CHAPTER 19: Introduction to Hematologic Malignancies

Just as studies of hematopoiesis have been at the forefront of stem cell biology, rapid advances in our understanding of the pathogenesis of hematopoietic and lymphoid tumors have brought us to the dawn of a new molecular era of cancer diagnosis and therapy. Remarkably, in the not-too-far distant future it will become possible to routinely sequence the entire genome of blood cancers, providing us with a detailed view of the genetic changes that underlie these tumors. These advances will undoubtedly revolutionize how we diagnose and treat patients with these diseases and other cancers as well.

This accelerating rate of progress is exhilarating for workers in the field, but it creates challenges for students who are trying to grasp the basics (and authors who are trying to help them do so!). With this in mind, this chapter and the following one are meant to serve as a primer for students new to these fascinating and clinically important diseases. Here, we provide an overview of the classification and diagnosis of hematologic malignancies. In Chapters 20, 21, 22, 23, 24 we delve into the pathogenesis of these malignancies, emphasizing molecular insights that have already impacted the diagnosis and treatment. In these chapters we provide a more detailed look at the pathophysiology, clinical features, and treatment of the major types of hematologic malignancy, which include some of the most common cancers of children and adults.

The names and descriptive terms used for the various hematologic malignancies reflect their origin and usual clinical behavior. Tumors composed of cells of the myeloid series (granulocytes, red cells, platelets, and their progenitors) are referred to as myeloid, myelogenous, or myeloproliferative, whereas tumors composed of lymphocytes or their progenitors are variously termed lymphoid, lymphocytic, lymphoblastic, or lymphoproliferative. Leukemia (literally, white blood) is applied to neoplasms that typically involve the bone marrow and the peripheral blood, whereas lymphoma is used to describe lymphoid tumors that commonly present as masses within lymph nodes or other soft tissues. Some lymphomas are named based on the resemblance of the tumor cells to a normal counterpart; for example, the malignant B cells of follicular lymphoma closely resemble the normal cells found within the B-cell follicles (also commonly referred to as germinal centers) of lymph nodes. Other descriptors relate to the natural history of the disease in question. Acute leukemias, if untreated, are lethal within weeks to several months, whereas chronic leukemias may be compatible with survival for many years without treatment.

Classification systems are meant to provide a common language for the diagnosis and treatment of specific disease entities. As recently as 20 years ago, few areas of medicine were as contentious and confused as the classification of hematologic malignancies. Diagnoses were then based largely on clinical features and the morphologic appearance of tumor cells, criteria that are inadequate for classifying the diversity of hematologic malignancies that we now recognize molecularly.

In 1994, this picture began to change with the advent of classification that subdivided lymphoid tumors into distinct entities based on clinical features, microscopic appearance, and objective markers, including lineage-specific proteins and tumor-specific genetic aberrations. This proved so successful that the same approach was subsequently extended to myeloid neoplasms. There is now wide acceptance of a classification system, shown in Table 19-1, that sorts hematologic neoplasms based on their lineage and presumed cell of origin. The "cell of origin" concept is particularly important in the classification of lymphoid neoplasms, many of which are composed of cells that appear to be the malignant counterparts of some normal stage of B or T lymphocyte differentiation.

The neoplasms recognized in the World Health Organization classification of hematologic malignancies fall into five broad clinicopathologic categories, each containing a number of entities:

• Acute leukemias are tumors in which early myeloid or lymphoid progenitors (blasts) accumulate in the bone marrow and to varying degrees spill into the peripheral blood and infiltrate other tissues. They are believed to arise from early myeloid or lymphoid progenitors. Leukemic blasts are so primitive as to be nonfunctional and tend to displace or suppress the production of normal hematopoietic elements in the marrow. As a result, most patients present with symptoms related to pancytopenia.

• Chronic myeloproliferative disorders are tumors that arise in the bone marrow and most commonly lead to the increased production of one or more types of mature myeloid cells. They may arise from early myeloid progenitors or hematopoietic stem cells. Most symptoms initially stem from the hyperproliferation of bone marrow progenitors and increased numbers of red cells, granulocytes, and/or platelets in the peripheral blood and the spleen.

• Myelodysplastic syndromes are a poorly understood, heterogeneous group of myeloid tumors in which the maturation of bone marrow progenitors is abnormal (dysplastic) and ineffective. The cell of origin is believed to be an early myeloid progenitor. Patients present with symptoms related to one or more cytopenias and follow a highly variable course.

• Lymphomas and related lymphoid neoplasms are a large, diverse group of tumors derived from mature lymphocytes or their progenitors. These disorders are the most common types of hematologic malignancy. In the United States, roughly 100,000 people per year are newly diagnosed with one of these disorders. The lymphomas are divided into two broad classes: 1) Hodgkin lymphoma, a group of B-cell neoplasms linked by the presence of highly unusual tumor giant cells referred to as Reed-Sternberg cells, and 2) all of the other lymphomas, which are called the non-Hodgkin lymphomas. Non-Hodgkin lymphomas may be derived from cells at any stage of lymphocyte development, which explains in part their remarkable diversity. Overall, tumors of B-cell origin make up 85% to 90% of non-Hodgkin lymphomas in the United States, with most of the remaining tumors being of T-cell origin; natural killer cell lymphomas also occur, but are very rare. The variation in clinical behavior among these tumors is enormous, ranging from some of the most rapidly growing, aggressive tumors of man to neoplasms that are so indolent that for years they were referred to as pseudolymphomas. As will be discussed in the chapters covering specific lymphoid neoplasms, the clinical approach to these diseases is dictated in part by their behavior; aggressive lymphoid neoplasms must be treated as quickly as possible, whereas indolent tumors may wax and wane for years even in the absence of therapy.

• Plasma cell neoplasms and related entities are tumors composed at least in part of terminally differentiated B cells (plasma cells). These relatively common tumors occur mainly in older adults and often present with symptoms related to the production of complete or partial immunoglobulins by the tumor cells. The most important of these tumors are 1) multiple myeloma, which characteristically causes destructive bone lesions and 2) lymphoplasmacytic lymphoma, in which the tumor cells often produce immunoglobulin M (IgM), sometimes in quantities sufficient to produce a hyperviscosity state called Waldenström macroglobulinemia.

These categories provide a very helpful means for organizing the hematologic malignancies, but it must be recognized that the boundaries between specific entities sometimes blur. All of the tumors that are called lymphomas and even plasma cell tumors occasionally present as leukemias with peripheral blood involvement; conversely, virtually all of the tumors called leukemias may on occasion present as a tissue mass without involvement of the bone marrow or blood. This variation in clinical presentation is reflected in the names of some of the entities of the classification (Table 19-1). For example, tumors of immature lymphoid cells (lymphoblasts) may be referred to as acute lymphoblastic leukemia or lymphoblastic lymphoma depending on their clinical presentation; the same is true of a tumor of mature B cells that is variously referred to as chronic lymphocytic leukemia or small lymphocytic lymphoma. For the student, the important thing to remember is that these pairs of names refer to different clinical manifestations of the same tumor.

A second complication that must be appreciated is that some indolent forms of leukemia and lymphoma commonly transform into more aggressive forms of disease that fall into a different diagnostic category. In hematologic malignancies where the cell of origin is a hematopoietic stem cell (eg, chronic myelogenous leukemia), the transformed tumor may not only have a different morphology but also have an immunophenotype that is very different from the tumor that preceded it. Specific examples of transformation events will be discussed in subsequent chapters.

Diagnosis of each of the hematologic malignancies requires a tissue biopsy or examination of a peripheral blood smear. In leukemias, the diagnosis typically can be made by examination of a peripheral blood smear and/or bone marrow biopsy and aspirate. For plasma cell neoplasms, the diagnosis is usually based on a bone marrow biopsy, whereas lymphomas are most commonly diagnosed by biopsy of a lymph node or another soft tissue site.

At the time of the initial biopsy, it is usually unknown whether a patient has a malignancy or a benign disorder leading to hyperplasia of myeloid or lymphoid cells, such as an infection or an inflammatory condition. The key to distinguishing between benign and malignant processes is clonality. Neoplasms originate from a single transformed cell and are therefore monoclonal. In contrast, reactive processes stem from the response of many different cells and are polyclonal.

Sometimes all that is required to make the distinction between benign and neoplastic processes is a glance at the tissue under the microscope; for example, replacement of normal tissue by a monomorphous population of abnormal cells is a hallmark of certain neoplasms (Fig. 19-1A). However, the make-up of other hematologic tumors is quite heterogeneous, mimicking the appearance of an exuberant reactive process (Fig. 19-1B). In such instances, molecular tests of clonality (described in a following section) play a critical role in establishing the diagnosis of a hematologic malignancy.

The mainstay of diagnosis remains the microscopic examination of tumor cells in smears or tissue sections, but morphology alone is insufficient to accurately diagnose and classify most hematologic malignancies. Thus, tumor samples are routinely examined by using several of the following complementary approaches:

• Morphology. The appearance of tumor cells usually allows a diagnostician (typically a pathologist with training in hematopathology) to generate a short list of diagnostic possibilities. For example, the presence of blasts containing Auer rods, abnormal needle-like azurophilic granules (Fig. 19-2A), is diagnostic of an acute myeloid leukemia and would trigger the performance of other tests needed to further classify this type of tumor. Similarly, replacement of the marrow by an abnormal population of plasma cells would strongly support the diagnosis of multiple myeloma (Fig. 19-2B). However, it is often impossible to determine the lineage of a hematologic malignancy with certainty by morphology. For example, many B-cell and T-cell lymphomas cannot be reliably distinguished by appearance alone, and (as mentioned earlier) it may not even be possible to tell a reactive hyperplasia from a malignancy without additional tests.

• Immunophenotyping refers to the staining of tumor cells with antibodies specific for antigens (mostly proteins) that are useful in the classification of hematologic malignancies. Some of the antigens that are particularly useful are summarized in Table 19-2. When fresh, unfixed tumor cells are available, antigens are often detected by flow cytometry (described in Chapter 1). In flow cytometry, cells are identified and enumerated based on their light scattering properties and their expression of various antigens, which are detected with antibodies tagged with fluorescent dyes. Flow cytometry is widely used in the diagnosis of acute leukemias and non-Hodgkin lymphomas. It is often also used to establish the clonality of mature B-cell tumors (Fig. 19-3). Mature B cells express only one type of immunoglobulin (Ig), which is composed of two heavy chains paired with either kappa or lambda light chains. In polyclonal B-cell proliferations, the ratio of kappa- to lambda-expressing cells is usually close to 1:1. However, because malignancies are derived from single transformed cells, in mature B-cell tumors, all of the cells express only kappa or lambda light chain. Due to its ability to quantify multiple markers on single cells, flow cytometry also can be used to detect cells expressing unusual combinations of antigens (another hallmark of certain hematologic malignancies), even within large populations of normal cells. Its chief limitation is that it must be done on fresh, unfixed cells, which are available only if a hematologic malignancy is suspected at the time the biopsy is obtained. A second method that avoids this limitation is immunohistochemistry (Fig. 19-4). Here, sections of tissues fixed in formalin and embedded in paraffin (the standard method of handling tissues in pathology departments) are placed on slides and incubated with specific antibodies. Staining is then developed by incubation with secondary antibodies linked to an enzyme such as horseradish peroxidase and a substrate that is converted by the enzyme to an insoluble colored precipitate. This method is widely used in the diagnosis of hematologic malignancies, particularly non-Hodgkin and Hodgkin lymphomas.

• Histochemistry is used mainly in the diagnosis of acute leukemias. It is performed by incubating smears of marrow or blood with chemicals that differentially stain bone marrow cells of various lineages. Some stains involve the use of dyes that change color in the presence of specific leukocyte enzymes. Examples include stains for myeloperoxidase and specific esterase (both expressed by myeloblasts and more mature granulocytes) and nonspecific esterase (expressed by monoblasts and monocytes). Other stains react with non-enzymatic cellular constituents. The periodic acid–Schiff stain is used to detect glycogen, which is frequently present in the lymphoblasts of acute lymphoblastic leukemias and occasionally found in erythroblasts in myeloid neoplasms. The Prussian blue stain is used to detect non-heme iron, which accumulates in the mitochondria of red cell progenitors in certain myelodysplastic syndromes.

• Cytogenetics. Tumor-specific cytogenetic aberrations can serve to establish the clonality of myeloid or lymphoid proliferations and in some instances are so characteristic of particular hematologic malignancies that they are required for the diagnosis. For example, all cases of chronic myelogenous leukemia are associated with chromosomal rearrangements that juxtapose the 3' portion of the ABL gene (normally on chromosome 9) with the 5' portion of the BCR gene on chromosome 22, thereby creating a chimeric BCR-ABL fusion gene with oncogenic activity. In most instances, this event occurs by way of a reciprocal translocation between chromosomes 9 and 22 that causes the formation of two abnormal chromosomes, one derived from chromosome 9 and the other from chromosome 22. The derivative chromosome 22 that bears the BCR-ABL fusion gene is often referred to as the Philadelphia chromosome (named for the city of its discovery). When fresh samples of marrow or blood are available, the Philadelphia chromosome usually can be directly visualized by determining the karyotype of tumor cells harvested in metaphase, the time in mitosis when chromosomes condense and can be recognized by their appearance (Fig. 19-5A). However, sometimes metaphases are not obtained, or complex rearrangements involving chromosomes 9 and 22 are present. In such instances, the BCR-ABL fusion gene and its reciprocal partner, an ABL-BCR fusion gene, can be detected directly by fluorescence in situ hybridization (FISH), in which DNA probes spanning the BCR and ABL loci are hybridized to interphase nuclei from tumor cells. In the example shown (Fig. 19-5B), the reciprocal rearrangement of BCR and ABL causes two abnormal fusion signals to appear, one due to the presence of the BCR-ABL gene on chromosome 22 and the second due to the nonfunctional ABL-BCR fusion gene on chromosome 9.

• Molecular genetics. Sequence analysis of DNA or RNA isolated from tumor cells is playing an increasing part in the diagnosis of hematologic malignancies. For example, the diagnosis of polycythemia vera, a type of myeloproliferative disorder, requires the detection of activating mutations in the gene encoding the tyrosine kinase JAK2. Another common use of molecular -genetics is the assessment of clonality in lymphoid proliferations. Rearrangement and assembly of the immunoglobulin genes in B cells and the T-cell receptor genes in T cells create unique DNA sequences that are specific for that cell and all its progeny. Polyclonal, reactive lymphoid proliferations are composed of cells with many different sets of rearranged antigen receptor genes, whereas in lymphoid neoplasms all tumor cells share the same clonal rearrangements. Polymerase chain reaction (PCR)-based tests that involve the amplification of rearranged antigen receptor genes can -distinguish between polyclonal and monoclonal lymphoid proliferations with a high degree of sensitivity and specificity and are now in wide clinical use (Fig. 19-6). PCR analysis of reverse-transcribed messenger RNA (mRNA) from chronic myelogenous leukemia cells offers a reliable and sensitive means of detecting the BCR-ABL transcript.

In addition to these tissue-based tests, other clinical and laboratory test findings weigh in on the diagnosis of certain hematologic malignancies. For example, tests that detect and quantify monoclonal antibodies or fragments of antibodies within the serum and the urine are important adjuncts in the diagnosis of plasma cell neoplasms, and radiologic studies to identify destructive bone lesions are part of the work-up of patients suspected of having multiple myeloma. The diagnosis of T-cell lymphoma is based in part on the pattern of tissue involvement, because specific subtypes of these diseases preferentially involve the skin, adipose tissue, gut, or spleen. It is believed that this behavior recapitulates the ability of normal T lymphocytes to home to specific tissues.