Rouleaux formation. Blood smear from a 56-year-old male with a 5.1 g/dL IgG monoclonal protein showing increased red cell rouleaux formation; × 50
Plasma cell leukemia. Blood smear showing numerous leukemic plasma cells in a 59-year-old man; a × 20, b × 100
The bone marrow examination is a key component in the diagnosis of plasma cell myeloma and should be performed in essentially all cases . A bone marrow examination is required to confirm the diagnosis of myeloma, even when there is substantial clinical, laboratory, and radiographic evidence. The marrow findings also provide prognostic information and are useful in following patients for response to therapy and identifying recurrent disease. In addition to the morphologic assessment, bone marrow is used for immunophenotyping, cytogenetics, molecular studies, and in some cases other ancillary testing. Morphologic criteria for the diagnosis of myeloma are listed in Table 1 .
Criteria for a morphologic diagnosis of plasma cell myeloma (from Dick )
Random bone marrow specimen showing:
1. Atypical plasma cells with morphologic appearance outside the range of a reactive process
2. Infiltrative sheets of plasma cells on sections
3. Nearly 100 % plasma cells on a hypercellular aspirate or section less-useful criteria include multinucleation and lack of predilection for vascular structures
Both aspirate smears and trephine biopsy sections are required for optimal assessment. In most cases, the two are independently diagnostic. However, in some patients it is the combined findings in the two preparations that validate the diagnosis. The mean number of plasma cells in the aspirate smears at diagnosis of plasma cell myeloma is 20–36 % (Fig. 3) [2, 3]. In about 5 % of cases of symptomatic myeloma, plasma cells comprise fewer than 10 % [2, 4]. This can result from a suboptimal marrow aspirate or sampling issues in cases with scattered focal lesions. Neoplastic plasma cells vary from morphologically normal appearing with mature features to blast-like cells, hardly recognizable as plasma cells [5–8]. Atypical features that characterize many cases of myeloma encompass changes in either or both the nucleus and cytoplasm. Myeloma plasma cells are often larger than normal but may be normal sized or even smaller than normal. Moderate to abundant basophilic cytoplasm is usual. A broad spectrum of cytoplasmic changes is observed that includes fraying and shedding of the cytoplasmic edges, vacuoles, granules, and cytoplasmic inclusions. The nucleus is larger than normal in most cases with less-dense chromatin; nucleoli are variable but often prominent .
Bone marrow aspirate smears in myeloma. 68-year-old female who presented with multiple vertebral body compression fractures was found to have an IgG lambda monoclonal protein. The bone marrow aspirate smear shows extensive involvement by neoplastic plasma cells with atypical features; a × 20, b × 100
Cytoplasmic crystals are found occasionally (Fig. 4); they have no obvious relationship to the type of M-protein produced by the neoplastic plasma cells. Multiple dark staining cytoplasmic inclusions are observed in rare cases of myeloma. These are often associated with large pleomorphic plasma cells. Multiple small Russell body-type hyaline intracytoplasmic and intranuclear inclusions are relatively common. In contrast to hyaline intranuclear inclusions, Dutcher type nuclear inclusions are pale staining, single, and generally large (Fig. 5). In some cases, cytoplasmic inclusions resemble the Bohot plasma cell structures found in patients with mucopolysaccharidosis. Phagocytic plasma cells are observed in a small minority of cases of myeloma; rarely, marked erythrophagocytosis is present .
Intracytoplasmic inclusions in myeloma plasma cells. Several intracytoplasmic inclusions in neoplastic plasma cells from patients with plasma cell myeloma including crystalline material (a, b) and myeloma plasma cells with intracytoplasmic inclusions resembling the “Bohot” cells seen in some patients with mucopolyssacharidosis (c, d); × 100
Intranuclear inclusions in myeloma plasma cells. Intranuclear inclusions (Dutcher bodies) at arrow tips (a × 100) and corresponding tissue section (b × 50)
Approximately 2 % of myelomas are distinguished by marked nuclear lobation and convolution [2, 10]. In some the convoluted cells are mixed with easily recognizable plasma cells, but in others they comprise a more uniform population and are difficult to recognize as myeloma plasma cells. Small plasma cells predominate in some myelomas and may have a distinctly lymphoid or lymphoplasmacytic appearance. In one study, 20 % of the cases with lymphoid morphology were IgD myelomas . Lymphoplasmacytic morphology has been associated with a t(11;14) chromosomal translocation (Fig. 6) . Attempts to relate morphologic features to the type of M-protein have generally failed except for a small number of cases of IgA myeloma with markedly pleomorphic, large multinucleate plasma cells, flaming plasma cells, and cells with pale, frayed, and fragmented cytoplasm. Intranuclear inclusions are found in about 20 % of cases of IgA myeloma, seemingly more frequent than for other immunoglobulin types (Fig. 5)  .
Myeloma with lymphoplasmacytic features and a t(11;14). 71-year-old male with an IgA kappa monoclonal protein and bone marrow involvement by plasma cells with lymphoplasmacytic cytologic features. Fluorescent in situ hybridization (FISH) studies on bone marrow biopsy touch preparations showed, t(11;14)(q13;q32); a H&E × 20; b Marrow aspirate smear × 100; c CD138 × 20; d kappa × 20; e lambda × 20; f CD20 × 20; g CyclinD1 × 20
Based on cytologic features myelomas can be classified into mature, intermediate, immature, and plasmablastic categories (Figs. 6 and 7) . Patients with plasmablastic myeloma have a median survival significantly shorter than for the other cytologic categories. There appears to be no significant difference in survival among the other three categories. Other morphologic classifications include three to six cytologic types [6, 13] .
Immature-type myeloma. Aspirate smear (a × 100) and tissue section (b × 50) showing immature morphologic features in plasma cells from a 60-year-old male with IgG kappa plasma cell myeloma; Plasmablastic-type myeloma. Aspirate smear (c × 100) and tissue section (d × 50) showing blastoid lambda-restricted plasma cells diffusely involving the bone marrow of a 50-year-old male; Immature myeloma with plasma cells exhibiting anaplastic features. Tissue sections (e × 10; f × 50) showing scattered plasma cells with anaplastic features
The diagnostic yield of trephine biopsies in plasma cell myeloma is usually excellent but can be affected by the size and quality of the specimen. Focal lesions may be irregularly distributed and widely spaced. Occasionally only one or two small myeloma lesions are found in a trephine biopsy with no evidence of a plasma cell infiltrate in the remainder of the section or in specimens from the contralateral posterior iliac spine. The pattern of the plasma cell infiltrate may be interstitial, focal, or diffuse (Fig. 8) [2, 6, 7]. The extent of marrow involvement varies from a small increase in plasma cells to complete replacement. The pattern of involvement is largely related to the extent of disease. With interstitial and focal patterns, there is generally considerable marrow sparing and preservation of normal hematopoiesis. With diffuse involvement, expansive areas of the marrow are replaced and hematopoiesis may be markedly suppressed. Typically, interstitial and focal disease in early myeloma progresses to diffuse involvement in advanced stages of the disease .
Patterns of bone marrow involvement in plasma cell myeloma with corresponding CD138 immunohistochemical stains: focal (a, b); interstitial (c, d); diffuse (e, f); × 20
Bartl and associates proposed a staging system based on percentage of marrow space replaced by neoplastic plasma cells in marrow trephine biopsies [6, 14]. In Stage I less than 20 % of the marrow is replaced, Stage II, 20–50 % and Stage III, more than 50 %. In most instances the extent of involvement in biopsy sections seems to reflect overall tumor burden. In the Bartl and associates’ study there was good correlation between histologic stage, clinical stage, and prognosis .
Myelomas with atypical plasma cell morphology may be difficult to recognize in trephine biopsies (Fig. 7d). Plasma cell myeloma with plasmablastic features, lymphoid appearing plasma cells, convoluted plasma cells, or markedly pleomorphic or anaplastic plasma cells are particularly problematic. Examination of the neoplastic plasma cells in aspirate smears is often essential for diagnosis in these cases. Numerous cytoplasmic inclusions in the myeloma plasma cells may be a distracting morphologic feature on the bone marrow section. The inclusions are often found in large plasma cells that are distorted by crystalline or globular material hiding the identity of the plasma cell neoplasm .
Reticulin or collagen fibrosis may be present in a minority (approximately 10 %) of cases of myeloma [2, 15]. In many of these there is extensive reticulin or collagen fibrosis. A disproportionate number of fibrotic myelomas produce monoclonal light chains only . Coarse fibrosis has been correlated with diffuse marrow involvement and aggressive disease . Fibrosis with osteosclerotic changes is found in plasma cell neoplasms associated with polyneuropathy, organomegaly, endocrinopathy, M protein, skin changes (POEMS) syndrome, discussed later in this chapter.
Bone changes are a frequent finding in trephine biopsies from patients with plasma cell myeloma. Marked osteopenia and evidence of increased osteoclastic activity with reabsorption of bone may be identified (Fig. 9) .
Osteoclastic activity. Bone marrow trephine biopsy from a 68-year-old female with IgA kappa plasma cell myeloma and hypercalcemia shows marked osteoclastic activity on the bone trabecular surface and sheets of adjacent plasma cells; × 20 HE
Immunohistochemistry is an important complement to the morphologic assessment of plasma cell neoplasms . Selected immunohistochemical stains can supplement flow cytometry immunophenotyping or provide the primary immunophenotypic assessment for plasma cell neoplasms when a specimen is not obtained for flow cytometry or contains inadequate numbers of plasma cells for flow cytometric analysis. The plasma cells in myeloma typically express CD79a, VS38c, CD138 , CD38, and monotypic cytoplasmic immunoglobulin and lack surface immunoglobulin. Unlike normal plasma cells they are nearly always CD19 negative, and CD56 is aberrantly expressed in about 75 % of cases (Fig. 10) [16–18]. Other aberrantly expressed antigens include CD117, CD20, CD52, and CD10, in decreasing order of frequency; occasionally, myeloid and monocytic antigens are found [16, 19, 20] .
Immunohistochemical stains for plasma cell myeloma. Diffuse bone marrow involvement with neoplastic plasma cells in a 31-year-old man with a 1.5 g/dL IgA kappa monoclonal protein. a H&E × 20, b H&E × 50, c CD138 × 50, d CD56 × 50, e Kappa × 50, f Lambda × 50
A more detailed description of the immunophenotype of normal and neoplastic plasma cells is found in the chapter on flow cytometry. The remainder of this discussion deals specifically with the utility of immunohistochemistry in diagnosis and in following patients with plasma cell myeloma.
Immunohistochemical stains on marrow biopsies or other tissues are valuable in the following:
Quantification of plasma cells in marrow biopsies,
Identification of a monoclonal plasma cell population,
Distinction of myeloma from other neoplasms.
Plasma cells may be difficult to recognize and quantify in sub-optimally prepared sections and when scattered interstitially in the marrow. Several plasma cell-associated antigens with variable sensitivity and specificity may be assessed by immunohistochemical stains to quantify plasma cells in biopsy specimens (Figs. 7 and 10) . Immunohistochemistry for CD138 , CD38, CD79a, kappa and lambda will usually stain plasma cells brilliantly on biopsy sections, allowing easy quantification. CD138 is a commonly used marker, expressed on normal plasma cells and in 60–100 % of myelomas [21, 22], and useful in identifying and quantifying plasma cells in biopsy sections. Generally, from 70 to 100 % of the cells within an individual neoplastic plasma cell population are CD138 positive . CD138 appears to be plasma cell-specific among normal hematopoietic cells in the marrow, however; other neoplastic B-cell diseases such as chronic lymphocytic leukemia and primary effusion lymphomas react with some anti-CD138 antibodies. CD79a is positive in the plasma cells in most cases of myeloma but is a pan B lymphocyte antigen and is found in most B-cell neoplasms. It may be helpful in distinguishing myeloma from non-B-cell hematopoietic neoplasms and metastatic tumors .
CD20 is positive in 15–20 % of myelomas, mostly those with a t(11;14) . CD79a and CD38, although not plasma cell-specific, are usually expressed on neoplastic plasma cells. CD56 is expressed in about 75 % myelomas and is a marker of aberrancy in plasma cell proliferations.
Immunohistochemical stains and in situ hybridization for kappa and lambda light chains are useful in characterizing plasma cell neoplasms and differentiating them from reactive plasma cell proliferations as may be found with connective tissue disorders, chronic liver disease, chronic infections, and metastatic tumors . Normal/reactive plasma cells and myeloma plasma cells are both rich in cytoplasmic immunoglobulin and generally react strongly with antibodies to kappa or lambda light chains. In cases of myeloma the plasma cells express a monoclonal pattern of reactivity [24, 25]. In normal marrow and in reactive plasma cell proliferations, there is a polyclonal pattern of kappa and lambda staining plasma cells, usually with a slight to moderate kappa predominance (Fig. 11) . In some cases of MGUS the kappa–lambda ratio is normal, in others it is skewed but generally less, than in myeloma. In one large study a kappa–lambda staining ratio of 16:1 or higher on marrow biopsies distinguished myeloma from MGUS in nearly all instances; other investigators found a ratio of 8:1 to be as effective [24, 26]. Neither the number of marrow plasma cells nor the quantity of M-protein correlates well with the light chain ratio . Kappa and lambda stains are particularly useful in cases with a relatively low percentage of marrow plasma cells, as it is often encountered when evaluating for residual/recurrent disease following chemotherapy or stem cell transplantation .
Reactive plasmacytosis. Staging bone marrow biopsy section from a 48-year-old HIV positive male with a CD4 count of 45 who was diagnosed with plasmablastic lymphoma of the rectum. Marrow plasma cells were increased a × 50 but without atypia of neoplastic plasma cells and were found to be polytypic by immunohistochemical stains for kappa b × 50 and lambda c × 50
Some plasma cell myelomas have extreme atypical morphologic features that disguise their identity. Several hematopoietic and non-hematopoietic neoplasms may occasionally exhibit features that mimic a plasma cell neoplasm. Immunohistochemistry is often helpful in differentiating these cases. Stains for plasma cell-associated antigens and a panel of appropriate antigens associated with other neoplasms are useful in differentiating myeloma from other hematopoietic neoplasms or a metastatic tumor.
Monoclonal Gammopathy of Undetermined Significance
Blood and Bone Marrow Findings
There are no specific blood findings associated with MGUS. In patients with M-protein levels on the high side of the range, rouleaux formation may be increased. Blood count abnormalities and other changes on blood smears when present are usually related to a coexisting disease.