Clinicians, scientists interested in cancer research, breast imaging, and MR Elastography.Breast cancer is the second most common cancer and one in eight women is expected to develop this type of cancer during her life. Treatment options include (neo-adjuvant) chemotherapy, radiotherapy and surgery, with many women opting for breast-saving therapy: excision of the tumour while the breast remains in place. To reduce breast cancer-related mortality, screening programs have been implemented in many countries to detect early cancers. Population screening is done with X-ray mammography, often with follow-up by (dynamic) contrast-enhanced MRI. Although MRI has high sensitivity, its specificity is only moderate due to overlap between benign and malignant lesions resulting in false-positive findings and hence, relatively low positive predictive value¹. MR Elastography (MRE) has shown to be a promising technique for distinguishing between benign and malignant tissues based on their viscoelastic properties². However, breast MRE has not found widespread use in daily clinical practice, probably due to the relatively complex MRE actuator set-ups necessary to generate shear waves for MRE. Many proposed breast MRE actuator set-ups use a coil as engine (engine-coil) to generate the shear wave, which is closely placed to the breast. The B-fields generated by these coils disturb the imaged FOV in close vicinity to the engine-coil, which often results in artefacts that reduce the valid image volume. In this study we present a compact clinically easy to handle breast MRE set-up where the engine-coils are placed at a distance from the breast, resulting in high quality artefact-free MRE images.

Breast MRE setup

The set-up, shown in Fig. 1, consists of (A) two bilaterally placed 8 Ohm engine-coils, wound on a custom 3D ABS plastic printed holder, fixed on a (B) in-house designed PEEK plastic mechanical spring to transmit mechanical waves to both breasts simultaneously. The latter is done via (C) splined nylon low friction rods (Sullivan Products, UK), movable in a shaft, coupled to (D) another PEEK spring part mounted inside a (E) 7-channel breast RF-coil (Philips Healthcare, The Netherlands). A (F) 3D printed C-shaped breast-holder is fixed to this (D) PEEK spring part while an (G) Urias® inflatable splint (Arden Medical Ltd, United Kingdom) allows custom cup-size fixation of the breast with an (H) oppositely placed C-shaped breast-holder. For patient safety the engine-coils were positioned in a (I) closed box. The (J) head support was designed to cover the (C) rods and to provide a comfortable place for the positioning of the head. Overview of the whole set-up positioned on top of the MR scanner bed, is shown in Fig. 2.

In vivo MRI

MRE was performed on two clinical 3T MR Scanners (Achieva and Ingenia, Philips Healthcare, The Netherlands) in both the Academic Medical Centre (AMC), Amsterdam, The Netherlands and King’s College London (KCL), London, United Kingdom to test the set-up. In total 5 healthy volunteers (2×KCL, 3×AMC) and 1 post-lumpectomy patient (1×AMC) were scanned. MRE acquisitions were performed according to Garteiser et al. with in-phase TE=9.6ms, f_MECH=30Hz and 60Hz, f_MEG=120Hz and 2.5×2.5×2.5 mm³ voxels in 3 minutes per motion-encoding direction³. A waveform generator and amplifier were used to drive the engine-coils. Raw phase images were visually inspected for artefacts and processed in dedicated software to obtain elasticity maps⁴.

Artefact free breast MRE data were acquired in all subjects. Fig. 3 shows a representative example of the curl of the displacement (f_MECH=60Hz) for the X, Y, and Z directions of both breasts of a healthy volunteer. Excellent shear wave coverage is observed through the differently sized breasts. The movable C-shaped breast-holder (Fig. 1H) for adjusting the fixation to the breast size assures good coupling. In Fig. 4 the wavelength image and the elastogram of the same subject as in Fig. 3 is shown together with a high-resolution anatomical scan. The regions with longer wavelength and higher shear stiffness accurately co-localize with the parenchyma in the anatomical image. Finally, Fig. 5 shows a same-breast comparison of the curls of the displacement for the X direction for the two f_MECH, demonstrating excellent coverage at both frequencies.A compact and easy to handle breast MRE set-up was successfully designed, built and tested in two different institutions on two different types of Philips 3T MR scanners. Artefact free high quality MRE images of both breasts were acquired simultaneously with the set-up.