Breast cancer is one of the cancer type most diagnosed. Its causes are unknown so there is not an effective way to prevent it, which increases the mortality rate. The early detection of breast cancer is the best practice to reduce this rate. The double reading of mammograms is a common practice to reduce the rate of missed cancer, but it has a hight cost. Computer Aided-diagnosis (CADx) Systems and Machine Learning Classifiers (MLCs) help to reduce these cost making automatic the second read of the mammograms.
The collaboration between CETA-CIEMAT, INEGI and FMUP-HSJ has generated a set of valuable resources to improve the breast cancer diagnosis process. We aim to achieve a reference repository for breast cancer diagnosis with BCDR, improving the existing implementations by storing a large number of annotated diagnosed cases reviewed by specialists, so researchers can have a reliable source of information for their researches. Using the BCDR data, the main MLCs algorithms are being tested in order to find the best configuration for obtaining accurate automatic diagnosis tool. MIWAD is a workstation which eases the specialist’s job in their diagnoses. It is a rich client for BCDR, and offers a set of tools that ease the breast screening and the integration of any MLCs to its workflow. All these resources has been built on top of DRI, a software platform aimed to ease the creation and management of digital repositories over heterogeneous storage. This work describes the advances made and the future work of the IMED project.