Local tumor control and neurological outcomes after surgery for spinal hemangioblastomas in sporadic and von Hippel-Lindau disease: A multicenter study

Abstract

Background: Spinal hemangioblastomas (sHBs) are rare vascular tumors with significant neurological implications. Their management, particularly in von Hippel-Lindau (VHL) disease, remains challenging due to recurrence and functional decline. Timely identification and intervention are critical for optimal outcomes.

Methods: This international, multicenter retrospective cohort study included 357 patients (199 VHL-associated, 158 sporadic) from 13 neuro-oncological centers. Clinical and imaging data were analyzed to assess progression-free survival (PFS) and functional outcomes using the modified McCormick Scale (mMCS) at 12 months. Secondary analyses identified factors associated with VHL disease in sHBs.

Results: Complete resection was achieved in 87.7% of cases, leading to significantly improved PFS at 72 months (sporadic: 95.1%, VHL-associated: 91.1%; hazard ratio: 0.18, 95% CI: 0.08-0.4). Multivariable analysis identified predictors of unfavorable outcomes at 12 months: preoperative mMCS ≥2 (odds ratio [OR]: 5.17, P = .008), intramedullary tumor location (OR: 9.48, P = .01), and preoperative bleeding (OR: 31.12, P = .02). Factors independently associated with VHL disease in sHBs included non-cervical tumor location (OR: 2.08, P = .004), intramedullary growth (OR: 2.39, P < .001), and age <43 years (OR: 3.24, P < .001). Functional improvements were observed in most patients, particularly those with sporadic sHBs.

Conclusions: Complete surgical resection is essential for long-term tumor control and favorable functional outcomes in both sporadic and VHL-associated sHBs. Early intervention, particularly in mild symptomatic and progressive cases, before neurological deterioration or hemorrhage, optimizes recovery. This study, the largest of its kind in a multicentric international setting, provides robust evidence to guide the management of both sporadic and VHL-associated sHBs.

Keywords: complete resection; multicenter study; neurological outcomes; progression-free survival; spinal hemangioblastomas; von Hippel–Lindau disease.

Graphical Abstract

Figure 1.

Schematic representation of the spinal column shows the distribution of spinal hemangioblastomas across different spinal levels. The dots represent individual tumors, categorized by their anatomical location: intramedullary (68.0%), extramedullary (16.0%), and intramedullary–extramedullary (16.0%).

Figure 2.

Multivariable Cox regression analysis and Kaplan–Meier survival curves for spinal hemangioblastomas. (A) Forest plot summarizing the multivariable Cox proportional hazards model for local progression-free survival (PFS). Incomplete resection was the only significant predictor of worse PFS with a hazard ratio (HR) of 6.2 (95% CI: 2.7–14.0, P < .001). (B) Kaplan–Meier curve demonstrating the impact of EoR on PFS in VHL-associated spinal hemangioblastomas. Patients undergoing CR had significantly better PFS compared to those with non-CR (P < .0001). Abbreviations: CR, complete resection; EoR, extent of resection; VHL, von Hippel–Lindau.

Figure 3.

Forest plot showing predictors of poor functional outcomes at 12 months post-surgery in patients with solitary spinal hemangioblastomas. Preoperative mMCS ≥2 (OR: 5.17, P = .008), preoperative bleeding (OR: 31.12, P = .02), and intramedullary tumor location (OR: 9.48, P = .01) were significant predictors, while other factors like EoR and tumor location approached significance. Abbreviations: EoR, extent of resection; mMCS, modified McCormick Scale; OR = odds ratio.

Figure 4.

Multivariable analysis and bubble plot of factors being correlated with von Hippel–Lindau (VHL) disease in spinal hemangioblastomas. (A) Multivariable analysis of pretherapeutic factors associated with VHL-associated spinal hemangioblastoma. The forest plot shows the adjusted odds ratios (ORs) for intramedullary component, non-cervical location, and age under 43 being correlated with the presence of VHL-associated spinal hemangioblastoma. The horizontal lines represent 95% confidence intervals (CIs). The table on the right summarizes the OR, 95% CI lower and upper limits, and the P-value for each factor. A red dashed line indicates the OR of 1 as a reference. (B) Bubble plot of spinal hemangioblastomas by age, spinal level, and intramedullary component among those with or without VHL. The plot categorizes patients based on the presence (green) or absence (orange) of VHL. Each bubble represents a spinal hemangioblastoma, with smaller bubbles indicating those with an intramedullary component. The y-axis indicates the spinal level (from cervical to lumbosacral), while the x-axis represents individual patients’ age.