Reproducibility of Leeds TOR(MAM) mammographic test object plates

Image quality test objects are essential tools for quality assurance in any branch of radiology. A major problem in test object construction is to make all examples of the test object closely comparable so that results from different examples can be compared. In mammography, the TOR(MAM) test object is considered to be the most sensitive of those generally available in the UK. This paper reports a comparison of 19 examples of this test object design, using five film scorers. Small differences in mean score were found to be statistically significant in about half of the test object plates tested. However, none of the plates differed from the overall mean at the 95% confidence level. Mean differences of score between single plates did not exceed 8%.