Breast Ultrasound Elastography Using Full Inversion-Based Elastic Modulus Reconstruction

Several cancer types, including breast cancer, are associated with tissue structural changes that yield tissue stiffening. Clinical breast examination (CBE) is a physical examination of the breast to find palpable breast tumors. This test lacks accuracy necessary for effective assessment and diagnosis of breast cancer. To develop an effective breast cancer diagnostic technique, an imaging method is proposed that maps the distribution of breast tissue relative elasticity modulus. Unlike CBE, this technique is quantitative; hence, it is expected that its accuracy is independent of the physician's experience. The proposed technique is a quasi-static elastography technique which uses radiofrequency data acquired through ultrasound imaging to determine both axial and lateral tissue displacements resulting from tissue mechanical stimulation. These displacements serve as input data for elastography image reconstruction. The reconstruction technique is developed using a full inversion framework where elastic tissue deformation equations are inverted using an iterative process. Each iteration in this process involves stress computation using finite-element analysis followed by updating elastic modulus until convergence is achieved. The proposed technique was validated by two tissue mimicking phantom studies before it was successfully applied to a clinical case. The two independent phantom studies demonstrated the robustness of the proposed method demonstrated by reconstruction errors of less than 12%. Elastic modulus images of the clinical case were compared to corresponding B-modes images where cancerous areas were identified as hypo-echoic areas. This comparison indicated marked tissue stiffening in those areas. Results obtained from the phantom and patient studies conducted in this study indicate that the proposed method is reasonably accurate; hence, the technique can be potentially used for quantitative assessment of breast cancer. The elasticity reconstruction algorithm developed in this work can be easily implemented on clinical ultrasound systems with no requirement to any additional hardware attachment for mechanical stimulation or data acquisition. As such, it can be applied as a low cost and potentially widely available technology for breast cancer diagnosis.