The Washington Manual of Oncology, 3 Ed.

Molecular Diagnostics

Ian S. Hagemann • Christina M. Lockwood • Catherine E. Cottrell

 I. ROLE OF MOLECULAR DIAGNOSTICS. Clinical application of insights derived from molecular genetic research studies has become increasingly important in patient care. Molecular analysis is useful for diagnosis, for prognosis or predicting response to various treatment options, for monitoring minimal residual disease (treatment efficacy), for identifying predisposition to disease, and for detection of therapeutic targets in gene-specific therapy. Clinical molecular diagnostic methods have been integrated into many laboratory disciplines (Table 2-1), and published guidelines and recommendations from both professional societies and regulatory agencies have been developed to assist in the development and performance of clinical molecular pathology testing.

 Molecular testing has been practiced for several decades in surgical pathology in the form of immunohistochemistry, in which antibodies are used to detect and quantify the expression of specific proteins of diagnostic, prognostic, or therapeutic importance. However, in current common usage, molecular diagnostics refers to the analysis of changes in nucleic acids, either DNA or RNA, and most frequently implies direct mutation testing. Genetic testing protocols can be designed to detect heritable DNA variations or, as is usually the case in neoplastic diseases, somatic or acquired DNA variations limited to abnormal cells.

 At the interface of genetic disorders and acquired mutations in sporadic cancers are familial cancer syndromes. Familial cancer syndromes are inherited disorders that place patients at increased risk for particular types of tumors. Tumors resulting from a predisposition syndrome differ from their sporadic counterparts in several respects: earlier age of onset, bilateral or multifocal tumors, occurrence in multiple family members, and occurrence in more than one generation of a family. The carriers of the increased risk inherit a genetic predisposition for tumor formation as an autosomal dominant trait; this risk factor is a gene-specific germline mutation that may be screened for in DNA isolated from circulating lymphocytes (or other nontumor tissue) of affected individuals rather than tumor tissue. The characterization of the specific genes underlying familial cancer syndromes has provided important insights into the nature of many tumor suppressor genes and oncogenes. Many examples of genes that confer a hereditary predisposition to cancer have been described and include RB1 (retinoblastoma, osteosarcoma, leukemia, and lymphoma), TP53 (Li–Fraumeni syndrome; bone and soft tissue sarcomas, breast cancer, leukemia, and brain tumors), BRCA1 and BRCA2 (breast, ovarian, colon, and prostate cancers), APC(adenomatous polyposis coli; colon carcinoma and osteomas), MSH2, MLH1, and MSH6 (hereditary nonpolyposis colorectal cancer/Lynch syndrome; colon, endometrium, ovary, and stomach cancers), RET (multiple endocrine neoplasia type 2 and familial medullary thyroid cancer; medullary thyroid carcinoma and pheochromocytoma), VHL (von Hippel–Lindau disease; renal carcinoma and pheochromocytoma), NF1 (neurofibromatosis; neurofibroma and optic gliomas), and many others. Notably, a subset of sporadic tumors also carries mutations in these genes.

1. Specimens. Specimen requirements are dictated by type of disease and type of test (analyte and methodology). Testing for constitutional disease can generally be performed on any tissue, while testing for acquired mutations requires submission of affected tissue. Since most, but not all, molecular diagnostic techniques are currently based on polymerase chain reaction (PCR) amplification, the amount of tissue required is relatively small, and so many routine pathology specimens can be analyzed effectively.

Regardless of specimen type, two general features of the tissue sample influence molecular pathology assays. First, there must be a sufficient quantity of the specific target cell, and therefore target DNA or RNA in the sample. If testing involves a clonal tumor admixed with normal cells, or a heterogeneous tissue specimen with many different gene expression patterns, a detection threshold must be defined. Second, since the size of the nucleic acid molecules after isolation from the tissue can dramatically affect the sensitivity of the detection of specific alterations, degradation (whether due to enzymatic, heat, pH, or mechanical forces) can reduce a specimen’s suitability for testing.

1. Tissue types. Peripheral blood, bone marrow, solid tissue biopsies, and enriched cell populations (e.g., from flow cytometry) are all adequate substrates for molecular analysis. They should be collected and transported to the molecular pathology laboratory using aseptic techniques. Transport on ice reduces cell lysis, minimizes nuclease activity, and reduces nucleic acid degradation. Specimens are stored at 4°C after receipt in the laboratory.
2. Tissue quality. Appropriate handling of the specimen maximizes detection limits. Hematologic specimens (peripheral blood, bone marrow) should be collected in the presence of an anticoagulant, preferably ethylenediaminetetraacetic acid (EDTA) or acid citrate dextrose (ACD). Heparin anticoagulation is discouraged, as heparin carry-over after nucleic acid isolation may inhibit subsequent PCR steps. Freezing solid tissue specimens after excision also yields good preservation of nucleic acids. However, frozen whole blood or bone marrow presents distinct obstacles to the preparation of good-quality nucleic acid and should be generally avoided.

 Formalin-fixed, paraffin-embedded tissue is suitable for many molecular diagnostic tests. Fixed tissue has several advantages, two of which are that fixation dramatically inhibits nucleic acid degradation and that fixed specimens can be easily stored and transported. However, a significant limitation of fixed tissue is that the quality of extracted nucleic acids is extremely variable because all fixatives, including formalin, chemically degrade nucleic acids to a greater or lesser extent.

 Specimens sent for cytogenetic testing require special handling techniques to preserve cell viability. Specimens should be transported to the cytogenetic laboratory as soon as possible, and should be drawn in sodium heparin collection tubes in the case of hematologic samples. Solid tissues may be transported in cell culture media. Cytogenetic specimens should never be frozen, but rather handled at room temperature for short periods or with a cold pack.

3. Tissue quantity. Minimum sample requirements are determined by the assay methodology and the extent of target cell involvement in the tissue submitted for analysis. Genomic Southern hybridization requires approximately 5 µg of DNA (approximately 10⁶ cells) per enzymatic digest for detection of single copy genomic DNA targets; hybridization signals to detect lower-molecular-weight DNA fragments (less than 1 kb in length) may require more DNA. PCR amplification has significantly reduced DNA requirements, and typically only 20 to 200 ng of DNA (approximately 10³ to 10⁴ cells) is needed per reaction for many applications, though multiplexed PCR may require additional DNA to equally represent all targets. The sensitivity of PCR for detection of a few target molecules in a large background of unaltered DNA molecules (1 in 10⁵) is one of the principal strengths of this methodology in molecular pathology.

1. Technical versus diagnostic aspects of testing. The analytical sensitivity and specificity of a molecular genetic test may be unrelated to its diagnostic sensitivity and specificity. Several factors influence the diagnostic sensitivity and specificity of a test, including the supposition that only a subset of cases of a specific tumor may harbor a characteristic mutation, more than one genetic abnormality may be associated with a specific tumor, and more than one tumor may share the same mutation. Thus, a molecular genetic method with perfect analytical performance can have a lower sensitivity and specificity when used for diagnostic or prognostic testing of patient samples. Differences between the analytical and diagnostic levels of analysis are often overlooked even though they account for many of the confusing or seemingly conflicting results regarding the utility of diagnostic molecular genetic testing in clinical practice.
2. Reported estimates of test performance. The experimental design of published studies regarding the utility of molecular diagnosis varies considerably. Many studies of test performance are complicated by the presence of selection bias (also called verification bias, posttest referral bias, and workup bias) or discrepant analysis (also known as discordant analysis or review bias), which hinders interpretation of the test results in the setting of routine clinical practice.
3. Discordant cases. Cases arise in which there is a lack of concordance between the diagnosis suggested by molecular genetic findings and the morphologic diagnosis. The debate over the best approach to resolve the ambiguity presented by these cases, especially those that are presumed false positives, reflects the fundamental impact of molecular genetics on the classification of disease as well as the power of morphology as the historic standard of pathologic diagnosis by which new methods of classification are measured. Rather than arbitrarily assuming that genetic testing or morphology is superior in all cases, the most reasonable way to handle discordant cases is to acknowledge the presence of the discrepancy, and to reappraise all the clinical data, pathologic findings, and therapeutic implications. For those cases in which the diagnosis suggested by morphology and genetic testing are different, prospective clinical trials are required to assess whether stage, prognosis, and response to treatment are more accurately predicted by the molecular test than by the morphologic findings on which current staging and treatment protocols are based.
4. Single-gene versus multigene testing. Despite the fact that molecular analysis of single markers has clinical value for many neoplasms, testing focused on individual loci is a reflection of the immaturity of molecular diagnostics rather than an optimized testing paradigm. Simultaneous analysis of multiple loci will likely provide more accurate diagnostic, prognostic, and therapeutic information. Consequently, clinical use of microarray and next-generation sequencing technologies that enable evaluation of thousands of markers from individual tumor specimens will likely expand.
5. CYTOGENETICS. A number of specific cytogenetic abnormalities uniquely characterize morphologically and clinically distinct subsets of hematopoietic malignancies and solid tumors.
6. Traditional karyotype analysis. Metaphase chromosome analysis can be performed on many different cell types, although different sample and handling procedures may be required. Inappropriate handling, as well as delay between specimen collection and culture initiation, can markedly decrease the likelihood that the sample will grow in vitro. Communication and coordination with the cytogenetics laboratory are essential.

Traditional karyotype analysis begins with a period of in vitro cell culture of the tissue sample and may be aided by the presence of mitogens to stimulate cell division. After a period of time in culture, cells undergo a process known as harvesting to produce a cell preparation suitable for downstream analysis. During the harvest, cells are exposed to a hypotonic solution, which causes cell swelling and assists in aiding proper chromosome spreading, followed by fixation using a methanol–acetic acid mixture. At times a mitotic inhibitor may be used during harvest to arrest cells in metaphase. Slides are prepared by dropping the fixed cell suspension onto a glass microscope slide, followed by staining to allow for visualization.

The most widely used staining technique is termed G-banding and relies on an enzymatic digestion (often with trypsin) followed by treatment with Giemsa or Wright’s stain to band the chromosomes. Chromosome resolution and quality are variable (dependent on the cell type, as well as the mode of preparation), but karyotype analysis is most often performed with a resolution of 400 to 600 bands per haploid set of chromosomes. Analysis is performed at a band-for-band level in order to identify structural and numeric chromosome abnormalities. Routine chromosome analysis of oncology specimens (bone marrow, involved peripheral blood, and tissue) requires band-for-band analysis of 20 metaphase cells. During analysis a karyogram is generated and serves as the representative image of the chromosomal complement of the cell displayed in a standard format. A standard clinical cytogenetics report will include information about the study performed, including banding technique, number of cells counted and analyzed, and banding resolution, as well as providing an interpretation, and a karyotype described according to the International System for Cytogenetic Nomenclature (ISCN).

1. Advantages. The power of conventional cytogenetic analysis lies in its ability to provide simultaneous analysis of the entire genome, albeit at low resolution, without requiring any foreknowledge of the chromosomal regions involved in the disease process. In many cases, the type and/or location of the identified chromosomal abnormalities can be used to aid in diagnosis or direct prognosis.
2. Limitations. The clinical utility of traditional chromosome analysis is restricted by two general features of the method. First, from a technical standpoint, analysis can be performed only on viable tissue specimens that contain actively dividing cells. Second, from a sensitivity standpoint, analysis has a resolution of only approximately 3 to 4 million base pairs (Mb) at an 850-band level, and only approximately 7 to 8 Mb at a 400-band level. Consequently, traditional chromosome analysis is suited only for the detection of numerical abnormalities and gross structural rearrangements; the method does not have the sensitivity to detect small copy number alterations, small insertion and deletion events, or single-nucleotide variation.
3. Fluorescence in situ hybridization. Use of nucleic acid probes labeled with fluorochromes and visualized through fluorescence microscopy has revolutionized in situ hybridization for detection of abnormalities in chromosome number and chromosome structure. Fluorescence in situ hybridization (FISH) probes may be categorically described as unique sequence probes, repetitive sequence probes, or whole chromosome probes. The most commonly used probe type is that which hybridizes to unique sequence in the genome. Unique sequence probes may be used to detect changes in the copy number of a specific locus (e.g., TP53), or to assay for rearrangements involving a specific locus (such as BCR-ABL or PML-RARA). Repetitive sequence probes hybridize to sequences that are present in hundreds to thousands of copies, and so produce strong signals; the most widely used probes of this type bind to α-satellite sequences of centromeres. β-satellite sequences, Y-chromosome satellite sequences, and telomeric repeat sequences may also serve as FISH probe targets. Repetitive sequence probes are most useful for the detection of chromosome aneuploidy. Whole chromosome probes, also known as chromosome painting probes, consist of thousands of overlapping probes that recognize unique and moderately repetitive sequences along the entire length of individual chromosomes. Probes of this type are used to confirm the interpretation of aberrations identified by traditional karyotype analysis, or to establish the chromosomal origin of structural rearrangements that are difficult to evaluate by other approaches.
4. Interphase FISH. In clinical oncology testing, interphase (nuclear) in situ hybridization is the method of choice for rapid assessment of aneuploidy or gene rearrangement. Interphase FISH may be performed as a stand-alone assay, or in combination with (often prior to) chromosome analysis. There is no requirement for actively proliferating cells for interphase FISH analysis, making this technique advantageous over metaphase FISH or chromosome analysis. The high sensitivity of interphase FISH also makes it useful for uncovering small rearrangements not detectable in standard karyotypes, as well as detecting cytogenetic alterations present in a very limited population of cells. However, only the number and relative position of the fluorochrome probe is obtained through this methodology.

 Since interphase FISH analysis eliminates the need for in vitro cell culture, the technique can be used to study a broader range of cell and tissue types (and therefore a broader range of tumors) than can be evaluated by traditional karyotype analysis. This characteristic allows archival material, frequently formalin-fixed, paraffin-embedded (FFPE) specimens, to be assayed by interphase FISH.

1. Advantages. Both repetitive sequence probes and unique sequence probes can be used for interphase FISH; concurrent hybridization with a combination of probes labeled with different fluorochromes permits simultaneous detection of two or more DNA regions of interest. Methodologic advances have made it possible to detect low copy-number gene sequences as small as 500 base pairs in length, although the efficiency of interphase FISH with small probes is quite low. Large numbers of interphase cells (200 to 500) are typically assayed per probe set allowing for low-level disease or subclonal cell populations to be detected.
2. Limitations. Even with optimized probes, interphase FISH lacks the resolution in routine use to detect small chromosomal copy number alterations less than several hundred kb in size. Notably, this technique cannot be used to evaluate some classes of mutations that are characteristic of many sporadic tumors and familial cancer syndromes, such as single base-pair substitutions or small insertions or deletions. FISH probes for structural rearrangements such as translocations and inversions are optimized to detect the most common molecular breakpoints, and may fail to detect complex rearrangements or those with atypical breakpoints.
3. Metaphase FISH. As with chromosome analysis, metaphase FISH requires actively dividing cells in order to obtain metaphase spreads for analysis. This technique provides additional information over interphase FISH in that the underlying structure of the chromosomes is revealed. Metaphase FISH data may be of particular use when deciphering complex chromosomal rearrangements as the probe can be visualized relative to chromosome position.
4. Multiplex metaphase FISH and spectral karyotyping (SKY). These two related techniques make it possible to hybridize metaphase chromosome spreads with a combination of probes. Both techniques are useful for the detection of aneuploidy as well as chromosome rearrangements; in many cases, the techniques can establish the chromosomal origin of rearrangements that cannot be defined on the basis of traditional karyotype analysis.
5. Copy number-based microarray analysis
6. Comparative genomic hybridization. Essentially a modification of in situ hybridization, comparative genomic hybridization (CGH) makes it possible to survey the genome for chromosomal deletions and amplifications. CGH was originally developed as a method to determine the relative copy number state in a tumor sample by hybridizing differentially labeled DNA derived from a tumor and a reference sample against metaphase chromosome spreads. Over time this technique evolved into array CGH, an approach that uses a microarray consisting of an ordered arrangement of DNA molecules linked to a solid matrix support. Genomic clones on the array, such as bacterial artificial chromosomes or synthesized oligonucleotide probes, serve as the substrate for the hybridization product. Array CGH relies on cohybridization of differentially labeled target DNA (patient or tumor) with that of control (reference) DNA against probes on the solid matrix. The DNA bound to each probe is quantified using a laser scanner allowing for relative copy number gains or losses between target and control samples to be determined. The resolution of array CGH is in theory limited only by the number of features in the array; commercially available arrays currently provide a resolution of under 10 kb.
7. SNP array analysis. The methodology for copy number evaluation has further evolved to include the use of probes for single-nucleotide polymorphism (SNP) detection in the assay. The inclusion of SNP probes is advantageous in that zygosity can be assessed, and therefore regions of the genome demonstrating a loss of heterozygosity (LOH), whether copy number-neutral or due to copy number alteration, can be defined. Also of note, SNP array analysis typically relies on the use of an in silico reference against which a target sample is compared to assess for relative copy number gain or loss; a modification that therefore only requires the target sample to be hybridized to the matrix support of the array. Chromosomal microarrays carrying probes for both SNPs and nonpolymorphic sites currently play a major role in the workup of patients with suspected constitutional disorders due to chromosomal copy number alterations. Increasingly, this methodology is being applied to oncology studies to assess the status of copy number and zygosity in a tumor sample.
8. Advantages. Both aCGH and SNP array analysis allow for a rapid, genome-wide assessment of copy number at a very high resolution using DNA derived from fresh or fixed tissue. Whole chromosomes gains and losses, as well as smaller copy number alterations in the kilobase range are detectable. SNP array analysis also allows for an assessment of LOH.
9. Disadvantages. Balanced rearrangements are not detectable by either methodology. Very small copy number alterations (intragenic) are often below the limit of detection of the assay.

 III. TISSUE IN SITU HYBRIDIZATION. Tissue in situ hybridization for mRNA is a versatile technique but is technically demanding. The approach can be used to correlate the expression level of a specific gene with tissue changes characteristic of a disease, with a tumor’s phenotype, or to monitor transgene expression in gene therapy regimens. However, even when optimized, the technique’s sensitivity is below that of PCR-based methods, and the method is therefore not currently in widespread use.

 IV. SOUTHERN BLOT HYBRIDIZATION. This filter hybridization method consists of DNA purification, digestion by a restriction endonuclease, size fractionation by gel electrophoresis, transfer and immobilization on a synthetic membrane made of nitrocellulose or nylon, and then hybridization to a specific nucleic acid probe. The probe is visualized by either radioisotopic or nonisotopic methods, and the location of the probe indicates not only the presence of the target sequence, but also the size of the enzyme-digested DNA fragment in which it is contained.

 Southern blot hybridization has historically been used to detect genomic rearrangements characteristic of specific tumor types, but can also be used to detect different alleles of the same gene, point mutations, deletions, and insertions. Since the method is quantitative, it may provide information about gene copy number.

 Southern blot hybridization has been a reliable and versatile method for sequence-specific DNA analysis, but requires at least 10 mm³ of fresh or frozen tissue in which there has been limited DNA degradation. Although the technique was widely used for molecular analysis of tissue specimens for several decades, PCR-based methods have replaced the routine use of Southern blot hybridization for many assays because PCR-based methods have similar reliability but increased sensitivity, are usually more rapid and less cumbersome, require less tissue, and can be performed on FFPE tissue.

1. POLYMERASE CHAIN REACTION. PCR makes it possible to detect a broad range of chromosomal abnormalities, from gross structural alterations such as translocations and deletions to single base-pair mutations, in individual genes. PCR can be performed on areas of tumor (or even individual cells) macro- or microdissected from routinely prepared tissue or cytology slides, or separated by flow cytometry. This is an advantage in that PCR methods enable correlation between tissue morphology and genetic abnormalities of specific regions of tumor, specific cell populations, or even individual cells. The clinical utility of PCR is due to the wide range of template DNA that can be amplified, the technique’s intrinsic extreme sensitivity, and the wide range of variations of the basic method that can be performed.
2. Basic PCR. A PCR mixture includes the DNA sample to be assayed, a thermostable DNA polymerase, the four deoxynucleotide triphosphates (dNTPs), and short oligonucleotide primers (typically approximately 20 base pairs in length) designed to be complementary to the regions that flank the DNA sequence of interest. With each PCR cycle, the DNA region of interest is doubled, resulting in exponential amplification.
3. Advantages. Basic PCR (and the variations described below) has many advantages over the techniques of Southern and Northern blot hybridization that have traditionally been used to analyze specific regions of DNA and mRNA. PCR-based analysis can be completed in hours instead of days, can be applied to a wider range of tissues (including fresh, frozen, or fixed tissue), requires much less template DNA or RNA, is able to detect target DNA or RNA even when present in a background of vast excess of nontarget sequence, and is highly versatile.
4. Limitations. When optimized, PCR can detect one abnormal cell in a background of 10⁵ normal cells, and can detect single copy genes from individual cells. However, many technical factors conspire to lower the clinical sensitivity of PCR. The most important practical limitation is degradation of target DNA or mRNA, especially when extracted from fixed tissue.

 The extreme sensitivity of PCR demands constant attention to laboratory organization and test design to avoid specimen contamination, the most troublesome technical issue. Contamination can be largely avoided by meticulous laboratory technique and maintenance, physical separation and containment of the various stages of the PCR procedure, and regular UV irradiation of laboratory workspaces and instruments to degrade any transient uncontained DNA.

 By design, PCR only analyzes the chromosomal region flanked by the primers employed. Therefore, unlike conventional cytogenetics, PCR does not survey the entire genome. Similarly, RT-PCR only analyzes the target mRNA, and provides no information about the presence of mutations or level of transcription of other related genes. Thus, design of the PCR assay is a critical factor affecting clinical sensitivity.

1. Reverse transcriptase PCR. In RT-PCR, mRNA is extracted from the sample and reverse transcribed into complementary DNA (cDNA). This cDNA then serves as the template in a subsequent PCR amplification. The use of RT-PCR permits straightforward amplification of the coding region of spliced, multiexon DNA sequences. This technique also permits detection of targeted translocations, including BCR-ABL.
2. Multiplex PCR. This approach involves the simultaneous amplification of multiple target sequences in a single reaction tube through the use of multiple primer pairs. It is used to evaluate a number of different sites for the presence of a mutation in a single reaction, and is therefore a practical screening method. The multiple sequences of interest must be widely spaced on the chromosome, or present on different chromosomes, to avoid cross-priming events.
3. Methylation-specific PCR. Pretreatment of the DNA sample with sodium bisulfite, which reduces unmethylated cytosines to uracil, makes it possible to evaluate the methylation status of individual CpG sites. This approach is used in tests to characterize specific gene silencing patterns that are correlated with tumor subgroups that are likely to respond to specific chemotherapeutic regimens.
4. Quantitative PCR. Also referred to as real-time PCR, this method permits more reliable quantification of the amount of input DNA than is possible by traditional endpoint measurement of the DNA product. Since the PCR product is quantified as the reaction progresses (i.e., in real-time) rather than after PCR completion, it avoids the potential artifacts of amplification efficiency. Quantitative PCR often provides more precise measurements of DNA or mRNA than can be achieved by filter hybridization or microarray-based methods.

When mRNA is the substrate, quantitative RT-PCR can be used to correlate changes in gene expression with the clinical features of a disease or a specific tumor type. One example in current use is the Oncotype DX breast cancer assay (Genomic Health, Inc.; Redwood City, CA), which measures the expression of 21 genes by RT-PCR and applies a formula to calculate a recurrence score.

 VI. DNA SEQUENCE ANALYSIS

1. Direct methods of sequence analysis. Although most molecular diagnostics involve querying DNA for sequence variants, a minority of assays involve DNA sequencing per se. Sequence data is somewhat cumbersome to integrate into a laboratory workflow, as the readout consists of alphabetic DNA sequence that requires manipulation or visual inspection to extract a test result, whereas tests with surrogate readouts, such as presence or absence of a band on an agarose gel, are more amenable to multiplexing and may be less expensive to perform. Sequencing is, however, invaluable in scenarios in which numerous different variants can occur, such as in BRCA1 and BRCA2 gene testing for hereditary breast–ovarian cancer.

Virtually all routine direct DNA sequence analysis is automated and performed on templates generated by the chain-termination method (Sanger sequencing). Even though these reactions require a very low quantity of template DNA, the amount of DNA present in patient specimens is seldom sufficient for analysis without a preliminary amplification step, usually PCR. The PCR-amplified template is used as input to a sequencing reaction whose product is a mixture of single-stranded DNA fragments of various lengths, each tagged at one end with a fluorophore indicating the identity of the 3´ nucleotide. These products are then resolved by capillary electrophoresis with fluorescent detection.

1. Single-nucleotide polymorphism (SNP) analysis. SNPs are common in the human genome, and once a set of SNPs is linked with an increased susceptibility to a certain disease, prognosis, or response to therapy, focused analysis of the SNPs (also known as haplotype analysis) may circumvent the need for more extensive DNA sequence analysis. More refined SNP catalogs, with identification of “tag SNPs” linked with traits of interest, will likely make systematic analysis of SNPs an important clinical tool.
2. Indirect methods of sequence analysis. Once normal and mutant alleles at a specific locus have been characterized by direct DNA sequence analysis, indirect methods can often provide enough sequence information to be of clinical utility. These indirect techniques include allelic discrimination by size, restriction fragment length polymorphism (RFLP) analysis, allele-specific PCR, single-strand conformational polymorphism (SSCP) analysis, heteroduplex analysis, denaturing gradient gel electrophoresis, cleavage of mismatched nucleotides, ligase chain reaction, the protein truncation test, and numerous variants, used either alone or in combination. These indirect methods are based on PCR and may be applied to a broad range of clinical specimens. In addition, many of the indirect methods are faster and less expensive than direct sequence analysis, and are ideally suited for screening large numbers of samples from patients.
3. Sequence analysis by microarray technology. High-density DNA microarrays have found their greatest use in hybridization-based evaluation of gene expression and as chromosomal microarrays for detection of karyotypic anomalies, but chip-based hybridization analysis is also a suitable platform for sequence analysis. Microarrays have been developed for simultaneous detection of numerous viral sequences and for genotyping multiple clinically significant positions in genes such as CFTR and TP53. Microarray-based sequencing has become less common in clinical practice as direct sequencing technologies have become more widely available.
4. Clinical next-generation sequencing. “Next-generation” sequencing (NGS) methods have revolutionized the life sciences by dramatically increasing the throughput of DNA sequencing. Many of the currently available NGS techniques have been described as cyclic array sequencing platforms, because they involve dispersal of target sequences across the surface of a two-dimensional array, followed by sequencing of those targets. The resulting short sequence reads can be reassembled de novo or, much more commonly in clinical applications, aligned to a reference genome. The Illumina (Illumina Inc., San Diego, CA), Ion Torrent (Thermo Fisher Scientific, Waltham, MA), and Roche 454 (Roche Diagnostics, Indianapolis, IN) platforms are the most widely used at present.

NGS has several attributes that are attractive in cancer testing. Since each library fragment is individually sequenced, the resulting data can resolve intratumor sequence heterogeneity, revealing the clonal structure of the tumor. NGS has the potential to detect all four of the major classes of genetic variation: single-nucleotide variants, insertions and deletions, structural variants, and copy-number variants. The technique can be performed upon DNA extracted from formalin-fixed, paraffin-embedded tissue, a commonly available specimen type. Finally, although these tests are currently not inexpensive, the cost per nucleotide sequenced is extremely low, and the cost of the test increases less than linearly with the size of the region analyzed.

Although NGS was initially developed as a research technique, it has found several prominent roles in clinical testing. Noninvasive prenatal testing (NIPT), based on analysis of cell-free fetal DNA in the maternal circulation, has resulted in a dramatic drop in invasive prenatal diagnostic procedures and is now recommended by the American College of Obstetricians and Gynecologists for women at high risk for fetal aneuploidy. NGS has also been adopted as a platform for simultaneous sequencing of multiple genes for constitutional disorders, including familial cancer syndromes. The most prominent application of NGS in clinical oncology, however, is the development of tests to detect actionable somatic variants in multiple cancer-related genes. These “cancer panels” leverage the fact that assessment of individual genes is no longer sufficient to guide selection of targeted therapies. Simultaneous detection of variants in multiple genes is potentially a more comprehensive and cost-effective approach than sequential testing of individual genes, yielding a large amount of predictive and/or prognostic data in a clinically relevant time frame.

 VII. MICROARRAY-BASED GENE EXPRESSION PROFILING. The morphologic features of disease are essentially reflections of altered gene expression within diseased cells. Characterization of disease-specific alterations in gene expression is therefore an area of intense interest. Given the limitations of Northern blot hybridization, a variety of new methodologies have been developed to identify differentially expressed genes, of which microarray technology is the most widely used.

 Microarray technology is based on the principle of nucleic acid hybridization. Fundamentally, a DNA microarray (or gene chip) employs multiple sets of DNAs or oligonucleotides complementary to the thousands of genes to be investigated, each attached at a known location on a glass or nylon membrane substrate the size of a computer chip. When RNA extracted from the clinical sample is used as the input, microarrays make it possible to rapidly measure the expression of thousands of genes in parallel. Fluorophore-labeled test (or target) RNA derived from the specimen is hybridized to the chip, and emitted light produced by laser scanning allows for quantitation of gene expression of even low-abundance transcripts.

1. Applications. Microarray-based techniques, although still largely experimental, are increasingly being used in clinical settings to demonstrate specific and reproducible differences in gene expression profiles that are of diagnostic utility, or that can be used to predict prognosis or response to specific therapeutic regimens. One such assay that has entered clinical use is MammaPrint (Agendia, Inc., Irvine, CA), an in vitro diagnostic multivariate index assay (IVDMIA) cleared by the FDA. The test measures expression of 70 genes on a microarray platform to classify patients into a low-risk or high-risk group.
2. Limitations. Microarray-based gene expression technology has until recently been applicable only to fresh or frozen tissue, but it has now been shown that it can potentially be validated in FFPE samples. Another major challenge facing microarray-based expression profiling (as well as other genome-wide techniques) is interpretation of the massive quantity of data generated by each sample tested.

VIII. EMERGING TECHNIQUES

1. Pharmacogenetics. In addition to tumor-specific genetic abnormalities, genetic variation has been estimated to account for 20% to 95% of variability in the metabolism, disposition, and effect of the drugs used to treat patients with cancer. Pharmacogenetics offers the opportunity not only to optimize the efficacy of therapy, but also to minimize toxicity. Many centers are now using prospective genotyping of a few targeted genes to guide antineoplastic treatment and dosing choices. These include thiopurine methyltransferase (TPMT) point mutations in mercaptopurine therapy for acute leukemia, thymidylate synthase enhancer region (TSER) polymorphisms in 5-fluorouracil therapy for colorectal cancer, and UDP glucuronosyl transferase (UGT1A1) promoter region variants in irinotecan therapy for metastatic colorectal cancer. Genome-wide analysis of the role of genetic variation to predict an individual patient’s response to treatment with a specific drug is known as pharmacogenomics.
2. Analysis of epigenetic modifications. Techniques for quantitative detection of the methylation status of every human gene individually have been developed. Given the effect of epigenetic changes on gene expression, it is reasonable to anticipate that the pattern of methylation at defined sets of loci may eventually become a component of the pathologic evaluation of individual tumors. At present, no test based upon methylation is associated with clear clinical benefit.
3. Proteomics. This technique is focused on the analysis of patterns of protein expression rather than the analysis of nucleic acids. Methods for high-sensitivity, high-throughput evaluation of protein expression profiles have been developed—methods that make it possible to identify patterns that correlate with specific malignancies, underlying mutations, epigenetic changes, transcriptional profiles, response to drug therapy, and so on.
4. Analysis of RNA interference. It has recently become clear that several classes of short single-stranded or double-stranded RNA molecules that are responsible for RNA interference (RNAi) have a profound role in the regulation of gene expression. The demonstration that the profile of miRNAs (one class of small RNAs that mediate RNAi) is different in normal and neoplastic tissue, that alterations in miRNA are oncogenic, and that RNAi-based approaches can have therapeutic benefit all suggest that analysis of specific small RNA molecules, either individually or on a genome-wide basis, may in the future have an important role in the molecular evaluation of tumors.

ACKNOWLEDGMENT

The authors thank Drs. John Pfeifer and Barbara Zehnbauer, authors of the previous edition of this chapter.

SUGGESTED READINGS

Barik S. Silence of the transcripts: RNA interference in medicine. J Mol Med 2005;83:764–773.

Bossuyt PM, Reitsma JB, Bruns DE, et al. Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative. Ann Intern Med 2003;138:40–44.

Chen EC, Miller SA, DeRisi JL, et al. Using a pan-viral microarray assay (Virochip) to screen clinical samples for viral pathogens. J Vis Exp 2011;50:2536.

Cottrell CE, Al-Kateb H, Bredemeyer AJ, et al. Validation of a next-generation sequencing assay for clinical molecular oncology. J Mol Diagn 2014;16:89–105.

Dieffenbach CW, Dveksler GS. PCR Primer: A Laboratory Manual. Plainview, NY: Cold Spring Harbor Laboratory Press, 1995.

Gersen SL. The Principles of Clinical Cytogenetics. New York, NY: Springer, 2012.

International HapMap Consortium. A haplotype map of the human genome. Nature 2005;437:1299–1320.

Ladanyi M, Gerald WL. Expression Profiling of Human Tumors: Diagnostic and Research Applications. Totowa, NJ: Humana Press, 2003.

Lewin J, Schmitt AO, Adorjan P, et al. Quantitative DNA methylation analysis based on four-dye trace data from direct sequencing of PCR amplificates. Bioinformatics 2004;20:3005–3012.

Marsh S, McLeod HL. Pharmacogenomics: from bedside to clinical practice. Hum Mol Genet 2006;15(Spec No 1):R89–R93.

Misek DE, Imafuku Y, Hanash SM. Application of proteomic technologies to tumor analysis. Pharmacogenomics 2004;5:1129–1137.

Pfeifer JD. Molecular Genetic Testing in Surgical Pathology. Philadelphia, PA: Lippincott, Williams & Wilkins, 2006.

Pfeifer JD, Hill DA, O’Sullivan MJ, et al. Diagnostic gold standard for soft tissue tumours: morphology or molecular genetics? Histopathology 2000;37:485–500.

Pritchard CC, Salipante SJ, Koehler K, et al. Validation and implementation of targeted capture and sequencing for the detection of actionable mutation, copy number variation, and gene rearrangement in clinical cancer specimens. J Mol Diagn 2014;16:56–67.

Pritchard CC, Smith C, Salipante SJ, et al. ColoSeq provides comprehensive Lynch and polyposis syndrome mutational analysis using massively parallel sequencing. J Mol Diagn 2012;14:357–366.

Shaffer LG, McGowan-Jordan J, Schmid M, eds; International Standing Committee on Human Cytogenetic Nomenclature. ISCN 2013: An International System for Human Cytogenetic Nomenclature. Basel, Switzerland: Karger, 2013.

Spencer DH, Sehn JK, Abel HJ, et al. Comparison of clinical targeted next-generation sequence data from formalin-fixed and fresh-frozen tissue specimens. J Mol Diagn 2013;15:623–633.

Vogelstein B, Kinzler KW. The Genetic Basis of Human Cancer, 2nd ed. New York, NY: McGraw-Hill, 2002.

Wu L, Williams PM, Koch W. Clinical applications of microarray-based diagnostic tests. BioTechniques 2005;39:S577–S582.