Cancer Biomarkers [11.3.2.3]
Writer and Curator: Larry H. Bernstein, MD, FCAP
Cancer Biomarkers
This discussion is extracted from the Special Section – Cancer Biomarker Update – in May Arch Pathology and Lab Med. It is not intended to be complete, but it has quite timely content. There are three articles I shall cover.
11.3.2.3 Cancer Biomarkers
11.3.2.3.1 Cancer Biomarkers in Structured Data Reporting
11.3.2.3.2 Cancer Biomarkers in Myeloid Malignancies
11.3.2.3.3 National Comprehensive Cancer Network Consensus on Use of Cancer Biomarkers
Cancer Biomarkers – The Role of Structured Data Reporting
Simpson, RW; Berman, MA; Foulis PR, et al.
Arch Pathol Lab Med. 2015;139:587–593
http://dx.doi.org:/10.5858/arpa.2014-0082-RA
The College of American Pathologists has been producing cancer protocols since 1986 to aid pathologists in the diagnosis and reporting of cancer cases. Many pathologists use the included cancer case summaries as templates for dictation/data entry into the final pathology report. These summaries are now available in a computer-readable format with structured data elements for interoperability, packaged as ‘‘electronic cancer checklists.’’ Most major vendors of anatomic pathology reporting software support this model. Objectives.—To outline the development and advantages of structured electronic cancer reporting using the electronic cancer checklist model, and to describe its extension to cancer biomarkers and other aspects of cancer reporting. Data Sources.—Peer-reviewed literature and internal records of the College of American Pathologists. Conclusions.—Accurate and usable cancer biomarker data reporting will increasingly depend on initial capture of this information as structured data. This process will support the standardization of data elements and biomarker terminology, enabling the meaningful use of these datasets by pathologists, clinicians, tumor registries, and patients.
Narrative Versus Structured Data Reporting Clinical laboratory reports typically consist of discrete data elements with structured qualitative or quantitative information, often using standardized laboratory methods, data elements, and units. When discrete data elements are electronically transmitted to external clinical information systems, the transmitted information may be annotated with one or more terminologies such as Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT)4 and Logical Observation Identifiers Names and Codes (LOINC),5 although the consistent application of such codes to structured laboratory data is not yet an interoperable standard. Because the structure of clinical laboratory data tends to be fixed and standardized before the point of data entry, reporting these data elements in a tabular synoptic format is a relatively simple process. The report output may not include all data collected (eg, methodologic details), but clinically relevant data can be easily extracted by computer algorithm and automatically reported in easily readable format (including custom text, result explanations, and test value trends).
Anatomic pathology reports, by contrast, have traditionally been narrative and recorded as unstructured or partially structured fields of text. Unfortunately, narrative reporting often lacks consistency in organization, content, units, terminology, and completeness.6–8 These structural inconsistencies create difficulties in finding and understanding clinically important data and increase the chance of omitting key data elements and misinterpreting information present in the narrative. This is particularly problematic when clinicians encounter reports from multiple pathology laboratories or when patients receive care at multiple institutions. Narrative reporting has equally negative effects on computer readability; the ability of computers to correctly parse and classify information contained in a narrative report is imperfect even when using advanced natural language processing software designed specifically for anatomic pathology.9,10 Natural language processing–parsed text must always undergo human review, editing, and signoff before release for patient care or research.
Ensuring consistency and readability of cancer and biomarker reports requires a reporting solution integrated into the pathologist workflow that supports entry of standardized data directly into a laboratory information system and/or electronic health record system. These systems can produce highly readable synoptic reports and can include computer-based report validation of standardized data elements to reduce or eliminate the chance of omitting required data elements. With the transition from narrative to synoptic reporting for cancer cases, many laboratories have been using modified CCPs or locally developed templates or macros, which may or may not contain all required data elements. This common mode of data entry fits well into the pathologist workflow and can result in organized, highly readable synoptic reports, but generally results in information stored as text in a single large data field. Even when results are entered as discrete data elements, subsequent storage mechanisms usually result in nonstructured text or nonstandard custom data fields in a local computer system. Unfortunately, narrative and nonstandardized data sets are very difficult to reliably aggregate and analyze for laboratory quality assurance, research, or cancer registry surveillance. Such aggregated data also remain relatively unreliable because of changes in information systems. Many of these issues can be eliminated by entering and reporting structured data with standardized electronic templates.
CAP eCC History, Development, and Adoption Efforts to bring structure to cancer pathology reporting began in the late 1980s and early 1990s1,11,12 with publication of templates that were the precursors of the current CCPs (Table). The primary goal was to improve the care of cancer patients by improving the reporting rate of clinically important data elements. The checklist approach was adopted to help standardize terminology and ensure all relevant data elements are reported. The 66 current CCPs, 3 new cancer biomarker templates,13–16 and 85 eCC templates represent the evolution of the original 1986 CCP model. The CCPs and templates are created and maintained through the ongoing work of the CAP Cancer and Cancer Biomarker Reporting Committees. The CCPs are widely adopted by laboratories and used for accreditation purposes by the American College of Surgeons–Commission on Cancer17 and CAP Laboratory Accreditation Program.18
During the past several years, the CAP has worked to create standardized pathology reporting templates that enable individual pathologists and software vendors to capture, store, retrieve, transmit, and analyze diagnostic cancer pathology information. Electronic versions of these templates, the eCCs, are based on consistent structured data representation, which enables simple yet robust computerization of cancer pathology data elements suitable for patient care, cancer registry transmission, and research. Synoptic reports are not the same as ‘‘structured data.’’ Although synoptic reports are formatted for optimum human readability and understanding, they consist of textbased questions and answers (ideally one pair per line) that present problems for computer readability and interoperability. Structured data, by contrast, refers to representation of data elements in a computer-readable data exchange format such as XML. The structured data model used by eCC XML templates assigns a unique identifier (a composite key) to every question and answer choice, template, section, and note listed in the template. Composite keys are used throughout the entire eCC life cycle to transmit the precise identity of each data element and its origin in a specific version of an eCC template.
The CAP eCC model has been implemented province wide by Cancer Care Ontario through their multiyear synoptic pathology reporting and change-management project.19–21 The Canadian Partnership Against Cancer is also currently working with several other Canadian provinces to implement population-level electronic synoptic reporting based on the CAP eCC. The Cancer Care Ontario project has shown that there is high acceptability among pathologists and clinicians22 and that data are usable for the secondary needs of tumor registries,20 but there remains room for improvement. For instance, both the Reporting Pathology Protocols project reports23–25 and the Cancer Care Ontario implementation reports22 suggest that pathologists require more time to complete the reporting task.
While this may meet quality and data reporting needs, it remains a potential barrier to acceptance. Automated human-readable report generation also could be improved, especially in terms of creating best practice guidelines for report output. Both data entry and report generation have traditionally been supported by laboratory information systems vendors, but the success of implementation has varied, and often significant effort is required to modify the resultant human readable report to satisfy local clinical needs. Because the final report remains most important for patient care, the CAP Diagnostic Intelligence in Health Information Technology and Pathology Electronic Reporting committees have initiated work on creating and promoting a standardized data structure within cancer pathology reports.
Abbreviated CAP eCC History and Milestones
2010–2012 CCO successfully implements population level electronic synoptic reporting in nearly all disease sites based on 2010 CAP eCC standards, which include AJCC 7th edition TNM staging; 97% of labs report using structured data from eCC.20
2010 CAP Laboratory Accreditation Program begins to survey institutions for inclusion of required CCP data elements in AP reports.38
2010 NAACCR Pathology Data Workgroup develops implementation guide to assist with CAP eCC-based transmissions of cancer data to central cancer registries.39
2011 CCO user acceptability data demonstrate high level of acceptance for eCC-derived synoptic reports among clinicians and pathologists.22
2012 CAP forms the multi-organizational Cancer Biomarker Reporting Workgroup, tasked to produce standardized reporting templates for breast, colorectal, and lung cancer biomarkers.40
2013 eCC-based reporting in Ontario is used to improve quality and practice performance.20
2013 The first cancer biomarker templates are produced for breast, colorectal, and lung cancer.13–16 They are available on the http://www.cap.org/cancerprotocols Web site in Word and PDF format (accessed April 28, 2014). The eCC versions are available through CAP.
2013 Launch of CAP eFRM, a software product to aid vendor integration of eCCs into AP-LIS systems or for use as a standalone product.
2014 By December 2013, CAP is maintaining current versions of 66 CCPs, 3 cancer biomarker templates, and 85 corresponding eCC templates.
References
13. Cagle PT, Sholl LM, Lindeman NI, et al. Template for reporting results of biomarker testing of specimens from patients with non–small cell carcinoma of the lung. Arch Pathol Lab Med. 2014; 138(2):171–174. http://dx.doi.org:/10.5858/arpa.2013-0232-CP
14. Cagle PT, Allen TC, Olsen RJ. Lung cancer biomarkers: present status and future developments. Arch Pathol Lab Med. 2013; 137(9):1191–1198. 15. Bartley AN, Hamilton SR, Alsabeh R, et al. Template for reporting results of biomarker testing of specimens from patients with carcinoma of the colon and rectum. Arch Pathol Lab Med. 2014; 138(2):166–170. 16. Fitzgibbons PL, Dillon DA, Alsabeh R, et al. Template for reporting results of biomarker testing of specimens from patients with carcinoma of the breast. Arch Pathol Lab Med. 2014; 138(5):595–601. 20. Srigley J, Lankshear S, Brierley J, et al. Closing the quality loop: facilitating improvement in oncology practice through timely access to clinical performance indicators. J Oncol Pract. 2013; 9(5):e255–e261. http://dx.doi.org:/10.1200/JOP.2012.000818 22. Lankshear S, Srigley J, McGowan T, Yurcan M, Sawka C. Standardized synoptic cancer pathology reports—so what and who cares?: a population-based satisfaction survey of 970 pathologists, surgeons, and oncologists. Arch Pathol Lab Med. 2013; 137(11):1599–1602. 38. College of American Pathologists. CAP cancer protocols frequently asked questions. http://www.cap.org/apps//cap.portal 41. Amin MB. The 2009 version of the cancer protocols of the College of American Pathologists. Arch Pathol Lab Med. 2010; 134(3):326–330. |
Figure 1. (not shown) Narrative versus synoptic versus structured reporting of breast biomarker testing (excerpts). The narrative row shows a portion of a dictated biomarker report. The synoptic row satisfies the College of American Pathologists (CAP) synoptic reporting requirements, but is not computer readable.
Figure 2. (not shown) The College of American Pathologists (CAP) Cancer Protocols (CCPs) are developed by the CAP Cancer Committee. Each CCP is reformulated as question/answer structures, entered into the CAP electronic Cancer Checklist (eCC) Template Editor (not shown), and stored in the eCC template database. The eCC files in XML format are produced from this database and delivered to vendors of anatomic pathology/laboratory information system (LIS) software systems. Vendors convert the eCC files into data entry form implementations using their local technologies. In addition, most vendors create eCC-based templates for creating synoptic reports. When pathologists enter data into the eCC-based data-entry forms, the vendor software is able to run validation checks such as assessing whether all CCP-required data elements are recorded. Synoptic reports are developed from the eCC-derived data and delivered to health care providers for patient care. The eCC-structured data is stored in the vendor database, where it can be transmitted in interoperable format to other computer systems. Secondary uses of eCC-based data include cancer registry reporting, quality assurance, biospecimen annotation, research, decision support, and financial reporting. The horizontal arrows involve the exchange of eCC composite keys, preserving the fidelity of the data as part of an eCC template, providing the foundation for interoperable data transmission formats, and enabling the regeneration of eCC datasets in the exact format in which they were recorded. Activity columns that directly impact health care activities are shaded in light blue. Abbreviation: EHR, electronic health record.
Future The use of standardized, structured data elements is foundational for the development of improved reporting and clinical decision support for biomarker results. Clinicians are currently faced with synthesizing data from multiple narrative reports to decide on treatment options. Often these narrative reports are from different laboratories with very different report formats and include variable methodologic details, all of which hinders understanding of important results. For biomarkers that determine a patient’s eligibility for specific drugs, a computer-generated report that presents test results in a tabular form, similar to antibiotic susceptibility testing, may be desirable. This reporting method would allow for display of biomarker test results over time and could also link to other databases.
Structured data allows for clinical decision support such that the report displays only eligible drugs, or the report displays a note stating that a test result suggests a patient is not eligible for a specific drug. Using standardized terminology allows these rules to be the same between institutions, even if electronic health record system vendors use different means of implementation.
Figure 3. (not shown) College of American Pathologists electronic Cancer Checklist lung cancer biomarker template—anaplastic lymphoma kinase (ALK). Abbreviations: EML4, echinoderm microtubule-associated protein-like 4; KIF5B, kinesin family member 5B; KLC1, kinesin light chain 1; TFG, tropomyosin receptor kinase–fused gene.
Figure 4. (not shown) Examples of tumor biomarker dashboards. Abbreviations: ALK, anaplastic lymphoma kinase; ROS1, ROS proto-oncogene 1, receptor tyrosine kinase.
This system would allow for more efficient, more accurate, and safer methods of providing data for optimizing patient care, with all of the discrete data transmitted electronically and linked to the original tumor report. In Ontario, Canada, this vision is rapidly advancing, as demonstrated by the Cancer Care Ontario successes with eCC implementation and current plans to implement the eCC biomarker templates across the province. Future challenges include the identity and tracking of related tumor samples over time and integration of testing from different laboratories. Because testing on a given specimen can be performed at different times and in different laboratories, a future standard must address the annotation of results with tumor source, procurement dates, and other biospecimen-specific data.30 The relationship of test results from multiple specimens from the same patient needs to be recorded in a standard format so that this parent-child hierarchical relationship can be analyzed over time.
Pathologists are increasingly asked to provide biomarker information for patient care, tumor registries, epidemiologic studies, translational research, and quality improvement activities.20 The eCC model provides a pathway to meet these demands, with efficient and error-free data entry, reporting, and transmission of data elements, and with the ability to produce output that is human readable, efficient to use, and easy to interpret. As the CCPs and eCCs have matured, Ontario pathologists and cancer registries have demonstrated success with large-scale implementations. However, continued improvements are needed. As the field of pathology grows, particularly in the area of biomarkers, structured electronic reporting will become critical to helping physicians provide optimal patient care and will facilitate secondary uses of pathology data.
- Robb JA, Gulley ML, Fitzgibbons PL, et al. A call to standardize preanalytic data elements for biospecimens. Arch Pathol Lab Med. 2014; 138(4):526–537.
Molecular Genetic Biomarkers in Myeloid Malignancies
Matynia AP, Szankasi P, Shen W, Kelley TW.
Arch Pathol Lab Med. 2015;139:594–601
http://dx.doi.org:/10.5858/arpa.2014-0096-RA
Recent studies using massively parallel sequencing technologies, so-called next-generation sequencing, have uncovered numerous recurrent, single-gene variants or mutations across the spectrum of myeloid malignancies. Objectives.—To review the recent advances in the understanding of the molecular basis of myeloid neoplasms, including their significance for diagnostic and prognostic purposes and the possible implications for the development of novel therapeutic strategies. Data Sources.—Literature review. Conclusions.—The recurrent mutations found in myeloid malignancies fall into distinct functional categories.
These include (1) cell signaling factors, (2) transcription factors, (3) regulators of the cell cycle, (4) regulators of DNA methylation, (5) regulators of histone modification, (6) RNA-splicing factors, and (7) components of the cohesin complex. As the clinical significance of these mutations and mutation combinations is established, testing for their presence is likely to become a routine part of the diagnostic workup. This review will attempt to establish a framework for understanding these mutations in the context of myeloproliferative neoplasms, myelodysplastic syndromes, and acute myeloid leukemia.
Pathways Affected by Recurrent Mutations in Myeloid Malignancies
Cell Signaling The ability of a cell to respond to diverse physiologic stimuli, including cytokines, chemokines, growth factors, and hormones, or to the presence of bacteria and other microorganisms is mediated via the interaction of specific ligands and their corresponding cell surface receptors. Ligand binding usually results in receptor dimerization and activation of a tyrosine kinase, either intrinsically present in the cytoplasmic domain or as an associated polypeptide. Further propagation of the signal from the cell surface to the nucleus involves the formation of macromolecular complexes and the activation or inactivation of various enzymes. The final outcome of the signal transduction is modulation of the expression of certain genes and their products, which ultimately produces a cellular response. In normal cells, this process is tightly regulated owing to the involvement of negative or inhibitory signals. In tumor cells these processes may be perturbed owing to mutations that impart inappropriate activation or deactivation of enzymatic function. Genes for receptor protein tyrosine kinases, such as FLT3 and KIT, or receptor-associated kinases, such as JAK2, are the most commonly mutated cell-signaling factors in myeloid malignancies. Activating mutations in these proteins occur in narrowly defined hotspots, resulting in ligand-independent dimerization or constitutive kinase activation. An example is the protein tyrosine kinase JAK2, which transduces signals from ligand-bound cell surface receptors for thrombopoietin (TPOR/MPL) and erythropoietin (EPOR).
Activating mutations in JAK2 are commonly found in the myeloproliferative neoplasms (MPNs): polycythemia vera (PV), essential thrombocythemia (ET), and primary myelofibrosis (PMF). These disorders appear to be driven, in large part, by the inappropriate activation of growth factor– signaling pathways, and JAK2 is central to signaling from EPOR and TPOR/MPL via STAT5, STAT3, the RAS-MAP kinase pathway, and the PI-3 kinase–AKT pathway. Many negative regulators of cytokine signaling, such as SH2B adaptor protein 3 (SH2B3, also known as LNK)1,2 and cytokine-inducible SH2-containing protein (CISH, also known as SOCS)3 keep the pathway in balance. Downstream RAS signaling is counteracted by NF1, a protein that stimulates the intrinsic RAS guanosine triphosphatase activity. An effect similar to JAK2 mutations may be achieved through mutations in other proteins involved in these pathways, including activating mutations of a surface receptor (MPL) or loss-of-function mutations in negative regulators (SH2B3, CISH), all of which promote survival, proliferation, and differentiation of committed myeloid progenitors.4 Example genes included in this group: JAK2, MPL, KIT, FLT3, CSF3R, PTPN11, KRAS, NRAS, and NF1.
Transcription Transcription is a tightly regulated process that depends on the formation and assembly of protein and protein-DNA complexes. These complexes, called transcription factors, bind to specific DNA sequences adjacent to the genes they regulate and promote (in the case of an activator) or block (in the case of a repressor) the recruitment of RNA polymerases to those genes. This regulatory activity controls the formation of messenger RNA (mRNA) transcripts. Mutations that block the activity of transcriptional activators may, in certain circumstances, lead to a block in cellular differentiation due to the lack of the necessary gene products. Many of the transcription factors that are recurrently mutated in myeloid malignancies, such as RUNX1, GATA1, GATA2, and CEBPA, are involved in fundamental aspects of myelopoiesis and it is believed that these mutations lead to a block in myeloid differentiation. Example genes included in this group: RUNX1, CEBPA, GATA1, GATA2, ETV6, and PHF6.
Epigenetic Modifiers: Regulation of DNA Methylation DNA methylation involves the addition of a methyl group to cytosine bases in the context of cytosine-guanine sequences (so-called CpG sites), leading to the creation of 5-methylcytosine. CpG islands are usually located in or near promoter regions. Their methylation is an important epigenetic mechanism for regulating gene expression and, in the context of heritable methylation patterns, underlies the process of genomic imprinting. Additionally, it has been hypothesized that aberrant DNA methylation may contribute to the pathogenesis of cancers,5–8 including myeloid neoplasms. Although cancer genomes tend to be globally hypomethylated, in comparison with normal tissues, hypermethylation of specific CpG islands at tumor suppressor genes, resulting in their inactivation, is common in many tumors.9
ormation of compact, inactive heterochromatin.10 Several factors regulate the process of DNA methylation. Mutations in some of these factors have been found recurrently in myeloid neoplasms.11–13 DNA methyltransferases catalyze the methylation at the 50 position of cytosine. DNMT3A and DNMT3B are involved in de novo methylation, whereas DNMT1 maintains hemimethylated DNA during replication. Once created, 5-methylcytosine can be further modified by a group of methylcytosine dioxygenases (Ten Eleven Translocation dioxygenases: TET1, TET2, and TET3) to 50-hydroxymethylcytosine, a presumed short-lived intermediary that may lead to demethylation of cytosine.
Mutations in both DNMT3A and TET2 likely lead to loss of function of the respective enzyme activities. Recently, mutations in IDH1 and IDH2 have been identified in myeloid neoplasms and other cancers. Interestingly, recurring mutations of an arginine residue in the active site (R132 in IDH1 and R140 in IDH2) prevent the normal catalytic function of the enzyme (conversion of isocitrate to aketoglutarate) and appear to induce a neomorphic enzyme activity resulting in the formation of the rare oncometabolite 2-hydroxyglutarate.14 TET2 belongs to a family of dioxygenases that requires a-ketoglutarate as a cofactor.15,16 It has been shown that 2-hyroxyglutarate acts as a competitive inhibitor of a-ketoglutarate–dependent dioxygenases, which include TET2 and members of the KDM family of histone demethylases, thereby inducing epigenetic changes at both the level of DNA methylation and histone modification.14,17
Therapeutic inhibitors of mutant forms of IDH proteins and the resulting 2-hydroxyglutarate are also under investigation.18 Example genes included in this group: DNMT3A, TET2, IDH1, and IDH2.
Epigenetic Modifiers: Mutations Affecting Histone Function Histone proteins are involved in the dynamic organization of DNA into zones of active euchromatin and inactive heterochromatin in a process that is regulated, in part, by a complex series of posttranslational modifications to histone tails, including acetylation and methylation. These modifications affect the recruitment of regulatory proteins such as transcription factors, corepressors, and coactivators, as well as histone-modifying enzymes themselves. Trimethylation of the lysine at position 27 in histone H3 (H3K27), one of the more common modifications, generally leads to reduced gene expression and it would be expected that mutations that reduce methylation at H3K27 would activate transcription. Recently, recurrent mutations in several genes encoding histone regulators, including ASXL1, EZH2, SUZ12, and KDM6A (also known as UTX), have been identified.
Perturbations in epigenetic pathways result in global, genome-wide effects and it is often difficult to identify which altered cellular function eventually leads to neoplasia. The same also holds true for perturbations in the RNA splicing machinery and the cohesion complex (see below). Example genes included in this group: ASXL1, EZH2, SUZ12, and KDM6A.
Cohesin Complex Genes The cohesin complex is a conserved multimeric protein complex that regulates cohesion of sister chromatids during cell division,21 postreplicative DNA repair,22,23 and global gene expression.24–28
Example genes included in this group: STAG2, RAD21, SMC1A, and SMC3.
RNA-Splicing Factors RNA splicing results in the formation of mature mRNA transcripts derived from exons, the protein coding portion of the genome. Splicing occurs in a macromolecular complex of small nuclear RNAs and proteins assembled de novo on each pre-mRNA strand in a multistep process. This complex is known as the spliceosome. Transcripts may undergo alternative splicing in a tissue, or context-specific manner and the protein products of alternatively spliced transcripts may have altered function.
Example genes included in this group: SF3B1, SRSF2, ZRSR2, and U2AF1.
Cell Cycle Regulators Example genes included in this group: TP53 and NPM1.
Genetic Biomarkers in Myeloid Malignancies
Myeloproliferative Neoplasms Myeloproliferative neoplasms encompass a group of clonal stem cell disorders characterized by expansion of 1 or more of the myeloid lineages resulting in bone marrow hypercellularity and increased peripheral blood myeloid cell counts. The MPN category includes chronic myelogenous leukemia, PV, ET, PMF, chronic neutrophilic leukemia, mastocytosis, and others. The underlying genetic landscape of some of these disorders is very well understood, as in the case of chronic myelogenous leukemia with t(9;22), but is much less well understood in many other entities.
The discovery of a recurrent codon 617 activating mutation (V617F) in exon 14 of the tyrosine kinase JAK237–40 and additional mutations in JAK2 exon 1241 provided the first genetic evidence of the importance of dysregulated growth factor signaling in these disorders. The prevalence of JAK2 mutations in classical MPNs varies from 95% to 99% in PV, 50% to 70% in ET, 40% to 50% in PMF,37–41,43 and molecular tests for their detection are available and widely used in clinical practice. Similarly, activating mutations in the MPL gene, encoding the thrombopoietin receptor, are present in approximately 4% of ET cases and approximately 11% of PMF cases.44–47 Recently, calreticulin (CALR), encoding an endoplasmic reticulum chaperone, has also been shown to be important. Somatic CALR mutations are found in 70% to 84% of patients with ET or PMF with wild-type JAK2 and MPL, 8% of MDS cases, and occasionally in other myeloid neoplasms.48 Clonal analyses suggest CALR mutations act as an initiating mutation in some patients.48 In ET and PMF, CALR mutations and JAK2 and MPL mutations are mutually exclusive,49 and CALR mutations appear to be absent in PV.49 CALR mutations appear to be primarily insertion or deletion mutations that result in a frameshift and the subsequent generation of a novel C-terminal peptide.48,49
Many other genes involved in intracellular signaling, such as negative regulators of the JAK2 signaling pathway, are mutated in MPNs. Among these is SH2B3, which negatively regulates JAK2 activation through its SH2 domain. Mutations in SH2B3 during the chronic phase are uncommon, fewer than 5% in ET and PMF50; however, their frequency increases during leukemic transformation, suggesting a role in disease progression.51 Another negative regulator found mutated in MPNs is the Cbl proto-oncogene, E3 ubiquitin protein ligase gene (CBL). CBL acts as a multifunctional adapter with ubiquitin ligase activity and by competitive blockade of signaling.
A shift from a simple, chronic myelogenous leukemia–like model for MPN pathophysiology to a more complex model occurred with the emergence of evidence of a ‘‘pre-JAK2’’ genetic event. This concept is based on the observation that mutations in signaling molecules are not sufficient for disease development and that several cooperating genetic hits appear to be required.4 Mutations in genes involved in epigenetic regulation, including EZH2, ASXL1, and TET2 (also found in many other myeloid neoplasms), are postulated to act as those initialing events, preceding JAK2V617F mutations.55 EZH2 mutations do not occur in ET, but are present in 3% of PVs and 13% of MFs.56 ASXL1 mutations are found in only approximately 7% of patients with ET and PV but more frequently in PMF cases (from 19%–40%).57,58 TET2 mutations occur in approximately 14% of MPNs, ranging from 11% in ETs to 19% in PMFs.59,60 Finally, there are mutations that are rarely found during the chronic phase but which may be present at transformation, and are therefore thought to play a role in disease progression. IDH1 and IDH2 mutations, for example, have a low frequency in the chronic phase (0.8% in ET, 1.9% in PV, and 4.2% in MF) but a much higher frequency in blast phase.61
Myelodysplastic Syndromes and MDS/MPN Overlap Disorders
The myelodysplastic syndromes are a group of clonal hematopoietic stem cell disorders characterized by ineffective hematopoiesis, morphologic evidence of dysplasia in at least 1 of the myeloid lineages, peripheral cytopenias, bone marrow hypercellularity, and an increased risk of development of AML.74 Clonal cytogenetic abnormalities, including large deletions and chromosome gains, as well as balanced translocations, are observed in approximately 50% of MDS cases by routine methods,74 and their identification aids in establishing the diagnosis and may provide prognostic information.
Acute Myelogenous Leukemia Acute myeloid leukemia is a genetically heterogeneous disease resulting in the accumulation of myeloblasts in bone marrow with a concomitant reduction in normal hematopoiesis. The diagnosis and subclassification of AML depends on detecting the presence of recurrent cytogenetic abnormalities.74 In many cases, particularly those that are cytogenetically normal (CN-AML), several single-gene mutations further aid in the stratification of disease outcomes. The significance of mutations in genes such as FLT3, NPM1, and CEBPA is well established but nextgeneration sequencing has led to the discovery of numerous additional recurrent mutations, including in TET2,12,59 ASXL1,104 IDH1/IDH2,13,105,106 DNMT3A,11,107 and PHF6.108
Among the most common mutations found in de novo AML are NMP1, FLT3, and DNMT3A mutations, present in 22% to 29%, 22% to 37%, and 15% to 26% of samples, respectively.31,110,111 Other genes less commonly targeted by mutations are IDH1/IDH2 (15%– 20%), KRAS/NRAS (12% combined), RUNX1 (5%–10%),TET2 (8%–14%), TP53 (2%–8%), CEBPA (6%–14%), WT1 (6%–8%), PTPN11 (4%), and KIT (4%–6%).31,110,111
The prognostic significance of a subset of recurrent mutations is well established. In CN-AML, biallelic CEBPA mutations127,128 and NPM1 mutations without FLT3-ITD mutations129–133 are associated with a favorable prognosis. In contrast, FLT3-ITD without NPM1 mutations112,113,129–132 and MLL–partial tandem duplication mutations134–136 portend poor outcome. KIT mutations in AML with t(8;21) or inv(16)131,137 are also associated with unfavorable outcome. The European LeukemiaNet panel138 first proposed a standardized classification according to both cytogenetic and molecular genetic data to allow a better comparison of prognosis among patients with AML. However, only mutations of NPM1, CEBPA, and FLT3 were included in their recommendations. The relevance of more recently discovered mutations, including IDH1, IDH2, WT1, TET2, ASXL1, among others, remains unclear.139–142 The presence of certain mutations may also allow for more targeted therapeutic regimens; for example, FLT kinase inhibitors may be useful in cases with mutations and IDH1 inhibitors are under investigation in patients with IDH1 mutations.143,144
An enormous amount of new information illuminating the genetic complexity of myeloid neoplasms has been generated during the last few years. Much work remains to be done but it is clear that the future role of the pathologist in collecting and interpreting this information will be an essential component of the management of these patients.
The Cancer Genomics Resource List 2014
Zutter MM, Bloom KJ, Cheng L, Hagemann IS, et al.
Arch Pathol Lab Med. http://dx.doi.org:/10.5858/arpa.2014-0330-C
Optimizing the Clinical Utility of Biomarkers in Oncology: The NCCN Biomarkers Compendium
Marian L. Birkeland, Joan S. McClure
Arch Pathol Lab Med. 2015;139:608–611
http://dx.doi.org:/10.5858/arpa.2014-0146-RA
The rapid development of commercial biomarker tests for oncology indications has led to confusion about which tests are clinically indicated for oncology care. By consolidating biomarker testing information recommended within National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology (NCCN Guidelines), the NCCN Biomarkers Compendium aims to ensure that patients have access to appropriate biomarker testing based on the evaluations and recommendations of the expert NCCN panel members.
Objectives.—To present the recently launched NCCN Biomarkers Compendium. Data Sources.—Biomarker testing information recommended within NCCN Clinical Treatment Guidelines as well as published resources for genetic and biological information. Conclusions.—The NCCN Biomarkers Compendium is a continuously updated resource for clinicians who need access to relevant and succinct information about biomarker testing in oncology and is linked directly to the recommendations provided within the NCCN Clinical Practice Guidelines.
Most recommendations contained within the NCCN Guidelines are based upon lower-level evidence and uniform NCCN consensus (category 2A).1 This is not a deficiency of the guidelines, but is rather because high-level evidence is not available for most decisions across the continuum of care. A deeper look at what constitutes a ‘‘recommendation’’ might begin to clarify that issue. A recommendation can include all of the recommended workup, surgical options, options for chemotherapy, and tests recommended for ongoing surveillance. Although many of these options are routinely used as standard of care in clinical practice, there is often not the available body of high-level evidence that supports category 1 recommendations, thus most are category 2A levels of evidence and consensus. In other instances, recommendations for chemotherapy regimens for which there is high-level, randomized clinical trial evidence are listed as category 1.
Several derivative products arise from the NCCN Guidelines. The NCCN Drugs & Biologics Compendium (NCCN Compendium) is a resource outlining appropriate use of drugs and biologics as recommended in the NCCN Guidelines. To be included in the compendium, an agent must first be recommended in at least 1 of the NCCN Guidelines. The compendium is typically used by clinicians and payors to determine appropriate use and as a standard to determine coverage. The rapid development and commercialization of biomarker and companion diagnostic testing in cancer gave rise to the NCCN Biomarkers Compendium, to be used by both payors and clinicians to facilitate identification of biomarker tests recommended for use by NCCN guideline panels. The NCCN uses a broad definition of ‘‘biomarker’’ for the purposes of this compendium. All tests measuring genes or gene products, which are used for diagnosis, screening, monitoring, surveillance, or for providing predictive or prognostic information, are included in the biomarkers compendium. This compendium focuses on the biology of the biomarker itself and its clinical utility in supporting clinical decision making. Information is organized by the biomarker being measured, and not by listing of commercially available tests or test kits. Close to 1000 biomarker testing recommendations are included in the NCCN Biomarkers Compendium.
The NCCN Biomarkers Compendium is presented on the NCCN Web site as a series of drop-down menus, allowing users to pick from menus listing Guideline, Disease, Molecular Abnormality, or Gene Symbol.3 Users can retrieve all recommendations for a particular disease, or can select a gene-based search in order to show which diseases have a validated use for a particular gene test. Additional fields can be displayed by selecting from a series of boxes to the right of the drop-down menus (see Figure 1, which shows default fields for RAS testing in colon cancer). Once records are displayed, the resulting table can be sorted by the information in any of the displayed columns. If a searcher chooses, all records can be displayed and then searched with any text term or sorted by any of the columns for a more comprehensive picture of the contents of the database.
Figure 1. (not shown) KRAS mutation testing recommendation from the National Comprehensive Cancer Network (NCCN) Biomarkers Compendium.3 Reproduced with permission from the NCCN Biomarkers Compendium [1] 2014 National Comprehensive Cancer Network, Inc. (NCCN.org; accessed February 21, 2014). To view the most recent and complete version of the NCCN Biomarkers Compendium, go online to NCCN.org. National Comprehensive Cancer Network, NCCN, NCCN Guidelines, and all other NCCN content are trademarks owned by the National Comprehensive Cancer Network, Inc.
Disease Description | Colon cancer |
Specific Indication | Metastatic disease |
Molecular Abnormality | KRAS/NKRAS mutation |
Test | KRAS/NKRAS |
Chromosome | 1p13.2, 12p12.1 |
Gene Symbol | KRAS/NKRAS |
Test Detects | Mutation |
Methodology | |
Category of Evidence | 2A |
Specimen Types | FFPE tumor tissue |
Recommendation | …Determination of RAS mutations. |
Test Purpose | Predictive |
Guideline Page | COL 4 of 5, COL 4, COL 9 |
Note | All patients with metastatic colorectal cancer should be genotyped for RAS mutations. At the very least … |
Figure 2. (Table) Example of PDF file generated from ‘‘print’’ command of National Comprehensive Cancer Network (NCCN) Biomarkers Compendium record. Reproduced with permission from the NCCN Biomarkers Compendium [1] 2014 National Comprehensive Cancer Network, Inc (NCCN.org; accessed February 21, 2014). To view the most recent and complete version of the NCCN Biomarkers Compendium, go online to NCCN.org. National Comprehensive Cancer Network, NCCN, NCCN Guidelines, and all other NCCN content are trademarks owned by the National Comprehensive Cancer Network, Inc.
Table 2. (List) Summary of Testing Types Included in the National Comprehensive Cancer Network Biomarkers Compendiuma,b
Protein expression |
Translocation |
Mutation |
Chromosomal defect |
Gene rearrangement |
Virus detection |
Antigen expression |
Serum proteins |
Amplification |
Short repeated sequences |
Promoter methylation |
Gene expression pattern |
Helicobacter pylori |
Table 3. Predictive Tests Used for Treatment Decision Making, Extracted From National Comprehensive Cancer Network Guidelines and Biomarkers Compendiuma,b
Test | Disease |
21-gene RT-PCR
BCR-ABL1 translocation |
Breast cancer |
ABL1 mutation | Ph+ acute lymphoblastic leukemia, chronic myelogenous leukemia |
ALK rearrangement | Non–small cell lung cancer |
BRAF mutation | Non–small cell lung cancer, melanoma, colon cancer, rectal cancer |
EGFR mutation | Non–small cell lung cancer |
ERBB2 amplification/overexpression | Breast cancer, esophageal and esophagogastric junction cancers, gastric cancer |
ESR1 expression | Breast cancer |
KIT mutation | Soft tissue sarcoma: GIST |
KRAS mutation | Colon cancer, rectal cancer, non–small cell lung cancer |
MGMT promoter methylation | Central nervous system cancers: anaplastic glioma/glioblastoma |
MLH1, MSH2, MSH6, PMS2 expression and/or mutation, MSI testing | Colon cancer, rectal cancer |
PDGFRA mutation | Soft tissue sarcoma: GIST |
PGR expression | Breast cancer |
ROS1 rearrangement | Non–small cell lung cancer |
A large number of tests were grouped for the purposes of this simplified table into the category of gross chromosomal abnormalities. Interestingly, the guidelines so far contain only a single recommendation for the use of a gene expression profiling test, and this is the 21-gene reverse transcription–polymerase chain reaction test recommended within the breast cancer treatment guideline, where the score for this test can be used as part of a decision-making process for chemotherapy recommendations in node negative, hormone receptor–positive, HER2-negative disease.
Table 3 summarizes the biomarker tests included in the NCCN Biomarkers Compendium that are predictive for either responsiveness (eg, BRAF mutation and vemurafenib sensitivity) or nonresponsiveness (eg, KRAS mutation testing and cetuximab or panitumumab insensitivity) to a particular type of therapy. As the number of companion diagnostics and targeted therapies grows, we expect this category of test to become one of the largest categories of testing contained within the Biomarkers Compendium, and it may be surprising to note that only 15 of these types of test are currently recommended within the NCCN Guidelines.
The NCCN Biomarkers Compendium generally avoids recommendations for particular methodologies or test kits to use to assess mutations and translocations. The choice of methodology and supplier for carrying out the recommended biomarker tests remains that of the treating oncologists and pathologists.
The NCCN Biomarkers Compendium may be used by payers as a reference for coverage decisions and by clinicians as a guide to which biomarkers are appropriate to test. The Biomarkers Compendium focuses on the clinical usefulness of biomarker testing rather than specific tests or test kits that identify the presence or absence of the marker. Other groups are continuing to assess clinical and analytic validity for specific biomarker test methodologies. Even the US Food and Drug Administration approval process is limited to clinical and analytic validity, and does not specifically address clinical utility. The NCCN Biomarkers Compendium is complementary to these other efforts. By providing biomarker testing information, the NCCN Biomarkers Compendium aims to ensure that patients have coverage and access to appropriate biomarker testing, based on the evaluations and recommendations of the expert NCCN panel members.
Leave a Reply