T1-weighted brain imaging with a 32-channel coil at 3T using TurboFLASH BLADE compared with standard cartesian k-space sampling.
Invest Radiol. 2009 Mar; 44(3):177-83.IR

Abstract

PURPOSE

Motion artifacts often markedly degrade image quality in clinical scans. The BLADE technique offers an alternative k-space sampling scheme reducing the effect of patient related motion on image quality. The purpose of this study is the comparison of imaging artifacts, signal-to-noise (SNR), and contrast-to-noise ratio (CNR) of a new turboFLASH BLADE k-space trajectory with the standard Cartesian k-space sampling for brain imaging, using a 32-channel coil at 3T.

METHODS

The results from 32 patients included after informed consent are reported. This study was performed with a 32-channel head coil on a 3T scanner. Sagittal and axial T1-weighted FLASH sequences (TR/TE 250/2.46 milliseconds, flip angle 70-degree), acquired with Cartesian k-space sampling and T1-weighted turboFLASH sequences (TR/TE/TIsag/TIax 3200/2.77/1144/1056 milliseconds, flip angle 20-degree), using PROPELLER (BLADE) k-space trajectory, were compared. SNR and CNR were evaluated using a paired student t test. The frequency of motion artifacts was assessed in a blinded read. To analyze the differences between both techniques a McNemar test was performed. A P value <0.05 was considered statistically significant.

RESULTS

From the blinded read, the overall preference in terms of diagnostic image quality was statistically significant in favor of the BLADE turboFLASH data sets, compared with standard FLASH for both sagittal (P < 0.0001) and axial (P < 0.0001) planes. The frequency of motion artifacts from the scalp was higher for standard FLASH sequences than for BLADE sequences on both axial (47%, P < 0.0003) and sagittal (69%, P < 0.0001) planes. BLADE was preferred in 100% (sagittal plane) and 80% (axial plane) of in-patient data sets and in 68% (sagittal plane) and 73% (axial plane) of out-patient data sets.The BLADE T1 scan did have lower SNRmean (BLADEax 179 +/- 98, Cartesianax 475 +/- 145, BLADEsag 171 +/- 51, and Cartesiansag 697 +/- 129) with P values indicating accordingly a statistically significant difference (Pax <0.0001, Psag < 0.0001), because of the fundamental difference in imaging approach (FLASH vs. turboFLASH). Differences for CNR were also statistically significant, independent of imaging plane (Pax = 0.001, Psag = 0.02).

CONCLUSION

Results demonstrate that turboFLASH BLADE is applicable at 3T with a 32-channel head coil for T1-weighted imaging, with reduced ghost artifacts. This approach offers the first truly clinically applicable T1-weighted BLADE technique for brain imaging at 3T, with consistent excellent image quality.

Links

Publisher Full Text
Aggregator Full Text

Authors+Show Affiliations

Attenberger UI
Department of Radiology, Scott & White Clinic and Hospital, Texas A&M University Health Science Center, Temple, TX, USA. ulrike.attenberger@med.uni-muenchen.de
Runge VM
No affiliation info available
Williams KD
No affiliation info available
Stemmer A
No affiliation info available
Michaely HJ
No affiliation info available
Schoenberg SO
No affiliation info available
Reiser MF
No affiliation info available
Wintersperger BJ
No affiliation info available

MeSH

BrainEquipment DesignEquipment Failure AnalysisHumansImage EnhancementImage Interpretation, Computer-AssistedMagnetic Resonance ImagingReproducibility of ResultsSensitivity and SpecificitySignal Processing, Computer-Assisted

Pub Type(s)

Journal Article

Language

eng

PubMed ID

19151605