Purpose

Todays Cardiac Magnetic Resonance Imaging (CMR) protocols can be extremely time consuming, expensive, and uncomfortable for the patient. Particularly, when including the realm of multi parametric MRI features that aid diagnosis. In order to substantially accelerate the image acquisitions, high-density receiver coil arrays^1-3 may be used. Increasing the density of coils coincides with reducing the coil size, which in turn will reduce tissue loading and therefore can increase the noise figure of the MRI system. But, as long as tissue load remains dominant, the density of the coils can be increased without substantially adding noise⁴. In this study, we have investigated the use of high-density coil arrays for accelerated^5-7 cardiac MRI while assuring tissue load dominance. In process of building the receiver chain, 16 sequential measurements with a repositioned high density 16-channel receiver array are performed to obtain coil sensitivity and noise correlations in a healthy volunteer to investigate the possible acceleration performance of a 256 Channel Cardiac Coil Array at 3 Tesla.

Methods

A 16-element (4 x 4) coil array was designed (MR Coils BV, Zaltbommel, the Netherlands) with loop sizes of 55 mm in length and 33 mm in width (Fig. 1). Individual Q-factor measurements were performed using an S12 measurement with two pick-up probes weakly coupled (<<-30dB) to the loop. Unloaded and loaded Q values were measured at different locations of the body. Measurements were performed on a Philips 3T wide bore system with digital receivers (Philips, Best, the Netherlands). Coil sensitivity maps were recorded using the standard Philips reference scan (resolution = 6 mm isotropic, two stacks: one obtained with 16 channels as receiver and one with the bodycoil as receiver). A separate noise scan was obtained by switching off the RF transmit power. To simulate the expected improvement in acceleration of the 256 Channel Cardiac Array, the measurements were repeated 16 times, each time with the array positioned at another location around the chest of the patient. The 16 data sets of the array were spatially registered and merged into a 256 by 256 array filling out only the diagonal 16 times 16 elements. The remaining part of the array is set to zero. G-factor maps are calculated with SENSE accelerations in FH and LR using the aligned sensitivity maps and the noise matrix.

Results

The unloaded Q-factor was measured to be 603, while the loaded Q-factor ranged from 53-98. The noise correlation matrix is shown in Fig. 2. No noise correlation could be obtained for the cross terms outside the 16 element wide diagonal of the 256x256 matrix, due to absence of sufficient receiver chains. Note the low noise correlation within the 16-element wide diagonal. When merging the 16 images obtained with the 16 times 16 elements, good penetration of the reception is obtained throughout the body (Fig. 3). SENSE acceleration can go up to 5 in one dimension, or 4x5 in 2D acceleration (Fig. 4).

Discussion

The results show that a 256 Channel Cardiac Coil allows acceleration factors of up to 20 when accepting a g-factor of less than 2. It should be noted that the results are obtained with only 16 elements physically present during each scan. Consequently, potential couplings in a full 256 Channel array have not been incorporated. However, it is expected that these potential couplings will not effect our results substantially as the elements within the 16 Channel array also show very low coupling.

Conclusion

We have demonstrated that the coil element that can be used for a high-density 256 Channel Cardiac Array remains in tissue load dominance (6 to 11-fold). Moreover, coil couplings within the 4 x 4 array can be very low. When neglecting potential noise cross correlation within the remaining elements of the 256 Channel Array Coil, image acceleration of over 20 can be obtained with SENSE encoding. Consequently, these calculations demonstrate the motivation to build the 256 Channel Cardiac Array Coil to facilitate a boost in acceleration of the multi parametric cardiac MRI.

Acknowledgements

No acknowledgement found.

References

1. Roemer P.B. et al., MRM 16 (1990): 192-225. 2. Schmitt M. et al., MRM 59 (2008): 1431-1439. 3. Etzel R. et al., ISMRM 2015:1780. 4. Schuppert M. et al., ISMRM 2015:1779. 5. Sodickson D.K. et al., MRM 38 (1997): 591–603. 6. Pruessmann K. P. et al., MRM 42 (1999): 952–962. 7. Griswold M.A. et al., MRM 47 (2002): 1202–1210.