Hematopoietic Tumors

Definition

Plasma cell myeloma is a monoclonal, neoplastic proliferation of plasma cells that involves the bone marrow and occasionally involves extraskeletal sites. Lytic bone lesions are a prominent feature of typical cases.

Clinical Features

Plasma cell myeloma is one of the most frequently occurring hematopoietic neoplasms and accounts for approximately 1% of all malignant neoplasms in white patients and approximately 2% in black patients. It is the second most common hematopoietic malignancy, accounting for approximately 10% of hematopoietic neoplasms, and the most frequent neoplasm presenting as skeletal lesions. The skull, vertebral bodies, pelvis, and proximal parts of the major long tubular bones are most frequently involved ( Fig. 12-2 ). More than 95% of cases are diagnosed in patients older than age 40 years, with a peak incidence between ages 65 and 74 years ( Fig. 12-3 ). Plasma cell myeloma is more common in the black population ( Fig. 12-4 ). Plasma cell myeloma is always preceded by monoclonal gammopathy of undetermined significance (MGUS), but MGUS may not be detected in this early subclinical phase of disease. Patients with MGUS progress to plasma cell myeloma at the rate of approximately 1% per year. The incidence of plasma cell myeloma is increasing, and there is currently no definitive cure; however, with greater understanding of the pathogenesis, patient survival is improving.

Plasma cell myeloma usually presents with multifocal osteolytic lesions, proliferation of plasma cells, and monoclonal gammopathy. Typically, patients have multifocal bone pain (especially in the weight-bearing sites), anemia, hypercalcemia, renal failure, proteinuria, and a history of recurrent infections. In the majority of patients, increased levels of monoclonal immunoglobulins (M-protein) can be detected by serum or urine electrophoresis. Recommended laboratory testing includes complete blood count, serum creatinine, serum calcium, serum protein electrophoresis with immunofixation, serum free light chain assay, and urine protein analysis ( Table 12-1 ).

Diagnostic criteria used to subclassify plasma cell neoplasms include laboratory and clinical features. MGUS is defined as the presence of an M-protein at levels below 3 g/dL in the absence of any other defining features of plasma cell myeloma. Two broad categories of plasma cell myeloma are designated smoldering (asymptomatic) myeloma and symptomatic myeloma . By the 2008 WHO criteria, a diagnosis of asymptomatic plasma cell myeloma may be made if there are more than 10% clonal plasma cells in the bone marrow or the serum M-protein is greater than 3 g/dL. A diagnosis of symptomatic plasma cell myeloma requires the presence of a clonal plasma cell population of any quantity (usually with associated detectable M-protein) and end-organ damage, as represented by the acronym CRAB (hyper C alcemia, R enal insufficiency, A nemia, lytic B one lesions) ( Table 12-2 ).

The main prognostic factors are related to cytogenetic findings, extent of disease, patient age, and renal function. Factors that can be evaluated in the laboratory include cytogenetic status, proliferation index, neoplastic plasma cells as a percentage of total plasma cells, free light chain ratio, serum lactate dehydrogenase, and β ₂ -microglobulin level.

With current therapeutic approaches, the progression-free survival and overall survival for patients with plasma cell myeloma has improved ( Fig. 12-5 ). Before 1962, in the absence of any known treatment, the median survival for newly diagnosed myeloma patients was less than 1 year. Although there has been steady improvement in survival over the years, median overall survival in the early 2000s was still only 2 to 3 years. In the current era of novel therapeutic agents, overall median survival is now approximately 8 years.

Patient prognosis and choice of therapy are most commonly determined by cytogenetic risk factors and age. An example of risk stratification as delineated by the Mayo group identifies high risk, intermediate risk, and standard risk groups based on cytogenetic features or gene expression profiling in the context of response to currently available therapies. High risk features are deletion of 17p, t(14;16), t(14;20), or a high risk signature by gene expression profiling. Intermediate risk features are t(4;14), deletion of 13, hypodiploidy, and a plasma cell labeling index greater than 2. Standard risk patients have none of the preceding features and may have t(11;14) or t (6;14). The current median overall survival for high risk, intermediate risk, and standard risk is 3 years, 4 to 5 years, and 8 to 10 years, respectively.

The most widely used therapeutic regimens include dexamethasone or prednisone and one or more novel agents, in some cases augmented with alkylating agents or anthracyclines. Patients under age 70 years are potentially eligible for autologous stem cell transplant, and currently cytoreductive therapy followed by transplant is the standard of care; however, treatment with novel agents may replace this approach in the future. Commonly used novel agents are immunomodulatory drugs such as lenolidamide or thalidomide, and the proteasome inhibitor bortezomib. Commonly used chemotherapeutic drugs include melphalan and cyclophosphamide.

Radiographic Imaging

The International Myeloma Working Group (IMWG) consensus guidelines for imaging studies were last updated in 2009. A skeletal survey, consisting of a series of plain radiographs, is recommended for every patient. If the skeletal survey detects no lesions, or if the initial diagnosis is solitary plasmacytoma of bone, magnetic resonance imaging (MRI) should be performed to detect possible occult bone lesions. Computed tomography (CT) scan or MRI are required for any patient with suspected spinal cord compression. Positron emission tomography (PET) scan is not routinely recommended.

Distinctive radiographically detectable changes are present in approximately 80% of patients with multiple myeloma. The axial skeleton and the trunk bones are preferentially involved. The most pronounced and earliest changes are typically seen in the skull, vertebrae, ribs, and pelvis and correspond to the predominance of hematopoietic marrow at these sites in patients older than age 50 years. They represent multifocal, sharply demarcated, lytic foci and are often called punched-out lesions ( Figs. 12-6 to 12-8 ). Erosions of the cortex are frequently seen, and prominent periosteal new bone formation is typically not present. Bones with a smaller diameter, such as the ribs, can exhibit expanded contours. Pathologic fractures are frequent and can be the presenting sign. In fact, collapse of the vertebral body is frequently a presenting symptom ( Fig. 12-9 ). The disease typically first appears in the axial skeleton and trunk bones, but with progression of the process, extensive involvement of the appendicular skeleton and multiple pathologic fractures can develop. Surprisingly, despite extensive bone destruction, the results of radioisotopic bone scans are frequently negative in patients with multiple myeloma. This is explained by the predominant bone destructive activity of myeloma cells with typically minimal new bone production. A small number of patients do not have lytic punched-out lesions at presentation and may show generalized osteoporosis that is sometimes quite pronounced. In some instances, the lesions in multiple myeloma may appear sclerotic on radiographs ( Figs. 12-10 and 12-11 ). Sclerotic lesions are typical for POEMS syndrome (discussed in the section on osteosclerotic myeloma in this chapter) .

Gross Findings

Antemortem gross examination is typically restricted to small biopsy or curettage samples from the site of pathologic fracture and show fragments of tan-gray soft tissue. At autopsy the principal growth patterns can be appreciated: diffuse involvement of bone marrow and growth in the form of discrete nodules ( Fig. 12-12 ). Most likely, the diffuse involvement represents a more advanced stage and is produced by the confluence of individual nodules. On the other hand, some patients have generalized osteoporosis at presentation, and nodular lytic changes develop later in the course of the disease. The individual nodules are best seen in the vertebral bodies. A more advanced process is associated with the collapse of one or several vertebral bodies. The ribs typically show multiple extended foci. The long bones in the appendicular skeleton show nodules or diffuse involvement with cortical thinning and disruption. Pathologic fractures are frequent, especially at the weight-bearing sites. The proximal parts of the appendicular skeleton are initially preferentially involved.

Microscopic Findings

Plasma cell myeloma consists of a proliferation of cells that usually include at least a subset of cells that resemble normal plasma cells. The cells form clusters or sheets within the bone marrow of involved tissue. These cells are round or oval and have an eccentric nucleus and a clumped “clockface” chromatin pattern with the chromatin clustered at the periphery of the nucleus near the nuclear membrane. The plasma cell cytoplasm has a prominent perinuclear hof, corresponding ultrastructurally to a prominent Golgi region, and rough endoplastic reticulum appropriate for active immunoglobulin production and packaging.

When multiple myeloma is suspected, smears should always be prepared because the plasma cell nature of the round-cell infiltrate is quite often more evident in cytologic preparations than in conventional histologic sections. The cytoplasm is dense and eosinophilic in hematoxylin-eosin–stained sections, and deeply basophilic in Wright-Giemsa or Diff-Quik (Romanovsky's) stained smears. As described by Bartl et al in 1987, the plasma cells may vary considerably in cytomorphology ( Fig. 12-13 ). The morphologic range includes cells with polymorphous or asynchronous features, as well as forms resembling normal mature plasma cells, small lymphocytes, blasts, or lymphocytes with cleaved nuclei. These morphologic variants may pose diagnostic challenges, raising the differential diagnosis of lymphoma, leukemia or other blastic neoplasms, or even carcinoma in cases with marked pleomorphism.

The degree of immaturity of the myeloma cells is a prognostic factor. Loss of the morphologic plasma cell phenotype with the presence of forms resembling blasts is associated with more aggressive behavior and a poorer prognosis. In most myelomas, clearly recognizable features of plasma cells can be found at least focally.

Numerous Russell bodies representing cytoplasmic spherical structures of polymerized immunoglobulins can be present. The accumulation of cytoplasmic immunoglobulins can be present in the form of morular or Mott cells, representing grapelike cytoplasmic structures. These structures are best seen in Giemsa stained cytologic preparations. Although Russell bodies are frequently seen in reactive plasma cells, intranuclear inclusions, or Dutcher bodies, almost always indicate neoplastic plasma cells ( Fig. 12-14 ). The cytoplasm of myeloma cells has ultrastructural features of maturity with well-developed rough endoplasmic reticulum and a prominent Golgi region ( Fig. 12-16 ). The latter can be seen by light microscopy as a perinuclear clearing.

Immunohistochemistry and Differential Diagnosis

Plasma cells represent the end stage of B-cell differentiation, but they typically do not express common pan–B-cell antigen CD20. They are also typically negative for most T-cell markers. A summary of microscopic ultrastructural and immunophenotypic features of myeloma cells is provided in Table 12-3 . Myeloma cells are of the plasma cell lineage and have many of the same features. The most characteristic feature of myeloma cells is their monotypic expression of κ or λ immunoglobulin. They typically brightly express normal plasma cell markers CD38, CD138, and MUM1. In the majority of cases, the cells show aberrant expression of CD56, an adhesion molecule, which may be lost in plasma cell leukemia and extramedullary plasmacytomas. The second most common aberrantly expressed marker is CD117.

The differential diagnosis includes reactive plasmacytosis, reactive inflammatory conditions with a prominent plasma cell component, and other neoplasms. Normal bone marrow plasma cells are located primarily in a perivascular distribution, in contrast to myeloma plasma cells, which form sheets and clusters in a predominantly interstitial pattern. Binucleate plasma cells without nuclear atypia can frequently be found in reactive infiltrations. Immunohistochemical stains or in situ hybridization for immunoglobulin κ and λ are very helpful in distinguishing reactive polytypic plasma cell infiltrates from plasma cell myeloma. The normal κ-to-λ ratio is approximately 2 or 3 : 1. Increased polytypic plasma cells may be seen in a number of reactive conditions, including viral and bacterial infections, autoimmune disease, cirrhosis, and solid tumors.

The presence of other inflammatory cells together with associated fibrosis and a prominent vasculature favors a reactive process. The clinical and radiographic data in such cases with the absence of bone marrow plasmacytosis, a lack of multifocal skeletal changes, and the absence of the M component are additional general clinical, radiologic, and laboratory features that support the diagnosis of a benign disorder.

An unusual pattern of immunoproliferation, including numerous polyclonal plasma cells and reactive immunoblasts, may mimic plasma cell myeloma. This proliferation has been termed systemic polyclonal immunoblastic proliferation and may present with systemic symptoms and plasma cell infiltrates mimicking plasma cell myeloma. The proliferation involves blood and bone marrow with variable involvement of other organs and tissues. The infiltrates may be composed by up to 50% of plasma cells. The polyclonal nature of the plasma cells may be demonstrated by immunohistochemical stains, in situ hybridization, or polymerase chain reaction (PCR) studies for immunoglobulin clonality.

Neuroendocrine tumors and melanoma may have plasmacytoid cytomorphology but can be differentiated from plasma cell myeloma by staining for neuroendocrine markers such as synaptophysin and chromogranin (CD56 will not resolve this diagnostic dilemma). The majority of melanomas express MUM1, and a subset of melanoma cases may show variable expression of CD138. Most plasma cell myelomas are strongly and uniformly positive for CD138 and MUM1, while the staining is typically more variable in melanoma. Κ and λ immunohistochemical stains will show monotypic light chain expression in the majority of plasma cell myeloma cases, whereas melanoma is negative for light chain expression and positive for S-100 in the majority of cases. HMB-45 or Melan-A may stain unusual S-100– negative cases of melanoma.

Pleomorphic variants of plasma cell myeloma may resemble carcinoma or lymphoma histologically. Care should be taken in selecting and interpreting immunohistochemical stains in tumors with pleomorphic or anaplastic morphology. Several types of carcinomas may show variable expression of CD138 and MUM1. Again, staining for CD138 and MUM1 is typically uniform in plasma cell myeloma and more variable in carcinomas. Κ and λ immunohistochemical stains will show monotypic light chain expression in the majority of plasma cell myeloma cases. Although carcinoma and plasma cell myeloma may be positive for EMA, a variety of cytokeratin stains will be positive in the majority of carcinomas and negative in plasma cell myeloma. Most T-cell lymphomas will express some combination of T-cell lineage markers (CD2, CD3, CD4, CD5, CD7, CD8), which are absent in plasma cell myeloma. Interpretation of B-cell lineage markers is more problematic, because the majority of markers expressed by plasma cell myeloma may also be expressed in various subtypes of B-cell lymphomas. Plasma cell myeloma is typically negative for CD20, but a subset of myelomas, including the small lymphocyte-like variant, may express CD20 and PAX5 ( Fig. 12-15 ). In difficult cases, staining for CD56, CD117, or uniform strong staining for cyclin D1 support a diagnosis of plasma cell myeloma. Clinical workup for myeloma may help distinguish B-cell lymphoma from plasma cell myeloma in cases with overlapping features.

Finally, the small lymphocyte-like variant of plasma cell myeloma has many morphologic and immunohistochemical similarities with mature B-cell lymphomas with small cell cytomorphology. In particular, this variant may be misdiagnosed as mantle cell lymphoma on the basis of staining with CD20 and cyclin D1 ( Fig. 12-15 ). Mantle cell lymphoma will be positive for CD5 in the majority of cases, and cyclin D1, while positive in the majority of cells, will show variable intensity of staining from cell to cell. Plasma cell myeloma will be negative for CD5 and typically shows strong homogeneous staining of the myeloma cells.

Genetic Features and Pathogenesis

The normal cell counterpart is a postgerminal center plasma cell with clonal immunoglobulin gene rearrangements and somatic hypermutation of immunoglobulin genes, typically secreting antibody. Following early molecular events, the abnormal plasma cells migrate to the bone marrow, where the cells are supported by their interactions with constituents of the marrow microenvironment. Subclones with slightly different mutations compete for survival within the marrow microenvironment. This stage of disease represents MGUS. Random mutations result in progression of disease and clonal evolution of multiple subclones. These subclones undergo natural selection, resulting in survival and proliferation of the best adapted clone or clones, at this stage representing plasma cell myeloma and in some cases plasma cell leukemia ( Fig. 12-17 ).

Plasma cell myeloma can be broadly divided into two major molecular categories: hyperdiploid and nonhyperdiploid. The majority of both hyperdiploid and nonhyperdiploid cases show features of dysregulation of D cyclins, driving cell growth. Overall, hyperdiploid cases have a better prognosis. Hyperdiploidy in plasma cell myeloma is the result of trisomies of the odd-numbered chromosomes 3, 5, 7, 9, 11, 15, 19, and 21. Nonhyperdiploid cases are associated with reciprocal translocations involving immunoglobulin genes. Hyperdiploidy and immunoglobin gene translocations represent early molecular events in plasma cell neoplasia, and the majority of cases of MGUS harbor one of these abnormalities. In addition to these major molecular categories, inherited single nucleotide polymorphisms and monosomy or partial deletion of chromosome 13 are recognized as initiating events ( Table 12-4 and Fig. 12-18 ).

Early chromosomal translocations in MGUS and plasma cell myeloma most frequently involve the immunoglobulin heavy chain gene IGH @ on chromosome 14, and less frequently involve lambda light chain gene IGL @ on chromosome 22. These translocations are thought to arise from aberrant class switch recombinations or somatic hypermutations mediated by activation-induced deaminase (AID). As a result of these translocations, expression of partner genes is placed under the control of the immunoglobulin gene enhancer, with associated enhanced expression of the partner gene product. The most frequent gene partners in these translocations are CCND1, CCND3, MMSET, FGFR3, MAF, MAFB, and MMSET/FGFR ( Table 12-4 and Figs. 12-19 and 12-20 ). CCND genes encode D cyclins involved in cell cycle progression. Overall, these translocations are associated with worse prognosis, with the exception of t(11;14) IGH@/CCND1 ( Fig. 12-19 ), which confers a better prognosis.

Mutations associated with disease progression and proliferation are additional genetic gains and losses, secondary translocations involving immunoglobulin genes, translocations involving MYC , epigenetic changes, and mutations resulting in dysregulation of various signaling pathways ( Table 12-4 and Fig. 12-18 ). NFκB is an example of a key pathway that is constitutively activated as a result of both activating and inactivating mutations of the regulators of this pathway. NFκB activation is important for cell survival.

Genome sequencing of 38 cases of plasma cell myeloma detected 23 translocations and 35 point mutations that resulted in an amino acid changes. These mutations were detected in several classes of genes with various functions and included genes known to be mutated in plasma cell myeloma as well as several genes in which mutations had not previously been detected. These mutations included oncogenes and genes involved in RNA processing, protein homeostasis, histone modification, and cell signaling. Frequently mutated genes included TP53, NRAS, KRAS, and genes involved in the NFκB signaling pathway. Although not common, mutations of BRAF were also detected in 4% of patients.

Several microRNAs are upregulated or downregulated in myeloma and other neoplasms. For example, microRNAs 15a and 16 are downregulated in myeloma. When present at normal levels, microRNAs 15a and 16 are thought to decrease BCL-2 expression, inhibit the NFkB pathway, and inhibit angiogenesis. Levels of circulating microRNAs have also been shown to have prognostic and theragnostic significance in myeloma.

In addition to being secreted, microRNAs, cytokines, and other proteins can also be directly transferred from bone marrow stromal cells and internalized by myeloma cells via exosomes, small membrane bound vesicles. In particular, a recent study found that bone marrow stromal cells in plasma cell myeloma produced exosomes containing increased interleukin-6 (IL-6) and decreased microRNA15a relative to normal bone marrow stromal cells. The study demonstrated transfer of these altered exosomes from the mesenchymal cells to the myeloma cells. The mechanism by which microRNA 15a is decreased is unknown; however, a recent sequencing study detected mutations of DIS3 , an exosome-based RNase, in approximately 10% of myeloma samples.

The neoplastic plasma cells in plasma cell myeloma share complex interactions with constituents of the bone marrow microenvironment, including bone marrow matrix, stromal cells, endothelial cells, T cells, dendritic cells, and osteoclasts and osteoblasts and their precursors ( Fig. 12-21 ). These interactions can be categorized as adhesion-mediated and cytokine-mediated.

Adhesion molecules play a role in homing and localization of myeloma cells to the bone marrow and mediate the direct cell interactions between myeloma cells and the constituents of the bone marrow microenvironment. Two adhesion molecules expressed by myeloma cells are syndecan 1 (CD138) and neural cell adhesion molecule (CD56). Myeloma cells that are bound to bone marrow constituents demonstrate resistance to chemotherapy. The binding of plasma cells to the stromal cells via adhesion molecules results in upregulation of the secretion of multiple cytokines that enhance plasma cell proliferation and survival, such as IL-6, insulin growth factor-1, B-cell activating factor, and a proliferation inducing ligand.

IL-6 is one of the key cytokines produced by both myeloma cells and the cells of the bone marrow microenvironment. IL-6 promotes proliferation and blocks apoptosis of myeloma cells. IL-6 is secreted by bone marrow stromal cells, osteoclasts, and osteoblasts. The normal homeostasis of bone formation and resorption is disrupted in plasma cell myeloma and is mediated by cytokines such as macrophage inflammatory protein-1α (MIP-1α) produced by plasma cells and receptor activator of nuclear factor κB ligand (RANKL) produced by marrow stromal cells. Increased MIP-1α and RANKL promote osteoclast differentiation and activation. Increased Dickkopf-related protein 1 (DKK1) produced by myeloma cells and decreased osteoprotegerin result in decreased osteoblast differentiation and activation. This disruption results in net loss of bone and the characteristic osteolytic lesions seen in plasma cell myeloma.

Increased vessel density supports tumor growth and is associated with poor prognosis in plasma cell myeloma. Vascular endothelial growth factor (VEGF) is produced by myeloma cells, marrow stromal cells, and vascular endothelial cells. VEGF promotes vascular proliferation and also upregulates secretion of IL-6 from marrow stromal cells. VEGF secretion is upregulated by MYC expression and hypoxia inducible factor-1α (HIF-1α).

Definition

Primary non-Hodgkin's lymphoma of bone is a proliferation of malignant lymphocytes classified according to its extraskeletal counterparts. Although there is no universal definition, by convention, an interval of 4 to 6 months between skeletal manifestation of the lesion and the development of extraskeletal disease is required for the lesion to be considered a primary tumor in bone. In every case of lymphoma presenting as a bone lesion, the appropriate staging procedure is required to rule out the presence of extraskeletal disease.

Historically, the description of lymphoma as a primary lesion in bone is credited to Oberling, although he did not make the distinction between true lymphomas and Ewing's sarcoma. The recognition of primary lymphoma of bone as a clinicopathologic entity distinct from other lymphomas and especially from Ewing's sarcoma was made by Parker and Jackson in 1939. Two parallel classifications based on the functional and phenotypic features of tumor cells as delineated by marker studies were proposed by Lukes and Collins in the United States and by Lennert in Germany (Kiel classification) in 1975 and 1974, respectively. The simplified Revised Working Formulation was published in 1982 in an effort to bridge the gap between morphologic and clinical aspects of lymphoma classification, combining some previous descriptive terms, mainly from the Rappaport system, with the biologic concepts and terminology proposed by Lukes and Collins.

The Revised European-American Classification System of Lymphoid Neoplasms (REAL Classification) proposed by the International Lymphoma Study Group was published in 1994 and represented a consensus effort to define the known distinct hematolymphoid neoplasms in terms of clinical, morphologic, immunophenotypic, and molecular features. The REAL classification was the starting point for 7 years of efforts to build the now generally accepted consensus classification, the World Health Organization (WHO) Classification of Tumours of Haematopoietic and Lymphoid Tissues.

Clinical Features

Primary lymphoma of bone is rare and accounts for 3% to 7% of bone tumors and less than 2% of adult lymphomas overall. Because of the infrequency of primary bone lymphoma and the disagreement as to its clinical definition, there have been few published series documenting patient outcomes, clinical characteristics, and biological aspects. The following discussion focuses on diffuse large B-cell lymphoma and is based primarily on three published large series of primary lymphoma of bone including several histologic subtypes; however, in these series, at least 70%, 79%, and 83% of the cases were classified as diffuse large B-cell lymphoma.

The peak age incidence and most frequent sites of primary bone lymphoma are shown in Figure 12-23 . The incidence of primary bone lymphoma is rising ( Fig. 12-24 ). More than 50% of primary lymphomas in bone occur in patients older than age 60 years ( Fig. 12-25 ) There is a slight male predominance in several major series, with a male-to-female ratio of approximately 1.2 : 1 ( Fig. 12-26 ). The vertebrae, femur, and pelvis are the most frequently involved bones, accounting for approximately 29%, 12%, and 13% of lesions, respectively. The humerus, ribs, and skull are next in frequency, and each is involved in approximately 10% of cases. In general, non-Hodgkin's lymphomas have a predilection for the trunk bones, including the ribs, sternum, and clavicle. The major long tubular bones, such as the femur and humerus, are the most frequently involved sites in the appendicular skeleton. Lymphomas of bone often present as multifocal lesions. The multifocality can be in the form of several foci within one bone, or several bones can be simultaneously affected. The rate of multifocal lesions is approximately 15%.

Pain is the most common symptom. Other symptoms, such as nerve compression, are related to the location of the tumor. Local tenderness, redness, and swelling may be present. Pathologic fracture can be the initial symptom. In contrast with symptoms of Ewing's sarcoma and generalized lymphoma, fever is uncommon. Features of generalized systemic disease, such as lymphadenopathy and hepatosplenomegaly, are not present at presentation by definition.

Radiation therapy and chemotherapy are the primary treatments for non-Hodgkin's lymphoma of bone. There is no consensus as to the optimal therapy for localized bone involvement. Conflicting results have been published, with no clear advantage to radiation versus chemotheraphy alone versus combined chemotherapy and radiation therapy.

The most important prognostic factors for primary lymphoma of bone are age at diagnosis and stage of disease. Patients under age 30 years have high rates of remission, with approximately 75% survival at 400 months versus 0% survival in patients age 60 years or older. Patients with lymphoma limited to the bone have better survival than systemic lymphoma. The 5- and 10-year survival rates for primary bone lymphoma are 65% and 53%, compared with 53% and 43% for patients with systemic lymphoma in addition to bone masses.

Radiographic Imaging

Radiographic features of non-Hodgkin's lymphoma are not specific and overlap with other small cell tumors. They usually do not permit a specific diagnosis. Typically, non-Hodgkin's lymphoma presents as a lytic destructive lesion with a permeative or “moth-eaten” pattern ( Fig. 12-27 ). Prominent periosteal new bone formation with multiple concentric or so-called onion-skin layers may be present but is usually less evident than in Ewing's sarcoma. In long bones, the shaft is preferentially involved, and disease may extend to the end of the bone. A lytic lesion with periosteal reaction at or near the end of a long bone, while not specific, raises the index of suspicion for lymphoma ( Fig. 12-28 ). Non-Hodgkin's lymphoma is usually lytic, but in some cases it provokes bone sclerosis and presents as a blastic lesion. Rarely, plain radiographs may show little or no abnormality.

Computed tomography or MRI are helpful in evaluating the extent of involvement within bone, to detect cortical erosion and soft tissue extension of tumor, as well as to evaluate the extent of involvement of vertebral and paraspinous lesions. MRI is particularly useful for detecting marrow involvement, which is characteristic in bone lymphoma. Bone marrow involvement may be detected by T1-weighted (low signal) or T2-weighted (bright signal) magnetic resonance images. A clue to the diagnosis is prominent marrow and soft tissue involvement with relatively minor cortical bone involvement.

Radionuclide scintigraphy is helpful in documenting additional foci of bone involvement, a common feature of non-Hodgkin's lymphomas in the skeleton. FDG PET/CT is useful in staging and confirming disease limited to the bone.

Gross Findings

On gross examination, diffuse large B-cell lymphoma of bone resembles its nodal counterpart and presents as a pink-tan or gray-white and fleshy lesion ( Fig. 12-29 ). Areas of necrosis, hemorrhage, and cystic degeneration may be present. The cortex and the bone at its periphery show patchy erosions and permeation of the medullary cavity. Complete cortical disruption and extension into soft tissue are frequently present. Large masses may show sharp demarcation at their peripheries. Reactive sclerosis presents as firm ivory-like areas and may occasionally occupy a significant portion of the tumor.

Microscopic Findings

The conventional diffuse growth pattern with solid proliferation of round tumor cells permeating the bone marrow spaces and haversian canals is most frequently seen. Sclerosing diffuse growth is often focally present in bone. In this pattern, the tumor cells are separated by dense fibrous bands and can form nests, cords, or organoid structures. Occasionally, they may have spindle cell morphology and grow in storiform structures ( Fig. 12-30 ). Compressed tumor cells may also develop a spindled morphology. The presence of reactive lymphocytes can alter the appearance of these lesions. Deviations from the classic lymphoma morphology can lead to an erroneous diagnosis of an inflammatory condition or nonlymphoid tumor, as in the case of the rare signet ring cell variant of diffuse large B-cell lymphoma ( Fig. 12-30 ).

Immunohistochemical Stains and Differential Diagnosis

A variety of special studies may be used to supplement the diagnosis of lymphoma of bone, including flow cytometry, classical cytogenetics, fluorescence in situ hybridization (FISH) studies for common translocations, PCR for detection of translocations, and clonal immunoglobulin gene rearrangement. Gene expression profiling is useful for subclassifying diffuse large B-cell lymphoma but is still not widely used in clinical practice. Exome or whole genome sequencing is becoming more widely available and likely will be one of the major tools used to identify actionable genetic alterations, allowing design and application of regimens, including specific targeted therapeutic agents.

Immunohistochemistry remains a useful tool in the diagnosis and subclassification of diffuse large B-cell lymphoma. Lymphoma must be distinguished from other hematopoietic neoplasms, such as plasma cell myeloma, myeloid sarcoma, Langerhans cell histiocytosis, and mastocytosis . The differential diagnosis also includes nonhematologic round-cell tumors such as Ewing's sarcoma, rhabdomyosarcoma, and metastatic small cell and poorly differentiated non–small cell carcinoma. The use of nonlymphoma markers, such as cytokeratins, desmin, smooth muscle actin, and FLI1 (Ewing's sarcoma), can help rule out other round-cell tumors.

A minimalist approach to suspected diffuse large B-cell lymphoma is initial evaluation of CD20 immunohistochemical stain; however, more often a panel including CD45 (leukocyte common antigen), CD20 (pan–B-cell marker), and CD3 (T-cell marker) is ordered, in addition to other lineage markers if the differential diagnosis is broad. The majority of untreated diffuse large B-cell lymphomas are positive for CD20. In cases of CD20-negative diffuse large B-cell lymphoma, PAX5 or CD79 will usually be positive. Once a diagnosis of diffuse large B-cell lymphoma is established, additional prognostic studies may be considered depending on the needs of the patient and treating physicians. A typical follow-up panel includes Ki-67 to evaluate proliferation rate, and CD10, BCL6, and MUM1 (with or without additional markers) to subclassify diffuse large B-cell lymphoma as either germinal center subtype or activated B-cell subtype (also see the discussion on genetic features and pathogenesis) ( Figs. 12-31 and 12-32 ). In addition, Epstein-Barr encoding region (EBER) in situ hybridization may be considered in patients older than age 50 years to confirm or exclude a diagnosis of Epstein-Barr virus (EBV)-positive diffuse large B-cell lymphoma of the elderly.

If the Ki-67 proliferation index is above 80%, some recommend FISH testing for translocations involving BCL-2, BCL-6, and MYC, to identify possible “double hit” or “triple hit” lymphomas, which have BCL2 and/or BCL-6 translocations in addition to MYC translocations. Double hit and triple hit lymphomas are particularly aggressive lymphomas with poor prognosis and poor response to conventional chemotherapy. Others question the use of a Ki-67 proliferation limit as a guide to consider FISH testing. An immunohistochemical score for double hit lymphomas has also been described. In this scoring system, patients with at least 70% of lymphoma cells positive for BCL2 and at least 40% of lymphoma cells positive for MYC had a worse prognosis when treated with conventional chemotherapy.

Hematologic neoplasms are frequently included in the differential diagnosis of diffuse large B-cell lymphoma. Multiple myeloma, especially the pleomorphic type, can resemble diffuse large B-cell lymphoma. Review of serum and urine protein electrophoresis, if available, may help distinguish myeloma from lymphoma. In addition, immunohistochemical stains for κ and λ immunoglobulin chains are negative in most diffuse large B-cell lymphomas and positive in the vast majority of plasma cell myeloma cases.

Myeloid sarcoma may mimic diffuse large B-cell lymphoma. The most sensitive immunohistochemical stains for myeloid sarcoma are CD43 and lysozyme. The majority of diffuse large B-cell lymphomas are also positive for CD43 but are negative for lysozyme, CD34, and myeloperoxidase. Eosinophilic myeloblasts, when present, are a clue to the diagnosis of myeloid sarcoma. The myeloid nature of the cells in question is often easier to recognize on touch preparations stained with Wright-Giemsa. Chloracetate esterase is inactivated by acid decalcification, and stains performed on decalcified tissue can provide false-negative results. Immunohistochemical stains for lysozyme myeloperoxidase are useful in identifying the myeloid nature of the tumor cells.

Langerhans cell histiocytosis can be easily distinguished from lymphoma by the presence of a mixture of histiocytic cells with a prominent eosinophilic infiltrate. Occasionally, when Langerhans cells predominate, it can be difficult to distinguish Langerhans cell histiocytosis from diffuse large B-cell lymphoma. The strong positivity of Langerhans cells for S-100, the younger age of the patients, and radiographic features usually help distinguish this disorder from lymphoma.

Mastocytosis can be suspected if the entire clinical presentation, the presence of skin lesions and systemic symptoms, is taken into consideration. The mast cell nature of the cells in question can be suspected if a round-cell infiltrate is negative for common leukocyte and epithelial markers. Mast cells are positive for CD117 and mast cell tryptase, and their granules can be revealed in preparations stained with Giemsa, toluidine, or alcian blue.

Diffuse large B-cell lymphoma may be differentiated from pediatric small round blue cell tumors such as Ewing's sarcoma/primitive neuroectodermal tumor (ES/PNET) by evaluating an immunohistochemical panel, including CD99, FLI1, and cytokeratin. CD99 is positive in ES/PNET; however, 55% of myeloid sarcomas may also be positive for CD99. ES/PNETs are negative for CD43 and lysozyme, positive for FLI1, and show cytokeratin expression in a subset of cases. Ewing's sarcoma is extremely rare in patients older than age 40 years, an age when most non-Hodgkin's lymphomas are diagnosed. Multifocal skeletal lesions are extremely rare in Ewing's sarcoma. On the other hand, multifocality is a frequent feature of skeletal lymphomas. In addition, detection of t(11;12) (q24,q12) by karyotyping, PCR, or FISH confirms a diagnosis of ES/PNET.

Metastatic small cell carcinoma occasionally can be very difficult to distinguish from lymphoma because it may have similar features, especially with sclerotic bone involvement by lymphoma. The use of appropriate markers to document the epithelial nature of cells and frequent neuroendocrine differentiation, as demonstrated by positivity for CD56, synaptophysin, crhomogranin, and TTF1, in small cell carcinomas are helpful distinguishing features.

Non-Hodgkin's lymphoma of bone is frequently misdiagnosed as chronic osteomyelitis . This error can be avoided if attention is paid to the clinicoradiographic features. The involvement of the shaft of long tubular bones in patients older than age 40 years is rarely a feature of chronic osteomyelitis. The presence of reactive lymphocytes in malignant lymphoma is frequently responsible for this error. In this instance, the recognition of atypical lymphoid cells and the identification of the phenotypic features consistent with lymphoma are the keys to the correct diagnosis.

Genetic Features and Pathogenesis

The normal cell counterpart is a mature B cell that has been exposed to antigen and undergone somatic hypermutation of immunoglobulin genes in the germinal center. The normal cell counterpart may be a germinal center B cell (GCB) or post-germinal center B cell. The post-germinal center subtype of diffuse large B-cell lymphoma is commonly referred to as activated B cell (ABC) subtype . The neoplastic B cells show clonal rearrangement of immunoglobulin genes and show somatic hypermutation of immunoglobulin variable chain genes, as is seen in normal germinal center and post-germinal center B cells.

The molecular profile of diffuse large B-cell lymphoma of bone is similar to that for the diffuse large B-cell lymphoma occurring at other sites. In 2000, Alizadeh et al. divided diffuse large B-cell lymphoma into two subtypes on the basis of gene expression profiles and showed that diffuse large B-cell lymphoma of GCB has a better prognosis than ABC type. The prognostic significance of these subtypes has been confirmed by some studies and have been challenged by authors of other studies. In 2004, Hans et al. described a simple immunohistochemical panel (CD10, BCL6, and MUM1), now commonly named the Hans classifier ( Fig. 12-32 ) that loosely reproduced cDNA microarray-based subtyping with a positive predictive value of 87% for GCB subtype and 73% for ABCsubtype. The panel also predicted better survival for the lymphoma classified as GCB subtype, similar to the gene expression profile classification. Subsequently, several groups have proposed algorithms based on more extensive immunohistochemical panels that slightly improve correlation between subgrouping based on gene expression-based and immunohistochemical features.

There is a tremendous degree of genetic variation among diffuse large B-cell lymphomas, resulting in highly variable response to current treatment regimens. It is estimated that an individual case of diffuse large B-cell lymphoma harbors between 30 and 100 different mutations. A comprehensive discussion of molecular genetics in diffuse large B-cell lymphoma is clearly beyond the scope of this chapter. Table 12-5 lists common genetic abnormalities in diffuse large B-cell lymphoma. In one large study, exome sequencing of diffuse large B-cell lymphomas detected recurrent mutations in 322 genes. The types of mutations included copy number alterations, intergenic mutations, and mutations of coding and regulatory regions of various genes. The types of genes mutated included genes implicated in apoptosis, cell adhesion, cell cycle, cell differentiation, metabolism, epigenetic modification, DNA repair, immune response, membrane transport, protein modification, signal transduction, and ubiquitin cycle.

Frequent mutations are being discovered in genes previously not known to be affected in lymphomas. These genes include epigenetic modifiers, genes involved in immune surveillance, and genes that regulate signaling pathways ( Table 12-5 ). Mixed lineage leukemia 2 ( MLL2 ) is an example of a gene encoding a histone methyltransferase that is mutated in more than 30% of diffuse large B-cell lymphomas. Inactivating mutations of β ₂ microglobulin ( B2M ) are seen in approximately 30% of diffuse large B-cell lymphomas, resulting in the inability of cytotoxic T cells to recognize lymphoma cells.

Constitutive activation of nuclear factor κB (NFκB) is a frequent finding in diffuse large B- cell lymphoma, particularly the ABC subtype. Mutations of several regulatory genes may be implicated in the constitutive activation of NFκB, including myeloid differentiation primary response 88 ( MYD88, mutated in 30% of ABC diffuse large B-cell lymphomas), TNFα-induced protein 3 ( TNFAIP3/A20, mutated in approximately 30% of ABC diffuse large B-cell lymphomas), and caspase recruitment domain-containing protein 11 ( CARD11, mutated in 9% of ABC diffuse large B-cell lymphomas).

B-cell lymphomas are thought to harbor many mutations at least in part due to mistargeting of physiologic DNA breakage that occurs during normal B-cell development. The mechanisms that produce immunoglobulin diversity via DNA breakage and reassembly include variable, diversity, and joining [V(D)J] immunoglobulin segment recombination, class switch recombination, and somatic hypermutation of immunoglobulin genes. The mediators of these processes, such as recombination activating genes ( RAG, which is required for V(D)J recombination) and activation induced cytidine deaminase ( AID, which initiates somatic hypermutation and class switch recombination) have been implicated in many of the translocations and point mutations seen in B-cell lymphoma.

The most common translocations in diffuse large B-cell lymphoma involve the B-cell lymphoma 6 ( BCL6 ) gene on the long arm of chromosome three at 3q27 ( Fig. 12-33 ). Translocations involving BCL6 are present in 30% to 40% of diffuse large B-cell lymphomas and are more common in the ABC type. In normal B cells, BCL6 is expressed only during the time the B cell is within the germinal center. BCL 6 is a transcriptional repressor that maintains germinal center B-cell functions until the cells mature into post-germinal center memory B cells or plasma cells. BCL6 protein also represses TP53, which is a checkpoint gene and activator of DNA repair mechanisms. The absence of TP53 promotes genetic instability, allowing somatic hypermutation of immunoglobulin variable segment genes. In normal B-cell maturation, somatic hypermutation produces mature memory B cells and plasma cells that produce high affinity immunoglobulin specific for a given antigen. In the setting of lymphoma, aberrant somatic mutation may be seen as a primary event or may be associated with transformation to a more aggressive lymphoma.

BCL6 can rearrange with more than 30 different gene partners. The translocation partner is an immunoglobulin gene in more than half of cases, with the remaining translocation partners represented by nonimmunoglobulin genes. The majority of the breakpoints for BCL6 gene rearrangement are located in major translocation clusters, including exon 1 and an intronic sequence between exon 1 and exon 2 ( Fig. 12-33 ). The result is translocation of a nearly complete BCL 6 gene under the control of the promotors of several different gene partners. The BCL6 breakpoints have sequences that are targeted by AID in normal immunoglobulin class switch recombination, and it is thought that these translocations occur in the germinal center under the influence of AID. In addition, the BCL6 gene in normal B cells and lymphoma undergoes somatic hypermutation, also thought to be mediated by AID.

Another common translocation, seen in up to 40% of diffuse large B-cell lymphomas, is t(14;18)(q32;q21) ( Fig. 12-34 ). This translocation involves the immunoglobulin heavy chain gene ( IGH ) on chromosome 14 and B-cell lymphoma 2 ( BCL2 ) on chromosome 18, resulting in overexpression of BCL2 under the control of the IgH enhancer ( Fig. 12-34 ). Increased expression of BCL2 increases cell survival by inhibiting apoptosis. The translocation alone is not sufficient to cause lymphoma. The t(14;18) is present in small numbers of B cells in normal germinal center B cells.

The t(14;18) IGH breakpoints are within the joining segments (JH). Approximately 50% of the BCL 2 breakpoints are clustered within the major breakpoint region, with the majority of the remaining breakpoints within the minor cluster region (25%) and intermediate cluster region. The breakpoints in the BCL2 major breakpoint cluster show an overrepresentation of sequences that are targeted by AID, and it is thought that t(14;18) occurs in bone marrow B-cell progenitors during VDJ recombination under the influence of AID.

Definition

Adult T-cell lymphoma/leukemia is a malignancy of mature CD4-positive T cells caused by infection with human T-cell lymphotrophic virus-1 (HTLV-1). It is characterized by the presence of numerous, mostly nonrecurrent, genetic abnormalities and atypical cytologic features with marked nuclear irregularity. Lytic bone lesions and hypercalcemia are common.