A review of methods of clinical image quality evaluation in mammography

Consistency in evaluation of mammography images in research and clinical practice is dependent on a standardised clinical image quality evaluation system. Currently two such systems are available-one developed by the American College of Radiology (ACR) and the other by the European Commission (EU guidelines). The purpose of this study was to review mammography clinical evaluation methods in research studies and their adherence to these systems.

A total of 23 research articles were reviewed from the period 2000-2006, 11 of these studies used digital images. The focus of the review was the criteria and rating scales used.

Only 5 studies used either the ACR (3) or EU guidelines (2). The remainder included aspects of these systems together with a range of other criteria and rating scales. Variation was found in the categories of criteria used, number of criteria, the descriptors of the criteria and the instructions used to evaluate the criteria. Instructions were frequently not specific and open to individual interpretation. Although breast density is an important criterion of image quality and contributes to perception of breast lesions, inclusion of this criterion was not universal, and even when used the area of breast density to be evaluated was not identified, thus enhancing inter-observer variability. Scales that were absolute or relative were used for evaluation, all of which incorporated inconsistent numbers of steps.

Low adherence to ACR and EU Guidelines has resulted in considerable variation in the evaluation methods used in research studies. The implications of this variability are considerable both for evaluation of image quality in research outcomes and clinical practice.