ShearWave™ Elastography imaging technique
Shear wave elastography is a novel technique for obtaining elastograms of soft tissue by tracking the transverse shear waves that spread laterally away from a mechanical disturbance of the tissue. They have special properties that differ from both the longitudinal waves of conventional ultrasound imaging (which relies on the bulk modulus of elasticity) and from conventional compression elastography. They travel at a few metres per second, depending on the visco-elastic properties of the tissue, and are rapidly attenuated, so that for practically achievable disturbances they have dissipated after travelling for a few centimetres (Fig 1 and Fig 2).
A particularly important feature is that their velocity is related to the Young’s modulus of elasticity so that a direct conversion to this fundamental figure is easily made. Though they travel relatively slowly through tissue, they traverse the typical 5 cm window of a small parts transducer in a few milliseconds and so are difficult to image using conventional scanning systems. In the Supersonic Imagine SWE system their progress is tracked using an ultrafast imaging system which insonates the entire field in a single plane pulse and then uses beam-forming to process the echoes on receive. This method can achieve frame rates of several thousands per second, whici is required to image the passage of the shear wave and to calculate its velocity. The SWE information is presented as a colour overlay on the B-mode information at frame rates of around 2 per second and the images are stable once they have settled around 2 seconds after turning the SWE on (Fig 3).
The shear wave is generated within the tissue using acoustic radiation force. In simple systems, the acoustic pressures required result in overheating of the transducer. In the Supersonic Imagine system a series of push pulses is sent faster than the velocity of the shear wave; this augments the effect by generating a supersonic front without transducer heating problems (Fig 4). The transmit power is well within the output limitations set by regulatory bodies (the MI is less than 1.5) and the resulting tissue excursion is great enough to be visualized while causing no risk of tissue damage.
Study protocol
A multinational clinical study, the BE1 (Breast Elastography 1) trial, aiming to recruit 2,300 patients with breast masses is at the half way mark. It started in mid-2008 and patients' enrolment is closed by the time of writing. Prototype systems of the Aixplorer® ultrasound system were placed at 16 clinical sites to capture both grayscale ultrasound and SWE images from a targeted two thousand three hundred lesions.
The study protocol was performed by licensed physicians or sonographers adequately trained in performing breast ultrasound examinations and interventional procedures. The protocol followed a methodology which fitted with the usual clinical workflow of breast ultrasound examination. Lesions that had been detected by palpation, mammography and/or ultrasound and/or MRI and for which an ultrasound examination (diagnostic or ultrasound guided procedure) was prescribed to characterize the lesion were included. Patients who met the inclusion criteria received both a conventional grayscale ultrasound exam and a SWE ultrasound exam. Images and data pertaining to the physical examination, mammograms, ultrasound and MRI studies were collected; when a biopsy and/or surgical excision were required, the histology results were made available to the study.
The main goal of the study was to assess the potential value of SWE for the characterization of breast lesions. More specifically, the protocol intended to assess if SWE information could potentially change the BI-RADS® classification with a benefit in specificity without degrading the sensitivity. Also, it would analyze the potential benefits of SWE images to differentiate cystic from solid lesions and benign from malignant lesions when FNA/biopsy results were available. Furthermore, we aimed at determining if quantitative SWE stiffness information as well as other SWE image features were reproducible.
A subset of 192 female breast lesions was analysed; 110 were benign and 82 were malignant. The reproducibility of SWE size and elasticity measurements was assessed using three still frames taken from the real-time sequence to measure intra observer reproducibility (IOR).
Logistic regression analysis was performed by adding features from the SWE data to predict the pathology findings, which were used as the gold standard. The reference model considered BIRADS®≥4 as a positive test for malignancy. One or two SWE features were added to BIRADS®≥4 cases sequentially to give models with two, three and more variables. These features were selected from the complete set of 8 features scored in the trial. They included three subjective features (Similarity between B Mode and SWE Mode shape of the lesions, SWE Mode shape of the lesions, Elasticity signal homogeneity) and five quantitative features (elasticity ratio between lesion over fat, Minimum SWE value of the lesion (kPa), Maximum SWE value of the lesion (kPa), Mean SWE value of the lesion (kPa), Ratio of the lesion area in SWE Mode over B Mode).
The impact of the adding the SWE features was assessed using the area under the receiver operator characteristics (ROC) curve, and the sensitivity, specificity, positive predictive value and negative predicted value for a given cut-off value of predicted probability of malignancy.