Physicists and physicians interested in quantitative tissue properties of the prostate measured by MR elastography (MRE).MRE is capable of generating image contrast based on the viscoelastic properties of tissue by inducing and detecting time harmonic shear waves in the body [1]. Different techniques to induce vibrations into in vivo prostate tissue were proposed in the past including transurethral [2] and endorectal [3] methods with excitation frequencies from 100 to 300 Hz, and transperineal methods [4,5] in the frequency range from 50 to 75 Hz. We identified two major challenges to bring prostate MRE into the clinic: 1) the need for specialized hardware and 2) limited spatial resolution of elastograms, especially for surface-based methods which suffer from low wave amplitudes at higher vibration frequencies. Therefore, we propose the use of pressurized-air actuators previously developed for abdominal MRE in the liver, spleen and kidney [6]. Moreover, we operated the actuators in a frequency range lower than reported in the literature (from 30 to 50 Hz) which better ensures that sufficiently high wave amplitudes reach the prostate. To achieve high resolution at the same time, multifrequency wave-number based inversion is proposed.To demonstrate the feasibility of in vivo prostate MRE using pressurized-air actuators in the low frequency range of 30 to 50 Hz.

An MRE setup was developed utilizing a single shot spin-echo EPI based sequence [7] with continuous vibration induced by three pressurized-air actuators. Two actuators (11 cm membrane diameter) were placed posterior and one actuator (7 cm membrane diameter) anterior to the pelvic region (see Fig. 1). The actuators were supplied by compressed air of approximately 0.5 bar from a 5-bar medical air pipeline and regulated by high-speed electromagnetic valves which were opened and closed with the desired drive frequency [6]. The prostate of eight healthy volunteers (mean age: 34, age range: 26-55) was scanned with a 12-channel phased array surface coil at 3 drive frequencies (30, 40 and 50 Hz). Further parameters: 1.5 T Siemens Sonata MR scanner, 21 transverse slices, 2.5 mm isotropic voxel size, 8 wave dynamics, 3 wave field components, MEG amplitude = 32 mT/m, TR = 2500 ms, TE = 55 ms, FoV = 260×260 mm², matrix size: 104×104, GRAPPA factor 2, 2 signal averages, measurement time per frequency: 2:04 min.

In order to reconstruct the parameter maps depicting the shear wave speed $$$c$$$, the acquired images were analyzed by multifrequency 2D-inversion of the wave number $$$k$$$. By this method, the shear wave speed $$$c(\boldsymbol{r})$$$ at location $$$\boldsymbol{r}$$$ is reconstructed via weighted summation of the wave numbers $$$k_{(j)}$$$ corresponding to the directionally filtered shear wave field component $$$u_j$$$ at drive frequency $$$\omega_i$$$ and directional filter index $$$n$$$:

\[ k_{(j)}(\boldsymbol{r},\omega_{i},n)=\left\Vert \nabla\left(\frac{u_{j}(\boldsymbol{r},\omega_{i},n)}{|u_{j}(\boldsymbol{r},\omega_{i},n)|}\right)\right\Vert ,\qquad\frac{1}{c(\boldsymbol{r})}=\frac{\sum_{i,j,n}\frac{k_{(j)}(\boldsymbol{r},\omega_{i},n)}{\omega_{i}}w}{\underset{i,j,n}{\sum}w} \]

The weighting factor $$$w=w(\boldsymbol{r},\omega_{i},n)$$$ is empirically set to $$$w = |u_j( \boldsymbol{r},\omega_{i},n)|^4$$$

As visible in Fig. 2, waves penetrate well into prostate tissue. The reconstructed $$$c$$$-maps show a clear separation between the prostate and the surrounding tissue. Moreover, a clear distinction between central zone and peripheral zone is visible based on the elastic properties. Values for the whole prostate, central zone and peripheral zone of each individual are given in in Table 1. In all $$$c$$$-maps, the central zone shows significantly higher $$$c$$$-values (1.71 ± 0.22 m/s in the central zone compared to 1.15 ± 0.11 m/s in the peripheral zone).We achieved high wave amplitudes throughout the prostate by using three pressurized-air actuators in the lower frequency regime between 30 and 50 Hz. Wave-speed maps with sufficiently high resolution are obtained to discriminate between stiffness values of central zone and peripheral zone which well agree to values reported in the literature [4,5]. This is an encouraging result of our study, since no specialized hardware in addition to the abdominal MRE setup was needed and longer wavelengths pertaining to lower vibration frequencies apparently not compromised the quality of our elastograms. Altogether, the proposed MRE setup for investigations of the prostate promises robust and easy-to-use applications in the clinic.