One of the most striking predispositions to cancer caused by a constitutional chromosome abnormality is the nearly 20-fold increased risk of leukemia in children who have trisomy 21 (reviewed by Rabin and Whitlock⁷). Trisomy 21 is also a common finding in the karyotype of leukemia cells from patients without Down syndrome (DS). Thus, presence of an extra chromosome 21 appears to be leukemogenic and may be acquired in the germline or somatically.

In children with DS, the ratio of leukemia subtypes is shifted to 60% lymphoid and 40% myeloid from the ratio in the general population of 80% lymphoid and 20% myeloid.⁷ This shift is principally due to the increased incidence of myeloid leukemias in children younger than 2 years of age. Approximately 30% of DS children with AML develop acute megakaryocytic leukemia representing a 400-fold excess (AMKL or M7) and the cytogenetic abnormalities within the tumor also differ as DS-AMKL is not associated with the characteristic t(1;22)(p13;q13) translocation seen in other AMKL patients.

In infancy, children with DS can develop transient myeloid proliferative syndrome that can appear similar to leukemia but that is self-limited.⁸ However, 25% of DS children with this syndrome eventually develop frank AML. Similarly, children with DS have a higher rate of myelodysplasia syndromes (MDS), which are characterized by thrombocytopenia, abnormal megakaryocytopoiesis, and an abnormal karyotype, most commonly trisomy 8.⁷

A recurring finding in DS-associated leukemia is that the pattern of somatic mutation found in the leukemia cells is distinct from that seen in the general population. For example, somatic mutations in the GATA1 gene are frequently detected in both TMD and AMKL samples from DS patients suggesting that it is an early event, probably occurring in utero^8,9 GATA1 encodes a transcription factor that is essential for maturation of erythroid cells and megakaryocytes.

Children with DS are also at increased risk to develop ALL. An international review of the therapeutic outcome for children with DS-ALL was compiled from 16 different international treatment trials (1995 to 2004).¹⁰ This retrospective analysis demonstrated that relapse was more common in DS-ALL patients despite the absence of other cytogenetics finding and the main contributor to increased mortality (74% ± 2% vs. 89% ± 1%, p < 0.0001) at 8 years for DS-ALL versus non-DS-ALL, respectively. There was also increased risk of infection in DS-ALL patients with both high- and low-risk forms of leukemia, suggesting the need for altered management of these fragile patients. As described in the chapter on ALL, the somatic alterations in DS-ALL are also distinct from those seen in children without DS who develop leukemia.

Despite the well-documented increase in the risk of leukemia in children with DS, a study based on exhaustive analysis of the Danish population found no increased risk of solid tumors in children or adults with DS including significantly fewer cases of breast cancer compared with an age-matched population.¹¹

Sex chromosome abnormalities comprise a large group of disorders that result from numerical and structural problems with the X and Y chromosomes. The overall incidence of sex chromosome abnormalities is high, with 47XXY and Turner syndrome (45X) each affecting approximately 1 in 2,000 individuals. Due to the lack of congenital anomalies, the diagnosis of these disorders, unlike DS, is often not made until late adolescence or young adulthood, when problems with the transition through puberty and fertility become apparent. However, children with these disorders are at increased risk for certain malignancies during childhood, arguing for earlier diagnosis. The large-scale implementation in the United States of noninvasive prenatal genetic testing methods that are sensitive to sex chromosome aneuploidy will likely significantly increase the early diagnosis of these syndromes.¹²

As cytogenetic techniques were improved in the 1970s, it became clear that many of the complex dysmorphic syndromes were the result of large cytogenetically visible deletions. During the next two decades, detection of deletions by Southern blot analysis and fluorescent in situ hybridization (FISH) permitted further progress in identifying the underlying cause of these syndromes. Starting around the beginning of the new century, array comparative genomic hybridization (arrayCGH), became the major technology used to detect copy number changes in tumor and constitutional DNA for both research and clinical applications.^19,20,21 ArrayCGH in current use allows the entire genome to be sampled for deletions or amplification in one experiment by using comparative hybridization of fluorescently labeled DNA from patient samples onto chips containing gridded arrays of oligonucleotides representing
the human genome. Some arrayCGH methods may be sensitive to copy number as well as the sequence (or genotype) of the patient at each point detected on the array. ArrayCGH is frequently employed as the frontline genetic test when faced with a child with multiple congenital anomalies or a child with cancer and either a congenital anomaly or intellectual disability.²¹ Whole-genome sequencing can also be used to detect copy number changes with breakpoint resolution²² although it is not yet being actively employed in the clinical arena. Conversely, it is important to remember that many sequence-based methods including Sanger sequencing and whole-exome sequencing, which attempts to read the open reading frame of every gene, are generally not able to detect deletions or duplications larger than 20 to 30 basepairs.

Children may present a complex phenotype when a chromosomal deletion large enough to remove multiple genes is present. Variation in the clinical phenotype may result from the size and number of contiguous genes that are missing. Deletion syndromes overlap with autosomal dominant disorders that are the result of smaller mutations affecting a single gene within the deleted segment. For example, retinoblastoma, can be transmitted as result of an autosomal dominant disorder due to single base changes in the RB1 gene or associated with a cytogenetically visible deletion of 13q14 in a small percentage of cases.²³

A small proportion of patients with WT may exhibit a spectrum of congenital abnormalities that result from a specific deletion of chromosome 11p13. The WAGR syndrome is named for the components of the disorder: WT, aniridia, genital abnormalities, including hypospadias, and the R referred to mental retardation, now reported as intellectual disability, as well as evidence of autistic features associated with cytogenetically detectable deletions at 11p13.²⁴ The PAX6 gene is contained within the deletion found in WAGR patients and is responsible for the aniridia phenotype. Surveys of children with WT in France revealed that 2% demonstrated WAGR.²⁵

WT1, the gene at 11p13 responsible for the WT phenotype encodes a zinc finger transcription factor (reviewed by Al-Hussain²⁶). All or part of WT1 is deleted in children with WAGR and WT. In contrast, constitutional point mutations in WT1 including missense mutations are found in children with the Denys-Drash syndrome, a disorder characterized by severe urogenital abnormalities and WT. This is an example where total loss of a gene product through gene deletion results in a less severe disease than production of a mutant protein due to a missense mutation. It is hypothesized that the mutant protein may have a dominant negative impact on genital development which may not occur when the WT1 gene is deleted.

ArrayCGH is performed for infants with aniridia to determine whether the deletion includes the WT1 gene in order to assess the risk of developing WT and need for surveillance. Screening for the development of WT in children with WAGR or Denys-Drash syndrome is often carried out by abdominal ultrasound examinations every 4 months until the age of 5 years, with decreasing frequency of examinations at later ages in order to decrease the stage of WT at diagnosis (see Table 2.2).²⁷

Screening for WT is further discussed in the section on management of patients with Beckwith-Weidemann syndrome. In addition to surveillance methods, parents should be counseled to bring the child in for evaluation if they suspect any change in abdominal girth or feel a mass, regardless of whether ultrasound screening is performed as interval tumors can develop. A long-term analysis of children with WT and either Denys-Drash or WAGR found a 62% and 38% rate of renal failure, respectively, 20 years after the diagnosis of WT.²⁸ Therefore, in addition to tumor surveillance, children with constitutional mutations in the WT1 gene require long-term follow-up for evidence of declining renal function.

The relationship between disorders of increased growth and predisposition to cancer are evident in Beckwith-Wiedemann syndrome (BWS) and hemihyperplasia (HH, previously termed hemihypertrophy) linked to a significantly increased risk of developing abdominal tumors, including WT and hepatoblastoma.²⁹ BWS is characterized by excessive intrauterine and postnatal growth, organomegaly, hypoglycemia at birth, macroglossia, and unusual linear ear creases and pitting. HH is defined as asymmetric growth due to overgrowth of one side relative to the other. It can be limited to a limb or the face, or can include the whole side. HH can be a feature of BWS or an isolated finding. Of 183 children in the BWS Registry, 13 had developed a tumor by age 4.³⁰ For children with isolated HH, the risk of WT is approximately 3%. Children with both BWS and HH had a higher risk of WT, than children with either condition alone. Nephromegaly seen on imaging is also a risk factor for WT.

The genetic basis of BWS and HH is complex.²⁹ Cytogenetically normal children with BWS may inherit two copies of a paternal chromosome 11 and no maternal copy, termed uniparental disomy (UPD). Alternatively molecularly visible rearrangements that result in paternal duplications of 11p15 are also seen. These unusual genetic alterations result from imprinting, which refers to the differential expression of genes, depending on whether they were inherited on the maternal or paternal chromosome. There are also families who have an apparent autosomal dominant pattern of inheritance mapping to 11p15. In these families, features of BWS are more likely to be seen in children who inherit the mutation from their mother.

Imprinted genes implicated in the etiology of BWS include the paternally expressed (maternally imprinted) genes insulin-like growth factor 2 gene (IGF2) and RNA transcript, KCNQ1OT1(LIT1) and the maternally expressed (paternally imprinted) genes H19 and CDKN1C.³¹ Among BWS patients, 25% to 50% have loss of imprinting (biallelic expression) of IGF2; 50% have an epigenetic mutation that results in loss of imprinting of KCNQ1OT1 and 5% carry mutations in the CDKN1C/p57^KIP2 gene.²⁹ Molecular testing for alteration in the genes implicated in BWS is performed by both methylation analysis (which detects altered imprints) and mutation analysis. Children who develop embryonal tumors such as rhabdomyosarcoma and hepatoblastoma are more likely to have epigenetic changes in domain 2 whereas WT is more strongly associated with epigenetic alterations in domain 1 or uniparental disomy.²⁹ The results also improve prediction of the likelihood of the parents having another child with BWS. For example, parents of a child with BWS due to uniparental disomy have a very low risk (much less than 1%) of recurrence in another pregnancy.

The risk of having a child with BWS is increased when the pregnancy was initiated by assisted reproductive technologies (ART) including in vitro fertilization (IVF).³² For example, a population-based case:control study from Australia confirmed a 17-fold relative risk of BWS in pregnancies initiated by IVF.³³ However, the absolute risk of BWS in an ART-associated pregnancy is still very low at 1 in 4,000 pregnancies.

Given the increased risk of WT and other abdominal malignancy in these conditions, screening by regular serial abdominal ultrasound examinations until age 9 and serial AFP levels until age 4 is recommended for children with BWS, HH, or both.³⁴ Children with BWS who were screened for WT were much less likely to present with advanced disease than those who were not screened (0 of 12 vs. 25 of 59).³⁵ Which health care professional manages cancer screening is quite variable in different centers including the child’s pediatrician, the medical geneticist making the initial syndromic diagnosis or an oncologist with a specific focus on hereditary cancer risk. Table 2.2 summarizes the WT
screening regimens implemented at the Childhood Cancer Prevention and Screening Clinic housed in the Texas Children’s Cancer Center.

Overall, bilateral WT or a family history of WT occurs in 1% to 5% of patients. Linkage studies have indicated that the gene for familial WT must be distinct from WT1 and from genes that predispose to BWS. Direct exome sequencing of WT kindreds revealed mutations in the CTR9 gene encoding a protein important for RNA polymerase in 3 of 35 kindreds.³⁶ In addition, some families carrying DICER1 mutations (described below) also demonstrate WT.

Several conditions are associated with overgrowth with or without an increased cancer risk (reviewed in Ref. 37). As exemplified here, the underlying genetic mechanisms are remarkably variable and include abnormal imprinting, mosaicism for specific activating mutations in proto-oncogenes, deleterious mutations in a dominant condition, and rare recessive mutations.

Proteus syndrome is a multisystem disease of asymmetric, disproportionate overgrowth, dysregulated deposition of adipose tissue forming lipomas, lipohyoplasia and focal fat depositions, a characteristic cerebriform connective tissue nevus commonly present on the soles of the feet, bony abnormalities, and intellectual disability or neurologic findings in some patients.³⁸ Mosaic mutation of the AKT kinase gene (typically the c.49G > A, p. Glu17Lys missense change) found in only a small proportion of the patient’s cells is the causal genetic lesion in patients with Proteus syndrome. Functional AKT disruption leads to cellular growth dysregulation, accounting for the characteristic multisystem overgrowth phenotype. Unlike BWS and other overgrowth syndromes, Proteus syndrome patients do not appear to develop malignant neoplasms, though benign lesions of adipose tissue are extremely common.

Sotos syndrome is characterized by macrocephaly, increased height, developmental delay, and facial dysmorphism.³⁹ Childhood cancers have been reported in children with Sotos syndrome, though no specific types appear to be overrepresented. The disease is caused by de novo mutations of the NSD1 gene (nuclear receptor binding SET domain protein 1) that encodes a ubiquitously expressed protein of the chromatin-modifying enzyme family whose precise function is not well characterized.

Simpson-Golabi-Behmel syndrome is an autosomal recessively inherited disorder characterized by skeletal, cardiac, renal, and craniofacial abnormalities. Generalized overgrowth is observed as well as susceptibility to a wide spectrum of tumors not dissimilar from that seen in BWS, namely, WT, gonadoblastoma, neuroblastoma, and hepatoblastoma. There are two main types of SGB syndrome, each associated with mutations of a particular gene. In SGB type 1, mutations of the glypican 3 gene (GPC3), which is ubiquitously expressed in mesodermal tissues, is observed. In SGB type 2, mutations of the CXORF5 gene, which forms part of the BRCA1/2 complex, have been suggested. The mechanism by which alterations of this gene might lead to the phenotype is not well described. Several other disorders classified as hereditary overgrowth syndromes have been described, but are generally either only very rarely associated with childhood-onset cancers or poorly molecularly characterized.

Much of our knowledge of autosomal dominant cancer families was gained from the study of retinoblastoma. In a landmark paper, Knudson hypothesized that bilateral retinoblastoma represented the hereditary form, and those patients had already acquired one “hit” or mutation.⁴⁰ The best statistical model consistent with his data indicated that the bilateral form required only one additional hit after birth but that the unilateral form required two hits.

The most striking clinical features of autosomal dominant cancer predisposition disorders are those initially observed by Knudson: hereditary forms of retinoblastoma present earlier and with a greatly increased percentage of bilateral and multiple primary tumors (reviewed in Lonser²³). Importantly, these patients are not exclusively bilateral, and variable expressivity results in approximately 15% of patients with unilateral retinoblastoma carrying a constitutional mutation. An even milder form, retinoma or retinocytoma, which spontaneously regress, can also be seen in apparently unaffected adults.⁴¹ Approximately 10% of people with a germline mutation in RB1 do not develop retinoblastoma, that is, incomplete penetrance.²³ However, the penetrance varies among families, with specific mutations (often missense changes or splice abnormalities) resulting in mutation carriers having a higher likelihood of not developing retinoblastoma or having unilateral (as opposed to bilateral) disease. These types of families are said to demonstrate attenuated or low-penetrant retinoblastoma.

Individuals carrying germline mutations in the RB1 gene are at increased risk for development of other primary tumors, including osteosarcoma and malignant melanoma (both within and outside the radiation field) in childhood and leiomyosarcomas in adults. In a UK cohort of long-term survivors of bilateral retinoblastoma, there was a 48% risk of developing a second neoplasm by age 50.⁴² Further follow-up of this cohort (up to age 84) identified a 68% cumulative incidence of second cancers including many epithelial cancers, for example, lung cancer, at later ages as well as leiomyosarcomas which occur almost entirely in the retroperitoneum, bladder, and uterus with equal distribution of the sexes.⁴³ Few individuals in the UK cohort received radiation therapy confirming that there is a significant risk of second primary cancers in all bilateral retinoblastoma patients. Data from a US cohort looking at cumulative cancer mortality (as opposed to incidence) identified 25% and 1% risk for hereditary and nonhereditary retinoblastoma, respectively.⁴⁴ Analysis of this long-term US cohort also demonstrates that the addition of alkylating agents in children <1 year of age increased the risk of leiomyosarcoma.⁴⁵ It is difficult to predict the outcomes from contemporary treatments. At 11 years of follow-up, 4% of patients with hereditary retinoblastoma receiving chemotherapy regimens consisting of carboplatin, vincristine, and etoposide demonstrated second malignancies.⁴⁶ A study comparing retinoblastoma patients who received proton therapy with those receiving photon therapy (follow-up of 5 to 10 years) demonstrates a decreased risk of second cancers.⁴⁷

The gene mutated in retinoblastoma, RB1, was identified on the basis of rare patients with cytogenetically visible deletions at 13q14.⁴⁸ Molecular studies confirmed Knudson’s two-hit hypothesis. Retinoblastoma requires loss of both copies (i.e., two hits) of the RB1 gene for a tumor to develop (Fig. 2.1). In the familial form, a mutation in one RB1 gene is inherited, and therefore all the cells in the body have only one normal allele. If during development that normal copy is mutated or lost, then cell cycle control is disrupted and retinoblastoma can develop. The most common mechanisms by which the second copy is lost are loss of the whole chromosome, large deletions, and gene conversion normally resulting in loss of heterozygosity (LOH) for markers near the RB1 locus. Silencing of the gene by epigenetic methylation of the RB1 promoter can also occur. In the sporadic form of retinoblastoma,
mutation, silencing, or loss of both RB1 genes must occur in the same somatic retinal cell for retinoblastoma to develop.

Although bilateral retinoblastoma results from constitutional mutations in the RB1 gene, 80% will have no family history of retinoblastoma and result from de novo mutation in the RB1 gene. Surprisingly, parents of a child with bilateral retinoblastoma who have normal eye exams retain a 6% risk to have a second affected child due to the de novo mutation occurring in either the mother’s or father’s germline with a variable percentage of the eggs or sperm carrying the mutation (germline mosaicism).⁴⁹ Therefore, all siblings of children with bilateral retinoblastoma should be evaluated by genetics and have ophthalmologic surveillance beginning at birth until the genetic status is clarified.

Several different molecular approaches are used to analyze the tumor and blood specimen of retinoblastoma patients to identify point mutations scattered throughout the gene, intragenic or whole-gene deletion, and promoter methylations.²³ With extensive testing, clinical laboratories can identify the causative mutation in about 95% of bilateral cases where a blood sample is tested directly. Recent studies suggest that the remaining 5% of patients are most likely mosaic for the causative mutation with too few blood cells containing the mutation to be detected.⁴⁹ Patients with a negative family history and unilateral disease have only a 15% a priori chance of having a germline RB1 mutation. A negative comprehensive RB1 mutation test from a blood sample from a unilateral patient will reduce this residual risk of hereditary disease to less than 1%. This risk can be further clarified by directly comparing RB1 results from tumor (when available) and blood.²³ Examples of genetic test results and interpretation for unilateral retinoblastoma patients are shown in Table 2.3.

Genetic evaluation and testing is recommended for all retinoblastoma patients to first determine the hereditary nature and RB1 mutation status of the proband’s disease (whether unilateral or bilateral) and the need for retinoblastoma surveillance of at-risk relatives. Family members are offered familial RB1 mutation testing, a simplified test that is targeted to the specific RB1 mutation identified in the family, for any at-risk relative. For example, in work from Texas Children’s Hospital, 48 at-risk relatives of patients with documented hereditary retinoblastoma underwent genetic testing that revealed six positive individuals who require retinoblastoma surveillance and 42 relatives who were negative and thus did not require retinoblastoma surveillance and are not at risk of having children at-risk, substantially decreasing costs.^50,51 If prenatal testing is not pursued, then any at-risk infant should have a careful eye examination within the first few weeks of life and a blood sample sent for analysis of the specific RB1 mutation found in the affected member of the family. Only those infants that carry the mutation need subsequent surveillance for retinoblastoma (see chapter on retinoblastoma). For adult long-term survivors of unilateral retinoblastoma, a positive test of a blood sample is found in approximately 12% of cases and is informative of a hereditary form of Rb and substantial risk to future children. Conversely, if comprehensive RB1 analysis of the blood is negative, then there is approximately a 0.5% to 1% residual risk for each offspring to develop retinoblastoma. Genetic testing of unilateral pediatric patients can be particularly informative for parents. If it can be documented that the child does not carry a constitutional RB1 mutation then (1) the child is not at substantial risk for secondary malignancies, (2) radiation therapy is associated with less hazard, and (3) the parents have a negligible risk of having another child with retinoblastoma. The child with unilateral retinoblastoma may carry some residual risk of having an affected child given the possibility of mosaicism.⁴⁹ A coordinated team of oncologists, ophthalmologists, pathologists, and genetics professionals facilitates the delivery of optimal care for families with a child diagnosed with retinoblastoma.

In 1969, an inherited cancer predisposition syndrome was reported by Li and Fraumeni on the basis of characterization of four families in which at least two cases of sarcoma occurred in early life.⁵² The list of LFS tumors includes premenopausal breast cancer, brain tumors, leukemias, adrenocortical carcinomas (ACC), gastric cancer, lymphoma, CPC, colorectal cancer, and possibly early onset lung cancer.⁵³ “Classic” LFS is defined by a proband with sarcoma diagnosed below age 45 years, who has a first-degree relative with any cancer below 45 years, plus another first or second-degree relative with either any cancer below 45 years or a sarcoma at any age. An example of an LFS pedigree is shown in Figure 2.2. The criteria for families that do not quite meet classic criteria, termed LFS-Like (LFS-L), are generally accepted to include those outlined by the revised Chompret criteria⁵⁴ to include all children with ACC or CPC regardless of family history; a family in which the proband has multiple tumors, two of which are classical LFS tumors and the first occurred before age 46 years; or a family in which the proband has a characteristic LFS tumor diagnosed below age 46 and has at least two first- or second-degree relatives with an LFS component tumor (other than breast cancer if the proband had breast cancer). TP53 mutation carriers exhibit greater than 80-fold increased risk of developing multiple synchronous or metachronous non-therapy-induced neoplasms.⁵⁵ In particular, the overall relative risk of occurrence of a second cancer was 5.3
(95% CI = 2.8-7.8), with a cumulative probability of second cancer occurrence of 57%.

In 1990, Malkin and colleagues detected heterozygous point mutations in the TP53 gene in constitutional DNA in all of five LFS kindreds.⁵⁶ However, numerous subsequent studies have shown that only 60% to 80% of “classic” LFS families harbor germline TP53 mutations (see, for example, in Ref. 57) while the majority of LFS-L families do not have detectable TP53 mutations (see, for example, in Ref. 54). Mutations occur throughout the TP53 gene, though they are primarily confined to highly conserved regions. Intragenic deletions of the TP53 gene have been reported in a subset of families that had negative sequencing studies.⁵⁸ In some patients, the tumors do not demonstrate LOH and the TP53 mutation is thought to act in a dominant negative form.⁵⁹

The cancer phenotype in LFS is quite diverse. While specific TP53 genotype:phenotype correlations have not been clearly demonstrated, several genetic modifier effects are reported. In particular, the mean age of onset of tumors is significantly less in TP53 mutation carriers who carry an MDM2 SNP309 G allele compared with those homozygous for the T allele.⁵⁸ Similarly, carriers of the TP53 codon 72 Arginine allele have an earlier tumor onset than those who harbor a homozygous Proline allele.⁵⁸ The cumulative combination of MDM2 SNP309 and TP53 codon 72 status, telomere length in peripheral blood cells, and possibly specific TP53 mutations may eventually be used as a predictive biomarker for cancer type and age of onset in LFS.⁶⁰ DNA copy number variation (CNV) is strikingly enriched in the constitutional DNA of TP53 mutation carriers, and these CNVs can be inherited and frequently encompass other cancer genes, suggesting that the genomic instability conferred by the TP53 mutation can be transmitted from generation to generation.⁶¹ The potential role of telomere attrition and aberrant CNVs is exemplified by evidence that demonstrates chromothripsis (chromosome shattering and reshuffling) in tumors, such as medulloblastomas, from patients harboring germline TP53 mutations.⁶²

A large number of studies have analyzed groups of patients with tumors characteristic of LFS, yet lacking characteristic family histories of cancer, for germline TP53 mutations. A few examples are provided here. For example, mutations have been identified in approximately 50% of children with ACC,⁶³ 10% of children with osteosarcoma,⁶⁴ and 10% of children with rhabdomyosarcoma.⁶⁵ In the latter study, the age average of onset (22 months) was lower for TP53 mutation carriers. Strikingly, 75% of children with anaplastic RMS harbor germline TP53 mutations.⁶⁶ Furthermore, one-third of children with sarcomas plus either multiple primary tumors, or a significant family history of cancer have germline TP53 mutations.⁶⁷ Studies reveal that more than 90% of ACC patients from Brazil carry the TP53 R337H missense mutation as a founder mutation.⁶⁸ ACC diagnoses are overrepresented in families carrying this mutation; however, many Brazilian families exhibit the classic LFS phenotype of multiple cancer types.⁶⁹ Evidence that germline TP53 mutations are associated with specific subtypes of childhood cancer was further validated by the recent finding that apparently germline TP53 mutations were identified in 50% of children with the rare hypodiploid form of ALL.⁷⁰

Presymptomatic molecular testing for TP53 germline mutations in members of LFS kindreds has been met with significant controversy. A major step forward came with the report of early tumor detection with favorable impact on survival in a small cohort of childhood and adult TP53 mutation carriers screened using a comprehensive tumor surveillance protocol.⁷¹ This protocol (often referred to as the “Toronto Protocol”) uses a combination of annual rapid sequence whole-body magnetic resonance imaging (MRI), dedicated brain and breast (for young adult) MRI, and frequent abdominal/pelvic ultrasounds and several biochemical assays in at-risk individuals (see Table 2.4). Relatively long-term follow-up has demonstrated that the approach is feasible and offers a significant survival advantage and reduced treatment morbidity when compared with carriers who do not undergo surveillance. The Toronto protocol does not include PET-CT evaluation, although a few anecdotal reports of detection of ACC with PET-CT imaging have been published in children with LFS.^72,73 With the implementation of the Toronto protocol, many additional centers are beginning presymptomatic testing of children at risk for TP53 mutations. Clinical evaluation of children at increased genetic risk for cancer is discussed further in the last section of this chapter.

The advent of novel gene discovery platforms as well as astute recognition of emerging phenotypes has led to the recognition of new syndromes. One of these, DICER1 syndrome (or pleuropulmonary blastoma familial tumor dysplasia syndrome), is characterized by a phenotypic association of distinctive dysontogenic, hyperplastic, or overtly malignant tumors. The most frequent of these is pleuropulmonary blastoma (PPB). Other manifestations include ovarian Sertoli-Leydig cell tumors (SLCT), nodular thyroid hyperplasia, pituitary blastoma, pineoblastoma, papillary and follicular thyroid carcinoma, cervical rhabdomyosarcoma, cystic nephroma, and WT.⁷⁴ Most of these tumors appear to manifest in childhood, although the thyroid tumor risk and multinodular goiter may persist through adulthood. DICER1 is an endoribonuclease that processes hairpin precursor microRNAs (miRNAs) into short, functional miRNAS. Mature 5′ miRNAs as well as other components of the RNA-induced silencing complex (RISC) down-regulate targeted mRNAs.⁷⁵ Germline DICER1 mutations have been identified in children and young adults affected with one or several of these tumors.⁷⁶