Algorithm for the fusion of ultrasound tomography breast images allowing automatic discrimination between benign and malignant tumors in screening tests

Abstract

This paper presents an algorithm which allows a multi-para-meter in vivo visualization of a breast structure by fusing two quantitative ultrasound tomography images and by automatically identifying areas of fat and glandular tissue, as well as areas of benign or malignant lesions, with the background of the reflection ultrasound tomography image of structures scattering ultrasonic waves. The threshold values for identifying particular tissue areas in quantitative images were adjusted to patient age using empirical linear functions. The viability of the algorithm was confirmed in preliminary in vivo breast ultrasound tomography screening tests. The obtained results allow a prediction that the ultrasound tomography scanner, which uses the here-presented algorithm to fuse the reconstructed images, and which is currently deployed by the private investor in cooperation with the authors of this paper, will provide a new standard in breast cancer diagnostics through fast and cheap screening tests.