Inherited Susceptibility for Pediatric Cancer
Sharon E. Plon,a Houston, Texas, Katherine Nathanson,b Philadelphia, Pennsylvania
ABSTRACT
The percentage of childhood cancers that are caused by a clearly inherited predisposition varies significantly from only a few percent to more than 50% with individual tumor types. Recent advances in genetic testing and studies of cohorts of cancer patients have demonstrated the likeli- hood of identifying a cancer susceptibility mutation for nu- merous childhood cancers. Inherited predisposition to cancer is frequently the result of dominant constitutional mutations in tumor suppressor genes, which can be in- herited from an affected parent or occur de novo during gametogenesis. In this article, we review the childhood ma- lignancies that are associated with at least a 10% likeli- hood of being caused by a genetic susceptibility to cancer and therefore warrant consideration for a genetic evalua- tion; these malignancies include retinoblastoma, adreno- cortical carcinoma, atypical teratoid and malignant rhabdoid tumors, optic pathway tumors, juvenile myelomonocytic leukemia, malignant peripheral nerve sheath tumors, vestibular schwannomas, endolymphatic sac tumors, he- mangioblastomas, medullary thyroid cancer, pheochromo- cytomas, and paragangliomas. Children with other malignancies may also warrant genetic evaluation if there is the co-occurrence of malignancy and two or more con- genital anomalies, or malignancy and a significant family history of related cancers. We also review the importance
From the ªTexas Children’s Cancer Center, Departments of Pedi- atrics and Molecular and Human Genetics, Baylor College of Medicine, Houston, Texas, and the Division of Medical Genetics, Department of Medicine, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania.
Received on June 14, 2005; accepted for publication June 29, 2005.
No benefits in any form have been or will be received from a commercial party related directly or indirectly to the subject of this article. S.E.P. was supported by grant U24-CA78142-02, and K.L.N. was supported by grant K08 - CA084030, both from the National Cancer Institute.
Reprint requests: Sharon E. Plon, MC 3-3320, 6621 Fannin Street, Houston, Texas 77030.
E-mail: splon@bcm.tmc.edu.
Copyright 2005 Jones and Bartlett Publishers, Inc.
of the correct genetic diagnosis in order to ensure appro- priate treatment and ongoing cancer surveillance for the child with cancer and closely related family members (e.g., parents and siblings).(Cancer J 2005;11:255-267)
KEY WORDS
genetic testing, cancer genetics, risk assessment, pediatric cancer
G eneticists focus on the pattern of congenital anomalies associated with an inherited predisposi- tion to cancer (e.g., Beckwith-Weidemann syn- drome) or the pattern of cancer in the family that places a healthy person at increased risk for cancer. However, the pediatric oncologist typically enters the scene once a child has been diagnosed with cancer. In addition, de novo mutations in the egg or sperm can result in a child with a hereditary predisposition to cancer in the absence of a family history of it. Thus, the pediatric oncologist should decide for each patient newly diagnosed with cancer whether there is a sufficient likelihood of cancer susceptibility for the patient and/or immediate relatives to warrant a ge- netics evaluation. Recent advances in genetic testing and studies of cohorts of patients with cancer have demonstrated the likelihood of identifying a genetic susceptibility mutation for certain childhood can- cers. The initial American Society of Clinical Oncol- ogy guidelines for the consideration of genetic testing suggested that if there were at least a 10% likelihood of identifying a mutation, testing was war- ranted.1 Although this is a somewhat arbitrary cutoff, in this article, we review the childhood malignancies that are associated with at least a 10% likelihood of being caused by a genetic susceptibility to cancer (Table 1) and therefore warrant consideration for a genetic evaluation. At the end of the article, we also summarize the other factors that indicate the need for a genetic evaluation, including the pattern of can- cer in the family and the association of cancer with congenital anomalies.
| TABLE 1 Family History | |
|---|---|
| Diagnosis | Genetic Loci |
| Retinoblastoma | RB1 |
| Adrenocortical carcinoma | p53 |
| Pheochromocytoma/ | VHL, NF1, RET, SDHB |
| paraganglioma | SDHC, SDHD |
| Retinal or cerebellar hemangioblastoma | VHL |
| Endolymphatic sac tumors | VHL |
| Optic pathway tumor | NF1 |
| Medullary thyroid cancer | RET |
| Atypical teratoid and malignant rhabdoid | INI1/SNF5 |
| tumor | |
| Acoustic or vestibular schwannomas | NF2 |
For each malignancy listed, there is greater than a 10% likeli- hood of identifying a constitutional mutation in the genetic loci indicated.
RETINOBLASTOMA
Because so much of our knowledge of inherited can- cer predisposition was gained from the study of retinoblastoma, it is a malignancy that most pediatric oncologists recognize as being due to an inherited mutation. Overall, approximately 40% of children with retinoblastoma carry a constitutional or germ- line mutation in the RB1 gene.2 This includes essen- tially all of the bilateral cases and approximately 15% of children with unilateral retinoblastoma. As Knud- son first noticed, bilateral retinoblastoma presents at an earlier age than unilateral.3 In contrast, recent data suggest that the age of onset of unilateral cases is not necessarily predictive of whether the child carries a constitutional mutation.4 Later-onset unilateral cases can be due to constitutional mutations that confer a lower cancer risk and are referred to as attenuated or low penetrant retinoblastoma.5 Thus, we recommend that all children diagnosed with retinoblastoma re- ceive a genetics evaluation, regardless of age of onset.
Surprisingly, despite the fact that essentially all children with bilateral retinoblastoma carry a consti- tutional mutation, only 10%-20% of these children have a family history of the disease.6 In the remaining 80% of cases, a de novo mutation in the RB1 gene oc- curred during gametogenesis, primarily during sper- matogenesis. This is a common phenomenon in dominant disorders affecting children, and physicians should not dismiss the possibility of a cancer predis- position syndrome based on a negative family history. Similarly, parents are often very surprised and con- fused about their child’s diagnosis of an inherited
form of cancer because the family history is negative; in these cases, the possibility of de novo mutations should be explained. In particular, the physician should emphasize to parents that these spontaneous mutations are not related to events that may have oc- curred around the time of pregnancy.
The gene mutated in retinoblastoma, RB1, was ini- tially localized based on cytogenetically visible dele- tions and subsequently cloned.7 Once the gene was cloned, molecular studies of patient samples allowed confirmation of Knudson’s two-hit hypothesis. In general, the tumor specimens reveals total loss of RBl function, often due to mutation in one copy and then either loss of the wild-type copy or silencing of the second copy by methylation.+
The most common mechanisms by which the sec- ond copy is inactivated are loss of the whole chromo- some, large deletions, and gene conversion, normally resulting in loss of heterozygosity (LOH) for markers near the RB1 locus. It is important to realize that in both sporadic and inherited forms the retinoblastoma tumor demonstrates loss of RB1 function. The key difference between the two situations is the timing of the first mutation. If the mutation was inherited or constitutional, it is present in all cells, leading to an increased risk of bilateral or multifocal disease. If it occurs sometime after conception, the mutation is a somatic event that may be limited to one eye. The normal function of the RB1 gene product is to nega- tively regulate the cell division cycle, a process re- ferred to as tumor suppression.8
Despite the gene name, the RB1 gene is expressed in most cells of the body, and therefore, it is not sur- prising that mutations in the RB1 gene result in an in- creased risk of other malignancies.9 It has now been clearly documented that individuals with constitu- tional mutations of the RB1 gene (primarily bilateral cases) have an increased risk of second primary ma- lignancies throughout their lives. For example, a re- cent long-term study of a British cohort of patients with retinoblastoma identified a 68% cumulative inci- dence of second cancers, including many epithelial cancers, such as lung cancer by age 84.10
There are several important reasons why a child di- agnosed with retinoblastoma should undergo a ge- netic evaluation. In particular, parents need to be informed about their risk of having another child with retinoblastoma, as well as the availability of pre- natal diagnosis and the need for appropriate screen- ing of siblings. Table 2 gives the risk of having a second child with retinoblastoma for parents of a child newly diagnosed with either unilateral or bilat- eral retinoblastoma.6 Unaffected parents of a child with bilateral retinoblastoma have a 6%-7% risk of having a second affected child. This residual risk is
| Clinical Scenario | Retinoblastoma Risk |
|---|---|
| Offspring of bilateral cases | 45% |
| Offspring of unilateral cases | 7.5% |
| Sibling of bilateral cases (with unaffected parents) | 5%-7% |
| Sibling of unilateral cases (with unaffected parents) | 1% |
due to the potential for germline mosaicism in a par- ent; typically resulting in a variable percentage of sperm or eggs carrying the mutation. If a mutation is identified in an affected member of the family, prena- tal diagnosis is available. If prenatal testing is not pur- sued, then siblings of the proband should undergo a careful eye examination at birth, and a blood sample should be sent for analysis of the specific mutation found in the affected child. This allows the clinician to identify the siblings who inherited the mutation and eliminate the need for examinations under anes- thesia in most of the siblings who did not inherit the mutation.11 If genetic testing is not possible, then all siblings of children with bilateral retinoblastoma should undergo ophthalmic surveillance, beginning at birth. Current recommendations for ophthalmic surveillance include examination in the first few days of life and then serial examinations under anesthesia every 3-4 months until the child is 3-4 years of age. Although some centers have recommended genetic evaluations only for bilateral cases, important infor- mation is also obtained for unilateral cases. If testing successfully documents a somatic mutation (an RB1 mutation in the tumor that is not found in the DNA from the blood), one can tell the parents that (1) the risk for subsequent malignancy in the child is not sig- nificantly increased, and (2) the parents and subse- quently the patient do not have an increased risk of having another child with retinoblastoma. However, it is recommended to continue surveillance of the contralateral eye in the proband because the somatic mutation may have occurred early enough in em- bryogenesis to affect both eyes.12
Molecular testing for mutations in the RB1 gene using various methods is now clinically available.13 Mutations are scattered throughout the gene. Thus, one should first test a sample from the affected child using methodologies that screen for the entire RB1 gene. In children with unilateral disease, most labora- tories first screen the retinoblastoma tumor specimen for mutations, and then if one is identified, the labo-
ratory looks for the same mutation in a blood sample. This is not necessary for bilateral cases, for which testing is performed directly from a blood sample. In either situation, mutation testing for retinoblastoma is not always informative, and it is recommended that the parents have detailed pre- and posttest genetic counseling in order to interpret and explain the results.
ADRENOCORTICAL CARCINOMA
Li Fraumeni syndrome (LFS) is associated with auto- somal dominant inheritance of cancer susceptibility for tumor types that include sarcomas, breast cancers, leukemia, brain tumors, and adrenocortical carci- noma (ACC).14 More than 80% of families meeting the classic criteria for LFS carry a mutation in the p53 gene.15 This observation has led to a large number of studies examining the p53 mutation status of children diagnosed with an LFS-associated cancer. Approxi- mately 5%-10% of children with sarcomas carry p53 mutations, and most of these mutation-positive chil- dren have some family history suggestive of LFS.16,17 In contrast, recent studies of children with ACC doc- ument that 50%-80% have constitutional p53 muta- tions, frequently without significant family histories of cancer.18-20 This is likely due to two factors, the first being a relatively high incidence of de novo p53 mu- tations in the child diagnosed with ACC. Second, there are specific p53 mutations that result in a signif- icant risk of ACC but a lower overall cancer risk, as documented in a study of patients with ACC in Brazil that demonstrated that most cases carried a specific mutation, R337H, but did not have family histories consistent with LFS.21 The R337H mutation repre- sents a lower-penetrance p53 mutation that imparts a distinct susceptibility to adrenocortical carcinoma. However, within other populations (including the United States and Europe), mutations in the p53 gene seen in patients with ACC are similar to those seen in other LFS families with significant risk of sarcomas, brain tumors, and breast cancer.22 Overall, given this high prevalence of p53 mutations in children with ACC, a genetic evaluation is warranted. If a p53 mu- tation is identified, it is especially important to deter- mine whether the mother (or other adult female relatives) carries the mutation, given the very signifi- cant risk of early-onset breast cancer (average age of diagnosis is 34 years in p53 mutation carriers) and the recommendation to institute breast cancer screening in the early twenties in those found to be positive.23 New screening modalities, including breast magnetic resonance imaging (MRI), should be considered.24 Similarly, the parents should be made aware of the es- timated 15% risk for diagnosis of a second primary
malignancy in mutation carriers.25 The patient’s pedi- atric oncologist must consider whether any new le- sion represents metastases from the primary ACC or a second primary malignancy. Similarly, in the future, the patient’s parents and primary physician should initiate a thorough evaluation of any new medical problem that could be consistent with a malignancy. Given the variety of different tumor types associated with LFS, there has not been a consensus on a screen- ing regimen for children with p53 mutations.26
ATYPICAL TERATOID AND MALIGNANT RHABDOID TUMORS
Malignant rhabdoid tumor (MRT) of the kidney is a rare, aggressive childhood cancer that histologically may resemble benign or malignant skeletal muscle cells.27 Approximately 10%-15% of rhabdoid tumors of the kidney in infants are associated with separate primary tumors of the central nervous system (CNS). The CNS tumors resemble primitive neuroectodermal tumors (including medulloblastoma or pineoblas- toma) but are now called “atypical teratoid/rhabdoid tumor” (ATRT).28
Cytogenetic analyses of ATRT of the CNS and MRT of the kidney revealed abnormalities of chromosome 22, including monosomy 22, deletion, or transloca-
tions centered around 22q11.2.29-32 In 1998, the hSNF5/INI1gene was isolated from chromosome band 22q11.2, and mutations in this gene were identified in MRTs and ATRTs.33 Analyses of tumor tissue revealed that virtually all MRT/ATRTs examined have mutations in hSNF5/INI1.34 Consistent with other tumor suppres- sor genes, the second copy of the gene is often lost or in- activated consistent with a two-hit mechanism. Figure 1 demonstrates monosomy 22 in a child with ATRT. The remaining chromosome 22 typically contains a point mutation in the hSNF5/IINI1 gene. hSNF5/INI1 encodes a protein that is part of a multiprotein complex involved in chromatin remodeling, an essential process for the regulation of gene expression.
Surprisingly, approximately 20% of patients with apparently sporadic tumors harbored constitutional hSNF5/INI1 mutations.35 Thus, these apparently spo- radic cases were actually due to constitutional muta- tions of the gene, an example of new dominant mutations. A familial pattern is not appreciated be- cause the cancer susceptibility derives from de novo mutations in the hSNF5/INI1 gene, and there is a high mortality in affected children diagnosed with MRT or ATRT, so the mutation is not passed on. Inheritance of an hSNF5/INI1 mutation results in “rhabdoid predis- position syndrome,” which includes renal and ex- trarenal MRT, choroid plexus carcinoma, central
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
-
19
20
21
22
x
Y
primitive neuroectodermal tumor, and medulloblas- toma.35 This same paradigm of de novo constitutional mutations in a gene that results in a significant risk of a tumor with high mortality may apply to other rare pediatric tumors.
Genetic evaluation of a child diagnosed with an ATRT and/or MRT is recommended to determine the presence of constitutional mutations, although we do not yet have long-term follow-up information on chil- dren with an hSNF5/INI1 mutation to predict sub- sequent cancer risk. Even in families in which unaf- fected parents are negative for the mutation, there is still some theoretical risk for having a subsequent child with the mutation. This risk derives from the potential for germline mosaicism in one of the parents that is similar to that seen in unaffected parents of a child with retinoblastoma. The identity of the specific hSNF5/INI1 mutation in the affected child provides the ability for prenatal or postnatal diagnosis of sib- lings to determine whether they have inherited an in- creased risk for malignancy.
OPTIC PATHWAY TUMORS
Optic pathway tumors (OPTs) are highly associated with a diagnosis of neurofibromatosis type 1 (NF1). Several studies of cohorts of children with OPTs have demonstrated that approximately 40%-50% meet the diagnostic criteria for NF1.36 Conversely, 20% of chil- dren with NF1 have evidence of an OPT on neu- roimaging.37 The onset of OPTs in NF1 is typically before 6 years of age, so it is not unusual for an OPT to be diagnosed before the recognition of a diagno- sis of NF1 in the child.38 Overall, NF1 is the result of a de novo mutation in approximately 50% of chil- dren, so again a negative family history for NF1 is common. The diagnostic features of NF1 also have different ages of onset.39 In very young children with an OPT, the only other diagnostic feature may be café au lait spots or macules. Pediatricians may over- look café au lait spots in an otherwise healthy infant or toddler. Therefore, it is incumbent on the pedi- atric oncologist who is newly diagnosing a child with an OPT to carefully evaluate the child and family for the question of NF1. This should include a careful skin examination, noting the presence and number of café au lait spots, evidence for axillary or inguinal freckling (which typically presents around age 3-5 years), plexiform neurofibromas, or pseudoarthrosis of the lower limb. Similarly, the parents should un- dergo a skin examination and an ophthalmic exam- ination, in which evidence of Lisch nodules is sought. Lisch nodules are pigmented lesions in the iris, which do not affect vision but are found in more than 90% of adults with NF1.39 Even though by adulthood, there are often multiple features of NF1 that are ap-
preciable on routine physical examination, it is not unusual for an affected parent to be first recognized to have NF1 when the child presents with a serious complication of the disease.
NF1 is the result of carrying a mutation in the NF1 gene that encodes the neurofibromin protein.40 For many years after the isolation of the NF1 gene, molec- ular testing was not readily available because of the gene’s large size and the variations in mutation type. However, with the development of new sequence- based techniques that efficiently screen the entire NF1 coding region, clinical molecular testing is now available and can identify the causative mutation in more than 90% of patients with the clinical diagno- sis of NF1.41 Thus, we would recommend perform- ing molecular testing for NF1 mutations in a young child with an OPT who has a negative family his- tory and does not yet have any other diagnostic fea- ture of the disease.
Making a diagnosis of NF1 has implications for the child and the family. NF1 is a pleomorphic disease that places the child at increased risk for a number of other medical problems, including skeletal prob- lems (pseudoarthrosis of the lower limb and scolio- sis), plexiform neurofibromas that can be deforming, other central nervous system tumors, and learning disabilities. Therefore, a child with NF1 requires yearly follow-up in a multidisciplinary neurofibro- matosis clinic in addition to appropriate treatment of the OPT. In addition, studies suggest that children with NF1 have an increased risk of vascular com- plications from radiation therapy in comparison to children without NF1 who have brain tumors.42 The natural history of OPT in NF1 appears to be more indolent in comparison with that of children with- out NF1, although this has not been carefully stud- ied. There are reports of regression, and many children with NF1 and OPT that is detected by rou- tine screening have stable disease over a number of years.38 Therefore, the diagnosis of NF1 should be taken into consideration when decisions about the appropriate treatment of a child with NF1 and OPT are being made.
In many tumors associated with a significant risk of heritable susceptibility, including retinoblastoma and ATRT, similar molecular mechanisms appear to be involved in both the sporadic and the inherited cases. This does not appear to be the case in OPT. Because most children are diagnosed and treated on the basis of imaging studies, there have been only limited molecular studies performed on OPT tissue from patients with and from those without NF1. LOH analysis at the NF1 locus was performed in one study. Surprisingly, the researchers found that there is a distinct difference in pattern of LOH be- tween the two groups of tumors, with more than
90% of NF1-associated tumors having LOH at the NF1 gene but only one of 24 sporadic tumors demon- strating LOH.43 This difference may result from two distinct possibilities: (1) different molecular pathways may be disrupted in the development of OPT in children with NF1 versus sporadic cases, or (2) the same pathway may be targeted, but different genes that regulate this pathway may be disrupted in the two groups, as seen for juvenile myelomono- cytic leukemia. The distinction between these two possibilities will aid researchers in determining whether these two groups of patients with OPT may respond differently to treatments targeted at the NF1 pathway.
If a child is diagnosed with NF1, then additional evaluation of the parents is indicated to determine whether this is a de novo case or whether the parents are affected and therefore other siblings may be af- fected as well. If both parents are documented to be unaffected (by eye and skin examinations), then the likelihood of other siblings having NF1 is very small because germline mosaicism appears to be an infre- quent event in NF1 (in comparison with neurofibro- matosis type 2 [NF2]). Of course, the parents should be counseled that when the child reaches reproduc- tive age, he or she should have genetic counseling be- cause of the 50% risk with each pregnancy of transmitting the disease.
JUVENILE MYELOMONOCYTIC LEUKEMIA
Juvenile myelomonocytic leukemia is a rare subgroup of leukemia that was previously referred to as juvenile chronic myelocytic leukemia. Many studies have dem- onstrated that almost 15% of patients with JMML have the diagnosis of NF1.44 However, in contrast to the sig- nificant risk of OPT in patients with NF1, the absolute risk of JMML in a child with NF1 is quite low (esti- mated at 1% or less). The age of onset is highly vari- able; in one series, 11 children with NF1 and JMML were diagnosed from 7 months to 7 years of age.45 In the younger children, the diagnosis of NF1 may not have previously been made and a careful examination for features of NF1 should be included at the time of diagnosis of JMML.
In the case of JMML, careful molecular studies have revealed that the same molecular pathway is affected in NF1-associated and sporadic cases. The neurofibromin protein has guanosine triphosphate activating protein activity and negatively regulates the activity of the Ras proteins. In sporadic cases, other aspects of the Ras pathway are mutated. The H-ras gene itself has activating mutations (in codons 12 and 13) in approximately 15% of tumors.46 Another 20%-30% have somatic activating mutations in the
SHP2 gene, which normally regulates the communi- cation between activated growth factor receptors and Ras.47 These mutations result in constitutive activa- tion of Ras in the absence of ligand signaling. In other cases of sporadic JMML, no specific mutation has been identified but there are a number of genes in the Ras pathway that have not yet been interrogated for mutations. The consistent mutation of the Ras path- way mediated by granulocyte-macrophage colony stimulating factor in JMML has resulted in phase I/II trials with targeted therapies for patients. 48
MALIGNANT PERIPHERAL NERVE SHEATH TUMORS
The clearest associations between NF1 and pediatric malignancies are the increased risk of optic gliomas and malignant peripheral nerve sheath tumors (MPN- STs).49,50 Population-based studies of children diag- nosed with cancer from Japan and the Netherlands revealed that approximately 50% of patients with MPNST had NF1.48,51 The age of diagnosis of MPNST is from adolescence through adulthood. For patients with NF1, the MPNST may result from the malignant transformation of a long-standing plexiform neurofi- broma or may be independent. Molecular studies of sporadic MPNSTs also demonstrate a high rate of dis- ruption of the NF1 gene in the tumor, suggesting that a similar mechanism is involved.52 The treatment op- tions are similar in the two groups and primarily focus on complete excision. As described earlier for the other NF1-associated tumors, the diagnosis of NF1 in a patient with MPNSTs allows appropriate counseling with regard to other NF1-associated med- ical problems, as well as recurrence risk for the af- fected patient and potential recognition of the disease in other family members.
VESTIBULAR SCHWANNOMAS
Bilateral vestibular schwannomas are the hallmark of NF2.39 NF2 is a much less common disorder then NF1. Most of the manifestations of NF2 occur in adulthood, but unilateral or bilateral vestibular schwannomas can be diagnosed during childhood, in particular, in children who carry a nonsense or trun- cating mutation in the NF2 gene.53 Other findings in NF2 include small numbers of café au lait spots, neu- rofibromas of the spine, and meningiomas.39 Thus, any child with either a vestibular schwannoma or multiple meningiomas should be evaluated for the presence of NF2. Treatment modalities include mi- crosurgery, radiosurgery, and radiation therapy.
NF2 is caused by mutations in the NF2 gene, which was isolated by positional cloning methods and
encodes the protein alternately called Merlin or Schwannomin.53 Typically, screening for NF2-associ- ated tumors begins in adolescence; however, any fam- ily that has demonstrated early-onset tumors in one member should initiate screening earlier. In addition, no features of NF2 may be identified on routine phys- ical examination, and so parents should be screened by MRI for NF2-associated tumors to determine whether they are affected.39 Genetic testing is avail- able for mutations in the NF2 gene, although the sen- sitivity of the test is only approximately 65%.54 Referral to a geneticist or genetic counselor is recom- mended because germline and somatic mosaicism is very common in patients with NF2, complicating the interpretation of test results.55
ENDOLYMPHATIC SAC TUMORS
Endolymphatic sac tumors (ELSTs) are locally inva- sive adenocarcinomas (also termed aggressive papil- lary cystadenomas), which arise in the temporal bone of the middle ear. ELSTs can cause hearing loss, tin- nitus, vertigo, and aural fullness.56 Occasionally, pa- tients can have paralysis of the facial (7th) cranial nerve. The hearing loss can be acute in onset and pro- found in nature, as seen in 43% of patients. ELSTs can present as either sporadic tumors or in association with von Hippel-Lindau disease (VHL); in VHL, ELSTs are more likely to be bilateral.57 Patients with VHL have a younger average age of onset (31.3 years) than patients with sporadic tumors (52.5 years).58 The youngest patient reported with ELST was 4 years old and carried a de novo missense mutation in the VHL gene so the presentation of an ELST may be the first feature of VHL syndrome and the presence of an ELST in both children and adults necessitates an evaluation for VHL disease. Conversely, an ELST was identified in 16% (21/129) of patients with known VHL when the patients were evaluated by history, physical examination, imaging (MRI and CT), and a comprehensive audiologic assessment.56
The VHL gene is localized to chromosome 3q25-26 and was identified in 1993 with the use of positional cloning.59 The VHL protein is expressed in most tis- sues and has been implicated in a variety of functions, in particular the regulation of hypoxia-inducible genes and angiogenesis, and fibronectin matrix as- sembly.60 Mutations in VHL are identified in virtually every case of VHL, making genetic testing the gold standard diagnostic test for VHL.61 Multiple types of mutations have been identified in VHL, with partial or complete gene deletions accounting for up to 40% of the mutations identified.61 In ELSTs from patients with VHL, somatic loss of the wild-type gene is found, consistent with the two-hit hypothesis, and somatic
mutations in the VHL gene have been identified in pa- tients with sporadic ELSTs.57
The identification of a VHL mutation in a child has important implications for both the child and family members. Standard screening beginning in childhood is recommended for patients with VHL; the screening evaluates for retinal, cerebellar, and spinal heman- gioblastomas; renal cancer; and pheochromocytomas (Table 3) so that the manifestations of VHL can be caught at an early stage and proper medical care, based on VHL guidelines, can be instituted.62 Patients with VHL can present at all ages, so parents of any child diagnosed with VHL should also be evaluated for the possibility of having undiagnosed VHL.
HEMANGIOBLASTOMA
Hemangioblastomas are well-encapsulated tumors composed of an extensive vascular network and stro- mal cells, typically filled with vacuoles composed of cystic and solid components; they produce symptoms from pressure on adjacent structures and hemorrhage.
| Tumor Type | Screening Recommendations | Age and Frequency |
|---|---|---|
| Retinal heman- giomas | Ophthalmoscopy | From infancy; annually |
| Pheochromocy- tomas | Plasma metanephrines or quantitative urinary cate- cholamines | From age 2; annually |
| Spinal and cere- bellar heman- gioblastomas | MRI with and without con- trast of brain and spine | From age 11; every 2 years |
| Renal cell cancer | Abdominal ultra- sound Abdominal CT | From age 11; annually From age 20; every 1-2 years |
| Endolymphatic sac tumors | MRI of inner auditory canals | Begin imaging if clinical signs or symptoms (hearing loss, vertigo, tinni- tus) appear |
Abbreviations: CT, computed tomography; MRI, magnetic reso- nance imaging.
Adapted from Lonser et al.62
Hemangioblastomas can arise in the retina, cerebel- lum, brainstem, and spinal cord.63 Hemangioblas- tomas are rare tumors, accounting for approximately 2% of all CNS tumors, and are found mainly in chil- dren and young adults. There are three types of clini- cal presentation of hemangioblastoma: multiple or single in association with VHL, multiple in a localized area without other signs of VHL, and solitary.
Approximately one quarter of patients with he- mangioblastomas have VHL; the remaining 75% are sporadic. CNS hemangioblastomas are the most com- mon manifestation of VHL, arising in 60%-80% of pa- tients.63 Retinal hemangioblastomas can present at any age, and screening for these lesions in patients with VHL starts at infancy.62 Although the mean age of presentation is 25 years, 5% of retinal heman- gioblastoma are diagnosed by age 10. However, CNS hemangioblastomas do not usually arise until late childhood (ages 9-12 years) in VHL, and therefore screening is not initiated until 11 years of age. The ge- netics of VHL have been reviewed in the previous sec- tion. Even after thorough evaluation revealing no other signs of VHL, patients with apparently sporadic hemangioblastomas still have a 4%-10% likelihood of carrying a constitutional mutation in the VHL gene.64
Because either retinal or CNS hemangioblastomas are commonly the presenting manifestation of VHL, the presence of a hemangioblastoma in a child necessi- tates evaluation for VHL, even though other manifes- tations of VHL may not be apparent. Genetic testing for VHL mutations should be performed, given the high sensitivity of the test. The complications of he- mangioblastomas in VHL are severe and can include acute hemorrhage, resulting in obstructive hydro- cephalus, visual loss, brainstem compression or herni- ation, blindness, and paralysis, so it is important that proper screening be instituted. Typically, retinal he- mangiomas are treated with laser therapy, and spinal and cerebellar hemangioblastomas are excised if they show progression and development of symptoms.62 As noted earlier, VHL disease also includes renal cancer and pheochromocytomas, among other manifesta- tions; again, the standard screening recommendations for VHL reduce the morbidity from these diseases, and so it is crucial to make the diagnosis of VHL.
MEDULLARY THYROID CANCER
The multiple endocrine neoplasia (MEN) Type 2 syn- dromes (MEN2A and MEN2B) are autosomal domi- nant cancer family syndromes that affect endocrine organs and can present in childhood. MEN2A is asso- ciated with medullary thyroid carcinoma, parathyroid adenomas, and pheochromocytomas. MEN2B is a re- lated disorder but with the onset of tumors in infancy,
ganglioneuromas of the gastrointestinal tract, and skeletal abnormalities. Additional families appear to show autosomal dominant medullary thyroid car- cinoma without the other features of MEN2A. Medullary thyroid cancer is a very rare malignancy in children, and the possibility of MEN2A or B should be considered in any child who presents with medullary thyroid cancer, parathyroid adenomas, and/or pheo- chromocytomas.65
The identification of mutations in the RET onco- gene on chromosome 10q11.2 as the cause for MEN2A, MEN2B, and familial medullary thyroid car- cinoma was the result of large-scale genetic mapping studies in MEN2 families.66 RET encodes a receptor tyrosine kinase gene, and mutations in the RET onco- gene in MEN syndromes was the first example of in- heritance of activating mutations in an oncogene as a cause for inherited cancer susceptibility.67 Analysis of the RET gene from constitutional DNA from multiple MEN2A families revealed a set of highly consistent mutations in conserved cysteine residues encoded in exons 10 and 11. In addition, specific mutations are associated with specific cancer risks. For example, mutations in codon 634 impart a high risk for pheochromocytoma.68 More than 95% of children with MEN2B carry one of two specific missense mu- tations in the tyrosine kinase domain.69
The identification of a RET mutation in an affected child has significant impact for the child and other family members. Children with RET mutations are clearly at increased risk for other MEN-associated tu- mors (parathyroid adenomas and pheochromocy- tomas) and require lifelong surveillance. It is also very important to evaluate other first-degree family mem- bers (including parents) to determine whether they also carry this mutation and may need prophylactic thyroidectomy and cancer screening. Thyroidectomy is recommended by age 5 for children with MEN2A and by age 1 for those with MEN2B.70 Therefore, all patients at risk for MEN2 should undergo molecular analysis of the RET gene. It is preferable to first per- form RET mutation testing on an affected individual to identify the specific mutation present in the family; subsequent family members can then be tested for this specific mutation.
PHEOCHROMOCYTOMAS AND PARAGANGLIOMAS
Tumors of the autonomic nervous system include (1) tumors of the sympathetic branch, termed pheochro- mocytomas, when they are located in the adrenal gland and extra-adrenal pheochromocytomas when they are not within the adrenal gland; (2) tumors of the parasympathetic branch, termed paragangliomas
named for their location. Pheochromocytomas are as- sociated with multiple genetic conditions, including NF1, MEN2A, VHL, and familial pheochromocy- toma/paraganglioma. Pheochromocytomas and para- gangliomas are rarely observed in children but when present are frequently associated with a genetic pre- disposition. Pheochromocytomas are seen at an in- creased frequency (0.1%-5.7%) in patients with NF1 and should be considered in the diagnosis of a child with NF1 and hypertension.71 Although pheochromo- cytomas associated with MEN2A are most commonly diagnosed between the ages of 30 and 40 years, they also have been rarely observed in children. In general, medullary thyroid carcinoma is diagnosed before pheochromocytoma; however, they are the first man- ifestation of MEN2A in 9%-27% of cases.72 Pheochro- mocytomas also can be diagnosed in children with VHL; in 64 patients with VHL with pheochromocy- tomas, the mean age of diagnosis was 29.9 years, with a range of 6-54 years.73 VHL is divided into subtypes based on the risk of developing pheochromocytoma.62 Patients with truncating or null mutations generally have VHL type 1 and a very low risk of pheochro- mocytoma. Patients with VHL type 2 have primarily missense mutations and a significant risk of pheochromocytoma. Families with type 2A do not demonstrate renal cell cancer; families with type 2B also have a risk of renal involvement, and those with type 2C only manifest pheochromocytoma (including in childhood). In particular, individuals with muta- tions at nucleotides 595 and 695 present with pheochromocytomas at a younger age than those with other mutations.72 Both isolated pheochromocytomas and/or paragangliomas are associated with germline mutations in members of the succinate dehydroge- nase (SDH) gene family, SDHB, SDHC, and SDHD,74 that comprise three of the four subunits of complex II (succinate:ubiquinone oxidoreductase) of the mito- chondrial electron transport chain. Mutations in SDHD are associated with disease when they are in- herited from the father, which may be the result of genetic imprinting.75
Among patients with nonsyndromic pheochromo- cytoma diagnosed under age 20 (excluding those with known clinical diagnoses of NF1, VHL, or MEN 2A), 31 of 57 (54%) had germline mutations: 23 in VHL, zero in RET, three in SDHD, and five in SDHB, high- lighting the need for all children with pheochromocy- toma to have a genetic evaluation.76 Consistent with the high frequency of germline mutations identified in children with pheochromocytoma, the tumors in children are more likely to be extra-adrenal, bilateral, and multifocal than those in adults. The importance and familial impact of identifying mutations in VHL and RET were reviewed in this paper. As with VHL
and MEN2, individuals who carry mutations in SDHD, SDHC, or SDHB are at risk for pheochromocy- toma and paraganglioma for the rest of their lives and need regular screening, which includes annual plasma metanephrines and head/neck and abdominal MRIs. Given the risk for multiple pheochromocytomas and paragangliomas, it is crucial that a skilled endocrine surgeon with experience in partial adrenalectomy and an anesthesiologist familiar with appropriate adrener- gic blockade treat the patient.
OTHER INDICATIONS FOR GENETIC EVALUATION
In the previous sections, we reviewed the cancer di- agnoses that in isolation are associated with at least a 10% likelihood of a constitutional mutation. In addi- tion, other items in the history and physical examina- tion should prompt the pediatric oncologist to initiate a genetics evaluation.
Congenital Anomalies
Certain genetic syndromes are associated with both multiple congenital anomalies (or physical findings) as well as increased cancer risk. Well-documented ex- amples include radial ray anomalies in children with Fanconi’s anemia, hemihypertrophy, and organo- megaly with Wilm’s tumor in Beckwith-Weidemann syndrome; poikiloderma and osteosarcoma associated with Rothmund Thomson syndrome; and palmar pits and odontogenic cysts with medulloblastoma in Gor- lin’s syndrome. If congenital anomalies or unusual physical findings are discovered on the initial physi- cal examination, then the pediatric oncologist should try to identify whether there is a relationship between these features and the cancer diagnosis. This can in- clude searching the medical literature for an associa- tion of the specific anomaly and the cancer diagnosis, as well as using free online databases of genetic con- ditions, including the Online Mendelian Inheritance of Man (http://www.ncbi.nlm.nih.gov/entrez/query. fcgi?db=OMIM&itool=toolbar).
Although much of this article has focused on auto- somal dominant conditions, many of the autosomal recessive cancer predisposition syndromes result in multiple congenital anomalies, dermatologic features, and skeletal findings. Because carrier parents are typ- ically healthy and in most Western cultures, family size is small and consanguinity is rare, it is not un- usual for the child with an autosomal recessive disor- der to be the only affected member of the family. Thus, even in the absence of family history, the diag- nosis of cancer in a child with several other physical findings or congenital anomalies should certainly
prompt a genetics evaluation to determine whether there is a specific underlying syndrome. Depending on the disorder in question, specific molecular or phenotypic testing can confirm the diagnosis. These tests are often specific to the disorder, such as diepoxybutane chromosome breakage assay in Fan- coni anemia, and thus, the clinician or genetics con- sultant needs to make a presumptive diagnosis based on clinical grounds in order to determine the appro- priate tests to order.
Although autosomal recessive disorders are rare, it is particularly important for pediatric oncologists to consider these diagnoses because of the potential for significantly increased toxicity to standard treatment regimens. Table 4 lists inherited conditions (primarily autosomal recessive conditions, except for the auto- somal dominant Gorlin’s syndrome), in which in- creased sensitivity/toxicity to cancer treatment has been reported and adjustment of the treatment regi- men is indicated to avoid complications or treatment- related deaths.
For children with multiple congenital anomalies, developmental delay, or physical findings that do not conform to a specific condition, it is often useful to per- form a peripheral blood karyotype to determine whether the child carries a novel translocation or dele- tion that might be responsible for the condition and to consider a genetics evaluation to more thoroughly eval- uate the etiology of the different clinical features and determine the need for further medical management.
Family History
Most cancer predisposition syndromes are autosomal dominant. Thus, one is more likely to see a positive family history in either the maternal or the paternal lineage, which may go back several generations, de- pending on when the mutation arose. Given that par- ents of a child diagnosed with cancer are often young adults and may be unaffected, it is worthwhile to briefly ask about the proband’s aunts and uncles, grandparents, and first cousins. We recommend
| Condition | Major Clinical Features | Diagnostic Test | Treatment Requiring Adjustment |
|---|---|---|---|
| Ataxia-telangiectasia | Cerebellar ataxia, telangiec- tasias, immunodefi- ciency, leukemias, lymphomas and solid tumors | Increased alpha-fetopro- tein, sensitivity to ioniz- ing radiation | Radiation therapy, chemotherapeutic agents that produce double- strand breaks |
| Nijmegen breakage syndrome | Microcephaly, immunodefi- ciency, developmental delay, lymphomas | Sensitivity to ionizing radia- tion, Polish founder mu- tation in NBS1 | Radiation therapy, chemotherapeutic agents that produce double- strand breaks |
| Ligase IV deficiency | Microcephaly, immunodefi- ciency, anemia, develop- mental delay, lymphomas | Sensitivity to ionizing radia- tion | Radiation therapy, chemotherapeutic agents that produce double- strand breaks |
| Fanconi anemia | Bone marrow failure, radial ray anomalies, microoph- thalmia, renal anomalies, bronzing of the skin | Chromosome breakage assay after exposure to diepoxybutane | Specialized conditioning regimen before bone marrow transplantation, sensitivity to cross-link- ing agents |
| Bloom syndrome | Short stature, butterfly rash on face, GI intoler- ance, immunodeficiency | Increased sister chromatid exchange | Increased sensitivity to chemotherapy, including 5-FU and cisplatin |
| Gorlin syndrome | Palmar pits, calcification of the falx, odontogenic cysts, BCCs, medul- loblastoma | Mutation analysis of the PTCH gene | Radiation therapy causes development of large numbers of BCCs in radi- ation field |
Abbreviations: BCC, basal cell carcinoma; 5-FU, 5-fluorouracil; GI, gastrointestinal.
specifically asking whether anyone has had cancer and the approximate age of diagnosis. Individuals with early-onset cancer or multiple malignancies are most indicative of cancer predisposition syndromes. In contrast, multiple individuals with diagnoses of cancer at more advanced ages (e.g., breast cancer in several women in their 70s) are less likely to indicate a genetic predisposition syndrome related to the child’s cancer diagnosis. The authors cannot empha- size enough that parents usually do not volunteer in- formation with regard to family history of cancer unless they are specifically asked because they may not realize that adult-onset cancers in relatives could be related to the cancer diagnosis in their child. As an example, a patient was treated and followed up for many years after the diagnosis of hepatoblastoma. It was only when the question of familial adenomatous polyposis was raised by an ophthalmologist who noted retinal lesions that the mother volunteered that her father had died of “colonic polyposis.” This prompted the diagnosis of familial adenomatous poly- posis in both the patient and her mother and urgent prophylactic colectomy in the mother. This case also highlights that it is the parents of a child who is newly diagnosed with a cancer predisposition syndrome who often benefit from the institution of the appro- priate screening or prophylactic procedures.
Despite the many different clinical and psychoso- cial issues that arise during the diagnosis of a child with cancer, it is worthwhile for the pediatric oncolo- gist to spend a few moments considering the possibil- ity of a genetic condition (based on the specific malignancy) and then examining the child and asking targeted questions with regard to family history in order to identify whether there is evidence for a can- cer predisposition syndrome and the need for further genetic evaluation. Similarly, approximately 5 years after the initial cancer diagnosis, the family history should be updated to determine whether intervening cancer diagnoses have occurred that now suggest a familial predisposition to cancer and the need for a genetic evaluation.
REFERENCES
1. Statement of the American Society of Clinical Oncology: ge- netic testing for cancer susceptibility. Adopted on February 20, 1996. J Clin Oncol 1996;14:1730-1736.
2. Gallie BL. Predictive testing for retinoblastoma comes of age. Am J Hum Genet 1997;61:279-281.
3. Knudson AG Jr. Mutation and cancer: statistical study of retinoblastoma. Proc Natl Acad Sci USA 1971;68:820-823.
4. Lohmann DR, Gallie BL. Retinoblastoma: revisiting the model prototype of inherited cancer. Am J Med Genet 2004;129C: 23-28.
5. Schubert EL, Strong LC, Hansen MF. A splicing mutation in RB1 in low penetrance retinoblastoma. Hum Genet 1997;100: 557-563.
6. Musarella MA, Gallie BL. A simplified scheme for genetic counseling in retinoblastoma. J Pediatr Ophthalmol Strabis- mus 1987;24:124-125.
7. Friend SH, Bernards R, Rogelj S et al. A human DNA segment with properties of the gene that predisposes to retinoblastoma and osteosarcoma. Nature 1986;323:643-646.
8. Sellers WR, Kaelin WGJ. Role of the retinoblastoma protein in the pathogenesis of human cancer. J Clin Oncol 1997;15: 3301-3312.
9. Eng C, Li FP, Abramson DH et al. Mortality from second tu- mors among long-term survivors of retinoblastoma. J Natl Cancer Inst 1993;85:1121-1128.
10. Fletcher O, Easton D, Anderson K et al. Lifetime risks of com- mon cancers among retinoblastoma survivors. J Natl Cancer Inst 2004;96:357-363.
11. Noorani HZ, Khan HN, Gallie BL et al. Cost comparison of molecular versus conventional screening of relatives at risk for retinoblastoma. Am J Hum Genet 1996;59:301-307.
12. Sippel KC, Fraioli RE, Smith GD et al. Frequency of somatic and germ-line mosaicism in retinoblastoma: implications for genetic counseling. Am J Hum Genet 1998;62:610-619.
13. Harbour JW. Overview of RB gene mutations in patients with retinoblastoma: implications for clinical genetic screening. Ophthalmology 1998;105:1442-1447.
14. Varley JM, Evans DG, Birch JM. Li-Fraumeni syndrome: a mo- lecular and clinical review. Br J Cancer 1997;76:1-14.
15. Birch JM, Hartley AL, Tricker KJ et al. Prevalence and diversity of constitutional mutations in the p53 gene among 21 Li-Frau- meni families. Cancer Res 1994;54:1298-1304.
16. McIntyre JF, Smith-Sorensen B, Friend SH et al. Germline mu- tations of the p53 tumor suppressor gene in children with os- teosarcoma. J Clin Oncol 1994;12:925-930.
17. Diller L, Sexsmith E, Gottlieb A et al. Germline p53 mutations are frequently detected in young children with rhabdomyosar- coma. J Clin Invest 1995;95:1606-1611.
18. Wagner J, Portwine C, Rabin K et al. High frequency of germline p53 mutations in childhood adrenocortical cancer. J Natl Cancer Inst 1994;86:1707-1710.
19. Varley JM, McGown G, Thorncroft M et al. Are there low-pen- etrance TP53 alleles? Evidence from childhood adrenocortical tumors. Am J Hum Genet 1999;65:995-1006.
20. Moutou C, Le Bihan C, Chompret A et al. Genetic transmis- sion of susceptibility to cancer in families of children with soft tissue sarcomas. Cancer 1996;78:1483-1491.
21. Ribeiro RC, Sandrini F, Figueiredo B et al. An inherited p53 mutation that contributes in a tissue-specific manner to pedi- atric adrenal cortical carcinoma. Proc Natl Acad Sci USA 2001;98:9330-9335.
22. Birch JM, Alston RD, McNally RJ et al. Relative frequency and morphology of cancers in carriers of germline TP53 mutations. Oncogene 2001;20:4621-4628.
23. Olivier M, Goldgar DE, Sodha N et al. Li-Fraumeni and related syndromes: correlation between tumor type, family structure, and TP53 genotype. Cancer Res 2003;63:6643-6650.
24. Leach MO, Boggis CR, Dixon AK et al. Screening with mag- netic resonance imaging and mammography of a UK popula- tion at high familial risk of breast cancer: a prospective multicentre cohort study (MARIBS). Lancet 2005;365:1769- 1778.
25. Hisada M, Garber JE, Fung CY et al. Multiple primary cancers in families with Li-Fraumeni syndrome. J Natl Cancer Inst 1998;90:606-611.
26. Malkin D. Predictive genetic testing for childhood cancer: tak- ing the road less traveled by. J Pediatr Hematol Oncol 2004;26:546-548.
27. Weeks DA, Beckwith JB, Mierau GW et al. Renal neoplasms mimicking rhabdoid tumor of kidney: a report from the Na- tional Wilms’ Tumor Study Pathology Center. Am J Surg Pathol 1991;15:1042-1054.
28. Rorke LB, Packer R, Biegel J. Central nervous system atypical teratoid/rhabdoid tumors of infancy and childhood. J Neu- rooncol 1995;24:21-28.
29. Fort DW, Tonk VS, Tomlinson GE et al. Rhabdoid tumor of the kidney with primitive neuroectodermal tumor of the central nervous system: associated tumors with different histologic, cytogenetic, and molecular findings. Genes Chromosomes Cancer 1994;11:146-152.
30. Besnard-Guerin C, Cavenee W, Newsham I. The t(11;22) (p15.5;q11.23) in a retroperitoneal rhabdoid tumor also in- cludes a regional deletion distal to CRYBB2 on 22q. Genes Chromosomes Cancer 1995;13:145-150.
31. Rosty C, Peter M, Zucman J et al. Cytogenetic and molecular analysis of a t(1;22)(p36;q11.2) in a rhabdoid tumor with a putative homozygous deletion of chromosome 22. Genes Chromosomes Cancer 1998;21:82-89.
32. Sawyer JR, Goosen LS, Swanson CM et al. A new reciprocal translocation (12;22)(q24.3;q11.2-12) in a malignant rhab- doid tumor of the brain. Cancer Genet Cytogenet 1998;101: 62-67.
33. Versteege I, Sevenet N, Lange J et al. Truncating mutations of hSNF5/INI1 in aggressive paediatric cancer. Nature 1998;394: 203-206.
34. Biegel JA, Zhou JY, Rorke LB et al. Germ-line and acquired mu- tations of INI1 in atypical teratoid and rhabdoid tumors. Can- cer Res 1999;59:74-79.
35. Sevenet N, Sheridan E, Amram D et al. Constitutional muta- tions of the hSNF5/INI1 gene predispose to a variety of can- cers. Am J Hum Genet 1999;65:1342-1348.
36. Janss AJ, Grundy R, Cnaan A et al. Optic pathway and hypo- thalamic/chiasmatic gliomas in children younger than age 5 years with a 6-year follow-up. Cancer 1995;75:1051-1059.
37. Lewis RA, Gerson LP, Axelson KA et al. von Recklinghausen neurofibromatosis: II. Incidence of optic gliomata. Ophthal- mology 1984;91:929-935.
38. Listernick R, Charrow J, Greenwald M et al. Natural history of optic pathway tumors in children with neurofibromatosis type 1: a longitudinal study. J Pediatr 1994;125:63-66.
39. Gutmann DH, Aylsworth A, Carey JC et al. The diagnostic evaluation and multidisciplinary management of neurofibro- matosis 1 and neurofibromatosis 2. JAMA 1997;278:51-57.
40. Gutmann DH, Collins FS. von Recklinghausen neurofibro- matosis. In: Scriver CR, Beaudet AL, Sly WS et al, eds. The Metabolic and Molecular Bases of Inherited Disease. New York: McGraw-Hill, 1995:677-696.
41. Messiaen LM, Callens T, Mortier G et al. Exhaustive mutation analysis of the NF1 gene allows identification of 95% of muta- tions and reveals a high frequency of unusual splicing defects. Hum Mutat 2000;15:541-555.
42. Guillamo JS, Creange A, Kalifa C et al. Prognostic factors of CNS tumours in neurofibromatosis 1 (NF1): a retrospective study of 104 patients. Brain 2003;126:152-160.
43. Kluwe L, Hagel C, Tatagiba M et al. Loss of NF1 alleles distin- guish sporadic from NF1-associated pilocytic astrocytomas. J Neuropathol Exp Neurol 2001;60:917-920.
44. Side LE, Emanuel PD, Taylor B et al. Mutations of the NF1 gene in children with juvenile myelomonocytic leukemia with- out clinical evidence of neurofibromatosis, type 1. Blood 1998;92:267-272.
45. Shannon KM, O’Connell P, Martin GA et al. Loss of the normal NF1 allele from the bone marrow of children with type 1 neu- rofibromatosis and malignant myeloid disorders. N Engl J Med 1994;330:597-601.
46. Flotho C, Valcamonica S, Mach-Pascual S et al. RAS mutations and clonality analysis in children with juvenile myelomono- cytic leukemia (JMML). Leukemia 1999;13:32-37.
47. Loh ML, Vattikuti S, Schubbert S et al. Mutations in PTPN11 implicate the SHP-2 phosphatase in leukemogenesis. Blood 2004;103:2325-2331.
48. Emanuel PD. Juvenile myelomonocytic leukemia. Curr Hema- tol Rep 2004;3:203-209.
49. Doorn PF, Molenaar WM, Buter J et al. Malignant peripheral nerve sheath tumors in patients with and without neurofibro- matosis. Eur J Surg Oncol 1995;21:78-82.
50. King AA, DeBaun MR, Riccardi VM et al. Malignant peripheral nerve sheath tumors in neurofibromatosis 1. Am J Med Genet 2000;93:388-392.
51. Matsui I, Tanimura M, Kobayashi N et al. Neurofibromatosis type 1 and childhood cancer. Cancer 1993;72:2746-2754.
52. Perry A, Kunz SN, Fuller CE et al. Differential NF1, p16, and EGFR patterns by interphase cytogenetics (FISH) in malignant peripheral nerve sheath tumor (MPNST) and morphologically similar spindle cell neoplasms. J Neuropathol Exp Neurol 2002;61:702-709.
53. Xiao GH, Chernoff J, Testa JR. NF2: the wizardry of merlin. Genes Chromosomes Cancer 2003;38:389-399.
54. Parry DM, MacCollin MM, Kaiser-Kupfer MI et al. Germ-line mutations in the neurofibromatosis 2 gene: correlations with disease severity and retinal abnormalities. Am J Hum Genet 1996;59:529-539.
55. Evans DGR, Wallace AJ, Wu CL et al. Somatic mosaicism: a common cause of classic disease in tumor-prone syndromes? Lessons from type 2 neurofibromatosis. Am J Hum Genet 1998;63:727-736.
56. Choo D, Shotland L, Mastroianni M et al. Endolymphatic sac tumors in von Hippel-Lindau disease. J Neurosurg 2004; 100:480-487.
57. Lonser RR, Kim HJ, Butman JA et al. Tumors of the endolym- phatic sac in von Hippel-Lindau disease. N Engl J Med 2004;350:2481-2486.
58. Bambakidis NC, Megerian CA, Ratcheson RA. Differential grading of endolymphatic sac tumor extension by virtue of von Hippel-Lindau disease status. Otol Neurotol 2004;25: 773-781.
59. Latif F, Tory K, Gnarra J et al. Identification of von Hippel-Lan- dau disease tumor suppressor gene. Science 1993;260: 1317-1320.
60. Sufan RI, Jewett MA, Ohh M. The role of von Hippel-Lindau tumor suppressor protein and hypoxia in renal clear cell carci- noma. Am J Physiol Renal Physiol 2004;287:F1-F6.
61. Stolle C, Glenn G, Zbar B et al. Improved detection of germline mutations in the von Hippel-Lindau disease tumor suppressor gene. Hum Mutat 1998;12:417-423.
62. Lonser RR, Glenn GM, Walther M et al. von Hippel-Lindau disease. Lancet 2003;361:2059-2067.
63. Wanebo JE, Lonser RR, Glenn GM et al. The natural history of hemangioblastomas of the central nervous system in patients with von Hippel-Lindau disease. J Neurosurg 2003;98: 82-94.
64. Olschwang S, Richard S, Boisson C et al. Germline muta- tion profile of the VHL gene in von Hippel-Lindau disease and in sporadic hemangioblastoma. Hum Mutat 1998;12: 424-430.
65. Hung W. Solitary thyroid nodules in 93 children and adoles- cents: a 35-year experience. Horm Res 1999;52:15-18.
66. Mulligan LM, Ponder BA. Genetic basis of endocrine disease: multiple endocrine neoplasia type 2. J Clin Endocrinol Metab 1995;80:1989-1995.
67. Santoro M, Carlomagno F, Romano A et al. Activation of RET as a dominant transforming gene by germline mutations of MEN2A and MEN2B. Science 1995;267:381-383.
68. Mulligan LM, Eng C, Healey CS et al. Specific mutations of the RET proto-oncogene are related to disease phenotype in MEN 2A and FMTC. Nat Genet 1994;6:70-74.
69. Thakker RV. Multiple endocrine neoplasia. Horm Res 2001;56[suppl 1]:67-72.
70. Szinnai G, Meier C, Komminoth P et al. Review of multiple endocrine neoplasia type 2A in children: therapeutic results of early thyroidectomy and prognostic value of codon analysis. Pediatrics 2003;111:E132-E139.
71. Walther MM, Herring J, Enquist E et al. von Recklinghausen’s disease and pheochromocytomas. J Urol 1999;162:1582-1586.
72. Modigliani E, Vasen HM, Raue K et al. Pheochromocytoma in multiple endocrine neoplasia type 2: European study. The Eu- romen Study Group. J Intern Med 1995;238:363-367.
73. Walther MM, Reiter R, Keiser HR et al. Clinical and genetic characterization of pheochromocytoma in von Hippel-Lindau families: comparison with sporadic pheochromocytoma gives insight into natural history of pheochromocytoma. J Urol 1999;162:659-664.
74. Baysal BE. Hereditary paraganglioma targets diverse paragan- glia. J Med Genet 2002;39:617-622.
75. Baysal BE, Ferrell RE, Willett-Brozick JE et al. Mutations in SDHD, a mitochondrial complex II gene, in hereditary para- ganglioma. Science 2000;287:848-851.
76. Neumann HP, Bausch B, McWhinney SR et al. Germ-line mu- tations in nonsyndromic pheochromocytoma. N Engl J Med 2002;346:1459-1466.