Endoscopic Endonasal Approach for Craniopharyngiomas with Intraventricular Extension: Long-Term Outcomes and Meta-analysis

Background: Traditionally, skull base tumors with intraventricular extension are approached via open craniotomy or transventricular approaches. However, craniopharyngiomas are increasingly resected via an endoscopic endonasal approach (EEA); yet those with intraventricular extension are of increased complexity and EEA with this degree of extension is not well studied.

Methods: Patients undergoing EEA for resection of craniopharyngioma with third ventricular involvement between 2002 and 2015 were retrospectively reviewed. Tumor characteristics and outcomes were reviewed and compared with previously published EEA and transcranial approach (TCA) studies for all craniopharyngioma locations. Comparison of these groups was done via meta-analysis and a random effects model.

Results: Sixty-two patients were included. Average tumor volume was 13.93 cm³ with an average 2.61 cm³ (17.3%) intraventricular volume. Patients presented with visual impairment (75.8%), panhypopituitarism (29.0%), headache (16.1%), and diabetes insipidus (16.1%). Gross total resection (GTR) was achieved in 47% of all cases (56% of cases where GTR was the goal); after 2012 the overall GTR rate increased to 77%. Of those presenting with visual impairment, 97.9% experienced improvement or stability of vision. The rates of postoperative cerebrospinal fluid (CSF) leak and meningitis was 19 and 8.1%, respectively. When nasoseptal flap reconstruction (since 2006) was used, the CSF leak rate dropped to 10%. Two patients had tumor bed/intraventricular hemorrhage requiring reoperation and one patient had an epidural hematoma requiring evacuation. There were no postoperative mortalities. Six (9.6%) patients required shunting prior to tumor resection. Overall, 25% underwent postoperative shunt placement with 7 out of 10 patients (70%) treated prior to introduction and routine use of the nasoseptal flap and only 7 of 46 (15%) required a shunt following routine use of nasoseptal flap reconstruction. Meta-analysis demonstrated overall similarity in presenting features and outcomes between the present cohort for intraventricular lesions, EEA, and TCA for all craniopharyngiomas. TCA demonstrated a higher GTR rate, however, with heterogeneity in the model (sizable study variation). EEA for craniopharyngioma with intraventricular extension showed improved visual outcomes compared with TCA but also increased incidence of CSF leaks without concomitant increase in rates of meningitis (again with heterogeneity in the random effects model).

Conclusion: EEA for craniopharyngiomas with intraventricular extension shows similar outcomes to those of TCA and EEA for all craniopharyngiomas and expands one of the potential anatomic limits of this approach. Given involvement of the ventricle, CSF leak/shunt rates are expectedly high. GTR rates increased and CSF leak/shunt rates decreased dramatically with time and the introduction of the nasoseptal flap. These outcomes reflect the importance of an experienced team and demonstrate a clear learning curve.