Vessel segmentation using MP2RAGE images at 7T MRI

Poster No:

T571 

Submission Type:

Abstract Submission 

Authors:

Uk-Su Choi1,2, Hirokazu Kawaguchi3, Tobias KOBER4, Ikuhiro Kida1,2

Institutions:

1Center for Information and Neural Networks, NICT, Osaka, Japan, 2Graduate School of Frontier Biosciences, Osaka University, Osaka, Japan, 3Siemens Healthcare K.K., Osaka, Japan, 4Advanced Clinical Imaging Technology, Siemens Healthcare AG, Lausanne, Switzerland

Introduction:

Magnetization prepared two rapid acquisition gradient echoes (MP2RAGE) sequence can provide not only more homogenous T1-weighted (T1w) image but also multiple contrast images simultaneously. These features have the potential to segment different brain structures, such as the brain tissues (Choi et al., 2018). Among the brain structures, a blood vessel is also an important brain structure that provides vascular system information in the brain and addresses specific anatomical territories (Tatu et al., 1998). Blood vessels in the brain possess higher signal intensity than other brain tissues in T1w image, but it is hard to segment the blood vessels in T1w image due to the low contrast or inhomogeneous intensity. Previous studies have attempted to visualize or segment the vessels by optimizing sequence parameters (Zwanenburg et al., 2008) or using their own algorithms (Penumetcha et al., 2008; Morgan et al., 2018). In this study, we proposed a new approach to segment not only the brain tissues but also the vessels simultaneously with submillimeter spatial resolution using MP2RAGE sequence at 7T magnetic resonance imaging (MRI).

Methods:

Four healthy normal subjects without a history of neurological disease or any other medical conditions participated in this study after providing their written informed consent. The experiments were performed on a 7T MRI scanner (MAGNETOM 7T, Siemens Healthineers, Germany) with a 32-channel head coil (Nova Medical, USA). MP2RAGE sequences (Marques et al., 2010), which are supplied by the vendor as work-in-progress packages, were used with the following parameters: voxel resolution = 0.7 mm3, repetition time (TR) = 5000 ms, echo time (TE) = 3.43 ms, inversion time (800 ms/2600 ms), and flip angle = 4°/5°. To evaluate the performance of our segmentation, we also acquired 3D Time-of-Flight (TOF) image with the following parameters: voxel resolution = 0.3 x 0.3 x 0.4 mm3, repetition time (TR) = 20 ms, echo time (TE) = 4.47 ms, and flip angle = 18°. We applied our calculations and Frangi filtering (http://www.vmtk.org, 2008) on MP2RAGE images. In addition, we applied the same filter to TOF images and registered individual native T1w images. Then, we masked both the images with the same mask image for direct comparison.

Results:

We segmented the three different brain tissues, including gray matter, white matter, and cerebrospinal fluid, as well as the blood vessels using MP2RAGE images (Fig. 1). The mean absolute voxel difference and dice coefficient scores between the vessel masks of MP2RAGE segmentation and TOF segmentation were 23.3% and 66.2%, respectively. The masks of the large arteries, such as the anterior and middle cerebral arteries, were well-matched between MP2RAGE and TOF segmentations but the smaller blood vessels were not (Fig. 2).
The blood vessels in the peripheral regions of the brain showed poorer segmentation owing to the relatively lower contrasts resulting from the B1 bias field. However, this limitation could be overcome by using B1 bias field correction technique (Marques and Gruetter, 2013) or by changing the acquisition parameters (Tanner et al., 2012). In addition, non-defined voxels in the mask of blood vessels using our approach could also be corrected using image spatial filtering, such as growth algorithm.
Supporting Image: Figure1.png
   ·Figure 1. Brain tissue and blood vessel segmentation
Supporting Image: Figure2.png
   ·Figure 2. Blood vessel segmentation
 

Conclusions:

We successfully segmented the blood vessels as well as the brain tissues using MP2RAGE multi-contrast images and Frangi filtering. The vessel segmentation using MP2RAGE sequence at 7T has the potential 1) to be acquired alongside other brain tissue segmentation from the same MR sequence, 2) to be used for the correction of white matter segmentation, and 3) to provide precise anatomical territory information with submillimeter spatial resolution.

Imaging Methods:

Anatomical MRI 2
Imaging Methods Other

Modeling and Analysis Methods:

Segmentation and Parcellation 1

Neuroanatomy:

Neuroanatomy Other

Keywords:

MRI
Other - Vessel segmentation

1|2Indicates the priority used for review

My abstract is being submitted as a Software Demonstration.

No

Please indicate below if your study was a "resting state" or "task-activation” study.

Other

Healthy subjects only or patients (note that patient studies may also involve healthy subjects):

Healthy subjects

Was any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.

Yes

Was any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.

Not applicable

Please indicate which methods were used in your research:

Structural MRI

For human MRI, what field strength scanner do you use?

7T

Which processing packages did you use for your study?

FSL

Provide references using author date format

Choi, U.S. (2018), ‘Brain tissue segmentation based on MP2RAGE multi-contrast images in 7 T MRI’, bioRxiv: 455576.
Morgan, L. (2018), ‘Validation of a bright-vessel removal algorithm to improve FreeSurfer segmentations’, OHMB: #2605.
Marques, J.P. (2010), ‘MP2RAGE, a self bias-field corrected sequence for improved segmentation and T1-mapping at high field’, NeuroImage, 49, no. 2: 1271–81.
Marques, J.P and Gruetter, R. (2013), ‘New developments and applications of the MP2RAGE sequence - focusing the contrast and high spatial resolution R1 mapping’, PLOS ONE 8, no. 7: e69294
Penumetcha, N. (2008), ‘Segmentation of arteries in MPRAGE images of the ventral medial prefrontal cortex’, Computerized Medical Imaging and Graphics, 32, no. 1: 36–43.
Tanner, M. (2012), ‘Fluid and white matter suppression with the MP2RAGE sequence’, Journal of Magnetic Resonance Imaging, 35, no. 5: 1063–70.
Tatu, L. (1998), ‘Arterial territories of the human brain: cerebral hemispheres’, Neurology, 50, no. 6: 1699–1708.
Zwanenburg, Jaco J.M. (2008), ‘MR angiography of the cerebral perforating arteries with magnetization prepared anatomical reference at 7T: comparison with time-of-flight’, Journal of Magnetic Resonance Imaging, 28, no. 6: 1519–26.