Magnetic resonance elastography predicts tissue stiffness and surgical outcomes in patients with posterior fossa brain tumors

Introduction:: The ability to non-invasively characterize the biomechanical properties of intracranial tumors prior to surgical resection could revolutionize surgery by providing insight into tumor structure and improving surgical planning. Magnetic resonance elastography (MRE) is a non-invasive imaging technique that uses propagating strain waves to measure the tissue’s viscoelastic properties which are dependent on its biophysical composition. MRE is used to demonstrate that stiffness correlates with patient outcomes in solid tumors such as breast cancer¹. However, it has not been well characterized in vestibular schwannomas (VS), one of the most common intracranial neoplasms with an estimated lifetime prevalence of 1 per 500 persons². VS also represents the hallmark tumor of neurofibromatosis 2, a devastating tumor disposition syndrome which severely affects one’s life expectancy and results in patients being unable to hear, see or swallow. Microsurgical resection remains the only treatment to remove VS. However, treatment associated morbidities are high and outcomes are variable, which are in part due to variations in tumor consistency and degree of adhesion to the brainstem and adjacent cranial nerves. Recently, we and others have identified dysregulated extracellular matrix (ECM) remodeling as a key feature in adherent VS. We hypothesize that VS display variations in tumor stiffness that can be detected preoperatively on MRE. The purpose of this work is to 1) explore how VS tumor stiffness measured on MRE relates to intraoperative assessment of tumor consistency and postoperative clinical outcomes, and 2) understand the molecular mechanisms underlying ECM remodeling in VS with increased stiffness.

Materials and Methods:: In a prospective study of adult VS patients undergoing microsurgical resection, we analyzed the association between preoperative MRE stiffness with: a) intraoperative findings, b) postoperative outcomes such as the extent of tumor resection and facial nerve (FN) function, a key determinant of patient’s quality of life, and c) expression of molecular markers of fibrosis in tumor sections. All patients underwent MRE imaging within 48 hours of surgery using a Siemens Prisma 3T MRI scanner with 60 Hz vibrations induced by a Resoundant pneumatic activation system and pillow-like driver. This system allowed for spin echo-echo planar imaging sequences to be obtained (3.0 mm isotropic resolution; 16 slices) with all MRE images taken prior to the surgery. After acquisition of the three orthogonal motion-sensitizing direction waves, the sagittal images were analyzed by a blinded reviewer to obtain a local-frequency estimation (LFE) inversion to estimate tumor stiffness (Fig 1A). The operating surgeons were blinded to the MRE results. Assessment of tumor consistency and degree of peritumoral adhesions to brainstem and cranial nerves were recorded on a quantitative scale by two independent operators. The extent of tumor resection was determined based on postoperative MRI obtained within 3 months postoperatively. Portions of the tumor collected from surgery were formalin fixed and paraffin embedded for histological analysis. Clinical information queried included demographics, tumor size, surgical approach, and FN function. Immunohistochemical staining of markers associated with ECM remodeling and fibrotic deposition was performed, scanned, and quantified using Fiji. Statistical analysis was performed using SPSS.

Results, Conclusions, and Discussions:: Ten patients were enrolled (Mean age 45 years, 20% female). The average tumor diameter was 29.7 mm (Range, 15.0 mm-46.0 mm), and mean tumor volume was 43,600 mm³ (Range, 10,200 mm³–149,000 mm³). Three patients (30%) underwent subtotal resection and three patients (30%) developed facial paralysis (House-Brackmann grade III or higher). Tumor stiffness measured from the average LFE of all tumor-containing MRI slices on pre-operative MRE positively correlated with tumor consistency measurements obtained intraoperatively (r²=0.66, p=0.05), validating MRE as a non-invasive imaging modality to predict VS consistency preoperatively. In addition, tumors that were adherent to the brainstem were stiffer than those that were non-adherent and easily dissected away (Mean stiffness: 5.18 kPa vs. 3.57 kPa, p=0.13). Surgical outcomes were analyzed by grouping tumors into “stiff” and “non-stiff” (Above or below the whole tumor mean stiffness value of 4.02 kPa). VS with high stiffness were significantly larger in volume (70,000 mm³ versus 17,000 mm³, p=0.048) and had significantly worse FN function postoperatively (HB Grade III or higher, 75% vs. 0%, p=0.035) (Fig 1B). Moreover, tumors that underwent subtotal resection were stiffer compared to those that were amenable to gross total resection (5.15 kPa vs. 3.35 kPa, p=0.055) (Fig 1B). To identify tumor stiffness-related biomarkers, histopathological analysis revealed that activated fibroblast (alpha-SMA) (r² =0.63, p=0.029) (Fig 1C & D) and Masson’s trichrome stain intensity (r² =0.63, p=0.029) were both significantly correlated with stiffness, while hyaluronan (HABP) trended towards significance (r² =0.56, p=0.051). When evaluated together, there was a significant increase in ECM remodeling factors (sum of staining intensity for alpha-SMA, Masson’s Trichrome, and HABP) in stiffer VS (p=0.036).

MR elastography predicts tumor stiffness intraoperatively for VS, a common intracranial neoplasm. In tumors that exhibited higher stiffness preoperatively, surgical outcomes were poorer (more likely to undergo subtotal resection and have poor FN function) than tumors with lower stiffness. Histopathological analysis identified several molecular biomarkers, namely alpha-SMA, trichrome, and HABP, that correlated with tumor stiffness. Taken together, these results suggest that MRE may predict tissue composition and consistency in intracranial neoplasms such as VS, assist in presurgical planning and improve surgical outcomes.


Acknowledgements (Optional): :

References (Optional): : References: [1] Balleyguier C, Lakhdar AB, Dunant A, Mathieu MC, Delaloge S, Sinkus R. Value of whole breast magnetic resonance elastography added to MRI for lesion characterization. NMR Biomed. 2018;31(1):10.1002/nbm.3795. doi:10.1002/nbm.3795 [2] Marinelli JP, Beeler CJ, Carlson ML, Caye-Thomasen P, Spear SA, Erbele ID. Global Incidence of Sporadic Vestibular Schwannoma: A Systematic Review. Otolaryngol Head Neck Surg. 2022;16Ma7(2):209-214. doi:10.1177/01945998211042006.