Limits...
Water Selective Imaging and bSSFP Banding Artifact Correction in Humans and Small Animals at 3T and 7T, Respectively.

Ribot EJ, Wecker D, Trotier AJ, Dallaudière B, Lefrançois W, Thiaudière E, Franconi JM, Miraux S - PLoS ONE (2015)

Bottom Line: In addition, ankles, full head coverage and legs of volunteers were imaged, demonstrating the possible application of the method also for large FOV.In conclusion, this robust method can be applied in small animals and humans at high magnetic fields.The high SNR and tissue contrast obtained in short acquisition times allows to prescribe bSSFP sequence for several preclinical and clinical applications.

View Article: PubMed Central - PubMed

Affiliation: Centre de Résonance Magnétique des Systèmes Biologiques, UMR 5536, CNRS/University Bordeaux, Bordeaux, France.

ABSTRACT

Introduction: The purpose of this paper is to develop an easy method to generate both fat signal and banding artifact free 3D balanced Steady State Free Precession (bSSFP) images at high magnetic field.

Methods: In order to suppress fat signal and bSSFP banding artifacts, two or four images were acquired with the excitation frequency of the water-selective binomial radiofrequency pulse set On Resonance or shifted by a maximum of 3/4TR. Mice and human volunteers were imaged at 7 T and 3 T, respectively to perform whole-body and musculoskeletal imaging. "Sum-Of-Square" reconstruction was performed and combined or not with parallel imaging.

Results: The frequency selectivity of 1-2-3-2-1 or 1-3-3-1 binomial pulses was preserved after (3/4TR) frequency shifting. Consequently, whole body small animal 3D imaging was performed at 7 T and enabled visualization of small structures within adipose tissue like lymph nodes. In parallel, this method allowed 3D musculoskeletal imaging in humans with high spatial resolution at 3 T. The combination with parallel imaging allowed the acquisition of knee images with ~500 μm resolution images in less than 2 min. In addition, ankles, full head coverage and legs of volunteers were imaged, demonstrating the possible application of the method also for large FOV.

Conclusion: In conclusion, this robust method can be applied in small animals and humans at high magnetic fields. The high SNR and tissue contrast obtained in short acquisition times allows to prescribe bSSFP sequence for several preclinical and clinical applications.

No MeSH data available.


3D human knee images acquired at 3T using standard bSSFP and WS-bSSFP sequences in combination with parallel imaging.The sagittal and axial images shown were either acquired On Resonance (OR + 0Hz), OR + (3/4TR) apart from the resonance (for TR = 7ms, the shift equals 107Hz), or summed using SOS technique (SOS2 or SOS4) and in combination with parallel imaging (SENSE 4). Arrows point at banding artifacts altering cartilage measurements. The arrowhead indicates a banding artifact in the muscle at the edge of the FOV.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4591352&req=5

pone.0139249.g004: 3D human knee images acquired at 3T using standard bSSFP and WS-bSSFP sequences in combination with parallel imaging.The sagittal and axial images shown were either acquired On Resonance (OR + 0Hz), OR + (3/4TR) apart from the resonance (for TR = 7ms, the shift equals 107Hz), or summed using SOS technique (SOS2 or SOS4) and in combination with parallel imaging (SENSE 4). Arrows point at banding artifacts altering cartilage measurements. The arrowhead indicates a banding artifact in the muscle at the edge of the FOV.

Mentions: The 3D WS-bSSFP sequence was then applied on a clinical MRI system at 3T in order to perform human musculoskeletal imaging (Fig 4, Table 4). Images acquired using the standard bSSFP sequence without the binomial pulse exhibited bone marrow and fat with hyperintense SNR (19.4±0.4 and 27.3±1.4, respectively). The presence of the chemical shift artifacts could alter the accurate measurement of cartilage thickness. Furthermore, as the synovial fluid engendered similar signal intensity than adipose tissue, its detection was impeded. However, as soon as the binomial pulse was used, the SNR from subcutaneous fat and bone marrow drastically decreased to 2.4±0.7 (p = 7x10-4) and 1.1±0.2 (p = 4x10-5), respectively. The SNR of other tissues (muscle, cartilage, joint fluid) remained similar (Table 4), except at the location of banding artifacts. For example, SNR of muscles affected by large bandings (example on Fig 4 arrowhead) decreased from 8.7±0.5 to 2.5±0.3 on the OR images. Fat SNR was inferior to 5 on the entire knee image even though the binomial pulse frequency was shifted by 107Hz (representing the maximum offset reached at TR = 7ms), whereas the signals from the other knee structures were significantly higher (Table 4).


Water Selective Imaging and bSSFP Banding Artifact Correction in Humans and Small Animals at 3T and 7T, Respectively.

Ribot EJ, Wecker D, Trotier AJ, Dallaudière B, Lefrançois W, Thiaudière E, Franconi JM, Miraux S - PLoS ONE (2015)

3D human knee images acquired at 3T using standard bSSFP and WS-bSSFP sequences in combination with parallel imaging.The sagittal and axial images shown were either acquired On Resonance (OR + 0Hz), OR + (3/4TR) apart from the resonance (for TR = 7ms, the shift equals 107Hz), or summed using SOS technique (SOS2 or SOS4) and in combination with parallel imaging (SENSE 4). Arrows point at banding artifacts altering cartilage measurements. The arrowhead indicates a banding artifact in the muscle at the edge of the FOV.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4591352&req=5

pone.0139249.g004: 3D human knee images acquired at 3T using standard bSSFP and WS-bSSFP sequences in combination with parallel imaging.The sagittal and axial images shown were either acquired On Resonance (OR + 0Hz), OR + (3/4TR) apart from the resonance (for TR = 7ms, the shift equals 107Hz), or summed using SOS technique (SOS2 or SOS4) and in combination with parallel imaging (SENSE 4). Arrows point at banding artifacts altering cartilage measurements. The arrowhead indicates a banding artifact in the muscle at the edge of the FOV.
Mentions: The 3D WS-bSSFP sequence was then applied on a clinical MRI system at 3T in order to perform human musculoskeletal imaging (Fig 4, Table 4). Images acquired using the standard bSSFP sequence without the binomial pulse exhibited bone marrow and fat with hyperintense SNR (19.4±0.4 and 27.3±1.4, respectively). The presence of the chemical shift artifacts could alter the accurate measurement of cartilage thickness. Furthermore, as the synovial fluid engendered similar signal intensity than adipose tissue, its detection was impeded. However, as soon as the binomial pulse was used, the SNR from subcutaneous fat and bone marrow drastically decreased to 2.4±0.7 (p = 7x10-4) and 1.1±0.2 (p = 4x10-5), respectively. The SNR of other tissues (muscle, cartilage, joint fluid) remained similar (Table 4), except at the location of banding artifacts. For example, SNR of muscles affected by large bandings (example on Fig 4 arrowhead) decreased from 8.7±0.5 to 2.5±0.3 on the OR images. Fat SNR was inferior to 5 on the entire knee image even though the binomial pulse frequency was shifted by 107Hz (representing the maximum offset reached at TR = 7ms), whereas the signals from the other knee structures were significantly higher (Table 4).

Bottom Line: In addition, ankles, full head coverage and legs of volunteers were imaged, demonstrating the possible application of the method also for large FOV.In conclusion, this robust method can be applied in small animals and humans at high magnetic fields.The high SNR and tissue contrast obtained in short acquisition times allows to prescribe bSSFP sequence for several preclinical and clinical applications.

View Article: PubMed Central - PubMed

Affiliation: Centre de Résonance Magnétique des Systèmes Biologiques, UMR 5536, CNRS/University Bordeaux, Bordeaux, France.

ABSTRACT

Introduction: The purpose of this paper is to develop an easy method to generate both fat signal and banding artifact free 3D balanced Steady State Free Precession (bSSFP) images at high magnetic field.

Methods: In order to suppress fat signal and bSSFP banding artifacts, two or four images were acquired with the excitation frequency of the water-selective binomial radiofrequency pulse set On Resonance or shifted by a maximum of 3/4TR. Mice and human volunteers were imaged at 7 T and 3 T, respectively to perform whole-body and musculoskeletal imaging. "Sum-Of-Square" reconstruction was performed and combined or not with parallel imaging.

Results: The frequency selectivity of 1-2-3-2-1 or 1-3-3-1 binomial pulses was preserved after (3/4TR) frequency shifting. Consequently, whole body small animal 3D imaging was performed at 7 T and enabled visualization of small structures within adipose tissue like lymph nodes. In parallel, this method allowed 3D musculoskeletal imaging in humans with high spatial resolution at 3 T. The combination with parallel imaging allowed the acquisition of knee images with ~500 μm resolution images in less than 2 min. In addition, ankles, full head coverage and legs of volunteers were imaged, demonstrating the possible application of the method also for large FOV.

Conclusion: In conclusion, this robust method can be applied in small animals and humans at high magnetic fields. The high SNR and tissue contrast obtained in short acquisition times allows to prescribe bSSFP sequence for several preclinical and clinical applications.

No MeSH data available.