1.1 Introduction

Bone and soft tissue sarcomas are relatively rare cancers that collectively account for approximately 1% of adult solid tumors and 12% of all pediatric malignancies (Burningham et al. 2012; Ries et al. 1999). Major bone tumors include osteosarcoma and Ewing sarcoma, whereas soft tissue sarcomas are largely categorized by rhabdomyosarcoma and non-rhabdomyosarcoma soft tissue sarcomas. In spite of their relative infrequency, these tumors remain a leading cause of cancer death in individuals <20 years of age in developed countries and are a significant source of cancer-related morbidity (Ries et al. 1999). Given the suboptimal outcomes seen among those with bone and soft tissue sarcomas, these malignancies have a substantial impact on public health as measured by the average number of productive years of life lost due to cancer (Burningham et al. 2012). Thus, it is important to characterize their clinical/biological behavior and their etiologies. Epidemiologic studies help this endeavor in two ways. Descriptive studies reveal the incidence of these sarcomas and their associated mortality and survival rates with respect to histologic subtype and demographic characteristics. Analytic studies compare the risk of bone and soft tissue sarcomas in people with and without certain characteristics (cohort studies) or compare the histories of people with and without sarcomas (case-control studies) to identify and assess a wide range of possible risk factors, including exposures to radiation and hereditary cancer predisposition syndromes. Combined consideration of epidemiologic insights and progress in the molecular classification of these tumors provides greater insight into the role of tumor biology in disease progression and sensitivity to radiation treatment and chemotherapy, ultimately providing a framework for improved assessment of sarcoma risk, early tumor identification, and eventually strategies for individualized prevention and treatment of sarcomas.

1.2 Descriptive Epidemiology

1.2.1 Osteosarcoma

Osteosarcoma (OS) in the young is rare worldwide, with rates between 4 and 7 cases per million children and adolescents <25 years of age, and shows minimal international variation (Mirabello et al. 2009). In the USA, an estimated 400 cases of OS are diagnosed each year among children 0–19 years of age and accounted for about ~3% of total cancer in this age group (Ries et al. 1999; Howlader et al. 2015).

The overall rate for OS is 5.3 cases per million children ages 0–19 years in the USA (Howlader et al. 2015), but this obscures substantial variation in incidence of the tumor by age (Fig. 1.1). Notably OS is very rare in early childhood but has a markedly peaked incidence in adolescence. Males have a somewhat higher incidence than females, and, interestingly, the peak incidence in females appears 2 years earlier than in males. Incidence among young adults (20–24 years) is also appreciable but less than at the peak (Wu et al. 2003). The incidence of OS in Surveillance, Epidemiology, and End Results (SEER) program data rose significantly by 1.6% per annum between 1975 and 1993 but remained steady between 1994 and 2012 (Ries et al. 1999). OS has a distinctive pattern of tumor location, with almost 80% of OS cases occurring in the long bones of the lower limbs and only about 5% in the central axis (Ries et al. 1999).

Fig. 1.1
figure 1

Incidence of OS by age and sex (SEER 2000–2012)

Full size image

1.2.2 Ewing Sarcoma

The incidence of Ewing sarcoma is 1 case per million for all ages and increases to 3 cases per million for those under the age of 20 years. Estimates vary, but approximately 80–90% of individuals who develop Ewing sarcoma are under the age of 24 years (Cotterill et al. 2000). For this reason, Ewing sarcoma is largely considered a childhood cancer, although it can also occur in adults. The most common primary site for Ewing Sarcoma is the pelvis and lower extremities (Cotterill et al. 2000). Males are at a slightly increased risk for developing Ewing sarcoma in comparison to females. The incidence of Ewing sarcoma also varies significantly by race. Caucasians have the highest incidence, followed by Asians/Pacific Islanders, and African Americans, with incidence rates of 0.155, 0.082, and 0.017 per 100,000 individuals, respectively (Fig. 1.2) (Jawad et al. 2009). The fact that Caucasians are 9-times more likely to Ewing sarcoma compared to African Americans is a very important observation. In populations worldwide, those of European ancestry exhibit the highest incidence rates, regardless of geography. Furthermore, those with African ancestry on different continents still exhibit the lowest incidence, suggesting racial disparities in incidence are due to differences in genetics (Jawad et al. 2009; Fraumeni and Glass 1970).

Fig. 1.2
figure 2

Incidence of Ewing sarcoma by race (Jawad et al. 2009)

Full size image

1.2.3 Rhabdomyosarcoma (RMS)

RMS is the most common soft tissue sarcoma in children and adolescents, with an overall incidence rate of 4.5 cases per million among those <20 years of age. In the USA, this equates to approximately 350 new cases per year. While adult cases of RMS make up 40% of all RMS diagnoses, these tumors are considered rare in adults and are often characterized by a different histologic subtype (pleomorphic) (Sultan et al. 2009). In fact, information regarding the clinical and biologic characteristics of RMS in adults is very limited. Large, multi-institutional trials have not been performed, and only reports from single institutions have been published.

Based on data from the SEER Program, we know the incidence of RMS differs both by age and histology (Fig. 1.3). Specifically, the incidence of embryonal RMS (eRMS) has a bimodal peak, with each peak occurring during key developmental periods (i.e., early development and puberty). This is not seen with alveolar RMS (aRMS), where the incidence remains constant throughout childhood and adolescence. Also using data from SEER (Ognjanovic et al. 2009), Ognjanovic et al. reported the incidence of eRMS for the period 1975–2005 has remained relatively stable. However, there has been a significant increase in the incidence of aRMS (annual percentage change (APC) = 4.20%, 95% confidence interval (CI): 2.60–5.82) for the same period. The authors noted that this might be attributable to changes in diagnostic criteria over time. There are also differences in incidence by gender. Males had a higher incidence of RMS compared to females, but this predominance was mainly among those diagnosed with eRMS (male/female ratio = 1.51, 95% CI: 1.27–1.80). Overall, there were no notable differences in incidence by race/ethnicity. This is consistent with a study pooling cancer registry data across five states (California, Minnesota, New York, Texas, and Washington), which indicated there were no significant differences in the risk of RMS by parental race (Chow et al. 2010). The only exception was the risk of RMS in offspring was significantly lower when both parents were of Hispanic ethnicity (OR = 0.65, 95% CI: 0.48–0.88).

Fig. 1.3
figure 3

Incidence of RMS by age and histology (SEER 1975–2011)

Full size image

1.2.4 Non-rhabdomyosarcoma Soft Tissue Sarcoma (NRSTS)

Non-rhabdomyosarcoma soft tissue sarcoma (NRSTS) incidence rates range between 1.8 and 5.0 per 100,000 individuals per year (Gustafson 1994; Levi et al. 1999; Ross et al. 1993; Storm 1994; Toro et al. 2006; Zahm and Fraumeni Jr. 1997). Differences in incidence rates between countries (e.g., 1.8 per 100,000 in Sweden (Gustafson 1994) and 4–5 per 100,000 in the USA (Toro et al. 2006; Zahm and Fraumeni Jr. 1997)) are likely due to racial differences, discrepant Kaposi sarcoma rates, and changing pathology definitions. Several reports documented increasing STS incidence rates over time (Toro et al. 2006; Zahm and Fraumeni Jr. 1997; Lahat et al. 2008), which are likely explained by improved registry systems and diagnostic tools, as well as the marked rise in Kaposi sarcoma incidence rates from 0.5 per 100,000 in 1975 to 3.3 per 100,000 during the HIV/AIDS epidemic of the late 1980s/early 1990s (Eltom et al. 2002). Incidence rates per year of specific NRSTS subtypes in the general population are summarized in Table 1.1 (Toro et al. 2006; UK CR 2008–2010).

Table 1.1 Incidence rates per year of specific NRSTS subtypes
Full size table

There is a modest peak in the incidence of NRSTS among young children under the age of 5 years with a larger peak observed at approximately 60 years of age (Burningham et al. 2012; Wibmer et al. 2010). Mean age of diagnosis is 58 years, and mean age of death is 65 years (Burningham et al. 2012). However, STS accounts for 7% of all cancer diagnoses in the pediatric age group and <1% of all malignancies in adults (Pappo and Pratt 1997). There are marked differences between histologies diagnosed in different age groups. Frequent NRSTS in the pediatric age group include synovial sarcomas (accounting for up to 42% of NRSTS (Dillon et al. 1992)); and fibroblastic/myofibroblastic/fibrohistiocytic tumors. In the adolescent and young adult (AYA) population, synovial sarcomas are relatively more common compared to the general adult NRSTS population and account for approximately 33% of AYA NRSTS (Herzog 2005).

Data on gender differences in STS risk are inconsistent, with male/female ratios ranging from 0.80 to 1.42 in different populations (Levi et al. 1999; Ross et al. 1993; Zahm and Fraumeni Jr. 1997; Wibmer et al. 2010). The highest NRSTS incidence rates were noted among black women, followed by black men, white men, and white women (Toro et al. 2006). Differences in black-white incidence rates were observed at all ages except childhood and driven largely by higher rates of leiomyosarcomas and dermatofibrosarcomas among non-Hispanic blacks (Toro et al. 2006). Race/ethnicity differences in STS incidence rates were also noted in another report, with incidence rates of 5.1 per 100,000 in blacks, 4.5 per 100,000 in whites, and 2.8 per 100,000 in Asians/Pacific Islanders (Burningham et al. 2012).

The epidemiology of certain NRSTSs is notable for specific features: classic gastrointestinal stromal tumors (i.e., GIST,which harbor C-KIT or PDGFRA pathogenic variants (Miettinen et al. 2005)) typically presents with a male/female ratio of 1.5 and affects individuals over 50 years of age (Tran et al. 2005), whereas its wild-type form (C-KIT/PDGFRA-negative) more commonly occurs in younger age groups and females (Nannini et al. 2013; Pappo and Janeway 2009). Finally, EBV-associated leimyosarcomas in immunodeficient individuals occur at any age, with a distinct age peak in children aged 0–9 years, whereas leiomyosarcoma rates in the general population increase with age and peak over the age of 50 years (Bhatia et al. 2012).

1.3 Environmental (Non-genetic) Risk Factors

1.3.1 Osteosarcoma

1.3.1.1 Growth and Development

The descriptive epidemiology of osteosarcoma strongly suggests an etiology related to growth and development. The age-incidence curve of OS closely follows that of the childhood growth curve (Fraumeni 1967), especially with respect to peak incidence in females reflecting their earlier growth spurts. Moreover, OS occurs most frequently at the metaphyseal region of long bones which are sites that contribute the most to vertical growth during the adolescent growth spurt (Price 1958).

Consequently, many studies have attempted to ascertain whether patients with OS attain greater height, grow faster, or reach puberty earlier than the general pediatric population. The largest analysis of height at diagnosis of OS pooled 1067 cases from seven studies to find that people in the 51st to 89th percentiles and those with ≥90th percentile of height had significantly increased risk of OS (odds ratio [OR] = 1.35, 95% confidence interval [CI]: 1.18–1.54 and OR = 2.60, 95% CI: 2.19–3.07, respectively) (Mirabello et al. 2011a). The same analysis pooled birth weight data from four studies to find that birth weight ≥4046 g resulted in a greater risk of OS (OR = 1.35, 95% CI:1.01–1.79), although a subsequent, population-based, fairly sizable Scandinavian study did not replicate the finding (Troisi et al. 2014).

Some investigations have sought to quantify the rate of growth in OS cases and controls (Buckley et al. 1998; Gelberg et al. 1997; Operskalski et al. 1987), while others have examined the age at appearance of secondary sexual characteristics (Buckley et al. 1998; Gelberg et al. 1997). No correlation has been found with risk for OS.

The epidemiology of canine OS also points to a role for growth in the etiology of the tumor. It has been known since the 1960s that OS is more common in large dog breeds than in small ones (Tjalma 1966). One study in Rottweilers indicated that endogenous hormone exposure, and not simply growth, drives OS (Cooley et al. 2002). Investigators found that earlier age of gonadectomy raised the risk of OS significantly both male and female dogs, independent of size.

1.3.1.2 Exogenous Exposures

Most investigations of exogenous exposures and OS have been unrevealing, with previous treatment for cancer being a prominent exception. Radiation and alkylating agents each raise risk of OS independently with risk increasing by dose (Hawkins et al. 1996; Tucker et al. 1987). However, only about 1% of childhood cancer survivors developed OS within 20 years of their primary diagnosis in one cohort (Hawkins et al. 1996).

Because fluoride and radium are deposited in the bones, and especially because the latter is radioactive, several ecologic or case-control studies have examined exposures to these elements, especially in drinking water, in relation to OS. One case-control study that comprehensively reconstructed fluoride exposure across the lifespan found risk increased at higher levels of exposure among males, but not females (Bassin et al. 2006); direct measurement of fluoride in bone samples from OS patients in this study showed no difference compared to controls with other tumors (Kim et al. 2011), suggesting the previous finding was by chance. Moreover, other studies of OS and fluoride have been mostly null (Moss et al. 1995; Blakey et al. 2014; Levy and Leclerc 2012; Gelberg et al. 1995; McGuire et al. 1991). Although high doses of ingested radium caused OS in an historical cohort of radium dial painters (Fry 1998), there is little evidence that radium at the levels present in drinking water raises the risk of disease (Moss et al. 1995; Finkelstein 1994; Finkelstein and Kreiger 1996; Guse et al. 2002).

Few population studies have directly addressed the topic of OS and infections although some evidence indirectly supports an association. Simian virus 40 (SV40), JC, and BK viruses comprise the polyomavirus family and have in common a T antigen which interferes with the function of the TP53 and RB1 tumor suppressor genes (Fanning 1998; Barbanti-Brodano et al. 1998), which are the frequent targets of somatic mutation in OS. SV40 induces OS in hamsters (Diamandopoulos 1973) and has been detected in OS tissue (Yamamoto et al. 2000; Mendoza et al. 1998; Carbone et al. 1996). On the other hand, studies in cohorts of children who received early batches of poliovirus vaccine contaminated by SV40 have not supported an increased risk of OS (Engels et al. 2003; Carroll-Pankhurst et al. 2001; Fisher et al. 1999; Strickler et al. 1998; Olin and Giesecke 1998). One case-control study of OS did not indicate an association with childhood or maternal infections. Little evidence supports the spatiotemporal clustering of OS cases (Ross et al. 1999; Silcocks and Murrells 1987; McNally et al. 2006, 2012; Basta et al. 2010), which can occur in the presence of a causative infection.

Lastly there have been several exploratory analyses of putative exogenous risk factors, which have generated few positive associations with OS. Risk factors examined to date include parental smoking (Hartley et al. 1988a), medications taken by mother or child (Hartley et al. 1988a), and in utero or postnatal diagnostic X-rays (Operskalski et al. 1987; Hartley et al. 1988a). One study suggested an association with bone fracture, particularly at the tumor site (Operskalski et al. 1987), but this was not replicated in another study (Buckley et al. 1998). Studies of parental occupation and OS have also produced only isolated reports of associations with OS (Buckley et al. 1998; Gelberg et al. 1997; Operskalski et al. 1987; Hartley et al. 1988a; Hum et al. 1998).

1.3.2 Ewing Sarcoma

To date, there are no well-documented environmental risk factors for Ewing sarcoma. A handful of environmental association studies have found a correlation between farming-related environments and an increased risk for developing the disease. Specifically, parental exposure to pesticides and other chemicals associated with agriculture have been linked to Ewing sarcoma (Holly et al. 1992; Stiller et al. 1991; Valery et al. 2002). A few studies have associated both inguinal and umbilical hernias with an increased risk for the disease (Cope et al. 2000; Valery et al. 2003, 2005). The possible association between these two factors makes analyzing them in the context of Ewing sarcoma somewhat challenging. Ewing sarcoma and hernias do share common embryological pathways. Therefore, it is possible that the two may share or have an overlapping genetic predisposition. However, this has not been demonstrated in a molecular context. More work is clearly needed to determine how farming related environments, and/or hernias may contribute to the development of Ewing sarcoma.

Patient stature and pubertal growth patterns have also been investigated as risk factors for Ewing sarcoma. Studies have been somewhat conflicting with no clear trends emerging (Cotterill et al. 2000; Fraumeni 1967; Bacci et al. 1992; Pui et al. 1987; Winn et al. 1992). Part of the difficulty in evaluating these variables likely comes from alterations in normal development due to treatment (Cotterill et al. 2000).

1.3.3 RMS

Several environmental exposures have been explored in relation to RMS risk among children, including paternal cigarette smoking (Grufferman et al. 1982), prenatal X-ray exposure (Grufferman et al. 2009), advanced maternal age (Grufferman et al. 2009), maternal antibiotic use (Grufferman et al. 1993), and parental recreational drug use (Grufferman et al. 1993). The majority of these assessments were based on the largest epidemiologic case-control study of RMS to date. Specifically, cases were collected from patients enrolled on the IRS-III study, coordinated by the Intergroup Rhabdomyosarcoma Study Group, whose protocols enrolled 80–85% of all childhood RMS cases in North America (Grufferman et al. 1984). Cases (n = 322) were 0–20 years old at the time of their RMS diagnosis from April 1982 to July 1988. Central expert pathology review confirmed all RMS diagnoses, as well as histologic subtype (i.e., embryonal, alveolar, or other). Controls (n = 322) were identified by random-digit dialing during the same period (1982–1988) (Grufferman et al. 1993, 2009). Controls were pair-matched to cases on race, sex, and age. Key findings from this study are presented in Table 1.2. While some notable associations were reported, few have been confirmed in independent assessments. Unfortunately, germline DNA was not collected as part of this study, making the evaluation of genetic susceptibility of gene-environment interactions impossible. Other epidemiologic studies of RMS evaluating the role of environmental exposures have been relatively small (<100 cases). Novel and larger studies are needed to validate previous findings and explore other potential risk factors. Additionally, much work is needed to identify risk factors for RMS among adults.

Table 1.2 Review of factors evaluated in IRS-III case-control study of RMS
Full size table

1.3.4 NRSTS

NRSTS have been associated with a number of non-genetic risk factors, including virus infections conferring NRSTS susceptibility in immunodeficient patients. Elevated rates of Kaposi sarcomas and leiomyosarcomas in transplant recipients and individuals with HIV/AIDS first linked these sarcomas with immunosuppression (Bhatia et al. 2012). HIV/AIDS patients and transplant recipients also experience an excess of leiomyomas and leiomyosarcomas (Bhatia et al. 2012). The majority of leiomyosarcoma and leiomyoma cells in these patients contain EBV (Bhatia et al. 2012), which was also detected in smooth muscle tumors from individuals with congenital immunodeficiency syndromes (Hatano et al. 2006). Yet, EBV does not appear to be a general requirement for the development of leiomyosarcomas, because tumors arising in immunocompetent individuals do not contain the virus (Bhatia et al. 2012; Fernandez et al. 2010).

Most studies aimed at evaluating NRSTS-relevant environmental exposures have employed small case-control studies covering groups of NRSTS subtypes and a wide range of exposures/oncogenic factors (Burningham et al. 2012). Consequently, findings are prone to bias and should be interpreted with caution. Gardeners (Balarajan and Acheson 1984; Wingren et al. 1990), farmers (Balarajan and Acheson 1984), building caretakers, and military personnel (Pukkala et al. 2009) appear to experience an increased risk for developing NRSTS. More specifically, high-intensity occupational exposures to chlorphenol and cutting oils (Hoppin et al. 1999) and dioxin exposure from incinerators (Comba et al. 2003; Viel et al. 2000) were associated with STS risk. There is no known association between STS risk and exposure to solvents, wood dust, asbestos, DDT, or benzene (Burningham et al. 2012; Hoppin et al. 1999).

Finally, excess risk of developing NRSTS has been reported in cancer survivors in relation to therapeutic radiation (Burningham et al. 2012; Rubino et al. 2005; Menu-Branthomme et al. 2004; Virtanen et al. 2006). STS risk in a large cohort of survivors of childhood solid cancer was increased 19-fold compared to the general population after radiation alone (Menu-Branthomme et al. 2004). The same study reported a 113-fold increase in STS rates after radiation plus chemotherapy (Menu-Branthomme et al. 2004). Of note, there is no known association between STS risk and birth weight (Schuz and Forman 2007) or growth/development in early adolescence (Burningham et al. 2012).

1.4 Germline Genetic Risk Factors

1.4.1 Osteosarcoma

As with most childhood cancers a small proportion of cases are due to inherited, high-penetrance genetic variation. Li-Fraumeni (Li et al. 1988), hereditary retinoblastoma (Wong et al. 1997; Hansen et al. 1985), and Rothmund-Thomson syndromes (Leonard et al. 1996; Wang et al. 2003), which result from germline variants in TP53, RB1, and RECQL4, respectively, each raise the risk of OS substantially in carriers. Somatic mutations of the former two genes are commonly found in sporadic OS (Miller et al. 1996), while those in the latter are not (Nishijo et al. 2004). The advent of next-generation sequencing is now bringing the prevalence of these syndromes in OS into greater focus. One study in which 765 OS cases were sequenced at the TP53 locus found that nearly 10% of harbored known or likely Li-Fraumeni variants (3.8%) or rare exonic variants of unknown significance (5.7%) (Mirabello et al. 2015a). Another recent study which performed whole exome sequencing on 39 cases found that 7 (17.9%) had pathogenic or likely pathogenic variants in one of 21 known autosomal dominant cancer-predisposition genes (Zhang et al. 2015).

The role of common variation on OS risk has also been evaluated. There have been a number of candidate gene studies of OS (Musselman et al. 2012; Mirabello et al. 2011b; Ruza et al. 2003; Savage et al. 2006); however none appeared significantly associated with OS in the sole genome-wide association study conducted to date (Savage et al. 2013). Rather, the only two variants which reached genome-wide significance were rs1906953 near the GRM4 gene and rs7591996 in an intergenic region of 2p25.2; the function of these variants has not been investigated to date. A second genome-wide study also identified the rs7034162 variant in the NFIB gene which more than doubled the likelihood of metastasis at diagnosis (OR = 2.43; 95% CI: 1.83–3.24); this finding was supported by in vitro experiments which showed cell lines with the variant behaved more aggressively (Mirabello et al. 2015b).

1.4.2 Ewing Sarcoma

At present, Ewing sarcoma is not considered to be part of any hereditary cancer syndrome. Given its rarity, the occurrence of Ewing sarcoma in siblings is slightly suggestive of an unknown genetic predisposition. Given this lack of evidence, the diagnosis of Ewing sarcoma cannot yet be considered to be part of a larger hereditary cancer syndrome requiring clinical genetic testing.

Parental age has been found to be a potential risk factor for some childhood cancers (Dockerty et al. 2001; Johnson et al. 2009; Olson et al. 1993; Yip et al. 2006). While unclear, the reason for this may be related to an increase in de novo germline mutations or epimutations (Johnson et al. 2009; Olson et al. 1993; Dryja et al. 1997). In the case of Ewing sarcoma, epimutations seem more likely given the paucity of mutations observed in tumor samples (Tirode et al. 2014).

A number of other studies have aimed at identifying a genomic region responsible for the racial disparities in incidence for Ewing sarcoma. One candidate that has been identified is an intron near a frequent breakpoint region in the EWS/ETS translocation. This region has been observed to be smaller in African American populations, making it an appealing possibility. Other work has identified polymorphic repeat region binding sites referred to as (GGAA) microsatellites for the EWS/ETS fusion protein as potential candidates (Beck et al. 2012; Gangwal et al. 2008; Monument et al. 2014; Zucman-Rossi et al. 1997). These regions also show variation that can be specific to races and therefore offer a potential explanation for Ewing sarcoma incidence. An important finding in recent years comes from the first genome-wide association study (GWAS) of Ewing sarcoma (Postel-Vinay et al. 2012). Two notable risk loci were identified. The first is located upstream of TARDBP (P = 1.4 × 10−20; (OR = 2.2) and the second upstream of EGR2 (P = 4.0 × 10−17; OR = 1.7). Interestingly, EGR2 also contains a (GGAA) microsatellite with a Ewing sarcoma-associated SNP that appears to alter EWS/FLI1 binding (Grünewald et al. 2015). The authors also showed that EGR2 knockdown induced regression of Ewing sarcoma xenografts, increasing its plausibility as a candidate for contributing to disease development. While it is somewhat unclear how these candidates would fit in with Ewing sarcoma development, the risk haplotypes were less prevalent in African Americans. Future studies will hopefully shed more light on these candidate genomic regions and their possible role in Ewing sarcoma development.

1.4.3 RMS

As opposed to OS (Savage et al. 2013) and Ewing sarcoma (Postel-Vinay et al. 2012), there has not been a genome-wide associations study (GWAS) of RMS. Additionally, while there have been whole exome and whole genome sequencing efforts to identify somatic mutations in RMS tumors (Shern et al. 2014), to date, there have few studies characterizing the role of germline DNA on disease susceptibility, especially among seemingly sporadic cases. There is, however, a great deal of literature to support the hypothesis that genetic susceptibility plays a role in RMS development. Numerous reports consistently highlight the fact that children with certain genetic syndromes develop RMS more frequently than their unaffected peers. The syndromes that are most commonly seen among those with eRMS are Li-Fraumeni (Diller et al. 1995), NF1 (Hartley et al. 1988b; Yang et al. 1995), Costello (Estep et al. 2006; Kratz et al. 2011); Noonan (Kratz et al. 2011), and DICER1 (Doros et al. 2012). The genes and syndromes previously identified among RMS cases are included in Table 1.3. DICER1 is particularly notable as this is a recent discovery in terms of germline genetic susceptibility to RMS. Based on smaller clinic-based studies, only about 5% of RMS cases are thought to be associated with these syndromes (Plon and Malkin 2010). Additionally, cancer predisposition syndromes appear to be more frequent in eRMS cases compared to those with aRMS (Yang et al. 1995; Estep et al. 2006; Kratz et al. 2011; Ognjanovic et al. 2012). However, there have been no large-scale population-based efforts to systematically characterize the prevalence of these variants among children with RMS.

Table 1.3 RMS predisposition syndromes and genes
Full size table

1.4.4 NRSTS

Several well-described cancer-predisposing germline mutations confer STS susceptibility (Table 1.4): NF1 germline mutations (associated with Neurofibromatosis type 1) confer a 10% cumulative lifetime risk of developing MPNST (Burningham et al. 2012; Pollack and Mulvihill 1997). RB1 germline mutations have been linked to STS, most notably leiomyosarcomas (Kleinerman et al. 2007). DICER1 germline mutations confer pleuropulmonaryblastoma susceptibility (Slade et al. 2011). SMARCB1/SMARCA4 germline mutations are associated with rhabdoid tumor predisposition (Sredni and Tomita 2015). Finally, TP53 germline mutations (associated with Li-Fraumeni syndrome (Gonzalez et al. 2009; Li and Fraumeni Jr. 1969)); and spontaneous chromosomal instability in Werner syndrome, an autosomal recessive disorder of premature age more commonly reported in Japan than elsewhere (Goto et al. 1996), are associated with a predisposition to develop various STS types, including RMS and NRSTS.

Table 1.4 NRSTS predisposition syndromes and genes
Full size table

NRSTS arising in the setting of a cancer predisposition syndrome may exhibit specific features that distinguish them from similar, spontaneous sarcomas. Familial forms of GIST, accounting for approximately 85% of pediatric and 15% of adult GISTs (Nannini et al. 2013; Corless et al. 2004), typically lack C-KIT or PDGFRA mutations (wild-type GIST (Miettinen et al. 2005)). Familial GISTs develop in individuals with neurofibromatosis type 1 (Nannini et al. 2013; Bajor 2009) and Carney triad, first described in 1977 as a triad of gastric leiomyosarcoma, extra-adrenal paraganglioma, and pulmonary chondroma (Carney et al. 1977; Matyakhina et al. 2007). Of note, familial GISTs in Carney-Stratakis syndrome, a familial predisposition to develop multifocal GISTs and multifocal paragangliomas, have been linked to succinate dehydrogenase B (SDHB) germline mutations (Carney and Stratakis 2002; McWhinney et al. 2007).

1.5 Conclusion

Even when using large population-based registries, the sensitivity of identifying an elevated risk of uncommon cancers depends on the categorization of cancer types, the magnitude of the association, and the incidence of these cancers in the general population. The rarity and heterogeneity of bone and soft tissue sarcomas represents a major challenge in epidemiologic studies aimed at illuminating etiologic factors. Nevertheless, descriptive epidemiology has already yielded important insights into the origins and manifestations of bone and soft tissue sarcomas and continues to provide a window into the etiologies of these malignancies. It is likely that the greatest gains in understanding the etiology of bone and soft tissue sarcomas in the near future will come from “omics” studies seeking to understand innate and exogenous factors that contribute to susceptibility. It will also be important to identify interactions between genetic and environmental factors and to conduct studies that integrate germline and somatic tumor data to determine how germline variation influences tumor mutation profiles and prognosis. For progress in these areas to occur, a coordinated investment in systematic collection of clinically annotated biological specimens (both tumor and normal) from a large number of bone and soft tissue sarcoma cases should be an international priority since cancer is a leading cause of death in children and bone and soft tissue sarcomas have very high rates of mortality and morbidity.