Update on Cancer Predisposition Syndromes and Surveillance Guidelines for Childhood Brain Tumors Free

Central nervous system (CNS) malignancies are the second-most common group of cancers in children and the leading cause of cancer-related morbidity and mortality. They are also the largest group of tumors associated with cancer predisposition syndromes (CPS; 15%–21%; refs. 1–5). The molecular understanding of CNS tumors has revolutionized neuro-oncology over the past decade. Simultaneously, the increased use of sequencing platforms has led to the recognition of previously unrecognized or underdiagnosed germline pathogenic/likely pathogenic variants (PV/LPV) as etiologies for many CNS tumors and related syndromes (e.g., ELP1-associated medulloblastoma; refs. 1–3, 5, 6). The American Association for Cancer Research (AACR) Cancer Predisposition Working Group (CPWG) for CNS tumors in 2023 elected to take a cancer-based, rather than a syndrome-directed, approach (7) to develop recommendations for management of broad CNS tumor types associated with CPS. As this is a rapidly evolving field, it is important to translate emerging associations into consensus guidelines that serve as a central reference for pediatric oncologists and other providers for the evaluation of CPS and implementation of surveillance recommendations. For CNS tumors that are part of syndromes covered by other expert groups, including Li–Fraumeni (LFS), neurofibromatosis type 1 (NF1), and DNA replication-repair deficiency (RRD) syndromes, we refer to the relevant manuscripts in this review series. Where strong evidence was lacking, we harmonized our guidelines with existing recommendations from other groups for standardization of practice to support the generation of future evidence.

The signs and symptoms of childhood brain tumors are variable and dependent on the age and location. Typical presentations include frequent headaches, seizures, personality or behavior changes, weakness, numbness or paralysis in part or side of the body, loss of balance, nausea, vomiting, dizziness, loss of hearing, or vision changes. In families with CPS, any such presentation must trigger immediate investigation, including high-quality, nonmotion degraded, dedicated brain and spine magnetic resonance imaging (MRI) scans as per institutional protocols. In addition, there should be a high index of suspicion in the presence of specific clinical stigmata, and those with a family history of relevant cancer types.

The review on genetic testing and counseling in this series addresses the many nuances and updates to testing children for CPS. Because treatment of the cancer, including decisions for radiotherapy and choice of chemotherapy, as well as future management may be directly affected by the CPS, we support early germline genetic testing for all children diagnosed with the specific CNS tumors and those with specific risks/stigmata discussed in this review and others in this series. The recognition of features associated with syndromes can help guide the differential diagnosis. For example, a child with pineoblastoma and a mother who had a Sertoli-Leydig cell tumor of the ovary has a high likelihood of having a germline PV in DICER1. While in cases such as this one, a strong suspicion may point to a specific gene of interest (DICER1), in other cases, timely diagnosis often requires testing a panel of potential causative genes in parallel. The group also recognizes that testing based only on personal and family history and clinical examination often misses clinically meaningful PVs (8, 9). Hence, the group agreed that, at a minimum, testing should include all genes with which the tumor has been associated and chromosomal microarray to detect genome-wide large deletions/duplications that cannot be detected by sequence analysis. For tumor types with multiple associated genes, such as medulloblastoma, we encourage the use of broad multigene, or preferably CNS tumor-targeted gene panels, as these provide a more efficient and cost-effective approach. Exome/ genome sequencing can be more informative and has become a first-/second-tier approach for many neurodevelopmental disorders (10), but in the context of CNS tumors, may not currently be available in a timely manner for therapeutic decision-making in most institutions around the world.

Identifying a CPS in a child with cancer can benefit additional family members through cascade testing. Although these issues are covered in more detail in the genetic testing and counseling manuscript in this series, in general, siblings and other family members who test positive for CPS but do not have cancer should be offered surveillance. For genes without a clear absolute risk of developing cancer that carry no formal recommendations, such as GPR161 and ELP1, counseling and the potential for imaging surveillance could be discussed based on the family history and shared decision-making with the family. At a minimum, it is helpful for the parents to understand the signs and symptoms of the associated tumors and for the child to get an annual, comprehensive, and CNS-targeted physical exam. Furthermore, it is the group's preference that follow-ups are also conducted in the context of research studies. Examples of studies that track predisposition include Zero (Australia; ref. 1), SickKids Cancer Sequencing Program (Ontario, Canada; refs. 3, 11), Precision Oncology for Young People (Canada; ref. 12), Molecular Characterisation Initiative (USA; ref. 13), and Cancer Predisposition Registry (Germany; ref. 14). Results identifying specific variants that are not completely clinically annotated can be tracked in these studies in families to both provide updated information and better understand the outcomes for those conditions with unclear long-term risks to date. In addition, there are predisposition-specific interest groups and consortia that can be a valuable resource for clinicians practicing in varied settings to help their patients and families. Select examples include the PPB/DICER1 registry in Minnesota, the International Replication-Repair Deficiency Consortium and Rare Brain Tumor Registry in Toronto, and the MAGIC consortium for medulloblastoma.

Reproductive genetic testing options available to families with CPS include preimplantation genetic testing, prenatal testing via chorionic villus sampling or amniocentesis, and postnatal testing, as detailed in the genetic testing and counseling manuscript in this series. We encourage discussion of all available options with families, as there are multiple personal and logistic reasons why one choice may be better for one family versus another. For early-onset tumors, families with a significant risk for the fetus to have inherited a PV should be counseled as they may choose to undertake third-trimester fetal imaging. Identifying a CNS tumor prenatally could affect labor and delivery decisions. If gene testing is not undertaken prior to or during pregnancy, we recommend testing infants at or shortly after birth, if the gene is associated with a high risk for early-onset tumors and there is a potential benefit to initiating surveillance in infancy.

Medulloblastoma (MB) is the most common malignant brain tumor in children. In a seminal paper on genetic predisposition including >1,000 patients with MB, Waszak and colleagues (8) reported 6% as having disease-associated germline alterations, of whom only 41% had clinical features and 46% had a family history, highlighting limitations of such clinical cues in diagnosing CPS. Furthermore, novel subgroup associations with CPS were defined in this and subsequent studies, as summarized below and in Table 1.

Up to 6% to 8% of all WNT-MB, and >70% of those who do not harbor a somatic CTNNB1 mutation, harbor germline loss-of-function PV of the tumor-suppressor gene APC, leading to a diagnosis of the autosomal dominant familial adenomatous polyposis (FAP) syndrome (15, 16). Hence, germline APC genetic testing is recommended in WNT-MB lacking somatic CTNNB1 mutations (or where somatic testing is unavailable; refs. 8, 16), as well as in patients with multiple polyps or a relevant personal/ familial history of FAP-related neoplasia. Surveillance protocols for FAP are available, including in a manuscript in this series (17–21). Because the incidence of MB in FAP is low (22) and prognosis is excellent (23, 24), primary neuroimaging is not currently recommended for these patients. The French cooperative group reported that despite the excellent survival for FAP-associated WNT-MB, there was a high risk of second neoplasms (desmoid, thyroid, pilomatrixoma, osteoma, hemangioma, triton tumors), many of which were treatment-related (radiation and/or surgery; ref. 15). This supports treatment deescalation for FAP-associated WNT-MB as in sporadic cases (8, 15, 16). While contextualizing polyposis syndrome and WNT-MB associations, it is important to note that a single adult patient with WNT-MB and germline biallelic MUTYH PV consistent with MUTYH-associated polyposis is reported (22). However, patients with MUTYH-associated polyposis syndrome do not currently have age-based surveillance recommendations.

SHH-MB are enriched for CPS, with current data suggesting a prevalence of up to 40% (6, 8, 25, 26). Among infants and younger children with SHH-MB, 20% have the autosomal dominant nevoid basal cell carcinoma syndrome (NBCCS), historically called Gorlin syndrome, caused by heterozygous germline PV/LPV in predominantly PTCH1 (most common) or SUFU (8, 27, 28). To date, PTCH2 has not been shown to be a relevant predisposition gene (29, 30). NBCCS is characterized by increased risk of both benign and malignant tumors, with an estimated absolute risk of SHH-MB development in PV/LPV carriers of SUFU (7%–9.2%) and PTCH1 (0.37%–1.1%; ref. 28). Non-CNS lesions include odontogenic keratocysts (common in PTCH1 but not described in SUFU carriers) that can cause severe discomfort but rarely undergo malignant transformation. In addition, fibromas, commonly in the heart and ovaries, can develop. Most importantly, affected individuals develop basal cell carcinomas, which increase in prevalence with age, but also after radiotherapy and with excessive ultraviolet light exposure. As most patients with NBCCS-associated MB are infants, radiation is usually omitted in first-line management. In older patients, prompt identification of an underlying CPS is crucial for informing discussions about the need for and/or dosing of any planned radiation to the craniospinal axis. Increased incidence of meningiomas is seen in patients with NBCCS, both primary and secondary to irradiation, more in the setting of SUFU than PTCH1. Individuals diagnosed with NBCCS should be counseled on sun protection and avoidance of ionizing irradiation unless absolutely necessary. For example, an annual orthopantomogram for PTCH1 PV carriers can be replaced with MRI if possible. Surveillance guidelines are tailored to the gene involved (Table 1; refs. 7, 28). Dermatological surveillance, echocardiogram for cardiac fibromas, and ultrasound for ovarian fibromas are recommended for all females, with an additional focus on detecting jaw cysts for germline PTCH1 carriers (7, 28). Surveillance for MB is done clinically for both PTCH1 and SUFU PV/LPV carriers, by frequent neurologic examinations, evaluation of head circumference, and high index of suspicion in case of neurologic symptoms. Routine brain imaging is currently advised for SUFU carriers only, during the first five years of life, with some consensus guidelines recommending intervals as short as every 3 months in the first years of life (7, 28). Screening for meningioma is currently advised only for adult PV/LPV carriers of SUFU (28) or following irradiation.

Germline G-protein–coupled receptor 161 (GPR161) gene PV/LPVs have been reported in up to 5% of infant SHH-MB (median age: 1.5 years). These patients can have an overlapping NBCCS-like phenotype (i.e., develop other associated tumors including BCC and meningioma in probands and family members and Gorlin syndrome-like stigmata). The MB is characterized by somatic copy-neutral loss of heterozygosity of chromosome 1q (25). Because the absolute risk of MB development in germline GPR161 is currently considered low (31), surveillance with brain imaging may not be warranted. Although surveillance is not recommended, a personalized strategy could be considered in specific situations after a thorough discussion with the family about limited data, as well as the benefits and risks of such a strategy. Most importantly, parents should be educated about the possible signs/symptoms of MB that warrant prompt assessment by their managing team and ideally documented in international trials and registries.

Germline biallelic PV/LPVs in Elongator (ELP1) have been recently identified in older children presenting at a median age of 6.3 years and account for 14% to 21% of SHH-MB. These tumors demonstrate somatic loss of 9q resulting in LOH of ELP1 and PTCH1 and biallelic inactivation of the remaining PTCH1 allele (6, 32). Family members with MB are reported and the utility of IHC has been recently described for identification of affected patients (32). Although more robust characterization of such patients is needed, currently the risk of MB development in germline PV/LPV carriers in ELP1 seems to be low (31). Surveillance with brain imaging is not currently recommended. However, as for GPR161 surveillance above, it could be considered in specific situations after a thorough discussion with the family about limited data, and the benefits and risks of a personalized strategy. Heightened clinical awareness and education of family around the signs and symptoms of MB may be warranted, especially for heterozygous siblings of patients with MB.

LFS is an extremely well-characterized CPS with patients at risk of developing multiple brain tumors including MB. Germline TP53 PVs were detected in 20% of SHH-MB ages 5 to 16 years, with only 4% having a family history suggestive of LFS (8, 33). On the other hand, 56% of patients with PV in SHH/TP53 MB had LFS (33). Survival is dismal (5-year OS <30%), but radiation remains paramount to MB survival (8, 33, 34). During and after MB therapy, surveillance for those with LFS should follow the respective guidelines as this affects survival (35), with heightened risk awareness for those receiving radiation.

A subset of TP53-mutant SHH-MB can arise in the context of other germline syndromes such as those related to DNA RRD (36), which includes constitutional mismatch-repair (CMMRD) and polymerase-proofreading deficiencies. These tumors have unique immuno-biology including hypermutation and genomic microsatellite instability and are amenable to immune-checkpoint inhibition (36). The patients are at risk for hematologic, colon, and other malignancies. Surveillance recommendations can be found in the review on CMMRD in this series.

A subset of SHH-MB can arise in the background of Fanconi anemia (FA) caused by homologous recombination repair deficiency (HRD; refs. 8, 37). The rare individuals who develop MB in the setting of germline HRD often do so at a very early age, often before a diagnosis of FA/HRD has been made. While undergoing treatment for cancer, these patients can be highly susceptible to irradiation- and chemotherapy- (especially alkylator-based) associated toxicities. Thus, for any patient with MB and with cutaneous, skeletal, or neurologic abnormalities, or those with severe, unexpected toxicity from chemotherapy, a possibility of an underlying HRD syndrome should be discussed (38). CNS imaging in FA is not currently recommended. Relevant family history of hereditary breast and ovary cancer syndrome cancers, or tumor sequencing revealing a pathogenic variant in or characteristic mutational signature profile suggestive of HRD, may prompt consideration for germline analysis, especially for BRCA2 and PALB2 mutations in patients with MB. Heterozygous BRCA2 and PALB2 variants are enriched in patients with SHH-medulloblastoma. (8) However, given that the estimated risk for medulloblastoma in individuals with a single PV/LPV in BRCA2 or PALB2 is presumed to be low, dedicated neuroimaging is not recommended at this time. Heterozygote carriers of typical adult-onset cancers are discussed in forthcoming AACR CWPG guidelines.

Unlike both WNT and SHH subgroups, groups 3 and 4 do not have as clear associations with known CPS. As discussed above, patients with heterozygous BRCA2 and PALB2 germline PV/PLV have been reported (8). One patient with Rubinstein-Taybi syndrome and a germline CREBBP PV was found to have a group 3 medulloblastoma, which is inadequate to support a true association or a recommendation for surveillance (39). Currently, evaluation for CPS in group 3/4 MB could be reserved for those with a family history of BRCA-associated cancers or HRD mutational signatures on sequencing.

Rhabdoid tumor predisposition syndromes (RTPS) are two distinct autosomal dominant disorders characterized by a predisposition to rhabdoid tumors in the CNS, kidneys, and other locations (7, 40). The CNS is the most common site (65%), where the tumor is called an atypical teratoid/rhabdoid tumor (ATRT). RTPS1 is caused by heterozygous germline PV/LPV in SMARCB1 (coding for INI1 or BAF1) and is more common than RTPS2, which is caused by PV/LPV in SMARCA4 (coding for the transcription activator BRG1; ref. 41). In addition, patients with malignant rhabdoid tumors and congenital deletion of distal 22q11.2 have been described (42). DNA methylation-based studies have confirmed biological differences between SMARCB1- and SMARCA4-altered tumors, and also reported associations between methylation ATRT subgroups, ATRT-SHH and ATRT-TYR, and germline predisposition (43, 44). SMARCB1 PV/LPV carriers are also at risk for developing schwannomas, malignant peripheral nerve sheath tumors, cribriform neuroepithelial tumors, meningiomas, and other rare tumors. Rare cutaneous manifestations have been reported (including congenital plaques and neurovascular hamartoma) and should prompt genetic testing (45). In contrast, SMARCA4 female PV/LPV carriers have a higher risk of developing small cell carcinoma of the ovary, hypercalcemic type (SCCOHT; ref. 46), and rarely uterine sarcomas (47, 48). Recently, neuroblastoma has been linked to RTPS2 (49).

Surveillance for known carriers of truncating germline PV/LPV variants in SMARCB1 includes MRI of the brain and spine, as well as an ultrasound of the abdomen including kidneys during early infancy through age 5. The frequency of investigations has varied in different guidelines based on age, varying from q 4 to 6 weeks at ages 0 to 6 months, to q 2 to 3 months between ages 6 months to 5 years (7, 41). Ultrasound remains the suggested first-line imaging modality in young patients given the logistical challenges of MRI of the abdomen for renal tumor surveillance. MRI is reserved for the characterization of detected lesions or for screening where renal ultrasound is considered nondiagnostic or suboptimal, e.g., due to the patient's body habitus or motion. Although RTPS is an extremely aggressive CPS with high early mortality, there are emerging clusters of patients surviving beyond early childhood and reported to develop second primary neoplasms after age 5, including schwannomas, extracranial malignant rhabdoid tumors, and sarcomas (50–52). Hence, clinical assessment every 6 months is recommended, and yearly whole-body (WB) MRI should be considered. The length of surveillance required for such patients remains unknown and will need further evidence to suggest how long this should continue, given the limited patient numbers. Clinical surveillance should also include pre-symptomatic carriers, especially when a first-degree relative has been diagnosed with a malignancy. This initially includes the affected patients’ parents, although identification of a PV/LPV in an unaffected parent is rare (41). Gonadal mosaicism may contribute to cases of familial occurrence despite unremarkable sequencing results in the parents and hence the group recommends testing young, at-risk siblings even if parents test negative. The recommendations are summarized in Table 2.

Although the penetrance of rhabdoid tumors is lower in probands with PV/LPV in SMARCA4 (53), the severity of cancer and outcomes at early ages are poor as those seen in RTPS1 (47). The current consensus by the European group is to offer surveillance for these individuals (41). As SCCOHT affects females from 18 months to over 50 years of age, surveillance for RTPS2 continuing after 5 years of age should focus on the genitourinary system and include discussions on the role of risk-reducing bilateral salpingo-oopherectomy (54). Regular screening for neuroblastoma is not currently recommended given that the lifetime penetrance of SMARCA4 PV/LPV for NBL is unknown (49). The consensus recommendations for surveillance are summarized in Table 2.

The majority of conditions linked to gliomas have guidelines in reviews in this series (neurofibromatosis type-1 (NF-1), tuberous sclerosis, LFS, PTEN hamartoma tumor syndrome, RRD syndromes, Ollier and Maffucci syndromes). Though covered elsewhere, optic pathway gliomas in particular deserve mention for CPS, given roughly half of NF-1–associated gliomas involve the optic pathway (55), with bilateral involvement being almost pathognomonic of NF1, and are more completely covered in the cousin manuscript. Importantly, CPS-linked glioma can be highly enriched in specific methylation subgroups (9). Such diagnoses should prompt referral for CPS testing and screening. A summary and links to surveillance guidelines are provided in Table 3. Below, we highlight some select entities not covered elsewhere.

Familial melanoma–astrocytoma syndrome is caused by inactivating germline deletion of the CDKN2A/B tumor suppressor gene. Surveillance guidelines are not established (56). The spectrum of disease in this newly described entity continues to evolve with increased knowledge of prevalence only recently published (57). Importantly, individuals with previously a clinical diagnosis of NF1 have been reclassified to this entity. Given the risk of aggressive cancers like melanoma, malignant gliomas, and pancreatic cancers, many in the group agreed that these patients may benefit from an annual full-body examination, as well as brain and WB MRI (56, 58). However, given the unclear evidence around the prevalence and impact of surveillance, we suggest it could be considered after a thorough discussion with the family about the limited data, and the benefits and risks of a personalized strategy similar to other gene strategies.

Ependymomas are rarely associated with CPS. Spinal ependymomas can develop in neurofibromatosis type-2 (59) or multiple endocrine neoplasia type-1 (60), and a single patient with intracranial ependymoma and CMMRD was reported recently (61). There is one case report of ependymoma in NBCCS (62).

Meningiomas and schwannomas in children and adolescents are frequently associated with CPS (63, 64), including neurofibromatosis type-2, RTPS-1, NBCCS/Gorlin syndrome, PTEN Hamartoma tumor syndrome/Cowden syndrome (65), Werner syndrome (66), and most recently, schwannomatosis in the context of a familial DGCR8 microprocessor defect (67). It is important that children and young adults developing meningiomas and schwannomas without other known risk factors are evaluated for an underlying CPS, independent of the presence of a personal or family history (63, 68, 69). Below, we focus on specific entities not covered in other reviews in this series.

Individuals with germline PV/LPV in SMARCE1 can develop clear cell-type meningioma (CCM; refs. 68, 69). This rare, aggressive tumor, arising either in the brain or spine in children and adults (median age = 17.5 years; range, 2–72 years), has up to 60% recurrence rates and significant potential for CNS dissemination (70). For CCM risk associated with germline SMARCE1 (71, 72), we recommend neurologic examination and MRI of the brain and spine yearly from diagnosis until the age of 30 and once every 2 to 3 years thereafter, or in between if there are clinical symptoms (72, 73).

LZTR1 (74) and BAP1 (75–77) are reported with meningiomas. For LZTR1 patients (74), baseline MRI brain and spine is recommended at diagnosis, then every 2 to 3 years, beginning at age 15 to 19 years, with a consideration for WB MRI and increasing surveillance frequency if symptomatic. Alterations in CDKN2A, widely implicated in cancer, have been rarely associated with a germline defect in patients with schwannomatosis (78, 79). To date, the prevalence of this PV does not meet surveillance criteria, and as such, no guidelines are suggested. For patients with the BAP1 CPS (75–77, 80), who can develop uveal and cutaneous melanoma, atypical intradermal tumors, mesothelioma, and renal cell carcinoma, we suggest following published guidelines (81). A summary is provided in Table 4.

Intracranial germ cell tumors encompass approximately 5% of all brain tumors with higher prevalence in pubertal patients and those of Japanese and East Asian descent (82). Evidence is evolving around CPS and this entity with newly identified changes in the neurodevelopmental gene JMJD1C and their close association with this entity (83). Although no specific advice is offered at this point for surveillance, given the incidence in particular in those of Japanese and East Asian descent, it would seem prudent for these patients to receive genetic counseling and consideration of CPS testing. As evidence continues to emerge it is likely given the early evidence, new recommendations on surveillance will be forthcoming. It is also important to note that a recent report from China found aberrations in germline FA pathway genes in 16% of patients with intracranial germ cell tumors (84).

Chordoma is a rare childhood malignancy, often presenting in children intracranially. Familial chordoma is definitively linked to heritable germline duplication of the Brachyury (T) gene (intracranial; ref. 85) as well as TSC1/2 (sacral; ref. 86). Again, genetic counseling should be sought, but no specific recommendations are made at this time.

As novel germline etiologies and biological associations are discovered, the impact of surveillance, early diagnosis (35, 87) and molecularly driven therapies will expand. Novel surveillance guidelines for CPS associated with childhood CNS tumors will continue to evolve over the next decade and will include sophisticated genomic diagnostic techniques including analysis of circulating tumor material and enhanced imaging technologies. An additional challenge involves how best to implement such time- and cost-intensive recommendations in resource-limited settings. Future AACR–CPWG meetings will continue to deliberate on how best to overcome such challenges.

J.R. Hansford reports personal fees from Bayer Australia, Alexion Australia, and Servier International outside the submitted work. M.-L.C. Greer reports other support from Alimentiv and grants from AbbVie outside the submitted work. R. Weksberg reports other support from Alamya Health outside the submitted work. K.W. Pajtler reports grants from the German Childhood Cancer Foundation (DKS2021.02) and the Federal Ministry of Education and Research (01GM2205A) during the conduct of the study. S.M. Pfister reports grants from several companies (IMI-2 Grant) and personal fees from BioScrib outside the submitted work. No disclosures were reported by the other authors.

J.R. Hansford is supported in part through the Hospital Research Foundation and the Jamie McClurg Foundation. D.R. Stewart is supported in part by the Intramural Research Program of the Division of Cancer Epidemiology and Genetics of the National Cancer Institute, Rockville, MD USA. K.W. Pajtler and S.M. Pfister have been supported by the German Childhood Cancer Foundation (DKS2021.02) and the Federal Ministry of Education and Research (01GM2205A). G.M. Brodeur is partly supported by a Consortium for Childhood Cancer Predisposition Grant from St. Baldrick's Foundation