Chemical exchange saturation transfer (CEST) is a type of magnetization transfer (MT) which employs the saturation transfer from low concentration endogenous and exogenous pools of exchangeable protons to the bulk water proton pool. There have two CEST effects apparent in vivo which are attributed to protons of mobile proteins: amide proton transfer (APT) and nuclear Overhauser enhancement (NOE). The NOE signals originate from mobile macromolecular components with a spectral range upfield relative to water. Glioma and meningioma, as the most common malignant and benignant tumors in human brain, which imaging diagnosis and the boundary description are significant to clinical treatment. NOE has been attributed to increased mobile protein concentrations in malignant cells. In our study, we evaluated the difference between glioma and meningioma in NOE signals to explore whether NOE* imaging can identify the differentiation of benign and malignant in human brain tumor.

All MR imaging was acquired on a 3T MR imaging system. CE-T1 weighted images were performed. In addition, MT-prepared GRE MRI sequence was used for CEST imaging. The MT saturation pulse was chosen as a 20 ms width Fermi pulse with B₁ of 0.6 μT. Forty-nine equidistant frequency offsets between 6 and -6 ppm and the additional S₀ image were acquired. Z-spectra were corrected for B₀ inhomogeneity employing a water saturation shift referencing map (WASSR).

Images were processed using the Matlab. Z-spectra were calculated from the normalized images for ROI. For NOE image mapping, three offset frequencies were selected: one at the center frequency of the NOE^* peak, while the other two were at the upper and lower bounds of the peak. In addition to the S₀ was the signal intensity without irradiation. Then NOE map can be obtained as equation: NOE^*=S{[(-2.5ppm)+S(-5ppm)]/2-S(-3.5ppm)}/S₀.

Eleven patients with recently diagnosed and histopathologically confirmed brain tumors and six normal controls were included in this prospective study.

Fig. 1A is a T₂WI of brain from a healthy volunteer. To quantitatively compare the NOE effect in the human brain, regions of interest (ROI) were analyzed. The ROIs enclosing the white matter and the gray matter were carefully drawn on the T2WI images. The average z-spectrums were drawn in Fig. 1B at B₁ = 0.6 μT. Both white matter and gray matter showed a board peak at -2.5 ppm to -5 ppm due to NOE effect.

Figure 2A-D show a patient with a grade-III astrocytoma in the left temporal lobe. In this case, the tumor core signal is enhanced area in the Gd-enhanced T1-weighted image (Fig. 2A). In Fig. 2B, we can observe that the tumor core is hyperintense while the NOE* is hypointense. In Fig. 2C and 2D, the average Z-spectrum and a linear fitting of date points were measured from tumor and contralateral normal-appearing white matter (CNAWM). We observed a drop in NOE signal strength (about 0.3 %) at -2.5 ppm to -5 ppm in tumor core region.

Figure 2E-H show an example of a patient who had a pathologically proven transitional meningioma in the right parietal lobe. The tumor core, as identified by the signal enhancement in the post-contrast T₁WI images (Fig. 3E). In the Fig. 3F, we can see there is no significant difference between tumor core and CNAWM in the NOE* image. However, we also can obtain the similar results in the average Z-spectrum and a linear fitting of date points was measured with the tumor core and CNAWM regions (Fig. 3G and H). NOE^*% is not significantly different between the tumor and CNAWM.

NOE^* contrast at -3.5 ppm was different in two types of human brain tumors, when we compared tumor core with CNAWM. Results demonstrated a positive NOE^*% at -3.5ppm in glioma (p ＜ 0.001) but an insignificant NOE^*% in meningioma (p = 0.112).

We investigated NOE signals in healthy human brain and human brain tumors at 3T. The NOE signals (-2.5 ~ -5 ppm) were found to peak around the B₁ of 0.6 μT, and were more abundant in the white matter than in the gray matter. However, in human brain tumors, we found that NOE signals were attenuated in glioma and not in meningioma. These findings may have implication for future NOE studies in the human brain tumors in clinical application. NOE imaging may help to distinguish the heterogeneity of benign or malignant tumors, and these capabilities of NOE imaging may be very important in diagnosis and treatment planning for patients with brain tumors.