The recent discovery of point mutations in isocitrate dehydrogenase 1 & 2 (IDH1/2) oncogenes and subsequent gain-of-function production of 2-hydroxyglutarate (2HG) has permanently altered the clinical landscape for patients with low-grade glioma (LGG). While the feasibility of utilizing 2HG as a magnetic resonance biomarker has been established ex vivo [1], several different approaches to obtaining in vivo data have been presented [2,3,4,5]. The asymmetric echo technique proposed by Choi et al [2] holds significant promise due to its ease of implementation and use of J-modulation to highlight the 2HG signal. This project aims to assess the concordance of 2HG detection using asymmetric echo PRESS MRSI with IDH1_R132H mutation as identified via antibody staining in patients with LGG and to investigate the relationship of other metabolites detected with this sequence to IDH status.

A total of 26 patients with newly-diagnosed (n=16) or recurrent (n=10) LGG were included in this IRB-approved study. MR imaging exams were either performed immediately prior to neurosurgery (n = 16) or as part of a serial imaging study (n = 10) at our institution.

Scans were conducted on a 3T EXCITE GE scanner (GE Healthcare) using an eight-channel phased-array head coil. Data was acquired using a 3D ¹H MRS asymmetric PRESS (TE = 97ms) with very selective saturation (VSS) pulses for lipid signal suppression (excited volume 80×100×40 mm³ or larger, overpress factor 1.2), TR/TE1/TE2 = 1140/32/65 ms, FOV = 16×16×16 cm³ (1×1×1 cm³ voxels) flyback echo-planar readout gradient in the SI direction, 712 dwell points and 988 Hz sweep width.

Immunohistochemistry (IHC) was performed to determine IDH1 mutation status. Because the IDH1_R132H antibody is specific for histidine variants of IDH-enzymes, samples with other IDH1 and IDH2 mutations, if present, would not have been detected.

MRSI data were reconstructed using in-house software [6]. T2 hyperintense lesions were defined on the axial FLAIR images. Spectra from voxels that were overlapping by ≥50% with the region of T2 hyperintensity were phase- and frequency-corrected, and then averaged to increase the SNR and improve the detection of 2HG (Figure 1). Metabolites in these averaged spectra were quantified via fitting to a simulated data set with 21 metabolites, using the LCModel software package [7]. The metabolite levels that are reported were normalized relative to total creatine (PCr+Cr), in order to improve the quality of comparison between different scans and patients. The detection of 2HG was defined by a positive nonzero linear combination fit with Cramér-Rao Lower Bounds(CRLB)≤20% of the total estimate. For all other metabolites, those with CRLB lower than 20% were involved in analysis.

The antibody staining technique was positive for the IDH_R132H enzyme in tissue samples from 19 patients and negative in 7 patients.

An example of 3D MRSI and averaged spectra from patient with IDH1+ oligodendroglioma is shown in Figure 1. The averaged number of T2 voxels from all the patients was 34, and the number of T2 voxels for IDH1- and IDH1+ groups is illustrated in Figure 2. The detection of 2HG associated with IDH1-mutation status as determined by IHC are presented in Table 1. There was a positive agreement of 2HG detection with IDH1 mutation of 32% (6/19 IDH1+ samples showed a statistically significant 2HG fit) and a negative agreement of 100% (7/7 IDH1− did not show any 2HG fit). This resulted in an overall concordance of 50%. Figure 3 illustrates 2HG/tCr for each subject. Comparisons between other metabolites demonstrated a statistically significant difference between groups in levels of 2HG and glutathione (GSH) (Figure 4).

The non-invasive detection of IDH mutation has the potential to add highly valuable information that will improve the clinical standard-of-care for patients with LGG. The results of this study show that the sensitivity for detecting quantifiable levels of 2HG was low, which is in agreement with another recent preliminary report [8]. Further research is required before routine clinical implementation of these methods is recommended. Correlations between 2HG detection and tumor size, grade, and type are being be explored and it is expected that IHC results for an additional 28 acquired scans will become available in the coming months. The observed elevation of GSH in the IDH1+ population was consistent with previous studies [9][10], and may reflect elevated oxidative stress in this type of LGG[11].