Robot-assisted identification of breast tumor biomechanics

There are increasing demands from physicians for providing modalities with second opinions, such as tumor hardness and mobility in the breast cancer diagnosis. Such system can improve the diagnosis accuracy and avoid unnecessary invasive biopsy through reducing the false positive cases. With such aim, we focused on complementing the ultrasound images with the biomechanics property of breast tumor so that the fused information can achieve this objective. In this paper, we utilize robot-automated inspection platform to acquire and characterize the mechanics data during palpating the breast tumor phantom. Subsequently a support vector regression model was constructed for describing tumor morphological mechanics model. The model has demonstrated its accuracy potentially to be applied in the future clinical applications.