Context.—Soft tissue pathology encompasses a remarkably diverse assortment of benign and malignant soft tissue tumors. Rendering a definitive diagnosis is complicated not only by the large volume of existing histologic subtypes (>100) but also frequently by the presence of overlapping clinical, histologic, immunohistochemical, and/or radiographic features. During the past 3 decades, mesenchymal tumor–specific, cytogenetic and molecular genetic abnormalities have demonstrated an increasingly important, ancillary role in mesenchymal tumor diagnostics.
Objectives.—To review molecular diagnostic tools available to the pathologist to further classify specific soft tissue tumor types and recurrent aberrations frequently examined. Advantages and limitations of individual approaches will also be highlighted.
Data Sources.—Previously published review articles, peer-reviewed research publications, and the extensive cytogenetic and molecular diagnostic experience of the authors to include case files of The University of Nebraska Medical Center.
Conclusions.—Cytogenetic and molecular genetic assays are used routinely for diagnostic purposes in soft tissue pathology and represent a powerful adjunct to complement conventional microscopy and clinicoradiographic evaluation in the formulation of an accurate diagnosis. Care should be taken, however, to recognize the limitations of these approaches. Ideally, more than one technical approach should be available to a diagnostic laboratory to compensate for the shortcomings of each approach in the assessment of individual specimens.
Soft tissue tumors constitute a rare, heterogeneous group of mesenchymal neoplasms that during the past 3 decades have been shown by cytogenetic analysis to have a remarkably high incidence of specific and primary alterations. These genetic alterations have not only guided molecular studies in establishing the underlying genes involved, thereby yielding important pathogenetic information, but also have provided clinicians with a valuable tool to add to their diagnostic armamentarium. The addition of molecular cytogenetic (fluorescence in situ hybridization [FISH], array-based comparative genomic hybridization [aCGH], and single nucleotide polymorphism [SNP] array) and molecular approaches (eg, reverse transcription–polymerase chain reaction [RT-PCR] and sequencing technology) has further enhanced the sensitivity and accuracy of detecting nonrandom chromosomal imbalances and/or structural rearrangements (eg, translocations), as well as specific gene mutations in soft tissue tumors, including assessment in formalin-fixed, paraffin-embedded (FFPE) tissues.
This review is divided into 2 major sections: (1) summary of recurrent or tumor-specific genetic events in soft tissue tumors, and (2) overview of the molecular approaches commonly used in clinical practice to identify them. The advantages and limitations of individual molecular diagnostic assays used in the management of soft tissue tumors are also underscored, with inclusion of select examples that serve as useful paradigms.
GENETIC EVENTS IN SOFT TISSUE TUMORS
Simplistically, 2 major groups of genetic events can be appreciated in soft tissue tumors. Many soft tissue sarcomas are characterized by tumor-specific anomalies. Most commonly, these anomalies are translocations that result in the production of chimeric genes encoding for abnormal, oncogenic proteins that are central to the causation of these tumors. Often, these tumors are in the setting of a simple karyotype. Alternatively, some soft tissue tumors feature specific activating or inactivating mutations within oncogenes or tumor suppressor genes, respectively. A second major group of sarcomas lack a tumor-specific abnormality and are associated with multiple and sometimes complex chromosomal changes and nonspecific genetic alterations. Thus, for most sarcomas in this group, the high degree of genomic complexity and instability (including large numbers of unidentifiable marker chromosomes and intratumoral heterogeneity) precludes the use of many routine clinical genetic approaches as a discriminating tool. For this reason, there will be no further discussion of this latter group in the current review.
Approximately one-third of all sarcomas exhibit a nonrandom chromosomal translocation (exchange of chromosomal material between 2 or more nonhomologous chromosomes), which leads to the juxtapositioning of 2 genes, one from each translocation partner, resulting in the formation of a fusion gene (Table 1).1–10 A soft tissue tumor translocation is often present as the sole karyotypic aberration and is presumed to be the initiating oncogenic event. Moreover, these tumor-specific translocations are retained throughout the clinical course, even as a given tumor metastasizes or becomes less differentiated, emphasizing a role in sustaining neoplastic transformation and aiding in the management of poorly differentiated sarcomas.11 Dominantly acting oncoproteins associated with chromosomal translocations in soft tissue tumors have multiple functions and may or may not require additional mutations for cell transformation.12–14
For soft tissue tumors, chromosomal translocations most commonly result in the production of a highly specific, novel chimeric gene, such as the EWSR1-WT1 fusion corresponding to the t(11;22)(p13;q12) of desmoplastic small round-cell tumor.15 In rare instances, the same translocation and gene fusion have been demonstrated in 2 or more unrelated neoplasms, for example, ALK-CLTC in inflammatory myofibroblastic tumor and anaplastic large cell lymphoma and ETV6-NTRK3 in infantile fibrosarcoma, acute myeloid leukemia, secretory carcinoma of the breast, and the mammary analogue secretory carcinoma of the salivary glands.16–21 These occurrences suggest that the type of cell the fusion oncogene is targeting may determine whether the neoplastic transformation is of mesenchymal, hematologic, or epithelial lineage.
Functionally, there are 3 types of resultant fusion oncoproteins in soft tissue tumors:
Most often, sarcoma-associated fusion oncogenes encode for chimeric transcription factors that cause transcriptional deregulation. Synovial sarcoma, myxoid liposarcoma, and clear cell sarcoma represent just a few of the sarcomas characterized by specific, aberrant transcription factors.
Inflammatory myofibroblastic tumor and congenital or infantile fibrosarcoma are examples of soft tissue tumors whose fusion oncogenes encode for chimeric tyrosine kinases that elicit deregulation of kinase signaling pathways.
Kinase signaling pathway deregulation is also the outcome of the rearrangements in dermatofibrosarcoma protuberans and giant cell tumors of the tendon sheath or diffuse-type giant cell tumors that feature chimeric autocrine growth factors as their characteristic fusion gene. Furthermore, dermatofibrosarcoma protuberans serves as an example of how the identification of sarcoma-specific rearrangements and associated fusion oncogenes may also be therapeutically important in addition to diagnostically useful. Activation of the platelet-derived growth factor receptor β (PDGFRB), a transmembrane tyrosine kinase, by fusion of the gene encoding for its ligand, PDGFB, with COL1A1 (collagen, type 1, α 1) renders dermatofibrosarcoma protuberans responsive to targeted therapy with tyrosine kinase inhibitors, such as imatinib mesylate.22
An important aspect of familiarizing oneself with the molecular diagnostics of sarcoma translocations is to recognize that both cytogenetic- and molecular genetic–variant translocations exist in these tumors and that new cytogenetic and molecular genetic variants continue to be discovered and defined. Therefore, depending on the type of molecular-pathologic approach used, some variant translocations and gene fusions may not be identified (false-negative result).
Cytogenetic-variant translocations arise from a rearrangement of one consistent gene with differing chromosomal translocation partners (Figure 1, A). For example, approximately 90% to 95% of Ewing sarcomas/peripheral primitive neuroectodermal tumors (pPNETs) exhibit an 11;22 translocation, resulting in the fusion of the EWSR1 gene to FLI1 (an ETS gene family member).23 Less-common cytogenetic variants, however, have also been identified and are characterized by the fusion of the EWSR1 gene with other members of the ETS family of transcription factors, including ERG (21q22), ETV1 (7p22), E1AF (17q21), and FEV (2q35–36).24–27 In addition, a fusion between EWSR1 and a gene from a transcription factor family other than ETS, the NFATC2 gene (encodes for a member of the NFAT transcription factor family), has been described.28 Rarely, fusions between EWSR1 and genes encoding for a member of the zinc finger family of proteins [PATZ1 (22q12) and SP3 (2q31)] have also been observed in Ewing sarcoma.29,30 Most recently, we identified a chromatin remodeling gene, SMARCA5 (4q31), as a novel gene partner in an extraskeletal Ewing sarcoma.31 Infrequently, a substitution of the FUS gene for EWSR1 in Ewing sarcoma–associated translocations has been observed [eg, FUS-ERG t(16;21)(p11;q22) and FUS-FEV t(2;16)(q35;p11)].32,33
In contrast, molecular variants are often the result of genomic breakpoint differences that lead to distinct fusion product exon combinations. For example, in Ewing sarcoma/pPNET, the chromosomal translocation breakpoints may arise within EWSR1 introns 7 to 9 and within introns 3 to 9 of FLI1, enabling the generation of several possible EWSR1-FLI1 chimeric transcripts. In most cases, exon 7 of EWSR1 is fused to either FLI1 exon 6 (type 1; 60%) or to exon 5 (type 2; 20%). In contrast to previous reports, recent studies have shown that the fusion subtype is not a predictor of clinical outcome for patients with Ewing sarcoma who receive contemporary therapy.34,35 However, the methodologic sensitivity and specificity of RT-PCR analysis has shown promise in research studies for monitoring micrometastatic disease in the blood or bone marrow of patients with Ewing sarcoma.36 Type I and type II molecular variants can readily be detected by their unique RT-PCR product-band size (Figure 1, B). The identity of less-common or unexpected product-band sizes should be confirmed using additional approaches, such as direct sequencing or digestion with specific restriction endonucleases. Cytogenetic analysis does not distinguish between molecular variants. Fluorescence in situ hybridization analysis may be useful for detecting rare cytogenetic variants that primer sets are not commonly designed to detect.
Many benign soft tissue tumors are also characterized by specific cytogenetic translocations or other rearrangements that may be useful in distinguishing these lesions from malignant neoplasms (Table 2). For example, distinguishing between the morphologic variants of the most common group of soft tissue tumors, lipomatous neoplasms, may be challenging in some clinical situations but is imperative for accurate therapy. Identification of C11orf95-MKL2, the resulting fusion oncogene of t(11;16)(q13;p13), is useful in establishing the diagnosis of chondroid lipoma, a benign soft tissue tumor that can be mistaken for a sarcoma, especially myxoid liposarcoma or extraskeletal myxoid chondrosarcoma.37 Cytogenetic studies of conventional lipoma are abundant and have revealed rearrangements of 12q14–15 and the underlying HMGA2 gene as most common (Figure 2, A and B).38 In contrast, atypical lipomatous tumor/well-differentiated liposarcoma is characterized by the presence of a supernumerary ring or marker chromosomes (often as the sole anomaly) containing amplification of the 12q14–15 region and corresponding tumor-associated genes (in particular MDM2 and CDK4) (Figure 2, C through E). For the histopathologic differential diagnosis of lipoma and atypical lipomatous tumor/well-differentiated liposarcoma, complementary molecular testing should especially be considered for “relapsing lipomas,” lipomatous tumors with questionable cytologic atypia (even if widely excised), and for large lipomatous tumors (>15 cm) without diagnostic cytologic atypia.39
Soft Tissue Tumors With Recurrent Alterations in Isolated Genes
Activating Oncogenic Mutations
A subset of soft tissue tumors not harboring specific translocations or fusion genes is characterized by identifiable, activating oncogenic mutations. For example, constitutive activation of specific kinases by oncogenic mutations with stimulation of downstream targets is an alternative mechanism for genetic deregulation of kinase signaling. This mechanism, in the form of KIT or PDGFRA mutations, is the molecular hallmark of gastrointestinal stromal tumor (GIST).40–42
Approximately 85% of all GISTs demonstrate a primary mutation at 1 of 3 sites: KIT exon 11 mutations (approximately 65%), KIT exon 9 mutations (10%), and PDGFRA exon 18 mutations (10%).43–45 Mutations in exons 13 or 17 of KIT or in exons 12 or 14 of PDGFRA collectively comprise about 5% of the mutations detected in GIST, and roughly 10% of GISTs are so-called wild-type GISTs, without any detectable mutations in the known hot spots.46 The KIT or PDGFRA exon involved and the mutation type often correlate with GIST anatomic location and/or prognosis. For example, KIT exon 11 mutations include in-frame deletions, internal tandem duplications, and substitutions (point mutations) and are observed in all clinicopathologic GIST subtypes; however, patients with GIST and KIT exon 11 single-nucleotide substitutions fare better prognostically than those with deletions.47 A less-aggressive clinical course is also observed in patients with GISTs exhibiting KIT exon 11 internal tandem duplications, which tend to arise in the stomach.47,48 ,KIT exon 9 mutations are more often found in small-intestinal GISTs and, in general, there appears to be a higher mortality for patients with small-intestinal versus gastric GISTs.42,48 Other phenotypic or prognostic correlates include KIT exon 17 mutations, which are more frequently identified in small-intestinal GISTs49; gastric GISTs with exon 13 mutations, which tend to be associated with more aggressive behavior49; and, KIT homozygous mutations, which often relate to GISTs with a more aggressive phenotype.50
Mutations in PDGFRA are almost exclusively localized to gastric GISTs, are frequently associated with epithelioid morphology, and exhibit more indolent clinical behavior.42,51 Substitutions in the second tyrosine kinase domain of PDGFRA exon 18 are most common; however, activating, in-frame deletions and, rarely, duplications, insertions, and complex structural mutations, have also been identified.42,51–53 Immunohistochemical expression of KIT (CD117) is seen in most GISTs, including PDGFRA mutated GISTs. (A common misconception is that KIT expression is caused by a KIT mutation).
KIT and PDGFRA, localized to 4q12, encode for the same class (III) of receptor tyrosine kinases. A schematic of the most commonly involved KIT and PDGFRA exons in relation to their coding regions within the receptor tyrosine kinase is shown in Figure 3, A. Patients with GIST have been successfully treated with tyrosine kinase inhibitors, such as imatinib mesylate and sunitinib. Routine tumor genotyping is recommended by many experts because the kinase genotype has been shown to correlate with different imatinib treatment outcomes and progression-free and overall survival.43,54–56 ,KIT exon 11 mutations appear to be most responsive; however, other mutations in KIT exons and select PDGFRA exon 18 mutations also demonstrate sensitivity.42,53,54 The presence or absence of a KIT exon 9 mutation may affect imatinib dose selection. Tumors wild type for KIT and PDGFRA have also shown partial response to tyrosine kinase inhibitors, indicating that other mechanisms involving tyrosine kinase activation also exist in GISTs.43 Primary PDGFRA mutations in exon 18 (primarily involving codon 842, especially p.Asp842Val) have shown resistance to imatinib.53 Testing is particularly advocated for GISTs that are malignant with metastatic risk or that have shown the development of resistance to imatinib.
A widely used approach for clinically evaluating GISTs for resistance mutations is by PCR amplification of the most commonly mutated exons, with subsequent direct sequencing analysis of the amplified exon (Figure 3, B and C). Alternatively, some laboratories use mutation scanning techniques, such as heteroduplex formation analysis with denaturing high-performance liquid chromatography, to first, broadly scan the PCR-amplified exons for potential mutations. If a profile suggests a possible mutation within an exon, direct sequencing of that exon is performed to identify the specific mutation. Although such scanning techniques entail additional technology, the extra time that otherwise would be required for directly sequencing multiple exons is avoided.
Inactivating Oncogenic Mutations
Inactivation of a tumor suppressor gene represents another form of aberration characterizing a subgroup of soft tissue tumors. Inactivation of tumor suppressor genes requires loss of both alleles or loss of one allele with an inactivating mutation of the other, either epigenetically (ie, methylation with inactivation of the tumor suppressor gene promoter) or genetically (eg, nonsense mutation, splice site mutation). Alternatively, there may be copy number–neutral loss of heterozygosity (LOH), also known as acquired uniparental disomy, caused by a mitotic segregation error with loss of one allele and reduplication of the remaining allele or by segmental mitotic recombination between pairs of high-identity, low-copy repeats, among other mechanisms.57 Inactivation of the tumor suppressor gene SMARCB (previously INI1) typifies extrarenal (and renal) rhabdoid tumors, proximal-type epithelioid sarcoma, and a subset of pediatric undifferentiated soft tissue sarcomas.58–61 Deletion and/or mutation of both copies of the SMARCB gene results in loss of INI1 protein expression demonstrable with an anti-INI1 antibody and immunohistochemistry.62
Historically, SMARCB molecular alterations have been detected in approximately 75% of rhabdoid tumors using conventional karyotyping, FISH, and direct sequencing analysis. Recently, a comprehensive molecular study63 of 51 rhabdoid tumors using a multimodality approach of FISH, PCR-based sequencing, multiplex ligation-dependent probe amplification analysis of SMARCB, and whole genome SNP-based array analysis revealed that a variety of events (deletions, mutations, and LOH) led to SMARCB inactivation in 98% of the rhabdoid tumors (50 of 51) examined. Therefore, although karyotyping, FISH, and/or sequencing are useful for identifying SMARCB alterations in most cases (Figure 4, A and B), a small subset of diagnostically suspicious cases may require higher-resolution techniques (multiplex ligation-dependent probe amplification or SNP-array analysis) to identify small deletions, duplications, or single base pair (bp) mutations within the tumors. Assessment by SNP-array analysis is also useful for revealing copy number–neutral LOH. This can be illustrated by a case recently encountered at our institution, where FISH analysis with a home-brew SMARCB probe was negative, but chromosome 22 copy number–neutral LOH was detected by SNP-array analysis (Figure 4, C). SMARCB molecular genetic testing can be used as an aid in the diagnosis and in treatment planning for most patients with rhabdoid tumors.63
MOLECULAR APPROACHES COMMONLY USED AS DIAGNOSTIC AIDS
Genetic approaches commonly used to identify mesenchymal tumor–specific abnormalities include conventional cytogenetic, molecular cytogenetic (FISH with novel technical variations), RT-PCR, and sequencing analyses. In this review, emphasis is placed on the practical applications of each of these techniques including their advantages and limitations.
Tissue submitted for cytogenetic analysis must be fresh (not frozen or fixed in formalin) because living, dividing cells are required. A soft tissue tumor sample submitted for cytogenetic analysis should be representative of the neoplastic process and preferably be part of the specimen undergoing pathologic study. Ideally, a 1- to 2-cm3 (approximately 0.5- to 1.0-g) fresh sample is provided for analysis. Although fine-needle biopsy or aspiration specimens (<500 mg) can also be analyzed successfully, a prolonged culture may be necessary to produce enough cells for examination.64 Notably, a limited sample size may also be more restrictive for a few neoplasms, such as benign adipose tissue tumors, which often have a low cell density per volume unit. Importantly, however, efforts to perform cytogenetic analysis are worthwhile even when material is limited because the presence of a single cell exhibiting a tumor-specific chromosomal abnormality provides strong diagnostic support.
The basic process of cell culturing and karyotypic analysis is the same for all soft tissue tumors. A short-term culture usually results in sufficient mitoses in 8 to 10 days or less. Lengthy culture times should be avoided because undesired overgrowth by common fibroblasts is likely.
An alternative to tissue culture is direct or same-day harvest. With this technical option, endemic dividing cells are arrested after a 1- to 12-hour incubation in a mitotic inhibitor, such as colchicine and culture medium. This method is useful for obtaining fast (<24 hours) or preliminary results but is constrained by the in vivo mitotic index. Effusions (eg, ascites and pleural fluids) and fine-needle aspirations of some tumors, such as small round blue cell neoplasms, can have a high mitotic index and are often well-suited for direct preparations or overnight culture. Receipt of the tissue sample in the laboratory within 1 hour after biopsy can significantly strengthen the probability for a successful direct harvest analysis.
G-banding, performed with Giemsa or Wright stains pretreated with trypsin or phosphate buffer, respectively, creates a unique pattern of bands for each chromosome, to facilitate karyotyping. G-banding is the most common form of banding because of the relative ease of performing the technique, the reliability of the results, and the permanence of the preparations. The number of alternating light and dark bands detectable with G-banding in the haploid genome varies with the level of chromosomal contraction in each metaphase cell, but preparations from most soft tissue tumors typically yield 400 to 600 bands per haploid set. Cytogenetic analysis is not considered a high-resolution technique because one band represents approximately 5 to 10 ×106 bp of DNA and, therefore, could potentially contain hundreds of genes.
A global assessment of primary and secondary, numerical and structural abnormalities in a single assay is perhaps the greatest strength of cytogenetic analysis. Moreover, in contrast to FISH or RT-PCR, knowledge of the anticipated anomaly or histologic diagnosis is not necessary. Additional advantages and limitations of conventional cytogenetic analysis are summarized in Table 3.
Molecular Cytogenetic Analysis
Fluorescence In Situ Hybridization
Hybridization refers to the binding or annealing of complementary DNA or RNA sequences that serve as probes. For this review, the discussion will be confined to DNA-based probes. Molecular cytogenetic assays typically are performed with chromosome-specific probes labeled with fluorescent dyes (FISH). Alternatively, hybridization signals can be detected with peroxidase or alkaline phosphatase, but these approaches are generally less sensitive and may be limited by the numbers of colors that can be confidently distinguished by standard bright-field microscopy.
A distinct advantage of FISH, in contrast to conventional cytogenetic analysis, is that this technique can be performed on nondividing (interphase) cells from fresh or aged samples (such as blood smears, touch imprint cytologic preparations, or cytospin preparations), FFPE tissue sections, and disaggregated cells retrieved from fresh, frozen, or FFPE material. Probe detection or labeling with fluorescent molecules of different excitation and emission characteristics permits simultaneous analysis of several different probes. Importantly, this procedure can provide results (such as identification of a tumor-specific translocation, amplification of an oncogene locus, or loss of a tumor suppressor gene locus) when the tissue is insufficient or unsatisfactory for cytogenetic analysis, when conventional cytogenetic analysis has failed to yield results, or when cryptic rearrangements are present.65
Cytologic preparations are usually air-dried and subsequently fixed in 3∶1 ratio of methanol to glacial acetic acid for 5 minutes. To visualize an anomaly within a specific region of a tumor or within a specific cell type, a 4- to 6-µm-thick, paraffin-embedded tissue section can be used. Analysis of thin sections, however, is limited because portions of most nuclei are lost during sectioning, and this may lead to false-positive results in the evaluation of chromosomal deletions or losses. For the most accurate assessment of subtle aneuploidy changes, the preferred approach is to obtain whole or intact nuclei by disaggregating and releasing cells from a much thicker (50 to 60 µm) section. Fluorescence in situ hybridization is a same-day or overnight procedure, depending on the probes used or the type of specimen analyzed (or both).
Chromosomal probes frequently used in clinical practice to examine soft tissue tumors can be divided into 3 categories: (1) centromere-specific (repetitive-sequence) probes, (2) locus-specific (unique-sequence) probes, and, (3) “paint” or whole chromosome probes. Although there are a variety of quality-controlled DNA probes intended for clinical purposes manufactured commercially and sold as analyte-specific reagents, there are relatively few of these probes designed specifically for the study of mesenchymal neoplasms. Rearrangements of some of these loci are seen in more than one soft tissue tumor type. For example, although the t(11;22)(q24;q12) is characteristic of Ewing sarcoma/PNET, rearrangement of EWSR1 (22q12) is not confined to Ewing sarcoma but is also seen in most or in smaller subsets of desmoplastic small round-cell tumor, clear cell sarcoma, extraskeletal myxoid chondrosarcoma, and myxoid or round cell liposarcoma, among others. Some laboratories also elect to develop “home-brew” probes for the assessment of specific research questions or for routine clinical use in analyzing rearrangements for which commercial probes are not available (Figure 5, A through D). Home-brew probes are developed and used exclusively in-house and are not sold to other laboratories. Home-brew probes are not currently regulated by the US Food and Drug Administration and, therefore, clinical laboratories using such probes must verify or establish, for each specific use of each probe, the performance specifications for applicable performance characteristics, eg, accuracy, precision, analytical sensitivity and specificity, among others.66
Advantages of FISH is that it can (1) be informative in both metaphase and interphase cell preparations, the latter to include FFPE tissue; (2) assist in deciphering the origin of marker chromosomes, ring chromosomes, and cryptic or complex chromosomal rearrangements; and (3) reveal tumor-specific anomalies that are diagnostically useful, especially in classifying mesenchymal neoplasms with atypical clinical or histopathologic features. Moreover, FISH provides cellular localization of DNA sequences in a heterogeneous population, including detection of low-level mosaicism. Additional advantages and limitations of FISH approaches are provided in Table 4.
Alternative Novel Molecular Cytogenetic Approaches
Alternative, novel FISH approaches, such as multicolor FISH,67 spectral karyotyping,68 combined binary ratio FISH,69 metaphase-based CGH,70 array-based CGH71,72 and SNP arrays73 are powerful genome-wide applications that have expanded the diagnostic and prognostic capabilities of molecular cytogenetics in the analysis of soft tissue tumors. The overall resolution of interphase FISH is approximately 50 to 100 kb compared with an overall resolution of 10 Mb for routine cytogenetic analysis, 2 to 3 Mb for multicolor FISH and spectral karyotyping, 2 to 10 Mb for metaphase-based CGH, and approximately 10 to 100 kb for aCGH.74 Note, some high-resolution CGH arrays can reportedly detect structural variations at a resolution of 200 bp.75 Similar to aCGH, SNP arrays are capable of identifying copy-number changes (gains and deletions) at a high resolution and throughout the whole genome but unlike aCGH, SNP arrays also have the means to detect allele-specific copy numbers and LOH. Loss of heterozygosity does not always refer to copy-number losses. For example, acquired uniparental disomy does not cause any genomic copy-number changes but, instead, reflects allele-based changes (in other words, the changes are copy-number neutral).57 Acquired uniparental disomy is emerging as a common event in some types of cancer.
Genomic imbalances or copy-number alterations can lead to tumor development through the loss of tumor suppressor genes or the amplification of oncogenes. Identification of recurrent copy-number imbalances or LOH using cytogenomic array (aCGH and/or SNP array) technology, respectively, provides valuable clues to potential cancer-related genes, points to potential new therapeutic targets, and contributes to tumor classification and diagnosis.76–78 Moreover, DNA copy-number changes identified with global genomic approaches may aid in predicting the prognosis of some soft tissue tumors.79–81 Notably, SNP array platforms may be supplanted with advancing next-generation sequencers which are able to generate all of the information that SNP arrays can produce but with (theoretically) greater resolution and accuracy.82
A sampling of some of the advantages and limitations of genomic and SNP array approaches are listed in Table 5.
Reverse Transcription–Polymerase Chain Reaction Analysis
The highly specific gene rearrangements that result from chromosomal translocations in soft tissue tumors can be identified with RT-PCR analysis. The PCR technique uses specific, synthetic primers to amplify a section of a gene in vitro. The PCR can be carried out on RNA following reverse transcription (messenger RNA→complementary DNA). Snap-frozen tissue is preferred for RNA extraction and RT-PCR analysis, but this procedure can also be performed on archival (paraffin-embedded) material. To ensure the integrity of the RNA is of sufficient quality for study (RNA degrades relatively quickly), RT-PCR analysis of a housekeeping gene is also conducted (RNA quality control). This prevents the possibility of reporting a false-negative result due to inadequate RNA.
Reverse transcription–polymerase chain reaction analysis is remarkably sensitive. It may allow for the detection of abnormalities present in cells too few to be identified with traditional cytogenetic or FISH methods. Reverse transcription-polymerase chain reaction analysis may be suitable for the detection or monitoring of minimal residual or minimal disseminated disease. For example, some authors using RT-PCR analysis for the detection of Ewing sarcoma or alveolar rhabdomyosarcoma–associated fusion transcripts have reportedly identified micrometastatic disease in bone marrow or circulating tumors cells at diagnosis or posttherapy and relate poorer outcomes for these patients.83–86 However, a study87 of Ewing sarcoma has suggested, at least at diagnosis, that bone marrow involvement may be more relevant than peripheral blood to clinical outcome. Additionally, the presence of tumor cells in the peripheral blood may be attributable to mobilization by tumor biopsies, rather than inherent tumor metastasis.88
Another advantage of RT-PCR analysis is that it is not dependent on successful cell culture, and, similar to FISH, it is rapid, with a short turnaround. Compared with cytogenetic analysis, the greatest disadvantage of RT-PCR analysis is the inability to detect chromosomal anomalies other than those for which the test was designed. With conventional cytogenetic analysis, all major chromosomal abnormalities, including those not initially anticipated by the clinician or laboratorian, may be uncovered. Additional advantages and limitations of RT-PCR analysis are listed in Table 6.
As the genetic characterization of soft tissue tumors is further refined with respect to specific gene alterations (either as the inciting primary genetic event or as secondary events in soft tissue tumors exhibiting tumor-specific translocations), other molecular methods will be increasingly used for identification of these mutations in oncogenes and tumor suppressor genes. DNA sequencing analysis of specific genes is helpful in the detection of activating or inactivating missense mutations, deletions, and insertions. Sanger sequencing (dideoxynucleotide sequencing), one of the originally developed sequencing options, allows for selection of either a full gene or specific exons and splice sites to sequence. The first step of Sanger sequencing requires amplification of the DNA fragments of interest by PCR. Sequencing can be performed on fresh or FFPE tissue if the DNA is of sufficient quality. Full gene sequencing can be time consuming, and therefore, knowledge of recurrent (hot spot) mutations or important functional exon targets (eg, tyrosine kinase domain) is helpful. A limitation of Sanger sequencing is that because of its lower analytical sensitivity, specimen areas rich in tumor content (≥20%–25%) must be identified and dissected for mutation detection. Assessment of specimens with lower tumor percentages may yield false-negative results, and therefore, only specimens with carefully ascertained calculations of the percentages of neoplastic cellularity should be studied.
Pyrosequencing is a more recently developed sequencing approach that permits the sequencing of short DNA fragments (10–100 bp) within a gene. In clinical practice, pyrosequencing is usually reserved for 10- to 20-bp gene hot spots in which a single site or a few codons may contain single point mutations. Pyrosequencing is more sensitive than Sanger sequencing (with a detection limit of about 5%–10% of mutant alleles).
Of great interest is the rapid and widespread growth of next-generation sequencing. Next-generation sequencing includes various high-throughput sequencing technologies that have the ability to rapidly process millions of sequence reads and generate large data sets that can be analyzed in different ways to answer a multitude of questions about genomic alterations in cancer. Although not currently used in clinical practice, next-generation sequencing is anticipated to represent the next frontier in tumor mutation analysis with the promise of an improved understanding of carcinogenesis and a more precise classification system with discovery of new therapeutic targets. For now, the expense and bioinformatics burden of this technology precludes its use in clinical molecular diagnostic laboratories.
Dramatic advances in cytogenetic and molecular biologic techniques have furthered our understanding of sarcomagenesis. Cytogenetic and molecular genetic assays are used routinely for diagnostic and prognostic purposes in molecular pathology laboratories and represent a powerful adjunct to complement conventional microscopy and radiographic assessment in the formulation of an accurate diagnosis. Care should be taken, however, to recognize the limitations of these approaches. Ideally, more than one technical approach should be available to a diagnostic laboratory to compensate for the shortcomings of another approach in different clinical situations.
Exciting new technologic innovations in molecular diagnostics are on the horizon and soon will allow for even higher-resolution analysis of the genetic anomalies underlying the pathogenesis of soft tissue tumors. Moreover, advancements in soft tissue pathology, not only at the DNA level but also at the RNA level through transcriptome analysis and at the epigenetic level through genome-wide analysis of promoter methylation or chromatin acetylation, can be anticipated. These advancements will likely not only affect diagnosis and classification schemes, but perhaps more important, personalized therapy based on identified targets within a given tumor can be employed for improved outcomes.
From the Departments of Pathology and Microbiology (Drs Bridge and Cushman-Vokoun), Pediatrics (Dr Bridge), and Orthopedic Surgery (Dr Bridge), University of Nebraska Medical Center, Omaha.
The authors have no relevant financial interest in the products or companies described in this article.