Objective

Tumor hypoxia is one of the key factors that affect the efficacy of radiotherapy^1,2. Thus, the ability to detect dynamic changes in oxygen saturation level within tumor is of great importance. Steady-state free precession (SSFP) sequence has recently been employed for functional MRI (fMRI)³. In our study, we aim to verify the reliability of SSFP fMRI for detecting oxygen saturation level changes. To achieve this, we compare image signal variations with direct measurements obtained from rat C6 glioma using MRI compatible fiber-optic microprobes.

Methods

A total of 30 male Sprague-Dawley rats were used to establish the subcutaneous C6 glioma model. MRI compatible OxyLiteTM fiber-optic microprobes were embedded in the tumor after anesthesia (see Figure 1). The rats were then examined in the prone position using a clinical 3.0 T whole-body MRI system (Ingenia, Philips Healthcare, the Netherlands). A dStream small extremity 8-channel coil was used for signal reception. Acquisition parameters were as follows: TR/TE=6.9ms/3.4ms; flip angle=70°; FOV=80mm×80mm; acquisition matrix=136×134; slice thickness=2mm (four slices); number of dynamic scans=100. SSFP fMRI scans were performed with Carbogen inhalation⁴ while a series of tissue pO₂ values were recorded simultaneously in real time using the fiber-optic microprobes. Image data was processed using non-hemodynamic response function (non-HRF) analysis provided by AFNI software⁵. A spherical region of interest (ROI) with a 2mm diameter was placed at the tip of the fiber-optic microprobe to determine the T value and percentage signal change (PSC). The mean difference of pO₂ (ΔpO₂) before (pO_2-air) and after (pO_2-car) Carbogen inhalation in the ROI was calculated. The difference between pO_2-air and pO_2-car was tested by Wilcoxon matched-pair signed rank test. Lastly, we investigated the correlation between ΔpO₂ and T value, as well as between ΔpO₂ and PSC.

Results

We obtained pO₂ values before and after Carbogen inhalation at 28 different locations in the tumor region during SSFP fMRI scan. The experimental results, shown in the format of [median (interquartile range)], were as follows: pO_2-air = 0.580 (9.020) mmHg, pO_2-car = 3.005 (19.830) mmHg, ΔpO₂ = 1.015 (13.088) mmHg, T value = 2.686 (6.736), PSC value = 1.267 (3.257)%. The mean value of pO₂ significantly increased after Carbogen inhalation (p=10^-3). The mean value of ΔpO₂ measured at all of the 28 tumor locations correlated positively both to T value (Spearman’ r = 0.420, p = 0.026) and PSC value (Spearman’ r = 0.465, p = 0.013).

Discussion

In this study, the real time changes of pO₂ in the tumor were measured directly using OxyLiteTM fiber-optic microprobes. The difference of pO₂ before and after Carbogen inhalation was statistically significant (p=10^-3). As illustrated in Figure 2, temporal variation of the pO₂ value measured by the fiber-optic microprobes resembled that of the PSC value extracted from the predefined ROI in the SSFP fMRI images. The mean value of ΔpO₂ correlated positively to both T value and PSC value. These results indicate that changes in the SSFP image signals are consistent with the local tissue pO₂ measurements.

Conclusion

To conclude, fMRI using steady-state free precession (SSFP) is able to track changes in oxygen saturation level in rat glioma, and hence has the potential to detect tumor hypoxia.

Acknowledgements

No acknowledgement found