Recent chemical exchange saturation transfer (CEST) MRI studies with administration of non-labelled D-glucose (Glc) and 2-deoxy-D-glucose (2DG) had demonstrated the capability of tumor detection [1-3]. However, the injection dose was relatively large in these studies, ranging from 1 to 3 g/kg, which will likely affect the normal physiology and limits its wide application. In this preliminary study, we evaluated the efficacy of an alternative technique, chemical exchange sensitivity spin-lock (CESL) [4,5], for the study of brain tumor. Brain MR images of Fisher rats bearing 9L glioma were acquired at 9.4 T in a total of 6 rats with detunable volume excitation and surface receiver coils. CESL data with spin lock time (TSL) of 0 and 50 ms were acquired in an interleaved manner for a total of 70 minutes, i.e., 25 min before and 45 min after an intravenous injection of 1 g/kg Glc (n = 4) and 0.25 g/kg of 2DG (n = 3). The 2DG was injected more than 2-hr after the Glc injection to make sure the blood and brain glucose level has returned to normal [4]. To gain a better understanding of the source of the glucoCESL signal in tumor, 0.25 g/kg of L-glucose (L-Glc) were injected to two different rats. The parameters for the image acquisition are: two-shot spin-echo EPI with a field of view 2.56 cm x 2.56 cm, matrix size 96 x 96, and one or two slices of 1.5-mm thickness. The repetition time was 2.5 s for each image, and the echo time was 8.5 ms. After the CESL studies, T1-weighted EPI images were acquired before and after an injection of Gd-DTPA to delineate the tumor region. For data analysis, time series of images with TSL = 0 and TSL = 50 ms were separated, and a time series of spin-lattice relaxation rate in the rotating frame (R1ρ) was calculated by pixel-wise fitting to R1ρ = ln(S_TSL=0/S_{TSL=50 ms})/50 ms [3]. Percent change maps were calculated by comparing the R1ρ data before and after the Glc or 2DG injection.Fig. 1A shows the Gd-enhanced T1-weighted images from two brain slices of a representative rat, in which the glioma can be identified as the hyperintense region. Fig. 1B and 1C show the relative R1ρ change maps of CESL with injection of Glc and 2DG, respectively. In both maps an increase of R1ρ can be seen in large portion of the brain, and the increase is higher within the tumor region (more yellow pixels). Comparable sensitivity was observed for the injection of 1 g/kg Glc versus 0.25 g/kg of 2DG. Fig. 2A shows the averaged time courses of the R1ρ change induced by glucose injection obtained from an ROI of normal contralateral tissue versus that from the tumor region. The onset-to-peak time is ~5 minutes and ~15 minutes for the tumor and the normal tissue ROI, respectively. R1ρ values from both ROIs return to baseline at similar time of 40 minutes post-injection, suggesting glucose is quickly metabolized in both regions. With 2DG injection (Fig. 2B), the R1ρ at normal tissue reaches a peak at ~15 min and sustain for the rest of the time course. In tumor region, R1ρ rises rapidly for the first 2 min or so, but the increase slows down after this initial period before it reaches a peak at ~15 min, and then start to decrease slightly. For both Glc and 2DG, the faster onset-to-peak time and larger magnitude of R1ρ response for tumor may be attributed to an elevated cellular glucose uptake, and/or an impaired blood brain barrier in the tumor vasculature. The difference between Glc- and 2DG-CESL results may indicate their different metabolic properties because Glc is fully metabolized but the metabolic product of 2DG (i.e., 2DG-6-phosphate) accumulates within cells. In Fig. 2C, injection of L-Glc (which has similar chemical exchange property as Glc, but cannot be transported by GLUT transporters) yields much smaller R1ρ change than 2DG for the same dose at 5-10 min after injection, suggesting that in the 2DG-CESL signal, the contribution from the BBB leakage should be small. Our results indicate that glucoCESL has high sensitivity and can be applied for the diagnosis and prognosis of brain tumor, offering a safe alternative to the conventional PET method. Further studies with administration of glucose and analogs will be helpful for the understanding of the source of the CESL contrast in brain tumor