NF1 results in loss of function of the tumor suppressor protein Neurofibromin. This large protein is a key negative regulator of the RAS pathway by catalyzing the hydrolysis of active guanosine triphosphate-bound RAS to inactive guanosine diphosphate-bound RAS [11]. Dysfunctional neurofibromin results in constitutive activation of downstream oncogenic pathways including MAPK and mTOR. Mutations or deletions in the NF1, gene can be identified in more than 95 % of individuals with NF1 [13]. However, since the gene is very large and difficult to analyze, diagnosis and management can be made based on clinical criteria. RAS/MAPK pathway activation is seen in almost all pediatric low-grade astrocytomas [13] and in 88 % of adult malignant gliomas [14]. Actual somatic mutations in NF1 occur in 20 % of adult gliomas.

Mouse models of gliomas frequently alter the RAS/MAPK pathway. However, additional alterations in major tumor suppressor pathways such as TP53, RB, and PTEN are required to generate tumors [15].

Indeed, NF1 deficient mice do not have tumors but hyperplasia mimicking the optic pathway gliomas (OPG) commonly seen in these individuals [16]. Taken together, the benign nature of tumors seen in the CNS in patients with NF1 support the concept of oncogene induced senescence as a mechanism to explain the spontaneous growth arrest of these tumors when the RAS pathway is constitutively active [17].

The most common CNS tumor in NF1 are optic pathway gliomas (OPG) affecting up to 15 % of individuals with the syndrome. Conversely, up to a third of children with OPG have germline mutations in NF1. Bilateral optic nerve gliomas exist almost exclusively in children with NF1 (Fig. 1.2).

NF1 related OPG typically have an indolent course with spontaneous growth arrest. Indeed, the vast majority of these OPG will not progress after initial diagnosis. Up to 15 % of these tumors, however, do progress, resulting in visual loss or other symptoms and requiring intervention. High-grade gliomas are relatively uncommon, but have been reported and should be considered in patients whose tumors arise in an uncharacteristic location or demonstrate particularly aggressive behavior [18, 19].

Patients with NF1 often exhibit multiple lesions, mainly in the basal ganglia and brainstem which are difficult to assess. These include T2 bright lesions on MRI, without significant mass effect, termed FLAIR (fluid attenuated inversion recovery) associated sub-cortical intensities or FASCI (Fig. 1.2). These lesions tend to disappear spontaneously after initial growth and rarely cause symptoms. Differentiating between FASCI and low-grade gliomas in NF1 patients may be challenging.

Since NF1 is a multisystem condition, careful monitoring is recommended in multidisciplinary clinics [6]. Due to the marked variability in clinical manifestations, including tumor occurrence, strategies for surveillance and follow-up must be tailored to each individual patient. Optic gliomas affecting both optic nerves and/or coexistence of FASCI should raise a suspicion of NF1 even in the absence of typical neurocutaneous findings such as café-au-lait macules, and genetic counseling is recommended.

Unfortunately, surveillance neuroimaging in asymptomatic children with NF1 has not been shown to reduce the incidence of visual loss in this population, and frequent neuro-ophthalmologic examination remains standard of care [20]. For FASCI and other atypical brain lesions, close monitoring is recommended and treatment should be reserved for patients with progressive clinical symptoms and in some cases, radiological progression.

Individuals with NF1 are particularly sensitive to the damaging effects of ionizing irradiation, leading both to an increased incidence of irradiation-induced cancers [21], as well as to cerebrovascular damage (Moyamoya syndrome) [22, 23]. Cranial irradiation in NF1 patients with OPG in particular and brain tumors in general, should be avoided as long as reasonable alternative treatment options exist.

Inhibitors targeting the RAS and mTOR pathways are of great interest for the potential treatment of NF1 and NF1-related tumors, and are being investigated as novel therapeutic approaches in preclinical and clinical studies. Furthermore, targeting the microenvironment believed to be necessary for NF1 tumor growth may allow for additional NF1 specific therapies for these patients [24, 25]. For example, inhibition of c-KIT has shown encouraging efficacy for peripheral NF1 related neurofibromas in subsets of patients [26, 27], although it remains to be seen whether c-kit represents a molecular target of value in NF1 related CNS tumors.

In 1990 the association between LFS and germline mutations in the tumor suppressor gene TP53 was made [35]. TP53 is located at chromosome 17p13.1 and has been called the “gatekeeper of the genome,” since it represents one of the key proteins that maintain genome integrity after DNA damage, hypoxia, and other stressors. TP53 activation results in cell cycle arrest, senescence, and apoptosis. TP53 is also involved in key metabolic pathways in the cell including cell metabolism and mitochondrial function [36]. TP53 represents one of the most commonly mutated tumor suppressors known. Molecular genetic evidence of TP53 pathway disruption can be found in more than 50 % of tumors from adult cancer patients, as well as in 80 % of adult [13] and 50 % of pediatric high-grade gliomas [37].

Development of faithful preclinical models of LFS has been hindered by the fact that TP53 alteration in animal models generally fails to recapitulate the tumor phenotypes of the human disorder, and brain tumors are not a part of the phenotype even in some of the newer models [38]. Nevertheless, many glioma and medulloblastoma mouse models utilize TP53 alterations in conjunction with other cancer genes to mimic the human disease [39].

Three types of brain tumors are associated with LFS: high-grade gliomas, choroid plexus carcinoma, and medulloblastoma. Choroid plexus tumors affect LFS carriers in the first decade of life and medulloblastomas usually in the second, while malignant gliomas can occur throughout childhood, but more commonly in young adults.

Gliomas have been recognized as part of LFS from the earliest reports [28]. Although TP53 expression is associated with worse outcome in childhood glioblastoma [40], currently, no data exist regarding the significance of germline TP53 mutations in pediatric high-grade gliomas. The surveillance protocol [41] developed by our group uncovered several low-grade gliomas suggesting that some of LFS associated glioblastomas arise as secondary glioblastomas and may benefit from early intervention.

Choroid plexus carcinomas are one of the most common presentations of LFS in young children and were recently added to the criteria for the diagnosis of the syndrome [33]. Furthermore, a significant number of patients with choroid plexus carcinoma will harbor germline TP53 mutations. Somatic mutations in TP53 are observed in up to 50 % of choroid plexus carcinomas, and this confers a poorer chance of survival for these patients [2]. This phenomenon may be caused by increased resistance of TP53 mutant tumors to radiation and chemotherapy [42].

Medulloblastomas harbor somatic TP53 mutations in 5–10 %, and appear strictly confined to the WNT (wingless-related integration site) and SHH (sonic hedgehog) subgroups of tumors [43]. Remarkably, TP53 mutations do not alter the excellent survival of patients with WNT medulloblastomas, while TP53 mutant SHH medulloblastomas are commonly seen in the second decade of life and have unfavorable outcome. Interestingly, these are commonly individuals with LFS. SHH medulloblastomas from LFS patients have a unique molecular genetic profile, suggesting chromothripsis (“chromosome shattering”) as the initiating event [44].

Current recommendation is to screen for germline TP53 mutations, i.e., LFS, in all individuals presenting either with a high-grade glioma and a family history of LFS tumors, or patients diagnosed with a choroid plexus carcinoma or medulloblastoma harboring somatic TP53 mutations [33]. Cancer surveillance protocols developed specifically for individuals with LFS have revealed a high rate of early tumor detection [45]. Recently, a striking survival benefit for children has been observed using these protocols, mainly due to improved early detection of brain tumors (Fig. 1.3). Although no molecular targeted therapy for TP53 mutated tumors is currently available, detection of a germline TP53 mutation has significant prognostic and therapeutic implications for the patient. Both children and adults with LFS have been considered to be at an increased risk for developing radiation therapy-induced secondary malignant tumors [46, 47], as well as secondary myelodysplastic syndrome following specific chemotherapies [48].

Germline mutations in MLH1, MSH2, MSH6, and PMS2 have been reported in association with CMMR-D. These mismatch repair genes are critical in repairing single base pair mismatches and misalignments [49]. In the absence of such genes, high mutation rates are observed, including in cancers which are described as “mutator phenotype” [14]. CMMR-D is inherited in an autosomal recessive fashion and is found mostly in consanguineous families. Some patients may have NF1 in addition to CMMR-D, which is thought to be caused by “secondary” early or germline mutations in the NF1 gene as a part of the mutator phenotype [53].

Interestingly, mouse models of mismatch repair deficiency recapitulate cancers of the gastrointestinal tract and lymphomas, but fail to develop brain tumors [54, 55].

Diagnosis of Lynch-related cancers can be made by evidence of microsatellite instability [58]. However, this method has not been shown to be sensitive in CMMR-D, especially in brain tumors and lymphomas. Immunostain of tumor tissue for the MMR proteins is almost universally negative in CMMR-D cancers, and has the unique diagnostic feature of negative stain in the corresponding normal tissue.

Any patient with gliomas, T-cell lymphoma and either café-au-lait macules, consanguinity or a family history of colon cancer should be screened for any of the four mismatch repair genes. Similarly, high index of suspicion should be raised for “NF1” patients with malignant gliomas and consanguinity.

Individuals with Lynch syndrome benefit from a strict surveillance protocol (http://​www.​NCCN.​org/​) and from preventive colectomy. Therefore, early and accurate diagnosis may benefit parents and other family members. Since the risk of gliomas and lymphoma is extremely high for biallelic MMR patients, following a surveillance protocol may be beneficial for children with CMMR-D [59]. There are several reports of MMR tumors responding to specific agents including retinoic acid [60], which may be exploited for the treatment of these tumors.

The gene responsible for BCNS is PTCH1 which is located on chromosome 9q22.3 [68]. PTCH1 is a protein that is a major suppressor of the sonic hedgehog (SHH) pathway by direct inhibition of SMO. Disruption of PTCH1 leads to constitutive activation of the pathway and induction of GLI target genes and cell proliferation and survival. SHH is involved in neural development and midline segregation, which can explain some of the syndromatic manifestations of BCNS.Germline mutations in Sufu, which is a direct inhibitor of GLI have been reported in familial and sporadic medulloblastoma [69] in up to 50 % of desmoplastic tumors [70].

Most mouse models of medulloblastoma utilize alterations in the SHH pathway [71]. While alteration of PTCH1 only results in a tumor incidence of 10–20 %, its combination with other alterations results in very efficient formation of aggressive tumors.

Sequence analysis of PTCH1 yields a mutation in 50–80 % of patients with Gorlin syndrome [72, 73]. Partial and whole-gene deletions are found in 6–21 % of patients [74]. In patients with mental retardation chromosome analysis and chromosomal microarray may reveal the 9q22.3 microdeletion. The 9q22.3 microdeletion syndrome is associated with cognitive impairment, metopic synostosis, obstructive hydrocephalus, macrosomia and seizures, in addition to the features of Gorlin syndrome [75].

The first report of a brain tumor, medulloblastoma, in association with this hereditable syndrome [76] was published in 1963. The development of medulloblastoma in the setting of Gorlin syndrome occurs earlier compared to sporadic tumors. Most patients are younger than 3 years of age, and almost all tumors have a specific pathological subtype termed desmoplastic variant [77]. More specifically, desmoplasia with extensive nodularity (MBEN) is almost pathognomonic for the syndrome in young children [78, 79]. Of note, desmoplastic medulloblastomas in young children usually have a favorable outcome even without radiation therapy [80].

Several reports have documented the development of intracranial meningiomas in patients with Gorlin syndrome with or without prior craniospinal irradiation [81], although strength of association remains unknown.

Individuals with the clinical manifestations of BCNS, a family history of basal cell carcinomas, or other manifestations of the syndrome should be screened for germline mutations in PTCH1 and SUFU.

Patients with medulloblastoma and any of the above should also be screened, since radiation therapy is associated with an increased risk of developing basal cell carcinomas within the irradiated fields [82] in almost all patients. Furthermore, since the rate of germline mutations in the SHH pathway is extremely high in young children with desmoplastic medulloblastoma (Fig. 1.4) [70], children less than 3 years of age with desmoplastic tumors should be screened even without clinical manifestations of the syndrome. The current consensus recommends yearly brain MRI scans for all patients with BCNS until the age of 8 years [83].

The recent development of novel SHH pathway inhibitors [84] may lead to future targeted therapies for individuals with both Gorlin and SUFU syndromes, possibly including primary tumor prevention strategies.

FAP is caused by germline mutations in the gene adenomatosis polyposis coli (APC) [87]. APC is located on chromosome 5q21-22 and is a major regulator of the WNT pathway, which plays a paramount role in controlling embryonic development, stem cell viability and proliferation. Hyperactivation of the WNT pathway is reported in 5–10 % of medulloblastomas, usually as a result of mutations in CTNNB1 [88, 89]. APC mutations are rare in sporadic medulloblastoma. Interestingly, no current mouse models exist for APC-driven medulloblastoma.

Since medulloblastoma is rare in FAP, carriers are not routinely screened for these tumors. However, a patient with medulloblastoma and FAP should undergo GI cancer surveillance, since the development of medulloblastoma in PBTS 2 may precede the development of colonic adenocarcinoma. Indeed, patients are reported with simultaneous diagnoses of medulloblastoma and colonic adenocarcinoma. It is important to note that APC mutated medulloblastomas are distinct from most WNT activated medulloblastomas. WNT tumors will have CTNNB1 mutations that can be diagnosed by nuclear staining of the gene product. Although WNT pathway activation generally confers favorable survival in sporadic medulloblastomas, the prognosis of APC mutated tumors is still uncertain, and FAP patients therefore should not be treated with less aggressive protocols.

FA is a genetically heterogeneous disorder associated with either biallelic mutations in any of the known 14 autosomal genes or a mutation in an X-linked gene [101]. Individuals in the FA complementation group FANCD1 are estimated to represent no more than 3 % of all individuals with FA, and it is this group in whom biallelic mutations with BRCA2 are found. BRCA2 mutations are well known to be associated with familial predisposition to breast and ovarian cancer. Brain tumors have also been reported in such families [102]. These individuals may present a more severe phenotype with early onset of cancer. In particular, the cumulative probability of developing a brain tumor (almost always medulloblastoma) could be high as 85 % in the first decade [99]. This knowledge has prompted a search for other genes in the pathway and familial childhood cancers. Recently, germline mutations in PALB2, another gene in the Fanconi pathway, were reported to be associated with medulloblastomas and other pediatric cancers [103].

Mouse models of FA usually fail to produce the cancers and other organ damage seen in humans [104]. However, an increased incidence of tumors has been observed in Fancd2 deficient mice [105]. This gene interacts with BRCA2 and PALB2.

The rare individuals who develop medulloblastoma in the setting of FA do so at a very early age, often before a diagnosis of FA has been made. Individuals with FA undergoing treatment for cancer are known to be highly sensitive to both irradiation and chemotherapy, with increased susceptibility for treatment-associated toxicities, especially from alkylator-based chemotherapy [106]. Thus, any early onset pediatric brain tumor with cutaneous, skeletal, or neurological abnormalities consistent with a diagnosis of FA or in case of severe unexpected toxicity from chemotherapy, genetic counseling is recommended.

The concept of synthetic lethality is being exploited in the use of PARP inhibitors in BRCA1/2 deficient breast, ovarian, and pancreatic cancers [107, 108], which could be of value in FA as well. To date, however, no data exist on the use of PARP inhibitors in the treatment of FA-related medulloblastoma.