The role of medical image analysis in the management of the breast cancer patient is increasing, with methods being developed and evaluated for use in computer-aided diagnosis (CADx), as well as for prognostic markers, for predictive markers, and in the assessment of cancer risk. Depending on the task, such "image-based biomarkers" may include morphological, textural, and kinetic characteristics. This lecture reviews the Giger lab's developments of various quantitative image analysis methods in multi-modality breast imaging (mammography, ultrasound, and MRI), including advances in CADx and prognostic markers as well as in data reduction techniques, use of unlabeled data, and intelligent interfaces. The interpretation and understanding of these characteristics of tumors is expected to ultimately improve the management of the cancer patient.

View full abstract