Identification of malignant breast tumors based on acoustic attenuation mapping of conventional ultrasound images

Although breast cancer imaging techniques continue to improve rapidly, about 75% of all breast biopsies turn out to be benign. These unnecessary biopsies are expensive and very stressful for the patients. In this paper we propose a new method for reducing the number of unnecessary biopsies. Our approach consists of transforming conventional ultrasonic images into corresponding attenuation maps. These maps are then analyzed, yielding automatic classification of malignant tumors. We provide a proof of concept for this approach by testing it on a benchmark of clinical images from three different image acquisition systems. Our tests show excellent sensitivity and specificity, indicating that up to four-fold reduction in the number of unnecessary biopsies may be possible. Moreover, we demonstrate the system robustness by working on all the images without any system-specific tuning.