Tumor biomechanical stiffness by magnetic resonance elastography predicts surgical outcomes and identifies biomarkers in vestibular schwannoma and meningioma

Abstract

Variations in the biomechanical stiffness of brain tumors can not only influence the difficulty of surgical resection but also impact postoperative outcomes. In a prospective, single-blinded study, we utilize pre-operative magnetic resonance elastography (MRE) to predict the stiffness of intracranial tumors intraoperatively and assess the impact of increased tumor stiffness on clinical outcomes following microsurgical resection of vestibular schwannomas (VS) and meningiomas. MRE measurements significantly correlated with intraoperative tumor stiffness and baseline hearing status of VS patients. Additionally, MRE stiffness was elevated in patients that underwent sub-total tumor resection compared to gross total resection and those with worse postoperative facial nerve function. Furthermore, we identify tumor microenvironment biomarkers of increased stiffness, including αSMA + myogenic fibroblasts, CD163 + macrophages, and HABP (hyaluronic acid binding protein). In a human VS cell line, a dose-dependent upregulation of HAS1-3, enzymes responsible for hyaluronan synthesis, was observed following stimulation with TNFα, a proinflammatory cytokine present in VS. Taken together, MRE is an accurate, non-invasive predictor of tumor stiffness in VS and meningiomas. VS with increased stiffness portends worse preoperative hearing and poorer postoperative outcomes. Moreover, inflammation-mediated hyaluronan deposition may lead to increased stiffness.

Keywords: MRI; Magnetic resonance elastography; Meningioma; Stiffness; Vestibular schwannoma.

Figure 1

Magnetic resonance elastography predicts stiffness in posterior cranial fossa tumors (A) Schematic of patient workflow. Prospectively enrolled patients diagnosed with either vestibular schwannomas or meningiomas scheduled for surgical resection underwent MRE 24–72 h prior to surgery. During surgery, the intraoperative experience and tumor consistency was recorded independently by the surgical team blinded to MRE results. Following surgery, tumors were collected and analyzed. (B) Representative T1-MRI (top row) and elastograms (bottom row) of a stiff (left column) and non-stiff (right column) vestibular schwannoma. Arrowhead denotes tumor; arrow denotes brainstem; * denotes cerebellum. An elastogram color scale bar is shown. (C) Preoperative DI MRE stiffness across the entire patient cohort. The black line represents the mean stiffness at 3.05 kPa. (D). MRE stiffness does not significantly correlate with tumor volume, p = 0.269 by analysis of the Pearson correlation coefficient. (E). MRE stiffness did not correlate with patient age, p = 0.388 by analysis of the Pearson correlation coefficient. (F) Preoperative MRE stiffness correlates with intraoperative survey measurements of tumor consistency determined by ultrasonic aspirator settings (higher settings indicate greater force), p = 0.020 by analysis of the Pearson correlation coefficient.

Figure 2

Posterior fossa tumor stiffness is elevated in cases with worse preoperative hearing and poorer postoperative outcomes. (A) Preoperative MRE DI stiffness is not significantly different in patients with worse preoperative hearing based on bone-conduction pure-tone average and word recognition scoring. Audiometric data is classified into grades (A–D) based on the 1995 AAO-HNS grading system, with grade D being the worst. (B) Representative T-1 MRI (top row), elastograms (second row), and audiograms (third row) for a patient with poor hearing and a stiff tumor (left column, 3.3 kPa) and a patient with good hearing with a soft tumor of similar size (right column, 2.4 kPa). Tumor volumes were similar (68 cm³ for the stiff tumor and 61 cm³ for the non-stiff tumor). Pure tone average (PTA) and word recognition scores in quiet (WRS) are noted below the respective audiograms. Arrowhead denotes tumor. Color bar denotes stiffness from 0 to 6 kPa. (C) In vestibular schwannomas, MRE stiffness is positively correlated with baseline bone-conduction PTA (db HL). (D) MRE stiffness is significantly elevated in VS that required a sub-total resection compared to those amenable to a gross or near-total resection. ** p < 0.01 by Mann-Whitney U-test. (E) MRE stiffness was higher in patients with lower cranial nerve dysfunction. * p < 0.05 by Mann-Whitney U-test. (F) MRE stiffness was elevated in patients with poor long-term facial nerve function (HB > III) compared to those with good facial nerve function (HB I-III). * p < 0.05 by Mann-Whitney U-test.

Figure 3

Histopathological features correlate with MRE predictive stiffness. Representative IHC images of stiff (left column) and non-stiff (right column) vestibular schwannoma tumors stained for hyaluronan (HABP)(A), Masson’s Trichrome (B), M2 Macrophages (CD163) (C), and activated fibroblasts (α-SMA) (D). Scale bar represents 100 µm. Positive staining were denoted by arrowheads. (E–H) Correlation between MRE DI stiffness and staining intensities of hyaluronan (HABP, (E), Masson’s Trichrome (F), M2 macrophages (CD163, (G), and activated fibroblasts (α-SMA, (H). All p < 0.05 by analysis of Pearson correlation coefficient. (I) Representative elastograms of a stiff VS tumor (left column, 4.9 kPa) with high HABP staining intensity versus a soft VS tumor (right column, 2.5 kPa) with low HABP staining intensity. The red line and white arrowhead detail the outline of the VS tumor determined by T-1 MRI.

Figure 4

Hyaluronan synthase activity is upregulated in vestibular schwannomas. (A) TNFα mRNA expression is upregulated in NF2 -/- schwannomatosis cells (HEI-193) versus wild-type human Schwann cells (HSC). *p < 0.05 by Mann Whitney U-test. (B) TNF-α is also actively secreted by human primary vestibular schwannoma cells (black dots) at greater levels than Schwann cell controls (red dot). (C) Representative images of HAS1 (top), HAS2 (middle), and HAS3 (bottom) expression in HEI-193 cells (left column) and in archived, formalin-fixed paraffin-embedded vestibular schwannomas (right column). Red, HAS enzymes; Blue, DAPI. The scale bar represents 50 µm (left column) and 100 µm (right column). (D) Relative expression of HAS1-3 mRNA normalized to GAPDH following TNF-alpha stimulation in HEI-193 cells. Data from three independent experiments are shown as a mean plus standard deviation. *p < 0.05, n.s., not significant, by one-way ANOVA with post-hoc Bonferroni correction.