Mammographic Quality Control Using Digital Image Processing Techniques

Breast cancer is the leading cause of cancer mortality among middle aged women. Survival and recovery depend on early diagnosis. At present mammography is one of the most reliable methods for early breast cancer detection. However relevance of diagnosis is highly correlated to image quality of the mammographic system. Hence periodic controls in mammographic facilities are necessary in order to make sure they work properly. In particular global image quality is evaluated from a mammographic phantom film. A phantom is an object with the same anatomic shape and radiological response as an average dense fleshed breast and in which are embedded structures that mimic clinically relevant features such as microcalcifications, nodules and fibrils. For each category of features, the targets have progressively smaller sizes and contrast so that the largest one is the most readily visible and the next is less visible and so on. Using a phantom makes it possible to free from the variable of differences in breast tissue positioning and level of compression from patient to patient. The process is as follows: the mammographic phantom film is analysed independently by several readers and a score is obtained by each of them depending on the number of objects they see. The independent object visibility scores are then averaged and the resulting score is assigned to the phantom film. Automating this score by using computer image processing of digitized phantom films should make the evaluation of mammographic facilities easier and less subjective. In addition image processing should enable us to take into account other parameters such as, for instance, noise, texture and shape of the targets that a reader eye cannot estimate quantitatively, and so to perform a more elaborate analysis. In collaboration with ARCADES (Association pour la Recherche et le Dépistage des Cancers du Sein et du col de l’utérus) which set, since 1989, a breast cancer screening program in South of France, a project aimed at automating phantom film evaluation is in progress. Such a project consists first in digitizing phantom films with the adequate spatial resolution and then in processing the obtained images in order to detect, segment and characterize the objects contained in the phantom. Little work has been done to automate quality control in mammographic facilities. Fast Fourier transform is used (Brooks et al., 1997) to establish some visibility criteria for the phantom test object. (Chakraborty et al., 1997) compares phantom images with a pattern image to obtain relations between some of the image parameters and the physical conditions