No Phase Wrap
How does phase oversampling eliminate the wrap-around artifact?
|
|
|
Phase oversampling, also known as "No Phase Wrap", is a technique to reduce or eliminate the wrap-around artifact. As described in the prior Q&A, phase wrap-around, a form of aliasing, occurs when the anatomic dimensions of an object exceed the defined field-of-view (FOV). The portions of the object outside the FOV are misidentified in terms of frequency and are folded over into the image from the periphery. Wrap-around is nearly exclusively seen in the phase-encode direction.
|
Wrap-around artifact
|
Several straightforward methods to reduce or eliminate phase wrap-around were discussed in the prior Q&A. These include: 1) increasing the FOV; 2) swapping phase- and frequency-axes; 3) using surface coils; and 4) applying saturation bands.
All vendors also offer phase oversampling methods to control phase wrap-around, the subject of the current Q&A. As expected, the nomenclature varies among the major manufacturers: "Phase oversampling" (Siemens), "No phase-wrap" (GE), "Fold-over suppression" (Philips), "Anti-wrap" (Hitachi), and "Phase-wrap suppression" (Canon).
With only minor variations, all vendors follow the same technique, illustrated below. The example illustrates the case of 100% phase oversampling, although the oversampling factor is adjustable to some extent on most scanners.
Phase-oversampling involves four steps, performed automatically in scanner software when this option is selected: (1) the field-of-view is doubled in the phase-encode direction, (2) the number of phase-encoding steps (Np) is doubled, (3) the number of excitations is cut in half, and (4) only the middle portion of the reconstructed image is displayed. Steps (1) and (2) maintain spatial resolution at a level identical to that before the phase-oversampling option was selected. Step (3) preserves signal-to-noise and imaging time.
Phase-oversampling involves four steps, performed automatically in scanner software when this option is selected: (1) the field-of-view is doubled in the phase-encode direction, (2) the number of phase-encoding steps (Np) is doubled, (3) the number of excitations is cut in half, and (4) only the middle portion of the reconstructed image is displayed. Steps (1) and (2) maintain spatial resolution at a level identical to that before the phase-oversampling option was selected. Step (3) preserves signal-to-noise and imaging time.
Although in general the phase-oversampling technique may be used without penalty, certain limitations and special considerations apply. First, phase oversampling cannot be combined with some specialized multi-section techniques (e.g., POMP), which also require increasing the defined FOV and altering phase offsets. Secondly, since the number of excitations is reduced to keep imaging time constant, the phase oversampling technique can be applied without time penalty only to scans in which at least two excitations were originally selected. Phase oversampling with a single excitation sequence can still be performed, but requires the use of partial Fourier ("½ NEX") imaging with a consequent loss in signal-to-noise and increase in phase errors.
Finally, phase-oversampling will not prevent noise and motion-induced phase shifts originating in tissues outside the FOV from propagating into the imaged volume (even though stationary tissues are effectively excluded). For example, a sagittal cervical spine examination phase-encoded in a superior to inferior direction may still be degraded by respiratory and cardiac pulsation artifacts, although the heart and chest are not visually wrapped over the spinal image.
Finally, phase-oversampling will not prevent noise and motion-induced phase shifts originating in tissues outside the FOV from propagating into the imaged volume (even though stationary tissues are effectively excluded). For example, a sagittal cervical spine examination phase-encoded in a superior to inferior direction may still be degraded by respiratory and cardiac pulsation artifacts, although the heart and chest are not visually wrapped over the spinal image.
References
Axel L, Morton D. Correction of phase wrapping in magnetic resonance imaging. Med Phys 1989;16:284-287.
Heiland S. From A as in aliasing to Z as in zipper: artifacts in MRI. Clin Neuroradiol 2008; 1:25-36.
Pusey E, Yoon C, Anselmo ML, Lufkin RB. Aliasing artifacts in MR imaging. Comput Med Imag Graphics 1988;12:219-224.
Zhuo J, Gullapalli RP. AAPM/RSNA physics tutorial for residents. MR artifacts, safety, and quality control. Radiographics 2006;26:275-297.
Axel L, Morton D. Correction of phase wrapping in magnetic resonance imaging. Med Phys 1989;16:284-287.
Heiland S. From A as in aliasing to Z as in zipper: artifacts in MRI. Clin Neuroradiol 2008; 1:25-36.
Pusey E, Yoon C, Anselmo ML, Lufkin RB. Aliasing artifacts in MR imaging. Comput Med Imag Graphics 1988;12:219-224.
Zhuo J, Gullapalli RP. AAPM/RSNA physics tutorial for residents. MR artifacts, safety, and quality control. Radiographics 2006;26:275-297.
Related Questions
Why does phase wrap-around occur?
Why does phase wrap-around occur?