoma patients. We obtained 93% accuracy in discriminating metastases from glioblastomas. This level of accuracy surpasses CNNs trained on other conventional DTI-based measures such as fractional anisotropy (FA) and mean diffusivity (MD), that have been used in other studies. Additionally, the CNN captures the peritumoral heterogeneity better than conventional texture features, including Gabor and radiomic features. Our results demonstrate that the extracellular water content of the peritumoral tissue, as captured by the free water volume fraction, is best able to characterize the differences between infiltrative and vasogenic peritumoral regions, paving the way for its use in classifying and benchmarking peritumoral tissue with varying degrees of infiltration.

Download Full-text