The Clinical Implications of Spontaneous Hemorrhage in Vestibular Schwannomas

Background: Spontaneous hemorrhage into vestibular schwannomas (VS) is a rare phenomenon that can render more rapid symptom onset and a seemingly poorer prognosis for an otherwise benign pathology. We describe our series of hemorrhagic vestibular schwannomas (HVS) and systematically reviewed the literature to better understand relevant clinical factors and outcomes.

Methods: A retrospective case review series and systematic review of the literature using PRISMA guidelines was performed. Data collected included patient demographics, blood thinning medication use, pre- and postoperative symptoms, tumor size, and extent of resection (EOR). Statistical methods analyzing clinical outcomes with perioperative variables including Fisher’s exact test, student’s t-test, and linear regression modeling were performed. In addition, all of our tumor samples underwent whole-exome sequencing (WES), and we conducted a histopathologic comparison of our HVS patients to matched nonhemorrhagic VS (NHVS) patients.

Results: Fifty-three patients with HVS met inclusion criteria. Compared with historical data for all VS, patients with HVS had astonishingly higher rates of perioperative mortality (9.7%), significant preoperative facial weakness (58.5%; defined as House–Brackmann grade 3) and harbored relatively larger tumors (mean tumor size 3.1 cm ± 0.9). Regardless of EOR, among patients with HVS, surgery resulted in significant improvement of facial weakness (p = 0.041), facial numbness (p < 0.001), vertigo (p < 0.001), headache (p < 0.001), and hearing loss (p < 0.001). Upon further analysis, patients who presented with facial weakness, in particular, tended to have larger tumors (p = 0.058) on average and demonstrated significant improvement after surgery, irrespective of EOR (p < 0.01). The use of blood thinning medications (17% of patients) did not affect outcome. While WES was unrevealing as all tumors were NF2 mutated, histopathologic assessment of our cohort, when compared with a matched NHVS cohort, showed an increased number of dilated/ectatic thin-walled vascular channels, which tended to form focal clusters but were also distributed throughout the tumor. The vessels were immunoreactive for CD31. Taken together, these findings were reflective of potentially increased vascular permeability and hypervascularity.

Conclusion: Our findings identify HVS to be an aggressive subgroup of VS as they are associated with a surprisingly high mortality rate (9.4 vs. 1% for all VS). When the features of HVS are identified on imaging studies, these patients should be treated expeditiously, especially given the fact that the facial nerve dysfunction, which is identified in more than half of patients presenting with HVS, appears to be reversible. We postulate EOR does not seem to affect clinical outcomes in these patients, presumably because the blood clot may comprise the bulk of the relatively larger tumor and as such, evacuation of mass effect from the clot may be considered the primary goal of surgery in HVS patients who are particularly ill. Overall, this study has significant implications in the management of VS, raising awareness of a small, but highly morbid subgroup. Given our histopathologic findings, further studies evaluating the underlying mechanisms of spontaneous hemorrhage in VS are needed to better elucidate risk factors, improved prognostication, and optimal management of these rare tumors.