Abstract
BACKGROUND: Isocitrate dehydrogenase (IDH)-mutant gliomas demonstrate metabolic reprogramming and remain challenging to assess using conventional MRI alone. (1)H MR spectroscopic imaging (MRSI) of steady-state metabolism and real-time hyperpolarized carbon-13 (HP-(13)C) MRI of dynamic metabolism are of interest for evaluating response to treatment. This study aimed to improve characterization of IDH-mutant glioma through multimodal analysis of (1)H/HP-(13)C MRI and histopathology in patients before surgery for suspected tumor progression. METHOD: Six IDH-mutant glioma patients (1F/5M, 46.2±7.6 years; 5 astrocytoma [2 grade 2, 1 grade 3, 2 grade 4], 1 grade 3 oligodendroglioma) received multiparametric (1)H/HP-(13)C MR examinations on a 3-Tesla MR scanner before surgery. All patients had confirmed tumor progression; one grade 2 underwent malignant transformation to grade 3. During surgery, 21 tissue samples were collected with imaging coordinates and analyzed for histopathology. Imaging parameters were reprocessed to center on recorded biopsy locations to provide normalized apparent diffusion coefficients (nADC)/relative cerebral blood volume (nCBV)/relative cerebral blood flow (nCBF)/peak height (nPH), %recovery, choline-to-N-acetylaspartate index (CNI), pyruvate-to-lactate conversion rate (k(PL)), ratios of lactate-to-pyruvate (Lac/Pyr) and bicarbonate-to-pyruvate (Bic/Pyr). Tissue samples were assessed for treatment effects, blood vessels (BV), carbonic anhydrase 9 (CA9), and Ki-67 expressions. Due to limited sample sizes, summary statistics and trends were assessed. RESULT: 3 tissue samples with pure treatment effects showed higher Bic/Pyr ratio (median [min max]: 0.05 [0.04, 0.15] vs. 0.03 [0.005, 0.31]) and lower CNI (0.20 [-0.24, 0.64] vs. 8.01 [0.58, 14.88]) compared to 15 recurrent tumor tissue samples. In recurrent tumor tissue samples, higher tumor proliferation (Ki67) was associated with higher k(PL), CNI, nPH, nCBV, nCBF, and nADC, while no clear association was observed between vascularity (BV) or hypoxia (CA9) indexes and imaging parameters. CONCLUSION: This study showcases our initial experience in using multiparametric (1)H/HP-(13)C MR imaging to characterize recurrent IDH-mutant gliomas and identify noninvasive imaging markers corresponding with tissue histopathology.