Selective Amide- and NOE-CEST- MRI in Prostate at 7T using a Multi-transmit system
Catalina S. Arteaga de Castro1, Hans J.M. Hoogduin1, Vitaliy Khlebnikov1, Peter R. Luijten1, Dennis W.J. Klomp1, and Moritz Zaiss2

1Imaging Division, University Medical Center Utrecht, Utrecht, Netherlands, 2Medical Physics in Radiology, German Cancer Research Center, Heidelberg, Germany

Synopsis

The feasibility of selective NOE- and amide-CEST detection in the prostate at 7T was investigated with a multi-transmit system. Both effects can be acquired simultaneously due to the increased sensitivity and spectral resolution available at 7T. Fitted NOE- and amide-CEST were reproducible within experiments. NOE-CEST was found to be more pronounced than amide-CEST in small and whole prostate ROIs and the peripheral zone showed the lowest amide- and NOE-CEST effects.

Introduction

Multi-parametric MR (spectroscopic) imaging of the human prostate has shown to be able to differentiate cancerous from normal tissue[1]. However, correlation of MRI features obtained from DCE, T2, DWI and MRSI to Gleason scores coincides with relatively low specificity, hence limiting their applicability in treatment guidance. Other contrast mechanisms such as chemical exchange saturation transfer (CEST) have been explored for differentiation of tumor types in prostate cancer. In particular, the amide-CEST and nuclear Overhauser effected (NOE) part of the Z-spectra have shown altered contrasts in tumor. Recently, CEST techniques in prostate cancer were explores at 3T[2], indeed showing altered amide-CEST effects in tumors versus healthy prostate tissue, albeit with very high standard deviations. In addition, due to limited spectral resolution, NOEs and downfield amide signals get mixed. In this work, we show that going to higher field strengths such as 7 Tesla, increased spectral resolution and sensitivity allows for selective detection of amide- and NOE-CEST independently. We present how a multi-transmit system enables reproducible measurements of amide-CEST and NOE-CEST selectively at 7T. This is to our knowledge the first selective CEST approach in the body and the first detection of a NOE-CEST in the prostate.

Methods

Four male volunteers were scanned at a 7T MR scanner (Philips, Best, The Netherlands) with an array of 8 transmit receive fractionated dipole antennas[3] (MR Coils, Drunen, The Netherlands) that were positioned symmetrically around the pelvis. A multi-transmit system was used for B1+ shimming. Amplitude and phase shimming was done to maximize the B1+ and homogeneity in the prostate. After B1 shimming, T2-weigthed images were obtained for anatomy localization. Whole prostate CEST was acquired using a single slice 2D gradient echo (TE/TR=1.23/15 ms, 220x4x393mm FOV, 4x4x4mm voxel, 135 sec acquisition time, SENSE 1) with 80 saturation pulses (gaussian shape, 25 ms duration and 1uT B1 amplitude each) and a 30 ms interval. At least 16 offset frequencies were acquired to sweep 8 ppm around the water frequency. Four measurements were acquired after each other to assess the stability and reproducibility of the measurement. An additional measurement with 4 averages was also acquired to investigate increased SNR influence. Data analysis was performed with home-built MATLAB (R2014b, The Mathworks, Inc. ©) scripts that include Lorentzian fitting of water, amides, NOEs. Pixel and regions of interest (ROI) analysis were compared: including the complete prostate and smaller ROIs in the central gland and the peripheral zone.

Results and Discussion

Figure 1 shows an axial T2-weighted slice for anatomy localization and the associated B1 map for that slice (the CEST measurements were also acquired in the same slice). After RF shimming the B1 values within the prostate had variations below 25%. Figure 2 shows the stability of the four repeated measurements and a fifth measurement with 4 averages. Z-spectra are reproducible and remain stable within experiments. The standard deviation for the repeated measurements stays below 10% and below 5% for the data set with 4 NSA as expected, for both small and big ROI (fig 2 a, b). The available sensitivity at 7T also made possible to create amide- and NOE-CEST maps per voxel to show the heterogeneity in the prostate (figure 3). Part of the heterogeneity in the CEST maps can be explained by the remaining B1+ heterogeneity (figure 1 b) which accounted for up to 25% changes. The remaining contrast in the amide- and NOE-CEST maps are inherent of the prostate. For the first time, fitted NOE- and amide-CEST effects were individually detected in the prostate for big (complete prostate) and small (central gland and peripheral zone) ROIs (fig. 4). Lower amide- and NOE-CEST effects are found for the peripheral zone when compared to the central gland.

Conclusion

Selective detection of amide- and NOE-CEST are feasible at 7T in combination with a multi-transmit system in the prostate. The CEST sensitivity at the 7T field strength enables generation of amide- and NOE maps. Fitted amide- and NOE-CEST show inherent lower effects in the peripheral zone when compared to the central gland, which may be used to increase the specificity of grading prostate cancer.

Acknowledgements

No acknowledgement found.

References

1. Heerschap, A., et al., In vivo proton MR spectroscopy reveals altered metabolite content in malignant prostate tissue. Anticancer Research, 1997. 17(3 A): p. 1455-1460.

2. Jia, G., et al., Amide proton transfer MR imaging of prostate cancer: A preliminary study. Journal of Magnetic Resonance Imaging, 2011. 33(3): p. 647-654.

3. Raaijmakers, A.J.E., et al., The fractionated dipole antenna: A new antenna for body imaging at 7 Tesla. Magnetic Resonance in Medicine, 2015.

Figures

a) axial T2-weighted MRI used for localization of the prostate, b) B1 map obtained after B1 shimming for that same slice and c) zoomed B1 map (from white rectangle in b), scaled to show the residual B1 heterogeneity up to 25% (all values are in uT).

Z-spectra from small and big ROIs of 4 repeated identical protocols to check for stability of the measurement (Exp1-4). The measurement acquired with 4 NSA is also shown (Exp5). Standard deviation of Exp5 is halved as expected. Note that both NOE- (-3.5 ppm) and amide-CEST (3.5 ppm) effects are visible.

Fitted amide-CEST and NOE-CEST per voxel in five measurements in the same subject. Within experiments there is good reproducibility, for both effects. Contrast differences are inherent of the prostate. Only 25% is due to B1 residual in-homogeneities (see figure 1c).

Selective NOE- (-3.5 ppm) and amide-CEST (3.5 ppm) for the central and peripheral zone ROIs (red and blue resp.) and of the whole prostate (green). The NOE effect is more pronounced in the prostate for all ROI sizes and the peripheral zone has the lowest amide- and NOE-CEST values.



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
1523