Findings From the Natural History of Central Nervous System Hemangioblastomas in von Hippel-Lindau Disease

The tumor suppressor von Hippel-Lindau (VHL) gene on chromosome 3 was identified by germline mutations in families afflicted by this disorder.1 VHL, transmitted in an autosomal-dominant manner, can result in the development of benign and malignant tumors throughout the viscera and nervous system.2 However, the most common manifestations of VHL are benign hemangioblastomas found along the craniospinal axis.3,4 Previous studies have estimated that 60% to 90% of VHL patients will develop multiple hemangioblastomas.5-12 Although hemangioblastomas are benign tumors, they are associated with significant morbidity and mortality based on their location, associated features, and multiplicity.13 We prospectively studied the natural history of central nervous system (CNS) hemangioblastomas14 to determine factors affecting tumor burden, development, and growth to better direct management for VHL patients with these tumors. The results are summarized below. STUDY FINDINGS Patient Characteristics Two hundred fifty VHL patients were enrolled in the study. At the study entrance, the mean age of the participants was approximately 39 years (range, 12.3-66.1 years). For many of the analyses, it was determined that at least 2 years of follow-up was best to power the statistical tests. Consequently, patients/tumors with >2 years of follow-up (225 patients, 90%) were used for analysis. Mean follow-up in these patients was approximately 7 years. Over the course of the study, 16 patients died (mean age, 42 years; range, 17.7-56.3 years). High-resolution imaging, craniospinal imaging, fluid-attenuated inversion recovery imaging, and T1- and T2-weighted magnetic resonance imaging (with and without contrast) were obtained at 6-month visits to assess hemangioblastomas. Hemangioblastoma Features At study entrance, participants harbored 1921 craniospinal hemangioblastomas (median, 7 tumors per patient; range, 1-33 tumors). At the last study visit, patients harbored 2336 craniospinal hemangioblastomas (median, 8 tumors per patient). The majority of the hemangioblastomas were in the cerebellum and spinal cord. However, they also arose in the cauda equina, brainstem, supratentorial compartment, and nerve roots, in descending order (Table). Over the course of the study, approximately 75% of patients developed new CNS hemangioblastomas (mean tumor development, 0.4 ± 0.4 new tumors per year per patient).TABLE: von Hippel-Lindau Disease--Associated LesionsHemangioblastomas have been classified primarily into 3 growth pattern categories: saltatory (exhibiting periods of growth and quiescence; 72% of growing tumors), linear (6%), or exponential (22%) (Figure 1). Hemangioblastomas that did not progress in size were classified as stable. Fifty-one percent of hemangioblastomas remained stable, and 49% grew during the study. Twenty-five percent of hemangioblastomas that were originally defined as stable eventually grew in 1.5 to 5 years, and if hemangioblastoma follow-up was extended to >5 years, 50% of hemangioblastomas grew (Figure 2). Tumor growth rate for a saltatory pattern was approximately 4 mm3/y; for the linear pattern, approximately 24 mm3/y; and for the exponential pattern, approximately 79 mm3/y (P < .001).FIGURE 1: Example of the different growth patterns of central nervous system hemangioblastomas in a patient with von Hippel-Lindau disease enrolled in the study. Left, axial contrast-enhanced magnetic resonance imaging of the cerebellum demonstrating 5 tumors (A-E) that were observed with 3 patterns of growth. Right, graphically described tumor growth patterns: saltatory (A-C; o, +, •), linear (D; X), and stable (E; Δ). Figure reproduction courtesy of the Journal of Neurosurgery.14FIGURE 2: Progression analysis of hemangioblastomas in patients with von Hippel-Lindau disease over an 8-year period. Fifty percent of (nongrowing) hemangioblastomas located in the brainstem progressed in 5.6 years. Twenty-five percent of hemangioblastomas located in the cauda equina progressed in 5.0 years. Fifty percent of hemangioblastomas located in the cerebellum progressed in 6.2 years. Fifty percent of hemangioblastomas located in the spine progressed in 8.9 years. Shading represents 95% confidence interval. Figure reproduction courtesy of the Journal of Neurosurgery.14Hemangioblastomas grew at different rates, depending on their anatomic location. The hemangioblastomas that arose in the brainstem (median, 3.1 mm3/y) and cerebellum (median, 1.2 mm3/y) grew significantly faster than those arising in the spinal cord (median, 0.3 mm3/y) or cauda equina (median, 0 mm3/y; P < .001). Hemangioblastoma can form cysts (peritumoral or intratumoral). In this study, cysts were associated with 295 hemangioblastomas (11.8% of tumors). The majority were peritumoral cysts (247 cysts, 9.9%). However, intratumoral cysts were also observed (37 cysts, 1.5%), and in some cases, hemangioblastomas developed both types of cysts (11 cysts, 0.4%). Many of the tumors with associated cysts became symptomatic. Therefore, it was interesting to observe that the ratio of the median growth rate of peritumoral cysts to the tumor growth rate for symptomatic cerebellar hemangioblastomas was 22:1 and the ratio of the median volume of the peritumoral cyst compared with tumor was 10:1. Although the overwhelming majority of patients with hemangioblastomas at the final evaluation (94%) were asymptomatic, 75 patients with 159 hemangioblastomas (6.4%) were symptomatic and required treatment. Approximately 10% of the symptomatic hemangioblastomas were not present on initial imaging at study entrance. Symptoms and signs correlated to mass effect related to hemangioblastoma and/or cyst size at their specific anatomic location. Clinical Progression The participants in the study underwent detailed neurological examinations during their visits, including assessment of Karnofsky Performance Scale scores. The majority (60%) of the patients’ Karnofsky Performance Scale scores remained unchanged; they improved in approximately 15% of patients, all of whom had surgical resection of symptom-producing hemangioblastomas; and scores worsened in 25%. Decline in the Karnofsky Performance Scale score during the study was accounted for by disease progression, postoperative morbidity, or unknown origin. Analysis of Factors Associated With Tumor Behavior Increased tumor burden or total tumors detected was associated with male sex (P < .01), longer follow-up (P < .01), and genotype (P < .01). Specifically, patients harboring partial germline deletions had more tumors per patient than patients with missense mutations (P < .01). Patients who enrolled with greater tumor burden developed more tumors over the course of the study (P < .01). The age of patient had a significant (P < .01) effect on tumor development. For example, patients 12 to 20 years of age developed more tumors per year than all other older age groups. Hemangioblastoma growth rate was increased significantly more in male than female patients (P < .01). Not surprisingly, the symptom-producing tumors grew faster than asymptomatic tumors (P < .01). Unexpectedly, the location of the hemangioblastoma influenced growth rate; symptom-producing tumors in the spine grew significantly more slowly than those in the cerebellum (P < .01). Likewise, asymptomatic tumors in the spine and cauda equina grew significantly (P < .01) more slowly than those in the brainstem and cerebellum. Regardless of the anatomic compartment, tumors associated with a cystic component grew faster than tumors without cysts (P < .01). In predictive analysis, cerebellar hemangioblastomas with a diameter of ≥5.6 mm, brainstem hemangioblastomas with a diameter of ≥4.0 mm, and spinal cord hemangioblastomas with a diameter of ≥4.5 mm would clinically result in symptomatic lesions with a sensitivity of 78% (specificity, 96%) for cerebellum, 53% (specificity, 92%) for brainstem, and 78% (specificity, 95%) for spine. DISCUSSION The natural history of VHL-associated hemangioblastomas has been inconclusive in previous studies because of their retrospective analyses, inadequate patient numbers, and limited follow-up.6,15-17 Therefore, timing for removal of VHL-associated CNS hemangioblastomas and the role of other radiotherapy and any chemotherapy regimen remain unclear. To fully understand the natural history of hemangioblastomas before intervention, we performed a long-term, prospective, natural history study in a large cohort of patients. This study indicated new characteristics that may aid in predicting tumor burden and development such as male sex, younger age, and germline mutation. These findings indicate that CNS hemangioblastoma development in VHL may be affected by biological features that may be related to developmental processes, hormonal factors, other systemic factors, and/or proteasomal processing. The differences in germline mutation are supported by a recently published study demonstrating that VHL protein containing a missense mutation can maintain intrinsic function.18,19 Additionally, missense mutation proteins specifically undergo degradation via proteasome control pathways.19 Studies have demonstrated an age-related decline in proteasome function,20 which may account for the differences we observed in hemangioblastoma development in terms of age. Taken together, these findings suggest that residual function of missense mutation VHL protein may underlie reduced disease severity. Lastly, the most common pattern of growth was saltatory (72% of growing tumors), followed by exponential (22%) and linear (6%). Tumors may remain in prolonged periods of quiescence (as seen by progression data of ≥5 years) but, after extended periods of follow-up, may grow. Thus, it is necessary in any future studies of therapy to include long follow-up times to define therapeutic success as tumor stability. Disclosure The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.