Quality Assurance: Accuracy, Precision, Controls and Phantoms 1

Quality assurance (QA) concepts for quantitative MRI measurements are explored, including frequency of scanning, healthy controls, phantoms and multi-centre studies. Uncertainty, accuracy, precision, and short- and long-term sources of error are considered. Parameter error can be modelled and minimised, using estimates of image noise. Sources of loss of accuracy and precision are considered. Measurement of precision is defined using Bland-Altman analysis of replicates. Healthy controls provide realistic QA, and normal values have small variation (<5%); values are given for T₁, MD and MTR. Controls can be used within-centre to monitor stability. In a multi-centre study, controls at each centre can be used to minimise between-centre variation. Phantoms (test objects) are considered in detail, including choice of material, temperature and field dependence, and stability. Doped water or agar remain useful, although organic liquids and iced water have their place. Large water phantoms are subject to dielectric resonance error. PD and MRS metabolite concentration standards required temperature control and correction. There are 4 tables, 13 figures and nearly 100 references in this chapter.