Quantitative breast elastography from B-mode images

Elastography images provide information about the mechanical properties of soft tissue in a noninvasive way and can be useful to identify abnormalities and ascertain differential diagnoses of suspicious prior findings obtained through mammography ultrasound. In this work we investigate, from a physics point of view, the feasibility of quantifying the Young's modulus of breast tissue from the autocorrelation of a diffuse acoustic field computed from a sequence of B-mode images acquired through conventional ultrasound scanners.

Inspired in the seismological approach of retrieving the Green's function by cross-correlation of diffuse fields, we obtained a quantitative expression that relates the local shear modulus of soft tissue to the autocorrelation of the displacement field generated by the presence of an acoustic diffuse field in the medium. In addition, we designed a mechanical prototype device adaptable to the breast anatomy, in order to create the necessary conditions in terms of diffuse field generation. The device is easy to handle, and its positioning does not interfere with the ultrasonic probe, being friendly to use within the clinical environment. The displacement field was measured from a sequence of B-mode images acquired at standard frame rates (30-50 Hz) with conventional ultrasound equipment. This method was tested in a breast tissue mimicking phantom using two standard ultrasound scanners (Toshiba Nemio NX and SIUI (Shantou Institute of Ultrasonic Instruments) Apogee 3800) and an open source research device (Verasonics V3.07 US). We also performed an in vivo measurement as a preliminary validation.

In the reconstructed Young's modulus maps the inclusions were identified and the obtained quantitative results for an inclusion and the background of the phantom were 60.0 ± 4.0 and 20.4 ± 0.5 KPa for the Toshiba equipment, 65.5 ± 6.9 and 22.6 ± 2.7 KPa for the SIUI equipment and 67.2 ± 7.3 KPa and 22.6 ± 2.8 KPa for the Verasonic scanner. These results are in good agreement with the values reported by the phantom's manufacturer of 60 and 20 KPa for the inclusion and the background, respectively. In the case of the in vivo measurement, the obtained images are in accordance with the patient known pathology (BI-RADS 5, Infiltrating Ductal Carcinoma, Score 6). The pathological breast showed a heterogeneous elasticity map with a mean Young's modulus of 98.1 ± 12.9 KPa, while the normal breast displayed a homogeneous map with a mean Young's modulus of 24.7 ± 8.1 KPa.

We successfully reconstructed the Young's modulus map of the breast tissue mimicking phantom and of a real breast tumor using B-mode images acquired with conventional ultrasound scanners. The results obtained in this work support that our technique can be developed as a medical tool to obtain quantitative breast tissue elasticity maps.