Astrocytomas have a spectrum of histologic features that correlate with the clinical course [2]. Grade I, pilocytic astrocytoma (PA), is a slow-growing, noninfiltrating tumor with a relatively benign course. Most PAs have a normal diploid karyotype; ∼30% show gain of one or more chromosomes. Approximately half of children younger than 15 years old show a single extra chromosome, while those greater than 15 years of age show gain of multiple chromosomes. Gains of chromosomes 5 and 7 are the most frequent, followed by chromosomes 6, 8, 11, 12, 17, 19, and 22. See Fig. 16.1. Tandem duplication of BRAF at7q34 is found in >50% of pediatric PA and in up to 66% of all PA [3, 4]. Tandem duplication produces a novel fusion gene, KIAA1549-BRAF, which has constitutive kinase activity [4, 5].

Grade II diffuse astrocytoma and grade III anaplastic astrocytoma show increasing cellularity and nuclear pleomorphism. Gliomas of World Health Organization (WHO) grade II or III are invasive, progress to higher-grade lesions, and have a poor outcome. As the tumor progresses, acquired genetic abnormalities include loss of 10q, 13q, 17p, and 19q, and TP53 and RB1 mutations. The majority of WHO grade II and III gliomas, including astrocytomas and oligodendrogliomas, have IDH1 (2q34) or IDH2 (15q26.1) mutations. Most also have TP53 mutation. Patients with mutations of IDH genes have longer median overall survival, 65 versus 20 months [6].

Grade IV, glioblastoma (GB), is the most malignant astrocytoma. In addition to increases in cellularity, nuclear atypia, and pleomorphism, mitotic activity, vascular proliferation, and necrosis are characteristic. Genetic changes in primary glioblastomas, which account for >90% of GBs, include loss of 9p (CDKN2A deletion), loss of 10q (PTEN deletion), PTEN mutation, frequent EGFR amplification, and infrequent MDM2 amplification. The secondary glioblastomas (5–10%) that result from progression of grade II–III tumors have mutations of IDH1 or IDH2 and TP53. Tumors with wild-type IDH1 and IDH2 have fewer TP53 mutations and frequent alterations of PTEN, EGFR, CKDN2A, or CDKN2B [1, 6–8]. Median overall survival in patients with IDH mutations is twice that of patients with wild-type IDH, 31 versus 15 months [6]. Double minutes (dmin) with EGFR amplification have been demonstrated in up to 50% of GBs [9]. Current clinical trials are evaluating several targeted molecular therapies [10].

Oligodendroglioma (OD), a diffusely infiltrating well-­differentiated (grade II) to anaplastic (grade III) glioma, is a relatively rare primary brain tumor, comprising 2–5% of all primary brain tumors. ODs characteristically show an unbalanced der(1;19)(q10;p10) with loss of 1p and 19q (Fig. 16.2). These aberrations may be detected by conventional chromosome analysis, fluorescence in situ hybridization (FISH), loss of heterozygosity (LOH) studies, or microarray comparative genomic hybridization (aCGH) [11, 12]. 1p/19q codeletion is associated with a longer median survival (∼10 years vs. 2 years) in patients with OD and anaplastic OD [13]. Anaplastic oligodendrogliomas show loss of 9p (CDKN2A/B) in additional to 1p/19q loss [14]. Recent studies have shown that IDH1 mutations are strongly associated with 1p/19q codeletion (∼85%) and that IDH1 mutation and 1p/19q codeletions are independent prognostic factors [6, 15].

Oligoastrocytomas (OA) are composed of a mixture of distinct cell types found in oligodendroglioma and astrocytoma. Genetic analysis reveals 1p/19q codeletion in ∼30–50%, TP53 mutations in ∼30%, and IDH mutations in 100% of OA. Progression-free survival (PFS) was 60 versus 30 months in patients with and without 1p/19q codeletion, respectively [1, 16]. Few studies have attempted to evaluate the genetics of the distinct cellular components. Rare tumors have been shown to have different genetic aberrations in the two histologic elements [17–19]. Anaplastic OA (grade III) acquire additional genetic alterations—e.g., 9p (CDKN2A/B) loss, which is also implicated in the progression of astrocytoma and OD. Recent studies of anaplastic OD and OA confirm that presence of 1p/19q codeletion correlates with significantly longer survival than presence of genetic abnormalities other than 1p/19q codeletion [20, 21].

Ependymomas are slow-growing neuroepithelial neoplasms that arise from the wall of the intracranial ventricles and the central canal of the spinal cord. Little is known about the genetics of myxopapillary ependymoma, a grade I tumor that arises predominately in the terminal spinal cord of young adults. The grade II ependymoma may arise at any site in the ventricular system and spinal canal with the posterior fossa the most common site in children. The most frequent cytogenetic aberrations include losses of 22q and 6q and gains of 1q and 9q (Fig. 16.3).

Alterations of 22q are more frequently identified in spinal tumors. NF2 may be a candidate tumor suppressor gene in spinal ependymoma. However, few mutations in candidate genes, including NF2, TP53, PTEN, MEN2, and CDKN2A, have been identified in pediatric intracranial tumors [1, 3, 22].

Recent studies have identified two clinically and genetically different posterior fossa subgroups [23, 24]. Group A patients are younger (median age 4 years), more often male, more often have higher grade tumors (grade III), have more metastases at recurrence, and have a diminished prognosis compared with group B patients. Group A tumors have predominately balanced karyotypes; some show gain of 1q. Group B tumors frequently show loss of 6q and 22q and gain of 9q, 15q, and 18q. Intracranial tumors are more common in children than adults, with spinal tumors more common in adults. Tumors with gain of 1q correlate with higher grade, tumor recurrence, and worse prognosis [25].

Medulloblastoma (MB) is a primitive small round cell tumor, which may show glial or neuronal differentiation. MB, also referred to as an infratentorial primitive neuroectodermal tumor (iPNET), is located in the cerebellum. MB is the most common malignant brain tumor in children, accounting for ∼20% of all pediatric brain tumors. MB is rare in adults, comprising ∼1% of primary adult intracranial malignant tumors. Supratentorial primitive neuroectodermal tumors (sPNETs) are histologically very similar to MBs, but are located in the cerebrum, and are clinically more aggressive.

Cytogenetically, i(17q) is the most common chromosome abnormality in pediatric MBs, present in ∼35% of cases, but it is not found in pediatric sPNET. Recent studies identified characteristic and prognostic cytogenetic subgroups [29–31]. Highest risk pediatric MBs are characterized by gain of 6q, MYC amplification, MYCN amplification, and ERBB2 amplification; intermediate risk MB by 17q gain; and low risk MB by loss of 6q [29, 32]. High-risk genetic abnormalities strongly correlate with the large cell/anaplastic histology found in ∼7% of MBs [29, 33]. MYC/MYCN amplification is rare (∼4–10% of MBs). Patients with low-risk genetic abnormalities, loss of chromosome 6 or 6q, and nodular desmoplastic histology have an excellent 5-year overall survival (OS) of ∼90%. Most (∼90%) MBs have classic histology with a 60% OS. Gain of chromosome 7, common in MB, is not associated with OS [34].

Cytogenetic abnormalities associated with poor OS in adult MBs include high-level amplification of the oncogene CDK6 at 7q21.3 (more frequent than MYC), chromosome 17 aberrations including i(17q), and loss of 10q [30]. Oncogene amplification frequently associates with aberration of chromosome 17. Independent significant predictors for poor prognosis are loss of 10q and gain of 17q. Combined 10q loss and 17q gain show the poorest OS of ∼16%; either 10q loss or 17q gain is associated with a 44% OS, and absence of both 10q loss and 17q gain is associated with the best OS (92%).

Supratentorial PNETs are less well characterized due to a smaller number of sPNETs as compared with MBs. The most frequent cytogenetic aberrations so far associated with sPNET are gain of 1q and losses of 14q, 16p, and 19p/q. Gain of 17q is not found in sPNET. Rare amplifications of 1q, 4q12-q13, 8q22-q24, 19q12-q13, and EGFR are reported [31, 35].

Atypical teratoid/rhabdoid tumor (AT/RT) is a highly malignant tumor in young (<5 years of age) children that may arise infratentorially (posterior fossa) or supratentorially. AT/RTs account for ∼10% of CNS tumors in infants. The tumor is defined by the rhabdoid cell, which resembles a rhabdomyosarcoma cell. Pathologically, AT/RTs may be difficult to recognize by histology alone due to variable components of primitive neuroectodermal, mesenchymal, and epithelial features. AT/RT may be distinguished from other poorly differentiated and anaplastic CNS tumors (e.g., sPNET, MB, or ependymoma) by a characteristic loss of chromosome 22 or 22q deletion [36]. The region of loss includes the SMARCB1/INI1 gene at 22q11.23. The overall prognosis for patients with AT/RT is poor, particularly in patients diagnosed at <3 years of age. AT/RT may be sporadic or part of the rhabdoid tumor predisposition syndrome [36, 37]. Germline deletion or mutation of SMARCB1 has been identified in up to 35% of patients with rhabdoid tumors [38]. Investigation of familial cases has revealed other affected relatives, unaffected carrier parents, and gonadal mosaicism. Individuals with germline mutations of SMARCB1 present at a younger age (5 vs. 18 months), may have multiple primary tumors, or affected siblings as expected with a germline tumor suppressor gene mutation [36, 38, 39].

Renal cell carcinoma (RCC) is the most common malignant tumor arising from the kidney. Prognosis is related to histologic subtype and tumor stage at diagnosis. Histologic subtypes include clear cell RCC (70%), papillary RCC (10–15%), chromophobe RCC (4–6%), Xp11.2 translocation RCC, and others. Different subtypes are characterized by different genetic abnormalities. Hereditary syndromes with RCC as a feature include von Hippel-Lindau, Birt-Hogg-Dube, tuberous sclerosis, hereditary papillary RCC, familial clear cell RCC, hereditary leiomyomatosis and RCC, and familial oncocytoma. Hereditary RCCs account for 4% of RCCs [7].

Prostate adenocarcinoma is the most common cancer in males, representing 29% of cancers and causing 9% of cancer deaths in men [74]. Prognosis for patients with prostate cancer correlates with stage and grade of disease at diagnosis. Frequently reported chromosome alterations include loss of 8p, 10q (PTEN), 13q (RB1), 17p (TP53), 7p, 16q, 6q, and gain of 8q24.1 (MYC) and 7q31. Hemizygous or homozygous deletion of PTEN at 10q23.3 correlates with disease stage, disease progression, and survival [75–77]. PTEN loss is among the most frequent recurring abnormalities in prostate cancer and is seen in preinvasive prostatic intraepithelial neoplasia as well as in invasive prostate cancer.

In 2005, TMPRSS2-ETS gene fusions were discovered in prostate cancer. TMPRSS2, transmembrane protease serine 2 gene, is androgen-regulated. The fusion of TMPRSS2 at 21q22.3 with an ETS gene results in overexpression of the 5′ truncated ETS oncogene. The most common fusion (TMPRSS2-ERG, present in >50% of prostate specific antigen-screened localized cancers) results from an interstitial ∼2.8 Mb deletion within 21q22; ERG is located at 21q22.2 [78, 79]. TMPRSS2-ERG fusion is associated with PTEN deletion and with earlier disease recurrence of localized prostate cancer after surgical resection. A study by Markert found that the poorest overall survival (57 months mean OS) correlated with TP53 and PTEN inactivation. TMPRSS2-ERG fusion positive tumors had a mean OS of 93 months compared with a mean OS of >103 months in the most favorable group. Data from this study and others suggests that a subset of TMPRSS2-ERG fusion-positive tumors with amplification of ERG have a poorer outcome [80].

Bladder cancer can generally be divided into superficial, invasive, and in situ categories, which correspond to the TNM (“tumor, nodes, metastasized”) staging scheme. Tumor stage is the most important independent predictor of patient prognosis. Most tumors are superficial at diagnosis. However, during the course of the disease, multiple recurrences are common, with a minority of superficial tumors progressing to muscle invasion or metastatic disease. Identification of genetic changes in exfoliated cells from the bladder has shown utility in detecting disease recurrence. Primary noninvasive (Ta) or superficially invasive (T1) transitional cell carcinoma of the bladder can be monitored using FISH with centromeric and locus-specific DNA probes. UroVysion®, a US Food and Drug Administration (FDA)-approved FISH probe set, detects aneusomy of chromosomes 3, 7, and 17, and 9p21 (CDKN2A) loss in patients with hematuria or a history of bladder cancer [81, 82]. Recent studies support FISH analysis for use in monitoring and predicting recurrence risk in patients with non-muscle-invasive bladder cancer and predicting residual tumor load after transurethral resection [83–86]. FISH analysis in combination with cytology and the telomeric repeat amplification protocol (TRAP) assay to detect telomerase activity increases the sensitivity of detection in low-grade and early-stage cancers [87]. See also Chap. 17, Fig. 17.16.

Endometrial stromal tumors (EST) are rare uterine mesenchymal tumors that occur in women of reproductive and postmenopausal age. In 2003, the World Health Organization (WHO) divided ESTs into three categories: endometrial stromal nodules (ESN), endometrial stromal sarcomas (ESS), and undifferentiated endometrial sarcomas (UES). ESNs are benign, circumscribed tumors. ESSs are low-grade, malignant tumors that invade the myometrium. Both can have variant histologic features, but most have a classic morphology. UESs are highly aggressive, malignant tumors with cytologic atypia and high mitotic activity. ESSs may be primary uterine, metastatic uterine, or primary extrauterine. The tumors are characterized by reciprocal translocations and gene fusions with t(7;17)(p15.2;q11.2)/JAZF1-SUZ12 in ESN, ESS, and rarely UES, and t(6;7)(p12.3;p15.2)/JAZF1-PHF1 or t(6;10)(p12.3;p11.22)/EPC1-PHF1 and variants in ESS [88–91]. Lee et al. recently identified a t(10;17)(q22.3;p13.3)/FAM22-YWHAE in high-grade ESS [68]. This same t(10;17) is found in clear cell sarcoma of the kidney (see earlier) [67].

Testicular GCTs (TGCT), divided into seminomas and nonseminomas, are the most common tumor of men in the second to fourth decades and are responsible for ∼10% of cancer deaths of men in this age bracket. The US incidence is ∼6/100,000 with a 5:1 white: black ratio. Cryptorchidism is associated with ∼10% of TGCTs. Seminomatous GCTs, ∼40–50% of GCTs, are composed of cells that resemble primordial germ cells. Nonseminomatous GCTs (NSGCT) may be composed of undifferentiated cells that resemble embryonic stem cells. Malignant cells can differentiate to generate yolk sac (endodermal sinus) tumor, embryonal carcinoma, choriocarcinoma, teratomas, and mixed malignant GCTs. Approximately 60% of TGCTs are composed of more than one of these cell types [95]. TGCT of infants are rare neoplasms occurring in boys 0–4 years of age, the majority of which are pure yolk sac tumors [96].

GCTs account for 15–20% of all ovarian tumors. Most are benign cystic teratomas. The remaining, primarily found in children and young adults, may be malignant with histologic types similar to those in the testis. The counterpart to the testicular seminoma is the ovarian dysgerminoma, which is always malignant. Dysgerminomas account for ∼2% of all ovarian tumors, but ∼50% of GCTs. They may occur in patients with gonadal dysgenesis.

Rare primary brain GCTs occur in the midline, pineal (male predominance), or suprasellar regions. They account for 0.2–1% of brain tumors in those of European descent, but up to 10% in those of Japanese origin.

Primary mediastinal germ cell tumors account for 10–15% of all mediastinal tumors [97]. Pediatric mediastinal GCTs represent ∼5% of all GCTs [98]. These tumors have the same histologies as gonadal GCTs, but have a worse prognosis. The mediastinum is the most common site of extragonadal tumors in young males. Mediastinal NSGCTs are associated with Klinefelter syndrome in ∼20% of cases [7]. Patients with Klinefelter syndrome have a 50-fold higher risk for a mediastinal GCT, but do not develop testicular GCT [2].

Sacrococcygeal GCTs, the most common extragonadal GCT in children, present prenatally to ∼4 years of age. Most (∼90%) external lesions are benign, while intrapelvic or intra-abdominal tumors are more likely to be malignant (60–90%) [99].

Cytogenetically, additional copies of all or part of 12p are the characteristic chromosome abnormality associated with GCTs in adults and postpubertal children [95, 100]. Additional copies of 12p are present as i(12p) in 80–90% of cases, while the remaining tumors show 12p amplification. Additional cytogenetic anomalies in this group include gains of an X chromosome and chromosomes 7, 8, 12, and 21 and loss of 1p, 11, 13, and 18 (Fig. 16.8).

GCTs in infants and prepubertal children characteristically show loss of 1p36, 4q, and 6q, and gain of 1q, 3p, 16p, and 20q. Gain of 12p is rarely reported in prepubertal children and infants [101, 102]. Array CGH has shown proximal 12p11.2-p12.1 gain associated with adult TGCTs and distal 12p12-pter gain in yolk sac tumors of very young children [102–104]. GCTs are rare in children between 5 and 9 years of age; thus there is a paucity of tumor genetic information in this age range.

Gastrointestinal stromal tumors (GIST), the most common mesenchymal tumor of the gastrointestinal (GI) tract, arise from the connective tissue of the GI wall. Approximately 90% of GISTs have activating mutations of the KIT or PDGFRA receptor tyrosine kinase genes [105]. In addition to KIT or PDGFRA mutations, other genetic events involved in tumorigenesis include primarily chromosomal losses (14q, 22q, 1p, 15q, 13q), nuclear/mitochondrial microsatellite instability, LOH at 9p21 (CDKN2A), methylation of CDKN2B (p15), homozygous loss of TLX2 (HOX11L1), and rare gene amplification (MYC, MDM2, EGFR, CCND1, KIT) [106–109]. The KIT and PDGFRA mutations are diagnostic, and tyrosine kinase inhibitors (TKIs) are used as targeted therapy. The molecular aberrations of KIT-PDGFRA are correlated with cell histomorphology, metastasis, prognosis, and efficacy of targeted therapy. Genotyping these tumors helps to guide therapy, as the effects of TKIs vary with the presence or absence and site of KIT-PDGFRA mutation. Metastatic GISTs often have secondary KIT kinase mutations, and some have KIT-PDGFRA genomic amplifications, which are responsible for therapeutic resistance [105]. Disease-free survival correlates with mutation (KIT  =  poor), site (stomach  =  best), cytogenetic complexity (≥3 abnormalities  =  poor), and losses of 1p and/or 22q and gains of 1q and 12q (shorter survival) [110, 111]. GISTs in patients with neuro-fibromatosis type I (NF1) lack KIT and PDGFRA mutations. Rare families have been reported with germline KIT or PDGFRA mutations [105, 110].

Dermatofibrosarcoma protuberans (DFSP) is an intermediate-grade soft tissue malignancy that usually arises in the dermis and subcutaneous tissue of adults and rarely in children [135]. DFSP is a slow-growing infiltrative dermal neoplasm with a propensity to recur locally after surgical resection, but is rarely metastatic (1–4%). Tumor-related deaths are very rare. There are several histologic variants of DFSP, such as giant cell fibroblastoma, Bednar tumor, and other fibrohistiocytic tumors, which should be considered in the differential diagnosis.

DFSP is characterized by a balanced or unbalanced form of a translocation t(17;22)(q21.3;q13.1) or by supernumerary ring chromosomes derived from this chromosome 22 [der(22)r(17;22)] that contain low-level amplified sequences from 17q21.31-qter and from 22q11.1-q13.1 [136]. The unbalanced form is usually a der(22)t(17;22) (Fig. 16.12). Both the ring and linear forms of the derivative chromosome 22 result in fusion of the α(alpha)-1 chain of type 1 collagen gene (COL1A1) at chromosome 17q21.3 with the second exon of the platelet-derived growth factor–β (PDGFB) gene at chromosome 22q13.1. Variability of the COL1A1 break point has no correlation with any clinical or histological parameter [135, 137–139]. However, cytogenetically, ring chromosomes are common in adult DFSP, while the translocation derivatives are seen in all childhood cases [140]. The COL1A1-PDGFB chimeric gene protein causes unregulated expression of platelet-derived growth factor leading to abnormal activation of the platelet-derived growth factor receptor beta (PDGFRB) tyrosine kinase through an autocrine loop. Tyrosine kinase inhibitors are the current therapy for recurrent, metastatic, or inoperable tumors [141, 142]. Demonstration of the COL1A1-PDGFB fusion is necessary for the diagnosis of DFSP or DFSP variants [137].

Hidradenoma, a benign sweat gland tumor, often presents as a solitary, slow-growing, solid, or cystic intradermal nodule. Malignant transformation is uncommon. t(11;19)(q21;p13.11) is characteristic of a subset of these tumors (Fig. 16.11). The CRTC1-MAML2 fusion was demonstrated in 50%, specifically in tumors with clear cells, representing the clear cell variant of hidradenoma [133]. As noted earlier, salivary gland Warthin’s tumor and mucoepidermoid carcinoma also demonstrate the t(11;19)/CRTC1-MAML2 fusion. The common glandular denominator for these different tumors suggests they may originate from a common progenitor cell in salivary, bronchial, and sweat glands [130, 133].

Melanomas are malignant lesions, primarily cutaneous, but may occur on mucosal surfaces and in the eye. Cutaneous malignant melanoma, a pigmented skin lesion, may be lethal if not recognized and completely excised prior to metastasizing. Malignant melanomas spread superficially before progressing to invade the deeper dermal tissues. Malignant melanoma may be recognized by visible changes of a pigmented lesion (mole) characterized by changes in size and color and irregular borders, and may evolve from dysplastic nevi. Individuals with dysplastic nevus syndrome have a 50% risk for developing melanoma by 60 years of age. Frequent aberrations found in melanomas include gains of 6p, 1q, 7p, 7q (BRAF), 8q, 17q, 11q, and 20q and losses of 9p (CDKN2A), 9q, 10q (PTEN), 10p, and 6q [143, 144].

Uveal melanoma, the most common form of primary eye cancer, is characterized by loss of chromosome 3 in ∼50% of tumors; metastasis is correlated with such loss. Chromosome 3 loss is often accompanied by i(8q); tumors without loss of chromosome 3 have 6p abnormalities.

Two regions of chromosome 3, 3p25 and 3q24-26, appear to harbor tumor suppressor genes. More than 80% of uveal melanomas have been found to have a constitutively active somatic mutation of one of two genes, GNA11 at 19p13.3 and/or GNAQ at 9q21.2. These genes appear to contribute to the development of these tumors [145, 146].

Breast carcinoma is the most common cancer in women and the second leading cause of cancer deaths in women. The lifetime risk for breast cancer for women in the general population is 1 in 8. A positive family history of breast cancer increases this risk. A germline mutation of one of the known breast cancer predisposing genes greatly increases risk [2]. Currently known breast cancer genes explain only ∼30% of the heritability. Mutations of the breast cancer predisposing genes BRCA1 and BRCA2 account for ∼16–20% of the familial risk of breast cancer in the general population [147, 148].

Lung cancer, the most common cancer worldwide, is largely due to tobacco products. Incidence and mortality rates of lung cancer have been declining since ∼1990 secondary to decreased smoking rates over the past 30 years. The most common types are non-small cell carcinomas (which include adenocarcinoma, squamous cell carcinoma, and large-cell carcinoma) and small-cell carcinomas [7].