Evidence of polyclonality in neurofibromatosis type 2–associated multilobulated vestibular schwannomas

Abstract

Background

Neurofibromatosis type 2 (NF2) is a tumor syndrome that results from mutation of the NF2 tumor suppressor gene. The hallmark of NF2 is the presence of bilateral vestibular schwannoma (VS). Though NF2-associated and sporadic VS share identical histopathologic findings and cytogenetic alterations, NF2-associated VS often appears multilobulated, is less responsive to radiosurgery, and has worse surgical outcomes. Temporal bone autopsy specimens and MRI of the inner ear performed on NF2 patients suggest that multiple discrete tumors may be present within the labyrinth and cerebellopontine angle.

Methods

Treatment-naïve ears in patients enrolled in a prospective NF2 natural history study (NIH#08-N-0044) were included for MRI analysis. T2-weighted and postcontrast T1-weighted MRIs were evaluated for the presence of multiple discrete tumors or a multilobulated mass. Peripheral blood (germline) and regional samples of tumor tissue were procured from consecutive patients enrolled in this study undergoing resection of a multilobulated VS (MVS). Histopathologic evaluation and genetic analysis (single nucleotide polymorphism array analysis, NF2 sequencing) were performed on each specimen.

Results

Over half of NF2 ears harbored either an MVS (60/139 ears) or multiple discrete masses (19/139 ears). For 4 successive MVSs, genetic analysis revealed an admixture of cell populations, each with its own somatic NF2 mutation or deletion.

Conclusions

These findings suggest that the majority of NF2-associated VSs are polyclonal, such that the tumor mass represents a collision of multiple, distinct tumor clones. This explains the characteristic lobulated gross appearance of NF2-associated VS, and may also explain the substantially different treatment outcomes compared with sporadic VS.

Neurofibromatosis type 2 (NF2) is an autosomal dominant tumor predisposition disorder. It is caused by a germline mutation of the NF2 tumor suppressor gene located on chromosome 22q.^1,2 The hallmark of NF2 is the development of bilateral vestibular schwannoma (VS), which is present in 90%–95% of patients. Subsequently, it has been the cardinal focus of research study in NF2, but little is definitively understood about why these tumors differ in gross appearance,^3,4 natural history,⁵ and surgical outcomes⁶ compared with sporadic VS, despite sharing a common pathobiologic etiology⁷ (NF2 tumor suppressor gene inactivation)⁸ and histology.

NF2-associated VSs are known to have worse surgical outcomes, including poorer hearing and facial nerve preservation rates⁹ compared with matched (size) sporadic counterparts. Adverse outcomes in NF2-associated VS resections have been attributed to their propensity to infiltrate the adjacent cochlear and facial nerves rather than displace these structures, as observed with sporadic tumors that respect this histologic cleavage plane.¹⁰ Furthermore, synchronous facial neuromas may be identified in at least 20% of cases and may contribute to surgical morbidity when not distinguished at operation.¹¹ A significant disparity in recurrence rates also exists. After complete resection of NF2-associated VS <1.5 cm in size, the long-term recurrence rate is >50%.¹² Comparable studies performed in patients with sporadic VS show essentially no recurrence.¹⁰ Finally, there is a significant difference in long-term control rate with radiosurgery of sporadic VS (over 95%) compared with NF2-associated VS (∼50%).^13,14

In addition to these notable differences in response to treatment, the posterior fossa component of NF2-associated VS is frequently multilobulated (botryoid) and appears phenotypically different from sporadic VS, which is typically ovoid in shape and has smooth contours.^3,15 Based on these observations, we hypothesized that the multilobulated VS (MVS) in NF2 arise from numerous, independent, somatic Knudson “second hits”' in NF2. We propose that these multiple, different clonal tumors merge over time in the cerebellopontine angle (CPA) to form a larger, polyclonal, and multilobulated mass (Fig. 1). To investigate this hypothesis, we evaluated the MRIs in a large cohort of untreated NF2-associated VS. To investigate the putative polyclonal origin of NF2-associated MVS, we sequenced NF2 and determined copy-number variation (CNV) in multiple regional samples from 4 tumors resected from 4 patients.

Fig. 1.

Hypothesis: MVSs in NF2 patients are polyclonal derivatives. Illustration depicting the hypothesis that MVSs in NF2 patients are derived from multiple discrete tumors (top panel) that have collided (middle panel). The premise underlying this investigation is that different regional samples obtained from an MVS should reveal multiple somatic second hits at the NF2 locus. VII, facial nerve; VIII, cochlear nerve; SVN, superior vestibular nerve; IVN, inferior vestibular nerve; GM, germline NF2 mutation; SM, somatic NF2 mutation.

Materials and Methods

Patients

Patients enrolled in a National Institute of Neurological Disorders and Stroke Institutional Review Board–approved prospective NF2 natural history study (NCT00598351) were included. MRI analysis was assessed in ears naïve to treatment with radiosurgery, chemotherapy, and/or previous operation. Tumor tissue and germline DNA (peripheral blood) were procured from consecutive patients enrolled in this study undergoing resection of an MVS. Informed written consent was obtained from all participants. All patients were diagnosed with NF2 by Manchester clinical criteria¹⁶ and/or were identified with a causative germline mutation in NF2.

Imaging

Patients underwent MRI on a 3T MR scanner (Phillips) at the time of entry into NCT00598351. MRI included 3D T2-VISTA (volume isotropic turbo spin echo acquisition) (which identifies a VS by hypointense “filling defect” surrounded by hyperintense CSF) and contrast-enhanced 3D T1 fast-field echo T1-weighted MRI (which identifies VS as a high signal due to the leak of contrast across the blood–tumor barrier). Images were inspected for the presence of multiple discrete tumors or for the presence of an MVS, identified by the presence of clefts within the outer contour (Fig. 2).

Fig. 2.

MRI of the CPA in patients with NF2. (A) T2-VISTA MRI reveals that multiple discrete tumors are identified in the CPA (arrows). (B) Postcontrast T1-weighted MRI from a different patient demonstrates bilateral MVS with significant intratumoral clefts.

Surgical Procedure and Tumor Procurement

Patients underwent resection of tumor via conventional surgical approaches, with an emphasis on tissue procurement. Samples (4 to 8 specimens per MVS) were taken from varying locations within the MVS and coded sequentially starting with MVS1. In each patient, a distinct facial nerve or facial neuroma was identified and excluded from the submitted specimen for tumor analysis. A portion of each specimen was kept for histopathologic confirmation. The remainder of each tumor specimen was snap frozen in isobutane, embedded in optimal cutting temperature compound (PSL Equipment), and stored at −80°C.

Blood Samples

Blood samples were processed within 2 h of collection. Patient blood was collected directly in a Vacutainer Cell Preparation Tube with sodium citrate (Becton Dickinson) and fractionated through centrifugation (30 min at 1500 g). Mononuclear cells and platelets were collected as a pellet, washed thrice with phosphate buffered saline (15 mL, 10 mL, 400 μL), and stored at −80°C.

Histopathology Analysis

Tumor specimens were fixed in 10% buffered formalin immediately after removal, processed overnight, and subsequently embedded in paraffin. Five-micrometer-thick sections were obtained from the paraffin blocks and subsequently stained with the standard hematoxylin and eosin method.

DNA Extraction

Frozen tumor tissue was minced, washed once in phosphate buffered saline, pH 7.4, and incubated in solution containing 100 mM TrisHCl, pH 8.0, 5 mM EDTA, 0.5% sodium dodecyl sulfate, and 200 μg/mL Proteinase K (Invitrogen) at 55°C for 2–3 h or at 37°C overnight. DNA was extracted by the phenol:chloroform procedure and precipitated by adding an equal volume of ice-cold isopropanol. DNA pellets were air dried, resuspended in 10 mM TrisHCl, pH 7.4, and 0.1 mM EDTA, divided into aliquots, and stored at –20°C.

Single Nucleotide Polymorphism Array Analysis

Genotyping by single nucleotide polymorphism was performed using the HumanOmniExpress-12v1.1 Illumina BeadChip arrays per manufacturer's instructions. The arrays were read using the iScan platform (Illumina) and visualized with GenomeStudio 2011.1 software. The call rate for all the DNA samples was >99%. Genomic coordinates are per hg19.

Sanger Sequencing

Sanger sequencing of NF2 was provided by Prevention Genetics, a DNA testing lab certified by the Clinical Laboratory Improvement Amendments. Briefly, PCR was used to amplify all NF2 coding exons as well as a flanking sequence of ∼20 bp. Sequencing was performed separately in both the forward and reverse directions, and all differences from the reference sequence were reported.¹⁷

Copy-Number Variation Analysis

Analysis of the data by allele-specific copy number analysis of tumors (ASCAT v2.1) was performed as previously described.¹⁸ CNV analysis of the tumors was performed using Nexus Copy Number software v.6.1 (BioDiscovery). The analysis settings were chosen based on the manufacturer's suggestions for the analysis of tumor samples. “Allelic imbalance” refers to a locus with a B-allele frequency class other than 0, 0.5, or 1.

Results

Clinicopathologic Findings

As of June 2013, 146 patients had enrolled in NCT00598351. Of 287 tumor/ears in these patients, 139 ears in 89 patients were untreated. Imaging evaluation revealed 79/139 (56.8%) ears harbored either an MVS (63 ears, 79.7%) or the presence of multiple discrete masses within the CPA (16 ears, 20.3%) (Fig. 2).

Four consecutive patients with treatment-naïve MVS underwent surgical resection. At operation, the botryoidal configuration of these masses was noted and specimens were obtained from distinct regions. The histopathology was consistent with a typical schwannoma in each specimen (Fig. 3).

Fig. 3.

Representative histopathology of resected MVS (20×). (A) Nuclear palisades (Verocay body formations) were identified in the hypercellular regions. The tumors were cellular with minimal intervening stromal connective tissue. (B) Pleomorphic schwann cells with elongated, spindled nuclei and scant cytoplasm were arranged in an admixture of hypercellular regions (Antoni A, asterisk) and hypocellular regions (Antoni B, arrow). The scale bar shown represents 1 mm.

Somatic Neurofibromatosis 2 Mutation Analysis

To investigate clonality in NF2-associated MVS, we determined the mechanism of biallelic inactivation of NF2 (discrete mutation or loss of heterozygosity [LOH] of chromosome 22) in 23 tumor samples from 4 VSs from 4 patients (Table 1). In all 4 patients, the NF2 germline mutation (patients 1 and 2) or partial deletion (patients 3 and 4) was detected in all samples from the corresponding tumor. In patient 1, two unique somatic pathogenic NF2 mutations were identified in 2 anatomically distinct samples (MVS2: splice site mutation; MVS5: frameshift mutation) from the same right MVS (Supplementary Fig. S1a). Although no NF2 mutation was found in MVS3 (anatomically separate from MVS2, MVS4, and MVS5), there was a deletion of the entire chromosome 22q arm, affecting a minority (27%) of tumor cells (Supplementary Fig. S1b). No NF2 somatic mutations or LOH events were detected in MVS4. The NF2 somatic mutations and 22q deletion in patient 1 were mutually exclusive and not detected in the same sample, suggesting that they each arose from an individual clone.

Table 1.

Mechanism of biallelic inactivation of NF2 in multiple samples from distinct regions of NF2-associated VS

Patient No.	Sample	Sample Location	NF2 Germline Mutation	NF2 Somatic Mutation	Chr22 LOH Analysis	Figure
1	GL1	Germline	Exon 13: c.1341–1G>A (splice site)	(Germline sample)	No LOH	Supporting Figure S1
	MVS2	Right VS	Exon 13: c.1341–1G>A (splice site)	Exon 12: c.1340+1 G>T (splice site)	No LOH	Supporting Figure S1
	MVS3	Right VS	Exon 13: c.1341–1G>A (splice site)	No somatic variant detected	Deletion of 22q in 27% of tumor cells	Supporting Figures S1 and S2
	MVS4	Right VS	Exon 13: c.1341–1G>A (splice site)	No somatic variant detected	No LOH	Supporting Figure S1
	MVS5	Right VS	Exon 13: c.1341–1G>A (splice site)	Exon 4: c.440delA* (p.Gln147ArgfsStop27)	No LOH	Supporting Figure S1
2	GL2	Germline	Exon 8: c.784C>T (p.Arg262Stop)	(Germline sample)	No LOH	Supporting Figure S3
	MVS7	Left VS	Exon 8: c.784C>T (p.Arg262Stop)	Exon 6: c.599+1 G>A (splice site)	No LOH	Supporting Figure S3
	MVS8	Left VS	Exon 8: c.784C>T (p.Arg262Stop)	Exon 14: c.1574+3_1574+11delATGTAGCCC* (splice site)	No LOH	Supporting Figure S3
	MVS16	Left VS	Exon 8: c.784C>T (p.Arg262Stop)	No somatic variant detected	MR of 22q in 56% of tumor cells	Supporting Figures S3 and S4
	MVS23	Left VS	Exon 8: c.784C>T (p.Arg262Stop)	Exon 2: c.115-1_115delGA* (splice site)	No LOH	Supporting Figure S3
	MVS24	Left VS	Exon 8: c.784C>T (p.Arg262Stop)	No somatic variant detected	MR of 22q in 83% of tumor cells	Supporting Figures S3 and S5
	MVS30	Left VS	Exon 8: c.784C>T (p.Arg262Stop)	Exon 7: c.675+5 G>A (splice site), same as MVS31	No LOH	Supporting Figure S3
	MVS31	Left VS	Exon 8: c.784C>T (p.Arg262Stop)	Exon 7: c.675+5 G>A (splice site), same as MVS30	No LOH	Supporting Figure S3
	MVS32	Left VS	Exon 8: c.784C>T (p.Arg262Stop)	Exon 7: c.651 C>A (p.Tyr217Stop); chr14: LOH	No LOH	Supporting Figure S3
3	GL3	Germline	Deletion chr22:30,051,433–30,216,982 (includes NF2 exons 6–17)	(Germline sample)	Germline partial NF2 deletion	Supporting Figures S6 and S7
	MVS10	Right VS	Deletion chr22:30,051,433–30,216,982 (includes NF2 exons 6–17)	Exon 10: c.988delG* (p.Ala330LeufsStop16), same as MVS11, MVS12	No somatic LOH	Supporting Figures S6 and S7
	MVS11	Right VS	Deletion chr22:30,051,433–30,216,982 (includes NF2 exons 6–17)	Exon 10: c.988delG* (p.Ala330LeufsStop16), same as MVS10, MVS12	No somatic LOH	Supporting Figures S6 and S7
	MVS12	Right VS	Deletion chr22:30,051,433–30,216,982 (includes NF2 exons 6–17)	Exon 10: c.988delG* (p.Ala330LeufsStop16), same as MVS10, MVS11	No somatic LOH	Supporting Figures S6 and S7
	MVS13	Right VS	Deletion chr22:30,051,433–30,216,982 (includes NF2 exons 6–17)	Exon 11: c.1021 C>T (p.Arg341Stop); exon 15: c.1624_1627delCTCA,* (p.Leu542ArgfsStop7), see also MVS14, MVS15	No somatic LOH	Supporting Figures S6 and S7
	MVS14	Right VS	Deletion chr22:30,051,433–30,216,982 (includes NF2 exons 6–17)	Exon 11: c.1021 C>T (p.Arg341Stop), see also MVS13, MVS15	No somatic LOH	Supporting Figures S6 and S7
	MVS15	Right VS	Deletion chr22:30,051,433–30,216,982 (includes NF2 exons 6–17)	Exon 11: c.1021 C>T (p.Arg341Stop), see also MVS13, MVS14	No somatic LOH	Supporting Figures S6 and S7
4	GL4	Germline	Deletion chr22:29,850,937–30,047,732 (includes NF2 exons 1–4)	(Germline sample)	Germline partial NF2 deletion	Supporting Figures S8 and S9
	MVS18	Left VS	Deletion chr22:29,850,937–30,047,732 (includes NF2 exons 1–4)	Sample discarded due to insufficient DNA	No data	Supporting Figures S8 and S9
	MVS19	Left VS	Deletion chr22:29,850,937–30,047,732 (includes NF2 exons 1–4)	Exon 12: c.1292delT* (p.Leu431ArgfsStop8), see also MVS20, MVS21, MVS22	No somatic LOH	Supporting Figures S8 and S9
	MVS20	Left VS	Deletion chr22:29,850,937–30,047,732 (includes NF2 exons 1–4)	Exon 2: c.169 C>T (p.Arg57Stop); exon 12: c.1292delT* (p.Leu431ArgfsStop8), same as MVS21, MVS22	No somatic LOH	Supporting Figures S8 and S9
	MVS21	Left VS	Deletion chr22:29,850,937–30,047,732 (includes NF2 exons 1–4)	Exon 2: c.169 C>T (p.Arg57Stop); exon 12: c.1292delT* (p.Leu431ArgfsStop8), same as MVS20, MVS22	No somatic LOH	Supporting Figures S8 and S9
	MVS22	Left VS	Deletion chr22:29,850,937–30,047,732 (includes NF2 exons 1–4)	Exon 8: c.771delG (p.Trp258GlyfsStop38); exon 12: c.1292delT* (p.Leu431ArgfsStop8)	No somatic LOH	Supporting Figures S8 and S9

Abbreviation: MR, mitotic recombination.

The degree of mosaicism in the LOH analysis was determined by allele-specific copy-number analysis of tumors software.

Similarly, in the left MVS from patient 2, five unique, pathogenic NF2 mutations were detected in 6 anatomically separate samples (MVS30 and MVS31, although separate, harbor the same somatic NF2 mutation) (Supplementary Fig. S2a). In MVS16 and MVS24, no NF2 mutation was detected, although a majority of cells in both samples underwent unique mitotic recombination events affecting all or nearly all of chromosome 22q (Supplementary Fig. S2b and S2c). Mitotic recombination is a common mechanism of LOH in tumors, including NF2-associated VS.¹⁹ Mitotic recombination is a type of genetic recombination that occurs between 2 homologous sequences (genes or entire chromosome arms) and results in the duplication of that sequence. In samples MVS16 and MVS24, mitotic recombination caused the loss of the wild-type NF2 sequence and generated 2 copies of the germline NF2 mutation. Functionally, this has the same consequence (development of a tumor) as the “second hit” mechanism generated by a unique somatic NF2 mutation. In the right MVS from patient 3, samples MVS10, MVS11, and MVS12, although anatomically distinct, each harbored the same NF2 pathogenic mutation (Supplementary Fig. S3a). Supplementary Fig. S3b shows the germline deletion of NF2 in patient 3. Samples MVS13, MVS14, and MVS15 all shared the same NF2 truncating mutation, although sample 13 had an additional NF2 frameshift mutation not detected in the other samples, suggesting a clonal evolution, or perhaps merger with another tumorlet. Similarly, all the samples from the left MVS from patient 4 shared the same NF2 frameshift mutation (c.1292delT*), although samples MVS20 and MVS21 carried an additional truncating mutation (c.169C>T), as did MVS22 (c.771delG) (Supplementary Fig. S4a). Supplementary Fig. S4b shows the germline deletion of NF2 in patient 4.

Copy-Number Variation Analysis

Supplementary Table S1 lists the CNVs detected by Nexus in both germline and somatic samples (P < 10⁻⁶). (Deletions and LOH at the NF2 locus, as listed in Table 1, are not included.) Germline CNVs were detected in patients 1, 2, and 3; however, since we lacked parental samples, we were unable to determine whether these were de novo CNVs. Neither of the MVSs from patients 1 or 3 harbored significant somatic CNVs within our stringent detection parameters. One tumor sample (MVS20) from patient 4 harbored an allelic imbalance of ∼194 kb. Tumors MVS24, MVS31, and MVS32 all featured large multigenic amplifications or deletions.

Discussion

Multiple Somatic NF2 Mutations and Chromosome 22 LOH Events Are Evidence of Tumor Polyclonality in NF2-associated MVS

In NF2, mutation or LOH of the NF2 gene is the initiating event of neoplastic transformation in several types of central and peripheral nervous system tumors, including VS.¹⁷ At surgery, NF2-associated VS often appears botryoid. We observed a multilobulated appearance or presence of multiple discrete masses in nearly 60% of the NF2-associated ears we examined. We hypothesized that the individual “grapes”' in the cluster were individual tumorlets that arose independently from distinct somatic mutations in the NF2 gene. To our knowledge, although there exists a perception that these tumors may be polyclonal, no report has supported this with conclusive genetic analysis.^20,21

From our data, the detection of multiple, independent pathogenic somatic mutations or LOH events of NF2 in anatomically distinct samples from a single MVS is consistent with our hypothesis that these tumors are polyclonal. It also supports our hypothesis that NF2-associated MVSs are not single tumors, but rather clusters of multiple smaller tumors, each arising from their own unique second hit event. In multiple patients with constitutional NF2 mutations, we found unique somatic pathogenic NF2 mutations and/or chromosome 22q LOH events (hereafter, “hits”') in multiple anatomically distinct samples from a single MVS (Table 1). In some samples (eg, MVS13, MVS20, MVS21, MVS22), we also found multiple somatic pathogenic NF2 mutations, suggesting the previous merger of even smaller tumorlets.

Second, from analysis of CNV in the MVS, we found evidence of intratumoral heterogeneity. In patient 2, we observed 2 samples from the same MVS (MVS31 and MVS32, each with its own unique pathogenic somatic NF2 mutation) that also harbored unique and mutually exclusive copy-number changes (Supplementary Table S1). These copy-number changes (MVS31: 19q13.33 amplification; MVS32: 14q amplification) serve as “signatures'” for a particular clone. Similarly, the 657-kb deletion at 7q11.21 from the MVS24 sample from the MVS from patient 2 is not observed in any of the other copy-number profiles from that tumor.

Third, we identified 18 distinct somatic NF2 hits in 4 MVSs. This is 4.5 hits per tumor (range 2–7 hits); 13 of the samples we analyzed contained more than one somatic mutation. By comparison, in a set of 30 sporadic VSs from young patients, Mohyuddin and colleagues²⁰ found on average 1 somatic hit per tumor. The exception was one patient with 4 separate hits in a VS; this patient was later determined to have NF2. Our data suggest that somatic inactivation of NF2 is a relatively frequent event in NF2-associated MVS and that each tumor comprises multiple independent cellular populations bearing distinct NF2 mutations.

Analysis of Copy-Number Variation in NF2-associated VS Reveals Limited Perturbation of the Genomic Architecture in Most Samples

Four MVSs in this study had statistically significant CNVs, which showed numerous whole-chromosome deletions (Supplementary Table S1). The most common copy-number change in NF2-asssociated MVS is deletion of chromosome 22q, affecting about 24%–32% of tumors.^19,22 A variety of other chromosomal deletions and amplifications, albeit without a consistent pattern, have been reported in sporadic and NF2-associated tumors.^22–25 Given the modest number of tumors that have been studied, it is unclear whether these associated CNVs are the cause or consequence of neoplastic transformation.

Clinical Implications

Imaging

It is evident that MVS associated with NF2 is polyclonal and that the multilobulated appearance on both imaging and gross inspection is the direct result of multiple tumors colliding. As the resolving capacity of MRI improves and earlier diagnosis due to genetic testing becomes more prevalent, the direct observation of multiple discrete tumors transitioning to a multilobulated mass will likely become more commonplace. At this time, this remains a rare observation (Fig. 4). Our determination of an MVS was conservative. We categorized VS as multilobulated only when there was clear anatomical evidence of deep intratumoral clefts suggesting the presence of multiple tumors. It is quite likely, though, that many of the NF2-associated VSs we characterized as “unilobulate” actually do represent the confluence of multiple independent clones.

Fig. 4.

The process of tumor coalescence is demonstrated over 7-year surveillance using contrast-enhanced T1- weighted MRI. At baseline, 4 distinct tumors are sited in the vestibule (black arrowhead), in the internal auditory canal along the vestibular (white arrow) and cochlear (black arrow) divisions of the cochleovestibular nerve, and in the CPA cistern (white arrowhead). Near complete fusion of the tumors is seen by 7 years. Note in particular the cleft (long arrow) at the site of tumor fusion. CB, cerebellum; P, pons.

Surgery

The unique morphologic finding of multilobularity in VS may in itself increase the risk of surgical morbidity. Deep intratumoral clefts may obscure visualization of surface vessels and make microsurgical dissection of the thinned facial nerve more difficult. Furthermore, patients with NF2 may harbor discrete facial neuromas juxtaposed to the main mass, thereby increasing the difficulty of preservation of facial function.^9,11,26,27 Histologic evaluation of the interface between VS and displaced cranial nerves reveals more frequent invasion in the setting of NF2.¹⁰ These factors have consistently led to worse facial nerve outcomes and hearing preservation rates in NF2-associated VS surgery.

Because most sporadic VS arises from the vestibular components of the eighth cranial nerve and displace the adjacent nerves of the internal auditory canal, it is possible to completely resect these tumors without insult to the cochlear or facial nerve. However, in the case of MVS associated with NF2, it is likely that some clones may arise from the cochlear and/or facial nerve. Therefore, the possibility of gross total resection of a multilobulated mass without hearing loss and/or facial nerve impairment is less likely in the NF2 cohort.

Furthermore, the long-term outcomes from complete resection of small VS (<2 cm) via the middle fossa approach in patients with NF2 have shown that ∼60% of patients have tumor within the operative surgical bed at 5 years.¹² Though it remains unclear whether recurrences are due to genetically distinct tumor formation, the marked difference in recurrence rate after complete resection of analogous tumors in the sporadic (monoclonally derived) population (0.3%) suggests that new tumor formation is occurring on residual vestibular nerve stumps or adjacent cranial nerves (cochlear, facial).²⁸ In aggregate, these data underscore the importance of tempered expectations for a long-term, local recurrence-free state following complete resection of VS in patients with NF2 and the need for strict adherence to a long-term MRI surveillance paradigm postresection.

Radiosurgery

The utility and safety of radiosurgery in the management of a VS in the setting of NF2 remains controversial.²⁹ Radiosurgery for the treatment of sporadic VS has resulted in local control rates exceeding 95% with long-term follow-up.³⁰ The use of radiosurgery for management of NF2-associated VS carries a significantly increased risk for failure. In some studies, this failure rate approaches 50% at 8 years,^14,31 but most studies show a defined decrement in efficiency of local control at any time point for NF2-associated VS.¹³

The polyclonal nature of MVS in NF2 may serve to explain the discordant rates of failure associated with treatment of VS in this population. Indeed, if the probability of successfully treating a sporadic VS (monoclonal derivation) is 0.95 at 5 years, then, if each clone in a polyclonal tumor is treated as an independent tumor, then the probability of achieving local control would be reduced by the product of these probabilities, that is, the probability of control is 0.95ⁿ, where n is the number of independent clones constituting the tumor mass. If the average multilobulated NF2 VS harbors 5 separate clones, we estimate a 77% 5-year local control rate that closely corresponds to published data in the NF2 population. Certainly, the failure of radiosurgery may be multifactorial and include undertreatment of foci of tumor outside the primary mass with marginal doses that are deliberately lower. The refined understanding of the biologic principles underlying worse outcomes with radiosurgery for treatment of MVS in the setting of NF2 begets review of the anticipated utility of radiosurgery in this population.

Taxonomy

Historically, nerve sheath tumors of the CPA were referred to as acoustic neuromas because of their frequent association with hearing loss. Because the majority of sporadic and NF2-associated tumors of the CPA arise from the vestibular nerves, the taxonomic nomenclature was modified to reflect this reality.²¹ Though they share a common histology and underlying pathobiology (NF2 gene inactivation), it is increasingly clear that the multilobulated mass located in the CPA of patients with NF2 does not share similar genetic constitution (polyclonal rather than monoclonal), surgical risks, or treatment outcomes compared with sporadic VS. Thus it may be prudent to include the descriptive term “multilobulated” when referring to the majority of NF2-associated schwannomas that assume a botryoid configuration within the CPA. The term “multilobulated vestibular schwannoma” may more accurately reflect and convey their differing etiology and prognosis.

Conclusion

These imaging findings and genetic analyses suggest that NF2-associated MVS is polyclonal and derived from the collision of distinct, independent tumors. This finding strongly correlates with their unique gross appearance and may explain their substantially different treatment outcomes compared with their sporadic counterparts. These findings have important implications when discussing the relative merits of various treatment strategies and will be an important consideration when establishing endpoints for future clinical trials.

Supplementary Material

Supplementary material is available online at Neuro-Oncology (http://neuro-oncology.oxfordjournals.org/).

Funding

This research was supported by the Intramural Research Program of the National Institute of Neurological Disorders and Stroke, the National Cancer Institute, and the National Human Genome Research Institute at the National Institutes of Health.