Key points are not available for this paper at this time.
The concept of quantitative MR (qMR) has existed for over three decades, offering direct access to biology and physiology, yet its implementation is still not straightforward or widespread. It has failed to translate into the clinic. Thus it is important to understand why this might be. Multi-centre studies show large differences between MR machines; good and convenient accuracy (closeness to the true value) and precision (repeatability) remain elusive. The development of suitable phantoms is a key stage in the evolution of qMR, and here a systematic categorisation is proposed. Currently there is much attention paid to creating phantoms containing materials with metrologically traceable values of MR quantities. However these simple phantoms are usually unrealistic; many of the disrupting phenomena present in clinical imaging are absent. These include RF B1 nonuniformity, and are a particular problem in body (as opposed to head) imaging, and at fields of 3T and above. Thus there is a premium on developing realistic phantoms. A proposal made for a realistic body phantom that includes RF B1 imperfections. It consists of lossy annuli placed around a standard head phantom.
Paul S. Tofts (Fri,) studied this question.