*2.1. Patient Information*

All procedures and protocols for this study were reviewed and approved by the Colorado Multi-Institutional Review Board (COMIRB 17-1136). Subjects included in this study were patients undergoing repeat resective surgery after radiologically defined tumor progression between August and November 2018 at the University of Colorado Hospital. The patient set consisted of 10 subjects who received prior resection(s) for recurrent glioma with detailed histopathology recorded for 2 or more biopsies (41 biopsies collected in total). Data were collected retrospectively from patient chart review. Two patients received two prior resections; all others received one prior resection. The patient set is divided into two groups: those that underwent radiation therapy prior to repeat resection (RT, *n* = 7) and those that did not (No RT, *n* = 3). For each patient, biopsy samples were collected during surgery from the radiologically-defined tumor region and examined by an expert neuropathologist (B.K.D.). The neuropathologist classified each sample and an expert neurosurgeon (D.R.O.) designated each classification as primarily consisting of abnormal, nontumor tissue (Abnormal), or tumor tissue (Tumor). Patient information is summarized in Table 1.


Abbreviations: M = male, F = female, MT = mutant, WT = wild type, NA = not available, lo = low expression, hi = high expression, + = methylated, − = unmethylated, IDH = isocitrate dehydrogenase, MGMT = O-6-methylguanine-DNA-methyltransferase, EGFR = epidermal growth factor receptor, BRAF = v-Raf murine sarcoma viral oncogene homolog B, RT = radiation therapy, CT = chemotherapy.

#### *2.2. Imaging Sequence Parameters*

All images were obtained using a 3.0-T whole-body MR imager (Signa HDxt; GE Medical Systems, Milwaukee, Wisconsin, USA) between 0–24 days prior to repeat surgical intervention. Acquisition times were 2.5, 5.4, 4.6, 7.8, and 9.0 minutes for non-enhanced T1-weighted (T1w), gadolinium-enhanced T1-weighted (T1ce), T2-weighted (T2w), T2-FLAIR (FLAIR), and diffusion-weighted (DW) images, respectively. For T1w, TE = 2.3 ms, TR = 5.5 ms, and flip angle = 8◦. Data were recorded as a 256 × 256 matrix with 1 mm × 1 mm pixel spacing, a slice thickness of 1.2 mm, and zero slice gap. For T1ce, TE = 2.5 ms, TR = 6.8 ms, and flip angle = 8◦. Data were recorded as a 512 × 512 matrix with 0.5 mm × 0.5 mm pixel spacing, a slice thickness of 1.2 mm, and zero slice gap. For T2w, TE = 6333 ms,

TR = 80 ms, and flip angle = 142◦. Data were recorded as a 512 × 512 matrix with 0.5 mm × 0.5 mm pixel spacing, a slice thickness of 2 mm, and zero slice gap. For FLAIR, TE = 6000 ms, TR = 128 ms, and flip angle = 90◦. Data were recorded as a 512 × 512 matrix with 0.5 mm × 0.5 mm pixel spacing, a slice thickness of 1.2 mm, and zero slice gap. For DW images, TE = 85 ms, TR = 16,000 ms, and flip angle = 90◦. The diffusion gradient was encoded in 32 directions at b = 1000 s/mm2 and an additional measurement without the diffusion gradient (b = 0 s/mm2). DW data were recorded as a 128 <sup>×</sup> 128 matrix with 0.9375 mm × 0.9375 mm pixel spacing. A total of 50 sections were obtained with a slice thickness of 2.6 mm and zero slice gap.
