Molecular Genetic Pathology Of Solid Tumors

Brain Tumor

Primary central nervous system (CNS) gliomas, originating exclusively from brain cells such as astrocytes, oligodendrocytes, and ependymal cells, account for only about 1.35% of all cancers, but rank second among the causes of death from neurologic disease. Glioblastomas are the most common primary CNS tumor and are categorized as World Health Organization (WHO) grades I to IV based on histologic characteristics. With the advent of new treatment modalities, the use of microscopic examination alone is insufficient for the histologic classification and grading of gliomas. Under the 2016 WHO classification schema, molecular and genetic data as layer 4 information supports histologic classification ( ).

Oligodendroglioma

Oligodendroglioma is an infiltrating glioma of the cerebral cortex diagnostically characterized by a triad of uniformly round to ovoid nuclei, perinuclear halos, and an even distribution of cells, together with a delicate chicken wire type of vasculature. However, in a significant number of these lesions, the microscopic morphology is not so clear-cut, and distinction from other diffuse glial lesions may be difficult. Combined loss of 1p and 19q ( Fig. 79.1A to 79.1C ) is typical in oligodendroglioma ( ), and loss of 19q occurs in astrocytoma and mixed oligoastrocytoma ( ). The incidence of 1p/19q loss varies from 50% to 80% in oligodendroglioma in different studies and is 1% to 10% in other gliomas, which indicates its utility in differentiating the diagnosis. Recent studies suggest that the combined loss of 1p/19q may follow a (1;19)(q10;p10) translocation, with subsequent loss of the derivative chromosome der(1;19)(q10;p10) ( ). Identification of 1p/19q loss is associated with two unique features of tumor biology with clinical indications: first, they grow slowly, even those that are anaplastic in nature; second, they correlate with better prognosis and response to chemotherapy ( ). In terms of treatment response, initial studies suggested that 1p/19q loss is a marker of response to PCV (procarbazine, lomustine/CCNU, and vincristine) or temozolomide chemotherapy ( ). Thus, the 1p/19q status has become an important part of diagnosis, prognosis, and predicted therapeutic response of oligodendrogliomas. In addition to 1p/19q alteration, LOH mutations in p53 and p16 may be associated with poor survival or tumor progression ( ).

Nearly all 1p/19q codeleted oligodendrogliomas are also mutated on isocitrate dehydrogenases (IDHs), IDH1 or IDH2 ( ). IDH1 has been found to be mutated in a vast majority of astrocytic, oligodendroglial, and oligoastrocytic gliomas (WHO grades II to III) as well as in secondary glioblastomas (WHO grade IV). However, IDH1 mutation is very rare in primary glioblastoma and is not involved in pilocytic astrocytomas. IDH1 and IDH2 are homologous, NADP ⁺ -dependent cytoplasmic and mitochondrial enzymes, respectively. The role of these enzymes is the conversion of isocitrate to α-ketoglutarate with the simultaneous reduction of NADP ⁺ to NADPH. The most common mutation is a heterozygous point mutation with substitution of arginine by histidine at residue 132 (R132H), located in the substrate-binding site. This IDH1-R132 mutation has a reported frequency of 50% to 93% ( ; ). IDH2 gene mutations affecting the amino acid R172 are much less common than the IDH1 isoform, 3% to 5%, yet have been identified in a small subset of gliomas that lack the typical IDH1 mutation with predominance in oligodendrogliomas ( ). IDH1 and IDH2 genes appear to behave dominantly and are mutually exclusive. IDH1 mutation has been shown to be a strong, independent positive prognostic biomarker in diffuse gliomas, glioblastomas, and oligodendroglioma ( ). IDH mutation was also identified as a marker for response to temozolomide in low-grade gliomas ( ).

Breast Cancer

Breast cancer is the most common cancer in women and the second most common cause of cancer death in women in the United States. Until recently, breast cancer was characterized according to tumor type (ductal vs. lobular carcinoma), histologic grade (I to III), steroid hormone receptors (estrogen receptor [ER] and progesterone receptor [PR]), and Her2/neu status (positive vs. negative), along with metastasis to lymph nodes and distant organs, for its prognosis and treatment. Gene expression profiling has also been utilized as an adjunct for the management of breast cancer. Gene panel study and germline testing using NGS have been introduced to search for actionable mutations and identify genetic predisposition for counseling and screening.

Sporadic, Nonhereditary Breast Cancer

Various genetic, epigenetic, and genomic changes have been associated with breast cancer. Traditionally, ER and PR status is used as a prognostic and predictive marker for breast cancer. Her2/neu (detailed in Chapter 77 ) status has been added in the past decade as a breast cancer prognostic and predictive marker. The FISH method has been used primarily to determine the copy number of the Her2/neu gene (see Fig. 79.1D and 79.1E ) for the purpose of selecting Her2-targeted therapies such as trastuzumab and lapatinib in both adjuvant and neoadjuvant settings. Enumeration of 20 interphase nuclei from tumor cells on a given case is reported as the ratio of average Her2/neu copy number to that of chromosome enumeration probe 17 for centromere, CEP17. Specimens with amplification showed a Her2/neu :CEP17 signal ratio ≥2.0 as abnormal and a ratio <2.0 as normal. A three-color FISH assay has been commercialized for assessment of standalone prognosis in ER-positive and ER-negative stage I breast cancers ( ). In addition, HercepTest (Dako Corp, Carpinteria, CA), an immunohistochemistry (IHC) test for Her2/neu using a polyclonal antibody, was the first approved by the U.S. Food and Drug Administration (FDA). Subsequently, Pathway (Ventana Medical Systems, Tucson, AZ) produced a monoclonal antibody (CB11) for Her2/neu for the same response indication, which has also been FDA-approved and is used in combination with FISH ( Fig. 79.2 ). IHC is widely used to measure prognostic factors, including ER, PR, Her2/neu , and the proliferation marker, Ki-67, and to predict response to hormonal and Her2-targeted therapies ( ).

Concerns associated with IHC-based assays for multiple markers include the nonlinear nature of IHC staining, the different subcellular localizations of different markers, and the impact of different slide-scoring thresholds for different antibodies. The introduction of whole-genome profiling technologies has greatly expanded our knowledge of the genomic pathways associated with the development and progression of breast cancer ( ). These research results have led to the development of several commercialized tests with multigenes/proteins as prognostic and predictive tools for clinical application using IHC, FISH, RT-PCR, and genomic microarray technologies with integrated bioinformatic/statistical algorithms designed to calculate disease recurrence and patient survival ( ).

Molecular classification of breast cancer has been proposed, based on gene expression profiling, to include luminal-like, normal-like, Her2-positive, and basal-like subtypes of invasive breast cancer ( ). It is important to note that all of the luminal groups of breast cancers are ER positive, and nearly two-thirds are of low or intermediate histologic grade, whereas 95% of all ( ) basal-like cancers are ER negative and 91% of these tumors are of high grade ( ). Studies of clinical outcomes based on molecular subtypes have shown that luminal A is more sensitive to antiestrogen than is luminal B. Her2-positive responds to trastuzumab, an anti–Her2 antibody, whereas basal is more aggressive (but is especially sensitive to anthracycline-based neoadjuvant chemotherapy) than are the luminal breast cancers ( ). Basal-like and Her2-positive subgroups are associated with the highest rates of pathologic complete response to neoadjuvant multiagent chemotherapy ( ).

Several commercially available assays have been developed to predict the risk for recurrence. Prosigna Breast Cancer Assay, developed by NanoString, is based on the prediction analysis of the microarray-50 (PAM50) gene expression signature. The PAM50 gene signature measures the expression levels of 50 genes in a surgically resected breast cancer sample to classify a tumor as one of four intrinsic subtypes (Luminal A, Luminal B, HER2-enriched, and Basal-like). This test outputs a risk of recurrence (ROR) score in these intrinsic subtypes. Prosigna uses multiplexed gene-specific fluorescently labeled probe pairs to measure gene expression in frozen or formalin-fixed paraffin-embedded (FFPE) tissues with equivalent ease and efficiency ( ). Clinical trials demonstrated its utility in prognostic information of distant recurrence in hormone receptor–positive, postmenopausal breast cancer patients treated with endocrine therapy ( ) and in node-positive patients ( ). The other two gene microarray-based clinical tests are the OncotypeDx and MammaPrint assays ( Fig. 79.3 ) ( ; ). Oncotype Dx (Genomic Health, Inc., Redwood City, CA) is a 21-gene multiplex prognostic and predictive RT-PCR assay performed on primary FFPE breast cancer samples. The original 16 informative genes ( ) that calculate the recurrence score (RS) were discovered on archived FFPE samples by transcriptional profiling and then were converted to the FFPE RT-PCR assay ( ). OncotypeDx determines the 10-year risk for disease recurrence in patients with ER-positive, lymph node–negative tumors by using a continuous variable algorithm and assigning a tripartite RS (17, low risk; 18–30, intermediate risk; >30, high risk). Of the multiple pathways assessed by this assay, the proliferation and ER pathways are the most influential in RS calculation, followed by the Her2 pathway. High relative levels of ER messenger RNA (mRNA) and low levels of Ki-67 proliferation gene mRNA have a low RS. Low levels of ER mRNA and high levels of Ki-67 mRNA have a high RS. The other 14 informative mRNA levels play their greatest roles in determining the RS in tumors with intermediate ER and Ki-67 mRNA levels. It should also be noted that OncotypeDx is best suited for detecting breast cancers with a low potential for recurrence. The MammaPrint assay (Agendia BV, Amsterdam, The Netherlands) was the first fully commercialized microarray-based multigene assay for breast cancer. This test is offered as a prognostic test for lymph node–negative breast cancer in women younger than 61 years of age with an ER-positive or ER-negative tumor. The prospective RASTER trial showed that MammaPrint testing could spare from chemotherapy 94 of 295 patients (32%) for whom chemotherapy would have been recommended by a standard (non–gene-based) decision tool ( ). The clinical utility of the assay supported the test to be the first assay approved by the FDA’s new in vitro diagnostic multivariate index assay classification. Like Prosigna and OncotypeDx, both fresh-frozen and FFPE tissues can be used for analysis. The 70 genes that make up the MammaPrint assay are focused primarily on proliferation. Additional genes are associated with invasion, metastasis, stromal integrity, and angiogenesis. Of note, the OncotypeDx and MammaPrint assays have only one gene in common—the SCUBE2 gene, which is a member of the ER pathway. Unlike the OncotypeDx test, which features a continuous RS result, the MammaPrint test uses a dichotomous “high risk versus low risk” result format. The MammaPrint test does not include ER, PR, or Her2 in the 70-gene microarray.

Multigene testing procedures such as disease classifiers and prognostic and predictive markers have been introduced with the use of slide-based methods (e.g., IHC, FISH) or molecular platform–based methods such as quantitative multiplex PCR and genomic array profiling. Morphologically based mRNA extraction performed with the use of tissue macrodissection or microdissection is recommended for analysis as it is highly enriched for invasive carcinoma and is not diluted with cells from benign tissues and in situ carcinoma areas ( ). The heterogeneous expression of important mRNAs—such as ER, Her2, and Ki-67, often reflected in the varying histologic grades seen in larger tumors—can influence the predictive accuracy of transcriptional profiling measurements. One caveat is that, although the number of genes that can be simultaneously assessed by multiplex qRT-PCR is significantly greater than that for IHC, multiplex qRT-PCR requires a more complex statistical evaluation of the gene expression profiles. Nevertheless, the RT-PCR technique has been used to predict overall prognosis and response to both hormonal and cytotoxic therapies.

High-throughput NGS has made it possible to sequence large numbers of genes to identify actionable mutations in the tumor. The diagnostic yield of breast NGS has been studied but is difficult to define and measure ( ). Sequencing of 46 cancer-related genes in 415 breast cancer samples (including some primary-metastatic pairs) revealed somatic nonsynonymous mutations in 220 of 354 patients (62.1%) ( ). A systematic multinational study of NGS profiling in breast cancer is underway ( ).

Thyroid Cancer

Thyroid cancer is the most common endocrine neoplastic condition. The two most frequent types of thyroid malignancy derived from follicular cells are papillary and follicular carcinomas, which constitute approximately 80% and 15%, respectively, of all thyroid cancers. These follicular cell–derived tumors are well differentiated, in contrast to poorly differentiated and anaplastic carcinomas, which constitute about 2% of cases. Another malignancy of the thyroid is medullary carcinoma, which originates from parafollicular C cells and accounts for approximately 3% of cases. Several genetic alterations in various thyroid tumors have been well documented, including translocations and point mutations. The genetic alteration in thyroid cancer involves BRAF , RAS , RET , and PAX8 .

Lung Cancer

Lung cancer remains the most common cause of cancer-related death among men and women worldwide. Epithelial tumors of the lung are of four principal types: squamous cell carcinoma, adenocarcinoma, large cell carcinoma, and small cell carcinoma. Clinically, the most important distinction is noted between small cell carcinoma and the others, which can be grouped together as non–small cell carcinoma. Small cell cancers respond to chemotherapy, whereas non–small cell lung cancers (NSCLCs) are removed surgically if possible. However, with the rapid advancements in the understanding of NSCLCs and the corresponding growth in available molecularly targeted therapy, recognition of specific molecular alterations is required for optimal patient care.

EGFR mutations have been identified in many human cancers, such as lung, breast, head and neck, colorectal, pancreatic, and bladder cancers; glioma; squamous cell carcinoma; Wilms tumor; and certain NSCLCs (adenocarcinoma). EGFR is a membranous oncoprotein that induces cell proliferation upon activation. Binding of a ligand to a receptor activates its tyrosine kinase activity, which phosphorylates several substrates on a number of signal transduction pathways, leading to increased signaling activity and activation of downstream effectors. This results in uncontrolled cell proliferation, tumor invasion, angiogenesis, and resistance to normal apoptotic signals (see Chapter 77 ). Increased EGFR signaling can be the result of EGFR gene amplification, protein overexpression, or specific activation mutations in the EGFR gene, as described in detail in Chapter 77 .

It has been shown that a subset of patients with NSCLC (10%–40%) ( ) have specific activating mutations in the EGFR gene ( Fig. 79.4 ) that are associated with increased sensitivity to TKIs, such as gefitinib (Iressa) or erlotinib (Tarceva), targeting EGFR in lung cancer. Common activating EGFR mutations occur most often in exons 18 to 21 and cluster in two major hot spots. The prototype mutation L858R in exon 21 (40% of EGFR mutations) encoding the tyrosine kinase domain and small in-frame deletions in exon 19 (>50%) are reportedly most often constituting up to 90% and are associated with responses to EGFR TKI therapy. Mutations in exons 18 and 20 account for the remaining 10% of EGFR mutations in NSCLC ( ). Recent guidelines suggest that routine EGFR assays should include EGFR exon 19 for deletions and insertions, EGFR exon 18 for E709 and G719 mutations; exon 20 for S768I, T790M, and insertions; and exon 21 for L858R, T854, and L861Q mutations. Clinical EGFR mutation testing should be able to detect all individual mutations that have been reported with a frequency of at least 1% of EGFR-mutated lung adenocarcinomas ( ). Studies have reported that EGFR gene mutations are more common among females, Asians, and nonsmokers with adenocarcinoma; these are the same groups that have the highest response rates to TKIs ( ). However, recent guidelines ( ) suggest that EGFR molecular testing should be done in all patients with lung adenocarcinoma component. T790M mutation detection is required by 2018 guidelines reversion for relapse cases ( ).

Immunohistochemistry for total EGFR and EGFR copy number analysis (i.e., FISH or chromogenic in situ hybridization) are not recommended for selection of EGFR TKI therapy according to recent guidelines. Recent studies have demonstrated that in patients with EGFR mutations, the EGFR locus is often concurrently amplified.

ALK gene chromosomal rearrangement is found in another ∼7% of lung adenocarcinomas, most commonly in the form of an intrachromosomal inversion leading to the EML4-ALK fusion product associated with ALK protein overexpression. This mutation is present more frequently in lung adenocarcinomas from younger patients with either no or only light smoking history ( ; ). However, recent guidelines suggest that ALK molecular testing should be done in all patients with a lung adenocarcinoma component. Some studies have reported associations with solid, mucinous cribriform and/or signet ring histology ( ; ). The translocation is infrequent in pure squamous cell carcinoma but has been reported in adenosquamous carcinoma ( ). Patients with this tumor type are responsive to therapy with the multitargeted TKI crizotinib ( ). FISH using break-apart probes is currently considered the gold standard for detection of ALK rearrangements ( ). Polysomy (multiple copies) at the ALK locus is common in lung adenocarcinoma and, when present, confirms that FISH has been performed in a tumor cell population. ALK immunohistochemistry has been included as an alternative to FISH by 2018 updated guidelines ( ). However, current evidence suggests that it does not predict response/resistance to targeted therapies. RT-PCR is not recommended as an alternative to FISH for selecting patients for ALK inhibitor therapy ( ).

Recent recommendations for EGFR and ALK testings are as follows: EGFR and ALK testing is not recommended in lung cancers that lack any adenocarcinoma component, such as “pure” squamous cell carcinomas, “pure” small cell carcinomas, or large cell carcinomas lacking any IHC evidence of adenocarcinoma differentiation. EGFR and ALK testing are the most important uses of the diagnostic sample after a diagnosis of adenocarcinoma is established. If specimens are insufficient for molecular testing, patients may need to undergo another invasive diagnostic procedure before they can be treated. NGS enables the use of small specimens for multiple molecular marker detection, avoiding the risks associated with obtaining surgical biopsies. Cell-free circulating DNA can be used for T790M detection in relapse ( ). EGFR and ALK results should be available within 2 weeks (10 working days) of receiving the specimen in the testing laboratory ( ).

Mutations in KRAS , frequently in codons 12 and 13, have been reported in up to 30% of cases of lung adenocarcinoma ( ). These mutations are usually found in cancers of smokers and are more common in adenocarcinoma (30%–50%) than in NSCLC (15%–20%). Conversely, oncogenic mutations in the KRAS gene were reportedly associated with resistance to EGFR-TKIs in metastatic colorectal cancer. Thus, testing for KRAS mutations as a negative predictor of response to EGFR TKI has become part of molecular diagnostic algorithms for lung adenocarcinoma in many centers. However, more recent data suggest that EGFR wild-type tumors have less favorable outcomes if they are treated with EGFR TKI than if they are treated with conventional platinum-based chemotherapy. Therefore, KRAS mutation testing is not recommended as a sole determinant of anti-EGFR TKI therapy ( ).

Lung adenocarcinomas contain a number of other less common alterations that may lead to treatment with targeted inhibitors. These include RET (∼2% of lung adenocarcinomas) ( ), increased copies of MET ( ), and sequence-altering mutations in ERBB2 ( HER2 ), BRAF , and PIK3CA ( ) as well as chromosomal rearrangements involving ROS1 (∼2% of lung adenocarcinomas; may respond to treatment with crizotinib). In fact, ROS1 mutation detection has been strongly recommended for all lung cancer patients regardless of clinical characteristics according to the updated guidelines. According to the updated guidelines from Association for Molecular Pathology (AMP), the College of American Pathologists (CAP) and the International Association for the Study of Lung Cancer (IASLC), multiplexed genetic sequencing panels (e.g., NGS testing with detection of BRAF , ERBB2 , MET , RET , and KRAS gene alterations) are preferred over multiple single-gene tests to identify other treatment options beyond EGFR , ALK , and ROS1 ( ).

Hepatocellular Carcinoma

Hepatocellular carcinoma (HCC) is the sixth most common malignancy and the third most common cause of cancer death worldwide. Major risk factors include chronic viral infection (hepatitis B virus [HBV] and hepatitis C virus [HCV]), alcoholic/nonalcoholic liver disease, environmental carcinogens (e.g., aflatoxin B1), and inherited genetic disorders (e.g., Wilson disease, hemochromatosis, α ₁ -antitrypsin deficiency, tyrosinemia) ( ; ; ; ). The development of HCC is a multistep process, comprising hyperplastic change, dysplasia, early HCC, and well-developed HCC in the setting of chronic hepatitis or cirrhosis ( ; ). A small percentage of HCC develops in patients with nonalcoholic steatohepatitis (∼14%), with 37% with no cirrhosis ( ).

The genomic abnormality of HCC is heterogeneous, largely as the result of various molecular mechanisms of carcinogenesis related to different causes and multifactorial processes ( ). Two major mutually exclusive subtypes of HCC are classified by CTNNB1 (the gene encoding β-catanine) and p53 mutations. Tumors with CTNNB1 mutations are featured with nuclear β-catenin staining and downregulation of the interleukin (IL)-6/JAK-STAT pathway. They are larger well-differentiated tumors with microtrabecular and pseudoglandular architecture. The tumors are cholestatic and are not associated with significant inflammatory infiltrate. Tumors with p53 mutations are related to activation of the PI3K/AKT pathway. They are poorly differentiated, highly proliferative, massive tumors with macrotrabecular architecture and pleomorphic histology, often with sarcomatous and multinucleated components. They have a high propensity for both macrovascular and microvascular invasion. In addition, somatic mutations with TERT are frequent genetic alterations in dysplastic nodules and HCC ( ; ).

Detailed molecular findings in various subtypes of HCC include the following. (1) Scirrhous HCC has been found to express the tuberous sclerosis genes, TSC1 and TSC2 , that promote epithelial to mesenchymal transition (transforming growth factor-β, VIM) and PI3/AKT pathway activation. CTNNB1 is not activated. (2) Steatohepatitic HCC is associated with IL-6/JAK-STAT involvement. It is also featured with lack of Wnt/β-catenin pathway activation. (3) A macrotrabecular massive variant of HCC has been found to express Tp53 mutations, FGF19 amplifications, ATM mutations, angiogenesis activation consisting of upregulation of the ANGPT2 gene, and vascular endothelial growth factor (VEGF) A expression. (4) Carcinosarcomas have been found to have overexpressed or mutant Tp53 involvement, which is found in both the carcinoma and sarcoma parts, PI3K involvement in the carcinoma part and FGFR3 involvement in the sarcoma sections. Activation of protein kinase A resulting from deletion on Chr 19 is reported in (5) fibrolamellar HCC, which leads to fusion between the DNAJB1 gene and the PRKACA gene ( ). Homologous recombination of telomeres has been involved in the pathogenesis of (6) chromophobe HCC. Wnt/β-catenin and transforming growth factor-β signaling are reported to be significantly activated in mixed (7) HCC–cholangiocarcinoma (CCA). (8) A stem cell subclass of HCC is associated with the MYC, mTOR, and NOTCH signaling pathways. Upregulation of the insulin-like growth factor 2 pathway and genes implicated in liver progenitor cells ( PROX1 , HNF-1β , FOXA1 , FOXA3 ) may also be involved ( ; ; ).

In addition, differential expression of miRNA in HCC has been reported to have prognostic value and pathogenesis significance ( ; ). The only curative therapy for HCC is surgical resection or liver transplantation. PD-L1 tumoral expression is observed in 16% of HCC from recent reports ( ). However, the identification of accurate predictive markers and/or immune classifiers to select ideal candidates for immunotherapy is still under investigation ( ).

Using genome-wide expression profiling analysis six, five, eight, and nine samples were analyzed for HBV, HCV, AC, and NAFLD HCC, respectively. The regulatory networks for HBV HCC revealed E2F1 as activated UR, regulating genes involved in cell cycle and DNA replication, and HNF4A and HNF1A as inhibited UR. In HCV HCC, interferon-γ, involved in cellular movement and signaling, was activated, while IL1RN, mitogen-activated protein kinase 1 involved in IL-22 signaling and immune response, was inhibited. In alcohol consumption HCC, ERBB2 involved in inflammatory response and cellular movement was activated, whereas HNF4A and NUPR1 were inhibited. For HCC derived from nonalcoholic fatty liver disease, miR-1249-5p was activated, and NUPR1 involved in cell cycle and apoptosis was inhibited. The prognostic value of representative genes identified in the regulatory networks for HBV HCC can be further validated by an independent HBV HCC data set established in our laboratory with survival data ( ).

Gastric Cancer

Stomach cancer represents 1.6% of all new cancer cases in the U.S. (Surveillance, Epidemiology, and Ends Results). The average annual percent change (AAPC) from 2012 to 2016 is significantly decreased by 1.7%. A majority of gastric cancers (GCs) are adenocarcinomas, historically classified into two major groups by Lauren’s classification: intestinal type with glandular pattern, and diffuse or poorly differentiated type with a signet ring cell morphology in some cases ( ). These two types of GCs have a distinct molecular basis of carcinogenesis, and clinical and genetic characteristics. The main causes of the intestinal type include dietary habits, environmental factors, and Helicobacter pylori infection ( ). Gastric cancer risk is also associated with hereditary tumor syndromes, such as hereditary diffuse-type gastric cancer syndrome, hereditary nonpolyposis colorectal cancer (HNPCC), Li-Fraumeni syndrome, and Peutz-Jeghers syndrome (PJS; i.e., familial adenomatous polyposis [FAP] and juvenile polyposis) ( ; ).

The development of intestinal-type GC is a multistep process that proceeds from gastritis (chronic or atrophic) and related changes (e.g., pernicious anemia) to intestinal metaplasia, dysplasia, and, eventually, carcinoma ( ; ). These processes are accompanied by a series of genetic alterations. The molecular alterations occur in the early stage of carcinogenesis, including molecules leading to genetic instability, inactivation of TSGs by methylation of CpG, telomerase activation, and p53 gene loss and mutation. Elevated telomerase activity was observed in precancerous lesions ( ). The inactivation mutation of the TSG p53 (see Chapter 77 ) is also seen in 38% to 71% of GCs ( ; ; ). Mutation in p73 , a member of the p53 family, is indicated in the development of GC associated with H. pylori infection in mouse model. A high level of microsatellite instability (MSI-H) is frequently seen in intestinal-type GC, largely caused by hypermethylation of the promoter regions of mismatch repair genes (most commonly, MLH1 and MSH2 ) or gene mutations (in a small percentage of GCs) ( ; ; ). The CpG island methylator phenotype, originally found in colorectal cancer, has been detected in 24% to 47% of GCs ( ; ; ). The oncogene product Her2/neu is frequently amplified in intestinal-type GC ( ).

The molecular pathogenesis of diffuse-type GC is less clear. H. pylori infection seems not to be related to the diffuse type of GC according to epidemiologic studies. It can be part of a hereditary gastric cancer predisposition syndrome ( ). The unique molecular abnormality of diffuse-type GC distinct from intestinal-type GC is deficiency in cell-cell adhesion caused by loss or downregulation of the E-cadherin gene ( CDH1 ) as a result of genetic or epigenetic alterations. The abnormality in CDH1 can be found in the early stages of diffuse GC development. Decreased CDH1 expression is detected in 50% of diffuse-type GCs, and more than 70% of those somatic CDH1 mutations consist of complete or partial deletion of exons ( ; ). The unique molecular genetics of GC might be useful for diagnosis in that CDH1 mutations can be detected by PCR on paraffin-embedded tissue. Detection of the germline mutation in CDH1 may help in identification of carriers of asymptomatic mutations involved in the GC family syndrome and may offer an opportunity for prophylactic total gastrectomy in patients carrying the CDH1 mutation ( ). Other frequently altered genes or gene products in diffuse-type GC include the met proto-oncogene, encoding the hepatocyte growth factor receptor, and the SC-1 antigen, an apoptosis receptor ( ; ).

Currently, the most valuable predictive and prognostic factor for GC is clinical staging independent of cancer type, but it does not reflect the heterogeneity of tumor biology. Molecular biomarkers have been investigated as alternatives or supplements to the current system and have shown very promising results in predicting prognosis and therapeutic response. For example, microsatellite instability (MSI) results in an accumulation of genetic abnormalities during the early stages of gastric carcinogenesis ( ). MSI-H is shown in only 4% of stomach carcinomas, especially in well-differentiated types in elderly patients. It is usually associated with intestinal-type GC, commonly seen in the distal stomach or antrum, and is associated with less frequent local lymph node metastasis. However, whether patients with MSI-H GC have more favorable long-term survival than those with MSH-L GC (novel tumor bands with one marker)/microsatellite stable [MSS], no novel tumor bands) remains controversial ( ; ). Expression in tumor tissue of caudal-type homeobox transcription factor 2 ( CDX2 ), in combination with normal E-cadherin and nonexpression of the transmembrane protein mucin 1 ( MUC1 ), is a favorable prognostic factor for patients with GC ( ; ).

Abnormal methylation is present in multiple TSGs, such as p16 , CDH1 ( E-cadherin ), hMLH1 , RAR-beta , RUNX3 , MGMT (O6-methylguanine methyltransferase), TSP1 (thrombospondin-1), HLTF (helicase-like transcription factor), and RIZ1 (retinoblastoma protein-interacting zinc finger gene-1). A 12-gene study has shown that methylation changes are more frequently observed in stages III/IV cancers compared with stages I/II cancers ( ). Overexpression of growth factors, such as EGF/TGF-α and EGFR, is associated with cancer progression and often indicative of poor prognosis ( ). VEGF overexpression enhances tumor angiogenesis; thus, it is associated with lymph node metastasis, hepatic metastasis, and shorter survival time.

In addition, mutation or abnormal expression of p53 may predict a reduced cumulative survival and is associated with lymph node metastasis and lower chemosensitivity ( ; ; ).

An alternative system, proposed by the WHO, divides GC into papillary, tubular, mucinous (colloid), and poorly cohesive carcinomas ( ). These classification systems have little clinical utility, making the development of robust classifiers that can guide patient therapy an urgent priority. Recently, The Cancer Genome Atlas ( ) project proposed a molecular classification dividing GC into four subtypes: (1) Epstein-Barr virus–associated tumors (∼9%, which display recurrent PI3KCA mutations, extreme DNA hypermethylation, and amplification of JAK2, CD274 [also known as PD-L1] and PDCD1LG2 [also known as PD-L2], EBV-CIMP, CDKN2A silencing, and immune cell signaling); (2) microsatellite unstable tumors (∼22%), which show elevated mutation rates, including mutations of genes encoding targetable oncogenic signaling proteins; (3) genomically stable tumors (∼20%), which are enriched in the diffuse histologic variant and have mutations of RHOA or fusions involving RHO-family GTPase-activating proteins, CLDN18-ARHGAP fusion, or mutation in the cell adhesion gene CDH1; and (4) tumors with chromosomal instability (∼50%), which show intestinal histology, TP53 mutation, RTK-RAS activation, marked aneuploidy, and focal amplification of receptor tyrosine kinases ( ). Notably, the chromosomal instability tumors were predominant (65%) in the gastroesophageal junction and the gastric cardia, whereas the EBV-positive tumors (62%) were primarily noted in the fundus and body of the stomach ( ).

Importantly, these molecular subtypes showed distinct salient genomic features, providing a guide to targeted agents that should be evaluated in clinical trials for distinct populations of GC patients. Through existing testing for MSI and EBV and the use of emerging genomic assays that query focused gene sets for mutations and amplifications, the classification system developed through this study can be applied to new GC cases.

The ToGA trial, an international multicenter phase III clinical study involving 24 countries globally, has shown that the anti-HER2/neu humanized monoclonal antibody trastuzumab (Herceptin) is effective in prolonging survival in patients with HER2/neu-positive adenocarcinoma of the stomach and the gastroesophageal junction ( ; ). Molecular therapy targeting HER2/neu is now approved for patients with inoperable locally advanced, recurrent, or metastatic adenocarcinoma of the stomach or esophagogastric junction if the tumor shows unequivocal HER2/neu overexpression by IHC (score 3+) or amplification by FISH or chromogenic in situ hybridization (CISH). Therefore, HER2 testing in gastric carcinoma is warranted for determination of treatment eligibility.

The prognosis of GC is relatively poor. Treatment of GC, especially late-stage tumor, usually requires a multidisciplinary approach that includes chemotherapy in combination with surgical resection. TCGA studies has developed a genetic profile of somatic gene alteration in GC using GNS ( ; ). With such high-throughput technology, potential novel biomarkers—that is, genes (genomics), mRNA (transcriptomics), proteins (proteomics), and metabolites (metabolomics)—are able to be discovered.

Colorectal Cancer

Colorectal cancer (CRC) is the third most common malignancy worldwide ( ). Prognostication of newly diagnosed CRC predominantly relies on stage or anatomic extent of disease based on the International Union Against Cancer and the American Joint Committee on Cancer staging classifications. However, CRC should be regarded as a heterogeneous, multipathway disease—an observation substantiated by the fact that histologically identical tumors may have neither similar prognosis nor similar response to therapy ( ). Mutations and/or epigenetic alterations in TSGs (e.g., APC , DCC , SMAD-2 , SMAD-4 , p53 ) and oncogenes (e.g., KRAS , p53 ) are molecular determinants during the development of sporadic CRC, which was first summarized in the adenoma-carcinoma sequence described by . Based on genetic background, sporadic CRCs can be subclassified into two major types: (1) chromosomal unstable and (2) microsatellite unstable, plus an additional group with CpG island methylator phenotypes. Various genetic syndromes are known to predispose to colorectal carcinoma and to have specific inherited gene defects. These molecular markers are currently being investigated as a means of improving the identification of patients likely to have poor clinical outcomes and, therefore, possibly benefiting from adjuvant treatment ( ; ; ).

EGFR (a receptor tyrosine kinase) overexpression is detected in up to 80% of CRCs ( ). Tumors with a high level of EGFR expression usually have a poor prognosis ( ; ). These data have led to the clinical use of EGFR blockade in the treatment of CRCs with anti-EGFR monoclonal antibodies such as cetuximab or panitumumab (see Chapter 77 ). Cetuximab is a mouse-human hybrid monoclonal antibody approved by the FDA in 2004 as second-line treatment of CRC. Panitumumab is a fully human immunoglobulin G ₂ monoclonal antibody approved by the FDA in 2007 as third-line monoclonal antibody therapy of refractory CRC. Both drugs are expensive and have significant side effects. Thus, it is imperative to use a biomarker to predicate the therapeutic response. However, use of EGFR as a biomarker to predict response to EGFR monoclonal antibodies has had limited success. Only detection of increased EGFR copy number by FISH has predictive value ( ), and it has been proven that EGFR expression levels detected by IHC are not as useful. In contrast to EGFR mutation in lung cancer, EGFR mutation in CRC is a rare event that occurs in fewer than 1% of cases ( ). Current studies are focused on finding the downstream effectors of the EGFR pathway that may serve as biomarkers for predicting the effectiveness of anti-EGFR therapy.

As summarized earlier in this chapter and in Chapter 77 Ras-p21 proteins are guanosine triphosphate (GTP)-coupled proteins that are important in many receptor tyrosine kinases that utilize signaling on the Ras/Raf/Erk/Map kinase signal transduction pathway. Mutations that result in single amino acid substitutions at critical positions in the amino acid sequence of this protein, such as Gly 12, Gly 13, and Gln 61, occur in the Ras protein and cause its constitutive activation. This leads to a series of oncogenic events that activate downstream signaling pathways, promoting tumor initiation and tumor growth and progression, including local invasion, angiogenesis, metastasis, and even immune response ( ; ).

The KRAS mutation occurs in at least 35% to 45% of cases of CRC and, in some series, in up to 75% of cases. Mutations of this gene occur almost exclusively in three hot spots (codons 12, 13, and 61) and rarely in others, including codon 146 ( ; ). Because KRAS is the downstream effector of EGFR and is frequently mutated in CRC, the mutation status of KRAS has become an important predictive marker for the effectiveness of cetuximab or panitumumab on metastatic CRCs. The American Society of Clinical Oncology (ASCO) has recommended that patients with diagnosed metastatic colon cancer who are candidates for anti-EGFR therapy should be tested for KRAS mutations in a Clinical Laboratory Improvement Amendments (CLIA)-accredited laboratory. If negative for KRAS mutations, these patients are deemed suitable for treatment with anti-EGFR antibody (cetuximab or panitumumab) therapy. This recommendation is based on the results of phase II and phase III clinical trials. Retrospective subset analyses of these trial data suggest that patients who have KRAS mutations detected in codon 12 or 13 do not benefit from this therapy. This was further confirmed by five randomized, controlled trials of cetuximab or panitumumab response in relation to KRAS mutational status (no mutation [wild type] and mutated [abnormal]) and another five single-arm retrospective studies undertaken to evaluate tumor response according to KRAS mutational status.

According to the CAP Perspectives on Emerging Technology (POET) guideline published in 2009, acceptable samples for KRAS mutation include fresh, frozen, and FFPE tissues. Acceptable assay types are RT-PCR and direct sequencing analysis. However, other methods, such as pyrosequencing or microarray (chip)–based tests, have also been developed ( ; ). As of the time of this writing, no type of KRAS testing has been approved by the FDA. Although recommended testing is limited to codons 12 and 13 of exon 2, which make up the vast majority of cases with KRAS mutations, many laboratories and commercial kits also include testing for other mutations, such as for codons 61 and 146.

The 2016 National Comprehensive Cancer Network (NCCN) guidelines state that patients with any known KRAS or NRAS mutation should not receive an EGFR inhibitor. This opinion is shared by ASCO ( ).

BRAF is an immediate downstream effector of RAS in the Ras/Raf/Map kinase pathway (see Fig. 79.1 ). Mutation of BRAF has been identified in approximately 10% of CRCs ( ). A vast majority of the point mutations are V600E, a mutational hot spot at nucleotide 1796 within exon 15, accounting for a T:A transversion mutation and a valine-to-glutamic acid substitution. All point mutations occur within the kinase domain of B-Raf protein, leading to constitutive activation of B-Raf kinase ( ). The BRAF mutation is almost mutually exclusive with KRAS mutations. Unlike the KRAS mutation, BRAF mutations frequently occur in female patients older than 75 years of age with proximal lesions. These tumors exhibit MSI with higher frequency than those in other patients with colon cancer ( ), which appears to be due to epigenetic effects involving hypermethylation of the promoter regions of mismatch repair genes. This finding helps in differentiating between sporadic MSI-H tumors with somatic mutations in the V600E hot spot in BRAF and Lynch-associated cancers with MLH1 or MSH2 mutations.

BRAF mutations in tumors have been characteristically associated with a worse survival than in those without mutations and with resistance to treatment with anti-EGFR monoclonal antibodies, cetuximab, or panitumumab for metastatic CRCs. Currently, molecular testing for BRAF mutation V600E has been performed in many clinical laboratories as a therapeutic indicator for anti-EGFR treatment in combination with KRAS mutation analysis ( ).

PI3KCA is a membrane lipid kinase that phosphorylates the 3-hydroxyl group of phosphatidylinositol. This critical enzyme, which binds directly to ras-p21 protein, can be activated by EGFR-mediated signaling in parallel with the K-ras/B-raf/MAP kinase pathway. Activation of PI3KCA counteracts PTEN, an antioncogene protein, and promotes AKT1 phosphorylation, stimulating cell proliferation, survival, and angiogenesis in cooperation with other pathways. The PIK3CA (phosphatidylinositol-4,5-bisphosphate 3-kinase, catalytic subunit α) gene is mutated in about 20% of CRCs ( ). The mutation of PIK3CA usually occurs in the hot spots in exon 9 (E542K, E545K) and exon 20 (H1047R), both of which are oncogenic in CRC cellular models. Unlike the BRAF mutation, PIK3CA and KRAS mutations may coexist. In addition to their association with poor prognosis (i.e., local recurrence and increased mortality of CRCs), PIK3CA mutations can independently predict resistance to anti-EGFR therapy with panitumumab or cetuximab for metastatic CRCs ( ; ; ). Testing for combinations of the PIK3CA/PTEN and KRAS mutations can detect up to 70% of patients who are unlikely to respond to anti-EGFR therapy, although more validation is required for clinical practice ( ).

Loss of PTEN has been found to co-occur with KRAS , BRAF , and PIK3CA mutations ( ; ; ; ). The recorded frequency of loss of PTEN expression varies from 19% to 36%, with some studies reporting an effect on response rate and survival, whereas others found an effect only on progression-free or overall survival. Moreover, data on the loss of PTEN expression are not concordant in primary and metastatic tissues ( ). It has been suggested that loss of PTEN expression, as determined by immunohistochemistry, is associated with lack of benefit from cetuximab in metastatic colorectal cancer ( ; ; ; ). Two recent retrospective studies reported an association between PIK3CA mutations and better outcomes among patients taking aspirin after a diagnosis of colorectal carcinoma ( and ).

The APC gene, a tumor necrosis factor-stimulated gene (TSG), encodes a 2843-amino acid protein and is located on chromosome 5q21-22, regulating Wnt/β-catenin signaling ( ). Mutations of this gene can be point mutations or small deletions or insertions. The latter cause frameshifts that generate stop codons. Germline mutations of this gene are responsible for FAP, whereas somatic mutations occur in the majority of sporadic CRCs. Somatic mutations of this gene may initiate and promote carcinogenesis of sporadic CRCs, and can be found in early neoplastic stages, even in dysplastic cryptic foci ( ; ; ). CRCs with APC mutation usually are those with chromosomal instability or MSS ( ). The APC gene can serve as a biomarker for early detection or as a risk factor for familial polyposis syndrome. Detection methods include various types of PCR, LOH, and others ( ). Screening for APC mutations may detect patients with relatively early CRC by purified DNA from routinely collected stool samples ( ).

p53 (or TP53) is a tumor suppressor that regulates the cell cycle and is involved in apoptosis and DNA repair. This protein is discussed at length in Chapter 76, Chapter 77 . It is located on the short arm of chromosome 17 (17p13.1) (Isobe et al., 1986). Mutation or homozygous deletion of TP53 is found in 40% to 50% of CRCs. Loss of wild-type p53 activity is thought to be a major predictor of failure to respond to radiotherapy and chemotherapy in various human cancers, including CRCs, and predicts lower survival ( ; ). Although TP53 mutation analysis has shown promising prognostic value, the ASCO Tumor Markers Expert Panel does not currently recommend its use in routine practice.

TGFBR1, SMAD2, and SMAD4 are proteins involved in the transforming growth factor-β (TGF-β) signaling pathway (see Chapter 77 ). SMAD proteins are the principal components mediating the TGF-β signaling pathway, which negatively regulates the growth of epithelial cells ( ). It has been found that germline allele-specific expression (ASE) of the gene encoding the TGF-β type I receptor (TGFBR1) occurs in 10% to 20% of CRCs in both familial (dominantly inherited and segregated) and sporadic forms ( ). The presence of multiple-mutated TGF-β has been linked with poor prognosis in a subset of CRC patients. Approximately 50% to 60% of CRCs have been found to harbor SMAD4 mutations ( ); in contrast, only a few have been found to have SMAD2 mutations. Both SMAD2 and SMAD4 are mapped to chromosome 18q21 ( ; ). Most mutations of SMAD2 and SMAD4 occur in their so-called MH2 domain. Inactivation of the SMAD4 gene through germline mutation and loss of the unaffected wild-type allele are responsible for familial juvenile polyposis, whereas inactivation of the SMAD4 gene through homozygous deletion or intragenic mutation occurs frequently in CRC and pancreatic cancer. SMAD4 mutation occurs at a later stage of colorectal carcinogenesis and is an indicator of advanced phenotypes. Loss or low levels of SMAD4 expression in CRCs are found to be associated with poor prognosis following surgery and 5-fluorouracil (5-FU)-based adjuvant therapy. However, the predictive value of SMAD4 mutation/inactivation is controversial ( ; ). Chromosomal loss at 18q encompassing both SMAD2 and SMAD4 has been reported in up to 70% of CRCs ( ). Patients with locally advanced stage II or stage III disease appear to demonstrate a significantly poorer prognosis with loss of 18q ( ).

MSI is involved in approximately 75% to 85% of CRCs, characterized by aneuploidy, allelic losses, amplifications, translocations, and mutations of APC , KRAS , and TP53 ( ). Although the prognosis of patients with MSI CRC is stage and grade dependent, tumors with identical morphologic features display considerable heterogeneity in terms of clinical outcome. The remaining 15% to 20% of CRCs have MSI-H characterized by inactivation of mismatch repair ( MMR ) genes ( ), as discussed in Chapter 77 . Sporadic MSI-H cases have been shown to arise predominantly through promoter hypermethylation and silencing of the hMLH1 gene in contrast to a hereditary form of MSI-H CRC, HNPCC, which typically shows germline mutations in one of the DNA MMR genes ( ). Tumors with MSI-H tend to be more proximally located, poorly differentiated, and of mucinous histologic type, and to show considerable lymphocytic infiltration; they have been linked with TGFBRII and BRAF mutations ( ). Patients with colon cancers demonstrating MSI-H have a stage-independent improved survival compared with patients with MSS tumors in most, although not all, studies. However, they do not benefit from treatment with 5-FU in randomized adjuvant therapy trials ( ). In stage II patients, when the tumor is MSI-H, no additional treatment is necessary because, based on trial, further chemotherapy or radiation treatment will not be beneficial and can potentially be harmful for this group of patients. Recently MSI is also defined as a predictor for a positive response to immunotherapy. Pembrolizumab is an anti–PD-1 (programmed cell death protein 1) immune checkpoint inhibitor that induces an objective response in 40% of patients with treatment-refractory metastatic, microsatellite unstable colorectal cancers. On the other hand, immunoexpression of PD-1 ligands on tumor cells does not seem to be a predictive marker of response to PD-1 blockade ( ).