Purpose: To evaluate the reliability of four CEST imaging metrics for brain tumors, at varied saturation power levels and magnetic field strengths (3−9.4 Tesla (T)). Methods: A five-pool proton exchange model (free water, semisolid, amide, amine, and NOE-related protons) was used for the simulations. For the in vivo study, eight glioma-bearing rats were scanned at 4.7 T. The CEST ratio (CESTR), CESTR normalized with the reference value (CESTRnr), inverse Z-spectrum-based (MTRRex), and apparent exchange-related relaxation (AREX) were compared. Results: The simulated CEST signal intensities using MTRRex and AREX were substantially increased at relatively high radiofrequency (RF) saturation powers at 3 T and 4.7 T, whereas CESTR and CESTRnr metrics remained relatively stable. There were tremendously high MTRRex and AREX signals around the water frequency at all field strengths because of the small denominators. In the rat tumor study at 4.7 T, both CESTR and CESTRnr showed clear contrasts in the tumor with respect to the normal tissue across all saturation power levels (0.5−3 μT), whereas the AREX showed negligible to negative insignificant contrasts. Conclusions: CEST metrics must be carefully selected based on the different experimental settings. CESTR and CESTRnr are more reliable at 3 T (a clinical field strength) and 4.7 T. Magn Reson Med 77:1853–1865, 2017.
- brain tumor
ASJC Scopus subject areas
- Radiology Nuclear Medicine and imaging