The aim was to optimally combine multichannel coil array data in sodium (23Na) MRI.Methods:
23 Na MRI was conducted on a 3 Tesla MR system using a 30-channel head receive coil array. The parameters used for the adaptive combination (ADC) reconstruction of the low signal-to-noise ratio (SNR) dataset have been optimized by finding the maximum mean SNR. A pseudo multiple-replica approach has been used to obtain SNR maps of the combined images. To prove reproducibility of the combination algorithm, the procedure was repeated for several measurements.Results:
For low SNR data, sum-of-squares (SOS) reconstruction leads to high background noise and a signal bias in the imaged object. The ADC reconstruction clearly reduces noise in the image and leads to an increase of the mean SNR in the range of 8% to 50%, compared to weighted SOS depending on the absolute SNR of the image. The evaluation of the effects of different noise scans showed that a small number of projections can be used to estimate noise statistics of the coil array without substantially decreasing the resulting SNR.Conclusion:
23 Na MRI can be markedly improved by using a 30-channel receive array and ADC reconstruction. The ADC reconstruction showed robust results for all measurements without the need for sensitivity maps. Magn Reson Med 75:527–536, 2016. © 2015 Wiley Periodicals, Inc.