Significant precision improvement for temperature mapping.

MR temperature measurements are important for applications such as the evaluation of thermal therapies and radiofrequency (RF) coil heating effects. In this work the spherical mean value (SMV) method has been applied to significantly improve the precision of MR temperature mapping in a homogeneous gel phantom. Temperature-increase maps of the phantom were obtained with three-dimensional (3D) MR phase difference mapping after heating with the RF coil. The temperature-increase distribution in most regions in the phantom is a harmonic function with the mean value property. Based on this property, the precision of temperature-increase maps was improved up to sixfold with the SMV method. Comparison of this method with conventional smoothing, further precision improvement, and the in vivo application of the SMV method are discussed. Copyright 2001 Wiley-Liss, Inc.