Solid tumors

Background

Once thought of as a simple clonal expansion of abnormal cells, cancer is now recognized as a complex, multifactorial, and polygenic syndrome. Many cancer types have been reclassified due to new understanding of molecular-based phenotypes and tumor-host interactions. Since the initial human genome reference sequence in 2004, cancer genomics research has focused on using this reference as a template for characterizing the somatic genomic variants that underlie cancer development and those germline genomic variants that underlie human susceptibility to develop cancer. The Human Genome Project also promised to identify novel therapeutic targets for treating cancer, and we are seeing those efforts come to fruition with new targeted small molecule- and immunotherapies. Major advances in technology, including massively parallel sequencing, have provided a comprehensive somatic landscape of most major cancer types that informs our understanding of the numbers and types of variants in the cancer genome.

Content

The diagnosis of human cancer is based mainly on the morphologic characteristics of tumor cells and some specific protein expression detected by immunohistochemical staining. More recently, genomic analyses have become an integral component of cancer patient management, guiding therapy selection based on somatic and/or germline variant profiles. This is possible due to the cumulative knowledge gained from large-scale cancer genomics discovery efforts in tens to thousands of human cancers across many tissues of origin using high complexity molecular-based technologies. This chapter will outline aspects of molecular assays of solid tissue malignancies in the clinical setting that have been informed by research-based discovery over the last 20 years.

Sampling and preservation methods

The assessment of human cancers is requiring that increased numbers of biomarkers be tested for diagnostic, prognostic, and therapeutic evidence. In this scenario, the most important of the preanalytical variables is sampling for adequate tissue quantity and quality. Solid tumors can be sampled by a variety of procedures that yield different amounts of tumor cells from which DNA, RNA, or both can be isolated for diagnostic assays. To that extent, tissue stewardship is critical to any biomarker testing. In the ideal setting, the number of assays required for pathologic evaluation of the tumor would dictate the amount of sampling done; however, the typical scenario involves a limited amount of sample that often restricts the comprehensiveness of assays that can be performed. Due to advances in detection and sampling technologies, tumors may be small at the time of diagnosis and the tissue obtained for diagnostic assays may be minimal. This scenario escalates the need for tissue stewardship due to the need for additional molecular tests beyond those required for traditional pathology evaluation. Initially, tissue samples must first satisfy the standard diagnostic repertoire of microscopy-based pathology which includes multiple stains and immunohistochemistry. One attempt to ameliorate some of the issues with low quantities of tumor tissue for molecular profiling has been to consolidate single gene/single variant assays that each require a specific amount of tissue, with more comprehensive gene panel testing using massively parallel sequencing and a single sample for the detection of multiple genes/variants.

Solid tumors are commonly sampled using either fine needle aspiration (FNA) or core biopsy procedures. ^, These approaches are often done in advance of surgery as the diagnostic procedure or in patients determined from imaging to have nonresectable tumors. The FNA (21- to 25-gauge needle) recovers a minimal amount of tissue and can consist of a few tumor cells in the fluid that is co-aspirated into the needle. ^, This sample type often is used to create cell blocks that are then processed similar to tissue samples. Core biopsies (18- to 21-gauge needle) can obtain intact and solid cores from the tumor mass. If imaging such as ultrasound or CT scanning is used to guide the biopsy needle to the tumor mass, several “passes” are made for better sampling. Alternatively, for patients with resectable cancer without plans for neoadjuvant therapy, the tumor mass can be removed at surgery and may be preserved in several ways, including processing and storage of formalin-fixed, paraffin embedded (FFPE) tissue blocks, frozen, or sampled and saved as core needle biopsies. Historically, pathology of solid tumors has focused on preservation methods that stabilize proteins and other cellular structural components, preserving tissue structure for microscopic visualization in the presence of specific staining or immunohistochemistry. Hence, tissue fixation by soaking in buffered formalin or formaldehyde was developed to stabilize the cellular proteins. Subsequent embedding in paraffin wax is performed to create an impervious, room temperature-stable substrate. Once the paraffinized tissue has solidified into a block, thin sections can be cut for subsequent characterization and diagnosis by staining and microscopic examination. Although this approach is facile and preserves tissues for long-term storage at room temperature, formalin causes crosslinking with cytosine residues in DNA and RNA. Subsequent oxidation results in breaks in the nucleic acid sugar-phosphate backbone, thereby degrading nucleic acids. The degradation is time-dependent, however, such that preserved tissues from FFPE of fewer than 3 years are often equivalent to fresh frozen tissues in terms of yield and quality of nucleic acids. As diagnostic assays of solid tumors have begun to include nucleic acids, alternative tissue preservatives and preservation methods are being used. These methods include flash freezing tissue at −80 °C (dry ice acetone bath) or immersion in nucleic acid stabilizers (e.g., OCT, PAXgene, RNAlater). Liquid cytology samples such as aspirates can be processed in methanol-based fixatives which are ideal for molecular studies. As described in this chapter and covered in-depth in Chapter 71 , newer approaches to evaluate bodily fluids for the presence of circulating tumor cells or circulating free tumor DNA are emerging, often requiring specific stabilization reagents in collection tubes (e.g., Streck or PAXgene tubes).

Staining and selection of tumor cells

Several staining methods have been developed to examine the preserved needle biopsy or resected tissues obtained in solid tumor sampling. The most basic assay is referred to as an “H&E,” or hematoxylin and eosin stain, and readily identifies tumor and normal cells in a tissue section under the light microscope ( Fig. 69.1 ). The H&E staining not only is used to diagnose cancer but also to enumerate tumor cells in a tissue section to provide an estimate of the percent tumor cell nuclei present. Since solid tumors are a mixture of tumor cells and various normal cells, this estimate of tumor “cellularity” indicates how tumor-rich that sample is relative to other needle biopsies or other portions of the bulk tumor. Tumor features such as necrotic areas may be identified in the microscopic evaluation as well. In sections with evident necrosis or focal areas of tumor cells, a process of macro- or micro-dissection to enrich for tumor cells and subsequent nucleic acid isolation can be pursued. For example, in tumor types such as prostate, thyroid, and pancreatic adenocarcinoma there is a low proportion of tumor nuclei present relative to surrounding normal cells (stroma and immune cells, for example). In these tumor types, laser capture microdissection (LCM) can be used to micro-dissect tumor cells away from surrounding normal cells prior to nucleic acid isolation ( Fig. 69.2 ). Here, the LCM imaging system produces an image of the tumor section, and an operator uses software to identify the tumor cells for harvest. A specific membrane is placed adjacent to the tumor section and the LCM fires an infrared laser pulse at each tumor cell identified for harvest, thereby affixing it to the membrane. Subsequent cutting of tissue and membrane is performed by an ultraviolet laser, completing the “LIFT” or laser-induced forward transfer process of cellular isolation. Once the desired number of tumor cells is harvested, the membrane goes through a series of processing steps to isolate DNA or RNA (or both) from the captured cells. Macro-dissection is used when the tumor cell region can be easily seen and removed away from other tissue using a scalpel blade or other device to scrape the cells off of the slide and into a receptacle for nucleic acid extraction.

A staining-based method to identify tumor-specific antigens is immunohistochemistry (IHC). In this technique, a protein or protein-epitope specific antibody is coupled to an enzyme (e.g., horseradish peroxidase) to identify tumor cells in a tissue section that are expressing that protein. Examples of IHC stains include those for estrogen and progesterone receptors, and the ERBB2 (HER2) receptor ( Fig. 69.3 ).

Background

While we have known for many decades that cancer’s origins lie in changes to the cellular genome, only recently have technologies become routinely available to obtain somatic mutation profiles of tumor genomes. Initial gene cloning efforts in the early 1980s identified the chromosomal locations and sequences of many oncogenes and tumor suppressors in the human genome. ^, The decoding of the human genome coupled with technological advances opened the door to genome-wide studies of cancer. For example, learning the sequences of human genes enabled the construction of microarrays to query RNA expression in tumors (described in Chapter 65 ). By using advanced bioinformatic analysis approaches such as clustering algorithms, similarities and differences in gene expression across tumors from a single tissue site (e.g., lung adenocarcinoma) were revealed. Clusters of gene expression results, when correlated with other pathology-based categories, revealed subtypes within a given tissue site, such as the intrinsic subtypes of breast cancer that also predicted a clinical correlate such as outcome. Similarly, SNP-based microarrays could be utilized to identify gross-scale chromosomal aberrations by comparing normalized signal strength between tumor and normal DNA. Newer approaches using whole genome imaging technology to detect large structural variations are discovering novel unbalanced translocations and deletions not detectable by other methods to date. As will be explored below, conducting research-based inquiries of cancer genomes with these tools provided the means by which the clinical utility (diagnosis, prognosis, treatment decision) of the resulting classifications was demonstrated, thereby indicating a rationale for translating these approaches into clinical laboratory assays. More recently, the introduction of massively parallel sequencing (MPS) platforms and associated methods for exploring cancer genes, genomes, and transcriptomes (expressed RNAs) have supplanted foundational technologies such as capillary-based Sanger sequencing and microarrays. MPS methods have the potential to produce more quantitative data while simultaneously expanding the identification of somatic alterations. The first application of MPS to decode a cancer genome was published in 2008 when Ley and colleagues sequenced and compared the whole genome sequence (WGS) data from a patient with FAB M1 (normal karyotype) acute myeloid leukemia (AML) to a matched normal skin sample. The second WGS cancer genome published was also from a single AML patient that revealed an unexpected mutation in the gene IDH1 (a glycolytic pathway enzyme). IDH1 mutations then were characterized as recurrent in a panel of 188 AML samples and associated with a poor prognosis. This unsuspected cancer driver was subsequently characterized as having a neomorphic function that could be therapeutically targeted. From these single patient beginnings, the field of cancer genomics has exploded to characterize hundreds of thousands of cases.