Integrating biological knowledge, novel imaging modalities, and modeling in breast cancer diagnosis

Abstract

Despite tremendous advances in modern imaging technology, both early detection and accurate diagnosis of breast cancer are still unresolved challenges. Today, a variety of imaging modalities and image-guided biopsy procedures exist to identify and characterize morphology and function of suspicious breast tissue. However, a clinically feasible solution for breast imaging, which is both highly sensitive and specific with respect to breast cancer, is still missing. As a consequence, unnecessary biopsies are taken and tumours frequently go undetected until a stage where therapy is costly or unsuccessful. Currently, the exact diagnosis of suspicious breast tissue is ambiguous in many cases. To resolve this, computer aided diagnosis methods are developed which use knowledge extracted from large multimodal case databases. Clinical workstations must be developed to allow clinicians to use additional image modalities in an optimal way. Dedicated tools are required to guide clinicians in establishing a diagnosis, and should ultimately lead to more specific and accurate diagnostic decisions.