Research article

Parallel imaging: is GRAPPA a useful acquisition tool for MR imaging intended for volumetric brain analysis?

Terri L Lindholm^1,2 , Lisa Botes^1,2 , Eva-Lena Engman² , Anders Frank^1,2 , Tomas Jonsson^1,2 , Leif Svensson^1,2 and Per Julin^2,3,4

¹Department of Diagnostic Medical Physics, Karolinska University Hospital Huddinge, Stockholm, Sweden

²SMILE Image Laboratory, Karolinska Institute, Stockholm, Sweden

³Astra Zeneca Research and Development, Södertälje, Sweden

⁴Department of Rehabilitation Medicine, Danderyd Hospital, Karolinska Institute, Stockholm, Sweden

author email corresponding author email

BMC Medical Imaging 2009, 9:15doi:10.1186/1471-2342-9-15

Published:	3 August 2009

Abstract

Background

The work presented here investigates parallel imaging applied to T1-weighted high resolution imaging for use in longitudinal volumetric clinical studies involving Alzheimer's disease (AD) and Mild Cognitive Impairment (MCI) patients. This was in an effort to shorten acquisition times to minimise the risk of motion artefacts caused by patient discomfort and disorientation. The principle question is, "Can parallel imaging be used to acquire images at 1.5 T of sufficient quality to allow volumetric analysis of patient brains?"

Methods

Optimisation studies were performed on a young healthy volunteer and the selected protocol (including the use of two different parallel imaging acceleration factors) was then tested on a cohort of 15 elderly volunteers including MCI and AD patients. In addition to automatic brain segmentation, hippocampus volumes were manually outlined and measured in all patients. The 15 patients were scanned on a second occasion approximately one week later using the same protocol and evaluated in the same manner to test repeatability of measurement using images acquired with the GRAPPA parallel imaging technique applied to the MPRAGE sequence.

Results

Intraclass correlation tests show that almost perfect agreement between repeated measurements of both segmented brain parenchyma fraction and regional measurement of hippocampi. The protocol is suitable for both global and regional volumetric measurement dementia patients.

Conclusion

In summary, these results indicate that parallel imaging can be used without detrimental effect to brain tissue segmentation and volumetric measurement and should be considered for both clinical and research studies where longitudinal measurements of brain tissue volumes are of interest.