8C-5 Full 3D Inversion of the Viscoelasticity Wave Propagation Problem for 3D Ultrasound Elastography in Breast Cancer Diagnosis

Abstract

An experimental 3D ultrasound elastography setup has been designed for breast cancer diagnosis improvement. 3D elastography assessment is generally based on the combination of adjacent 2D elasticity maps, obtained through simple 2D inverse problem resolution. Meanwhile, 3D sonoelastography is based on simple inversion approaches to the viscoelasticity problem. The system presented here is based on the resolution of a full 3D inverse problem, from the complete ultrasound-based measurement of the three components of the 3D displacement field. The 3D information considerably improves the accuracy and reliability of the quantitative measurements and circumvents the operator-dependent aspects of 2D echography diagnosis. The combination of 3D echography and elastography could be a very promising tool for in vivo breast cancer diagnosis. The X-ray system of a commercial mammographic bed was replaced by an ultrasound device. Shear waves were generated using a low frequency vibrator. Resulting displacements in tissues were imaged using an echographic probe moving stepwise around the breast. Advanced techniques such as compounding echographic probe sub-apertures and 2D vector Doppler algorithms were used to assess the three components of the displacement. Shear elasticity, viscosity and anisotropy were quantified using a 3D elastic properties reconstruction algorithm. A 3D finite difference simulation algorithm based on the viscoelastic propagation equation was used to model the 3D forward problem, and validate the inverse reconstruction algorithm. Simulated displacements in a numerical phantom were used as inputs for the inverse problem resolution, allowing the reconstruction of elastic properties similar to that of the numerical phantom. Similarly to MR-elastography, the inverse problem was solved in the Fourier domain. However, overcoming the data acquisition limitations of MR-elastography, the ultrasound-based approach enables the implementation of frequency compound methods based on averaging the data at different shear frequencies, increasing the measurement accuracy. In the present study, the experimental setup was optimized using numerical simulations and validated in vitro. In vitro experiments were conducted on a calibrated phantom exhibiting harder inclusions. Its 3D elastic properties were reconstructed and found consistent with that given by the manufacturer. This study allowed the numerical and experimental validation of the complete 3D elastography protocol.