Update on Cancer and Central Nervous System Tumor Surveillance in Pediatric NF2-, SMARCB1-, and LZTR1-Related Schwannomatosis Free

The term schwannomatosis (SWN) describes a group of inherited cancer predisposition syndromes characterized by the development of multiple schwannomas (Table 1). NF2-SWN is an autosomal dominant, highly penetrant cancer predisposition syndrome, previously known as neurofibromatosis type 2, and is caused by pathogenic variants (PV) in the NF2 gene. Individuals with NF2-SWN have a distinct tumor risk, including the development of multiple cranial nerve schwannomas (predominantly bilateral vestibular schwannomas; VS), intradermal schwannoma, meningioma, and less commonly spinal ependymoma (1). Mosaic NF2-SWN is common and has been identified in as high as 60% of individuals with de novo NF2-SWN (2). In recent decades, the identification of additional genes, SMARCB1 and LZTR1, was found to be responsible for overlapping, but distinct, clinical phenotypes in individuals with negative germline, negative postzygotic somatic mosaic NF2 testing, and absence of bilateral VS. SMARCB1-SWN is due to germline, nontruncating, partial loss-of-function PVs in SMARCB1 in individuals without germline PVs in NF2 (3, 4). SMARCB1-SWN is an autosomal dominant–inherited condition that presents with nonvestibular and nonintradermal schwannomas. Meningiomas are less common, and malignant peripheral nerve sheath tumors (MPNST) have been observed even in the absence of radiation (5–8). LZTR1-SWN arises due to germline PVs in LZTR1 and is an incompletely penetrant, autosomal dominant condition that increases an individual’s risk for peripheral schwannomas and, less commonly, unilateral VS (9, 10). We aim to succinctly aggregate current recommendations to update previously published diagnostic and surveillance guidelines with a focus on tumor surveillance during childhood (1, 11–13).

NF2-SWN is confirmed in an individual with a germline PV in NF2, whereas mosaic NF2-SWN is diagnosed when two anatomically distinct tumors share an identical somatic PV in NF2. Non-NF2 familial SWN individuals have negative germline NF2 findings and multiple tumors without common somatic variants in NF2. Further germline analysis of non-NF2 familial SWN individuals has identified PVs in SMARCB1 or LZTR1 (3, 10, 14). There remains a small subset of individuals with familial SWN in which a germline PV has not been identified, and additional unknown SWN genes likely exist. PVs in DGCR8 have been identified in individuals with multiple schwannomas, but more data are needed to better understand the clinical presentation of individuals with germline PVs in DGCR8 and the overlap with SWN (15). Molecular and clinical diagnostic criteria for SWN were recently updated, and protocols for genetic testing have been developed (1, 13). Owing to overlapping clinical presentations, high de novo and mosaic rates, and multiple allelic conditions detailed below, individuals with concerns for a familial SWN diagnosis should be seen by a genetics provider with expertise in SWN syndromes.

NF2-SWN is a highly penetrant syndrome caused by PVs in NF2 and is inherited in an autosomal dominant manner. NF2-SWN has an estimated birth incidence of 1:28,000 and a prevalence of 1:50,000 (16). High de novo rates contribute to more than 70% of cases, with the remaining cases being inherited from a parent (17). Clinical presentation is typical in adulthood, with an average onset age of 18 to 24 years. However, early presentation in childhood is often seen, particularly with high-risk genotypes (18–24). The earliest clinical features of NF2-SWN include ophthalmologic findings, intradermal plaques, and meningioma (23, 25). Unlike other SWN syndromes, NF2-SWN is associated with bilateral VS development and a high risk of complete hearing loss, with nearly all individuals (88%) developing VS by age 30 (22). Frequent ophthalmologic findings include juvenile subcapsular or cortical cataracts, retinal hamartoma, and childhood epiretinal membrane and should raise suspicion for NF2-SWN even prior to the development of hallmark tumors (1). Similarly, intradermal schwannomatous plaques and, less commonly, intracutaneous neurofibromas may be present in childhood, and individuals presenting with these lesions should be monitored for additional clinical signs concerning for NF2-SWN (11). Café-au-lait macules may be more prominent in individuals with NF2-SWN compared with the general population but typically do not reach the diagnostic criteria for neurofibromatosis type 1. Genotype–phenotype correlations have been determined by multiple studies with truncating PVs conferring a more severe and earlier presentation of the syndrome (18–24).

In addition to bilateral VS, cranial nerve, spinal, and peripheral nerve sheath SWN are frequent and lead to increased morbidity and early mortality for these individuals. Meningiomas are the second most prevalent tumor and can often be the first presenting sign of NF2-SWN in childhood. Spinal ependymomas are present in 20% to 40% of people but are classically nonprogressive (12).

Familial SWN caused by PVs in SMARCB1 was initially described in 2007, and further cohort analysis continues to elucidate clinical features and genotype–phenotype correlations with an overall estimated prevalence of 1 in 1.1 million (3, 4, 7, 17, 26). SMARCB1 encodes a subunit of the SWItch/Sucrose Non-Fermentable (SWI/SNF) chromatin remodeling complex. Alterations in SMARCB1 are causative of multiple allelic disorders: SWN, Coffin–Siris syndrome, and rhabdoid tumor predisposition syndrome 1 (RTPS1). Coffin–Siris syndrome may be due to gain-of-function variants in SMARCB1, among other genes, but has minimal-to-no phenotypic overlap with SWN (27). RTPS1 is characterized by truncating variants in SMARCB1 that occur predominantly in the central exons 2 to 9 and result in complete loss of the SMARCB1 protein, whereas SMARCB1-SWN is more frequently seen with nontruncating variants at the 3′ or 5′ end of the gene, particularly in exons 1 or 8, and presents with a hypomorphic SMARCB1 protein (28). Truncating frameshift variants in exon 1 have been described; however, rather than a complete absence of a protein product, as seen in RTPS1, the near full-length protein is produced because of the creation of a reinitiation codon downstream, and individuals demonstrate a phenotype consistent with SMARCB1-SWN (29). Individuals with RTPS1 have been diagnosed with peripheral nerve sheath tumors although the risk for rhabdoid tumors remains the hallmark of this syndrome (30). Deep intronic PVs in SMARCB1 have been reported, and RNA sequencing should be considered for individuals with negative DNA testing with strong clinical concern for SWN (31).

Clinical presentation of SMARCB1-SWN is most common in the second to third decades of life and is characterized by peripheral nerve schwannomas. VS are uncommon, and if the cranial nerves are involved, it is most often the trigeminal nerve. Meningiomas have been seen in SMARCB1-SWN but are less common than NF2-SWN (32, 33). Four individuals with SMARCB1-SWN have been reported in the literature to develop an MPNST, even in the absence of radiation (5–8). Additional malignancies have been reported, but a strong association has not been determined (34). Furthermore, rhabdoid tumors and atypical teratoid/rhabdoid tumors have been diagnosed in patients with a SMARCB1-SWN diagnosis, and counseling should be provided about potential overlap between these allelic disorders (7, 35, 36). Prompt evaluation is needed for clinical changes concerning for an underlying malignancy.

In 2014, germline PVs in LZTR1 were identified in families with SWN (10, 14, 37). Further population studies indicate an incidence of LZTR1-SWN of 1 in 527,000 live births (17). Heterozygous variants in LZTR1 demonstrate loss of function and are also identified incidentally in the general population without individuals ever developing clinical features (10). Therefore, LZTR1-SWN is only diagnosed in individuals with a PV in LZTR1 and at least one schwannoma or nerve sheath tumor (1). Penetrance is incomplete, but clinical features most commonly present in the second to fourth decades of life, with a painless mass or neuropathic pain due to a peripheral nerve sheath tumor/schwannoma. Increased pain has been identified in patients with LZTR1-SWN with a somatic SH3PXD2A–HTRA1 fusion (12). Meningiomas and ependymomas have not been reported in individuals with LZTR1-SWN, but unilateral VS occurs (38, 39).

PVs in LZTR1 are causative of allelic disorders: Noonan syndrome (NS), constituting about 8% of NS diagnoses (40), and isolated multiple café-au-lait macules (41). Variants demonstrating autosomal dominance in NS because LZTR1 are typically gain of function, whereas LZTR1 variants with missense and loss of function contribute to autosomal recessive NS and may have overlap with PVs concerning for LZTR1-SWN. Moreover, as germline genetic testing is being expanded, particularly through the wider availability of hereditary cancer gene panel testing, PVs in LZTR1 are being increasingly identified in individuals with no family history or personal history of SWN. In these individuals, the LZTR1 PV should be considered incidental and is not diagnostic of LZTR1-SWN; therefore, these individuals should not undergo surveillance for LZTR1-SWN (12). Prospective follow-up studies will facilitate accurate penetrance estimation in which LZTR1 PVs are identified off-target. Clinical evaluation and genetic counseling are recommended because of the heterogeneity of clinical presentations (9).

In NF2-SWN, tumor development follows the well-recognized Knudson two-hit hypothesis of tumorigenesis with initial germline NF2 PV followed by somatic inactivation of NF2 in the second NF2 allele. LZTR1-SWN and SMARCB1-SWN, however, require PVs in multiple genes to drive tumorigenesis (1, 3, 4, 10). Individuals carry an initial germline heterozygous precipitating PV in SMARCB1 or LZTR1. Tumors in these individuals demonstrate loss of the second allele of SMARCB1 or LZTR1 through copy number deletion of a portion of chromosome 22 that also includes NF2. A subsequent independent somatic PV of the remaining NF2 allele also occurs, resulting in biallelic NF2 (and SMARCB1 or LZTR1) inactivation in the tumor (1, 3, 4, 10). Importantly, each NF2 somatic variant will be unique in different schwannomas from the same individual and is not a sign of undiagnosed mosaicism. In this regard, SMARCB1-SWN and LZTR1-SWN remain inherited disorders but with reduced penetrance, likely due to the multistep, “3-event, 4-hit” process required for tumorigenesis (3, 10).

Initial evaluations for individuals with an SWN diagnosis are comprehensive and often require multidisciplinary evaluation. Some imaging exams may be performed as part of the diagnostic workup but, if not completed, should be done for baseline screening (Tables 2 and 3). Although symptomatic presentation occurs in the second to fourth decades of life, individuals with a genetic diagnosis at a young age due to clinical presentation or familial inheritance should initiate surveillance by age 10 years. Surveillance is then recommended to continue lifelong with a goal to diagnose and manage tumors and counsel patients before severe symptoms are present (12). This imaging includes brain MRI with gadolinium-based intravenous contrast, including a dedicated high-resolution MRI (with fine cuts of 1–3 mm) of the internal auditory canals, preferably on a 3 Tesla MRI scanner and a spine MRI (42). The dermatology exam should assess for schwannomatous plaques. A thorough ophthalmic exam is needed to evaluate for retinal hamartomas, lens opacities, and orbital schwannomas. Finally, whole-body MRI (WBMRI) is now also recommended for baseline tumor burden assessment but should not be completed in lieu of the initial detailed brain, internal auditory canal, and spine imaging (12, 43–47). If an individual is diagnosed at a young age, WBMRI should be considered starting at age 10 years if sedation is not expected to be required. If WBMRI is not available, spinal MRI and a detailed physical exam with complete neurologic evaluation are recommended to guide more focused imaging. These initial evaluations are the same for individuals regardless of the predisposing genetic variant, NF2, SMARCB1, or LZTR1, as the identification of certain tumors or skin and eye findings should prompt reexamination of the diagnosis if unexpected for the genetic phenotype.

In individuals with NF2-SWN who are asymptomatic or stable-symptomatic, a brain MRI should be undertaken annually due to increased meningioma and other central nervous system tumor risk (Table 2). When tumors are detected, brain MRI should be repeated at 6 and 12 months to gain an idea of the growth rate. If stable or slow asymptomatic growth is observed, brain MRI may return to routine annual surveillance. For asymptomatic children and adolescents with SMARCB1-SWN and LZTR1-SWN, brain MRI does not require annual monitoring and can be repeated every 3 years, even if they demonstrate tumors on their baseline imaging (Table 3; ref. 12). Due to increased peripheral schwannoma risk in SMARCB1-SWN and LZTR1-SWN, spine MRI and WBMRI should be undertaken, alternating every 3 years, and completed concurrently with brain MRI (12, 48). Audiology and ophthalmology evaluations should also occur annually in individuals with NF2-SWN. All individuals should have a thorough annual physical exam with complete neurologic and pain evaluations. For patients with clinical symptoms, imaging may be obtained earlier and dedicated to the region of concern. This imaging should predominantly be MRI unless there is a contraindication as CT has not been shown to be particularly useful for SWN evaluation and carries a radiation exposure risk. Due to the increased malignancy risk in SMARCB1-SWN, a growing or changing tumor, especially if painful or causing a functional deficit, should prompt careful and expedient assessment for malignant transformation. In this regard, nuclear medicine imaging like 18F-fluorodeoxyglucose–PET may be considered in addition to MRI for symptomatic, rapidly changing lesions in a patient with SMARCB1-SWN due to the increased risk of MPNST. Nuclear medicine imaging is not beneficial for patients without SMARCB1-SWN and concern for high grade malignancy. If prostheses, cochlear implants, or implantable pain stimulators are required, their placement should be coordinated with radiology and those managing surveillance to minimize the impact on routine imaging. Knowledge of the specific implant or device used is essential to determine MRI safety once in situ.

All individuals with a diagnosis of SWN should have multidisciplinary management by neurology, neurosurgery, otolaryngology, audiology and ophthalmology (NF2-SWN in particular), clinical/medical genetics, and oncology. Recommended services also include pain management, social work, and psychology. Surgical resection should be considered for the treatment of painful or growing peripheral or spinal schwannomas not responsive to initial pain management although pain is not always resolved with tumor resection. Surgery is not often recommended for VS in children with NF2-SWN due to the risk of further cranial nerve damage. However, for individuals with LZTR1-SWN, VS management options should follow the routine recommendations for sporadic VS with a role for surgical resection. Radiotherapy should not be used routinely due to the risk for malignant transformation—which has been reported in multiple cases of NF2-SWN—but the benefit may outweigh the risk in rare individual cases such as recurrent and atypical meningiomas in NF2-SWN (12, 34, 49). Medical therapy is increasingly available for tumor control in SWN, with advances in bevacizumab improving schwannoma outcomes, brigatinib reducing meningioma tumor size, and VEGF vaccine therapies demonstrating early safety and response in schwannomas. More clinical trials are underway to improve tumor control and symptom management for this patient population (50–57).

Guidelines for surveillance will continue to develop as novel germline associations are elucidated, identifiable somatic drivers are discovered, and treatment modalities improve. There is a need to ensure that treatment and prevention strategies being investigated in the adult population are considered and, where appropriate, offered in the childhood setting. Although young adulthood is the typical age of onset for SWN tumors, childhood presentation often presents a surveillance and management dilemma for providers and families left without evidence for effective therapies. In this regard, further genotype–phenotype correlations and the natural history of at-risk tumors will continue to advance surveillance recommendations. Currently, we do not recommend altering surveillance based on genotype within each syndrome, but for high-risk NF2-SWN genotypes or individuals presenting with clinical features in young childhood (<age 10 years), one could consider surveillance imaging at the time of initial clinical presentation. Improvement in medical therapies to treat patients with identified tumors would further the case for early surveillance. Future American Association for Cancer Research Cancer Predisposition Working Group meetings will continue to be convened for up-to-date surveillance and intervention recommendations.

M-L.C. Greer reports grants from AbbVie and nonfinancial support from Alimentiv outside the submitted work. J.R. Hansford reports other support from Servier Pharmaceuticals, Alexion Pharmaceuticals, and Bayer Pharmaceuticals outside the submitted work. K.A.P. Schultz reports other support from Children’s Minnesota Foundation and Pine Tree Apple Classic Fund during the conduct of the study. A. Das reports other support from Servier Canada outside the submitted work. S.M. Pfister reports nonfinancial support from Epignositix GmbH and personal fees from BioSkryb Genomics and PMC outside the submitted work. E.R. Woodward reports personal fees from Cancer Research UK outside the submitted work. C.P. Kratz reports grants from Deutsche Kinderkrebsstiftung during the conduct of the study. No disclosures were reported by the other authors.

The authors wish to thank many children, young adults, and families for contributing to this research. This work was supported by the Intramural Research Program of the Division of Cancer Epidemiology and Genetics, NCI, Rockville, MD (to G.M. Ney and D.R. Stewart); the American Lebanese Syrian Associated Charities (to M.R. Perrino); the NIHR Manchester Biomedical Research Centre (grant number NIHR203308; to E.R. Woodward); the Hospital Research Foundation and the McClurg Foundation (to J.R. Hansford); and the Deutsche Kinderkrebsstiftung (DKS2024.03; to C.P. Kratz). The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.