BACKGROUND

Human body function is highly related to perfusion. Malignant lesions generally have a higher flow rate that is required to support the increased metabolic activities around tumors. This increased flow provides us a way to detect lesions and further characterize tumor types. Arterial spin labeling (ASL) is a technique in which endogenous water spins within the arterial blood are “labeled†at the tagging location before entering the imaging slices.

Magnetization transfer (MT) is a phenomenon that changes the MRI signal when magnetization exchange occurs between bound and free water. Because the perfusion signal differences between control and tagging in ASL are usually 1–2%, the incidental MT effect can lead to inaccurate perfusion estimates. Specifically, the tagging RF pulse applied adjacent to the imaging slice acts as an off-resonance excitation with respect to the water resonance frequency of the imaging slices. The MT effect can cause signal loss that is substantially larger than the perfusion-induced signal change [1]. It is thus essential to equalize the MT effect between the control and tagging scans. Using the same tagging strategy as followed in Edelman et al. [2], we test for equivalence of the MT effect in tagging and control pulse implementations.

METHODS

Six phantoms with different agar concentrations were imaged on a 3T Siemens system using turbo-FLASH as imaging sequences. A 20.48ms hyperbolic secant pulse with a nominal flip angle of 360° (peak voltage 428.5 V) and two 180° inversion pulses (peak voltage 303.0 & 214.3 V) were applied in tagging and control scan respectively. Other imaging parameters were: TR = 10 s, TE = 1.53 ms, Matrix size = 1.6 x 1.6 x 8.0 mm3. Post-labeling times = [0, 50, 100, 200, 400, 800, 1200, 1600, 2000, 2500, 3000] ms.

RESULTS

From figure 1 and 2, it is demonstrated that the tagging pulses are applied correctly in terms of thickness and offset. Figure 3 shows the differences between images acquired at multiple post-labeling times and 3sec. An overall larger difference can be observed from agar phantoms than that from water phantoms. Highest MT effect is found in phantom with 6% agar concentration. The slice profiles of phantoms with tagging pulse, two control pulses and no pulses are presented in Figure 4. These results demonstrate that the MT effects caused by tagging and control pulse are identical. No obvious difference is detected from the control pulses with two amplitudes.

CONCLUSIONS

It is demonstrated that the MT effects caused by the tagging and control pulses are identical. That is to say, the perfusion signal in ASL will not be affected by the MT effects because it is fully compensated. Future work includes in vivo study where the perfusion map from the signal difference between tagging and control is calculated.