Introduction

Cancer cells exhibit up-regulated lactate dehydrogenase (LDH) resulting in increased levels of lactate production in tumors (Warburg effect)¹. Consequently, non-invasive imaging of lactate is of enormous significance, particularly in oncologic diseases or metabolic disorders². The ability of Lactate-weighted Chemical Exchange Saturation Transfer (Lactate-CEST) magnetic resonance imaging (MRI) to measure LDH activity in vivo was recently demonstrated by DeBrosse et al. in human skeletal muscle². In this work, we applied Lactate-CEST MRI at 7 Tesla (7T) to newly-diagnosed glioma patients in order to assess the methods potential to detect areas of increased lactate production in brain tumors.

Methods

Ten patients (5 female, 5 male) with newly diagnosed and histologically proven glioma (8 patients with glioblastoma WHO grade IV and 2 patients with glioma WHO grade II) were enrolled in this prospective institutional review board-approved study after written informed consent was obtained. Lactate-CEST MRI was performed at a 7T whole-body scanner (Magnetom 7T; Siemens Healthcare, Erlangen, Germany) employing a centric-reordered 2D-GRE-CEST sequence (three slices with resolution of 1.71x1.72x5 mm³) and a 24-channel head coil for reception. Pre-saturation consisted of 10 Gaussian–shaped pulses (pulse length: tp = 90 ms, duty cycle: DC= 80%, and saturation power B1 = 1.1 µT) applied at 31 equidistant frequency offsets in the range from -1.5 to 1.5 ppm with an additional M0-scan at -300 ppm. The WASABI approach was used to correct for B0 field inhomogeneities³. Magnetization transfer asymmetry (MTR_asym) calculation, defined as MTR_asym= Z(Δω = -0.6 ppm) - Z(Δω = +0.6 ppm), yields a lactate-weighted MRI contrast comparable to deBrosse et al². T1-weighted gadolinium contrast-enhanced (T1gdce) images were obtained along the clinical 3T protocol and used for data segmentation. High resolution T2-weighted images (matrix= 512 x 400, resolution=0.4 x 0.4 x 2 mm³, TR=12340 ms, TE=54 ms) were additionally acquired at 7T MRI. Data registration and segmentation was performed using the MITK software⁴. Three regions of interest (ROI) were selected on T1-gdce images for quantitative analysis: solid tumor, necrosis, and contralateral normal appearing white matter (CLNAWM). Mean signal intensities in the investigated ROIs of all patients were illustrated as boxplots and compared for statistical significant differences using the Wilcoxon rank-sum test after normality testing (Shapiro-Wilk-test) failed.

Results

Lactate values were significantly increased both in solid (3.94 ± 1.22) and necrotic (4.06 ± 1.29) tumor regions compared to CLNAWM (3.45 ± 1.21) overall patients (p<0.05) (Figure 1). There were no statistically significant differences between the solid tumor and necrosis (p>0.05). Figure 2 shows the lactate-weighted CEST MR imaging of one exemplary glioblastoma patient. Increased signal intensity displays in all tumor areas (necrosis, contrast enhanced region = CE-t1 tumor, and peritumoral edema) compared to CLNAWM (Fig 2). Visual observations are confirmed by Z-Spectrum analysis (Fig 2D) and MTRasym calculation (Fig 2E).

Discussion & Conclusion

In this work we show that lactate-weighted CEST-MRI reveals increased levels of lactate production in brain tumors of patients with glioma. Our results are in agreement with proton MR spectroscopy (¹H-MRS) studies reporting increased lactate concentrations in human glioma, particularly in glioblastoma⁵. However, DeBrosse et al. reported that lactate-weighted CEST MRI has significantly higher sensitivity than conventional ¹H-MRS². Therefore, Lactate-CEST MRI could serve as an additional imaging biomarker providing complementary information about the heterogeneity of tumors, with implications for biopsy targeting, patient therapy and response monitoring.