Diagnostic and Clinical Considerations in Concomitant Bone Marrow Involvement by Plasma Cell Myeloma and Chronic Lymphocytic Leukemia/Monoclonal B-Cell Lymphocytosis: A Series of 15 Cases and Review of Literature Open Access

After obtaining respective institutional review board approval, the electronic databases of surgical pathology reports at 4 institutions were searched, covering a 10-year period between 2000 and 2009, to identify reports of bone marrow biopsies seen in which monotypic B cells and monotypic plasma cells were both identified. From the cases flagged by the automated searches, a manual review of pathology reports was conducted to identify cases of concurrent occurrence of CD5⁺ monotypic B cells and monotypic plasma cells. All available cytologic and histologic preparations were reviewed along with flow cytometric, cytogenetic, and molecular diagnostic data. The available hematology laboratory data and radiographic and clinical information were reviewed from the electronic hospital information systems at the respective institutions.

Additional immunohistochemical stains were performed as required. Five-micron-thick sections of the bone marrow core biopsies were stained for CD56 (clone TB01, Leica Microsystems, Bannockburn. Illinois) and cyclin D1 (clone SP4, Thermo Fisher Scientific, Fremont, California) by using established labeled polymer immunohistochemical staining technique. The sections were cut and mounted on adherent glass slides, dewaxed in xylene, and rehydrated in graded ethanols. Endogenous peroxidase activity was blocked by immersion in 0.3% aqueous peroxide for 15 minutes, followed by 2 washes in 1× Tris-buffered saline (TBS) for 5 minutes each. Slides were treated with a cocktail of 10 mmol/L Tris and 1 mmol/L EDTA pH 9.0 and heated in a pressure cooker for 2.5 minutes at 125°C, followed by a 20 minute cooling period. The sections were then incubated for 1 hour at room temperature with the primary antibody diluted in TBS. The negative control consisted of substitution of the primary antibody in selected cases with mouse immunoglobulin (Ig) G. This was followed by 2 washes in TBS and then incubation with anti-mouse horseradish peroxidase–labeled secondary antibody (Envision Plus, Dako Corp, Carpinteria, California) for 30 minutes. The bound complexes were visualized by the application of diaminobenzadine (Dako) containing 0.3% hydrogen peroxide as a substrate. After incubation, the sections were washed, lightly counterstained with hematoxylin or light green solution, washed again, dehydrated, cleared, and cover slipped.

The online database PubMed, maintained by National Center for Biotechnology Information, was searched for English language articles describing case reports or case series of patients with concurrent or sequential diagnosis of PCM and CLL. Complete versions of all selected articles were obtained and the clinical and pathologic information was abstracted from each and tabulated.

All clinical and pathologic information from the selected cases, as well as from the published articles, was tabulated with Excel (Microsoft Corporation, Redmond, Washington). The same software was used to apply the t test. The Web-based software provided at http://www.quantitativeskills.com/sisa/index.htm (accessed December 25, 2011) was used to perform the χ²  and Fisher exact tests, and the GraphPad Prism (GraphPad Software, Inc, La Jolla, California) was used to perform a Kaplan-Meier survival analysis.

Fifteen cases were identified in which the bone marrow showed simultaneous presence of monotypic and/or atypical plasma cells and a CD5⁺/CD23⁺ monotypic B-cell population. The clinical features in these patients are summarized in Table 1. There were 13 male (86.6%) and 2 female patients. The median age was 74 years (range, 56–91 years). In eleven of 15 cases (73.3%), the initial presentation was related to the plasma cell dyscrasia; but in all 11 of these, the first bone marrow specimen obtained showed the presence of both abnormal plasma cells and CD5⁺ monotypic B cells. The plasma cell dyscrasia was further characterized as symptomatic plasma cell myeloma (PCM) in 8 patients and as smoldering (asymptomatic) myeloma in 3 patients. Among the patients with symptomatic PCM, the primary clinical findings were related to pathologic fractures/lytic bone lesions in 6, while amyloid nephropathy and polyneuropathy with weight loss and anemia were the presenting features in 1 patient each. In 2 patients with smoldering myeloma, an incidentally detected monoclonal gammopathy prompted the bone marrow examination, while the third patient had a 14-year history of stable monoclonal gammopathy with more recent leukocytosis and an increasing level of the monoclonal paraprotein. Only 1 patient with symptomatic PCM had a documented 5-year history of monoclonal gammopathy preceding the diagnosis of PCM. In 4 of 15 patients (26.7%), the initial clinical and pathologic diagnosis was CLL, preceding the diagnosis of plasma cell dyscrasia by 2 years or longer (average, 34 months), and the bone marrow examination was performed to rule out disease progression or to investigate a new finding of monoclonal paraprotein in serum or urine. In 1 of these 4 patients (case 14), the CLL was in an advanced stage and required treatment, while the other 3 patients with CLL had not received specific treatment for this condition. Except for case 14, clinically detectable lymphadenopathy was not noted in any patient.

The peripheral blood findings at the time when the bone marrow biopsy showed the 2 monotypic processes concomitantly are summarized in Table 2. Peripheral blood counts are not available for 4 cases (cases 1, 2, 11, and 14) because the bone marrow biopsy was obtained at an outside institution. For the remaining 11 cases, 6 patients (54.5%) were anemic at the time of diagnosis. The mean hemoglobin level for the 11 cases was 12.1 g/dL (range, 10.1–15.5 g/dL). The platelet count was largely unaffected with only 1 case having a platelet count of less than 150 000/μL. Two of 11 patients (18.2%) had a circulating leukocytosis (cases 8 and 13) with an absolute lymphocyte count greater than 5000/μL. One of these patients had a diagnosis of CLL for 3.5 years, while the other had a stable IgG monoclonal paraprotein for 14 years and recent lymphocytosis. Three more patients (cases 5, 6, and 7) had lymphocytosis of greater than 1500/μL. Peripheral blood flow cytometry was performed in 3 cases, including the 2 cases with greater than 5000/μL lymphocytosis (cases 8 and 13), which showed 52% and 59% circulating abnormal lymphocytes with the CLL phenotype. The third case (case 3) showed only 2% abnormal circulating lymphocytes but the marrow showed 10% abnormal B cells (see below).

The salient bone marrow findings are summarized in Table 3. The bone marrow cellularity varied from 30% to 90% (average, 53%) in the whole group. The average bone marrow cellularity was similar in those diagnosed first with PCM (54%) or with CLL (50%). The marrow was moderately to markedly hypercellular for age in 7 of 15 patients, most (6) of whom were initially diagnosed with PCM. The proportion of abnormal plasma cells varied from 2% to 50% of total marrow cells (average, 23%), while the CD5⁺/CD23⁺ monotypic B cells accounted for 0% to 50% of total cells (average, 17%). Neither type of cells occurred at significantly different proportions when comparing cases initially diagnosed as plasma cell dyscrasia or CLL. The proportion of either type of abnormal cells or overall proportion of abnormal cells did not correlate with marrow cellularity (linear regression, r² < 0.01). Reticulin fibrosis was present around the lymphoid cells in 5 of 10 cases initially presenting as plasma cell dyscrasia (reticulin stain was not performed in 1 case), but only in 1 of 4 cases with the initial diagnosis of CLL. Cytogenetics and/or fluorescence in situ hybridization abnormalities were detected in 7 of 10 patients for whom data were available. Abnormalities including deletions of 13q or 17p, and numeric abnormalities of chromosomes 11 or 12, were observed in 5 patients, generally in a small proportion of total cells (3% to 15%). One case showed fusion of cyclin D1 and IGH gene and 1 patient had deletion of the Y chromosome.

In every case, the abnormal B cells coexpressed CD5 and CD23, either by flow cytometry (14 cases) or immunohistochemistry (1 case). In 10 of 15 cases, the abnormal plasma cells were restricted for immunoglobulin κ light chain and in the remaining 5 for λ. For the CD5⁺/CD23⁺ abnormal lymphocytes, flow cytometric characterization of light-chain expression was available for 13 patients, while 1 case had essentially no surface immunoglobulin light-chain expression and flow cytometry was not performed in 1 case. The CD5⁺/CD23⁺ B cells were κ restricted in 8 of 13 cases and λ restricted in 5 of 13. The distribution of light-chain restriction between the plasma cells and lymphocytes is not significantly different (P = .29, Fisher exact). In these 13 cases for which light-chain restriction for both abnormal cell types was known, the expressed light chains were different for the 2 populations in 6 cases and the same in 7 (both were κ in 6 and λ in 1). Both immunohistochemistry and flow cytometry results are required for correct diagnosis when the light-chain restriction by the 2 abnormal populations is different but CD56 or cyclin D1 are not aberrantly expressed by the plasma cells, as illustrated in Figure 1, A through H, and Figure 2, A through D. Cyclin D1 overexpression occurred in the plasma cells in 8 of 14 cases (tissue was exhausted in 1 case and could not be tested). CD56 was aberrantly expressed in the plasma cells in 6 of 12 cases, along with cyclin D1 expression in all but 1 case. In 4 of 6 cases for which both abnormal B cells and plasma cells were κ light-chain restricted, the plasma cells were cyclin D1 positive and a fifth case had CCND1-IGH fusion by fluorescence in situ hybridization in all plasma cells. In the sixth case, the plasma cells aberrantly expressed CD56. This immunoreactivity pattern identified the plasma cells as constituting a plasma cell dyscrasia separate from the CD5⁺ monotypic B cells. In the case illustrated in Figure 3, A through F, the abnormal plasma cells and B cells both showed κ light-chain restriction (not shown), but the 2 hematologic malignancies can be clearly identified by the distinct nodular infiltrates of CD5⁺ B cells and CD138⁺/cyclin D1+ plasma cells. Overall, in 12 cases the B cells and plasma cells were shown to be distinct monotypic populations (6 by virtue of different light-chain restriction and 6 by virtue of cyclin D1 and/or CD56 expression in plasma cells). In 3 cases, the additional immunostaining was negative and the light-chain restriction did not clearly differentiate the plasma cells and B cells as separate clonal processes. In 1 case, the restriction was concordant (λ) and in 2 cases the B-cell light-chain restriction was unknown. In these 3 cases, flow cytometric or immunohistochemical demonstration of the CD5⁺/CD23⁺ phenotype of the abnormal B cells was important to rule out other differential diagnostic considerations (see discussion).

Five of 15 patients did not receive any specific treatment, 4 of whom were diagnosed first with plasma cell dyscrasia. For these 4 patients, treatment was withheld per patient request either owing to advanced age and/or because the disease was clinically stable. Seven of 11 patients who first presented with symptoms related to plasma cell dyscrasia were treated with myeloma-specific therapy (melphelan, bortezomib, thalidomide, lenalidomide, steroids, or localized radiation in different combinations), even though CD5⁺ monoclonal B cells were documented in the marrow. Only 1 patient received additional therapy directed at the monoclonal B cells (anti-CD20 antibody rituximab). Two of 11 patients underwent stem cell transplant, 1 with allogeneic and one with autologous cells. Only 1 of 4 patients initially presenting with CLL received disease-specific treatment (cyclophosphamide and fludarabine). In another patient from this group, anti-PCM treatment was given when abnormal plasma cells were later detected. Follow-up information was available for 14 patients and duration of follow-up varied from 1 month to 122 months (median, 6.5 months; average, 27.3 months). Three patients died; for 2, death was disease related and for the third it was due to a third malignancy (pancreatic adenocarcinoma). Eight of 11 patients initially presenting with PCM and 3 of 4 presenting with CLL had a reduced amount of marrow disease and/or clinically stable disease at last follow-up.

Plasma cell myeloma and CLL are the 2 most common B-cell malignancies occurring in the older adult population.¹  Despite largely overlapping epidemiologic features and origin from mature B cells, simultaneous occurrence of both diseases in a patient is rare. A study from the Mayo Clinic,¹⁵  spanning 10 years, found 31 cases of CLL with a second malignancy, but none of these was PCM. The authors also conducted a literature review and noted that none of the 107 cases of a second malignancy found in patients with CLL was PCM. In a large epidemiologic study spanning 38 years and including 54 159 patients with hematolymphoid malignancies from Sweden, only 3 cases of plasma cell myeloma and lymphoid leukemia were found together.¹⁶  This low incidence is remarkable because greater than 8000 patients each with myeloma and lymphoid leukemia were included in this series and the total number of second malignancies noted was 475 and 659 in myeloma and lymphoid leukemia, respectively. In a recent German study⁴⁷  involving 589 consecutive patients with myeloma seen at a single institution in 11 years, 59 (10%) had a second neoplasm before, during, or after the diagnosis of myeloma. Of these cases, 5 (0.84%) were CLL, while 6 cases of MDS and 46 cases of solid tumors were seen in the same cohort. All CLL cases occurred in males and the condition was diagnosed nearly 2 years before the diagnosis of myeloma. With the recognition of MBL, the proportion of older adults having detectable monoclonal B cells with CLL-like phenotype in the blood has dramatically increased⁴⁸  but, to our knowledge, data on actual incidence of concurrent occurrence of PCM and MBL are not available.

We have identified 30 case reports or case series published in the English language literature describing CLL and PCM together in the same patient.^17–46  During the span of 50 years covered by these reports, 51 total cases showing the presence of CLL and PCM in the same patient are described. The salient findings from these studies are summarized in Table 4. In comparing the published data to our findings, the demographic profile of patients is similar (most are males in their sixth or seventh decade of life), but in our series only 4 of 15 (26.7%) were first diagnosed with CLL and then developed PCM, compared to 31 of 51 such cases (60.8%) in the published literature (P = .02, Fisher exact). This difference is probably due to the case selection bias: we selected cases from surgical pathology reports of bone marrow, while many published reports have selected cases by clinical criteria. The published literature reports a significantly higher proportion of abnormal cells in the bone marrow than was observed in our series (46.6% versus 16.5% CLL cells and 34.1 versus 23.1% plasma cells, P < .001 and P = .08, respectively, t test). Relevant data are missing in 12 published reports for CLL cells and in 7 reports for plasma cells. The differences in the level of marrow involvement by PCM are difficult to explain because they are unlikely to be the result of previous treatment, which was given in a similar proportion of cases in our series (73.3%) and published reports (66%).

On the other hand, the lower level of marrow involvement by CD5⁺ B cells seen in our series appears to be a consequence of the method of case selection and the inclusion of patients with MBL in our series. In the absence of extramedullary tissue involvement, the diagnosis of CLL, according to the World Health Organization (WHO) classification,⁴  requires the presence of greater than 5000 monotypic B cells per μL in the blood, with a CD5⁺ and CD23⁺ phenotype for at least 3 months. Notably, this definition differs from earlier definitions that required the presence of 5000 lymphocytes per μL (rather than monotypic B cells).⁴⁹  According to the current definition, the presence of fewer than 5000 monotypic B cells per μL is now categorized as MBL, provided there are no clinical or laboratory features of lymphoma/leukemia.

It is now widely accepted that MBL is an extremely common occurrence in older adults.^5,6,50  In about 75% of these cases, the monoclonal B cells have the same phenotype as CLL cells. Only a minority of MBL cases have absolute lymphocytosis of 1500/μL or greater but less than 5000/μL and these are designated as “clinical MBL” (cMBL). The incidence of cMBL is estimated at 0.6% to 0.9% in persons older than 50 years (which is about 100 times higher than the incidence of CLL, estimated at 5 to 6 cases per 100 000). Presence of very low levels of monoclonal B cells (median, 1 monoclonal B cell per μL) without significant lymphocytosis is much higher, estimated to affect 20% of otherwise healthy older adults in some European countries such as Spain. These cases are detected by population screening with high-sensitivity flow cytometry methods and are designated “screening MBL.” ^7,51 Perhaps all cases of CLL are preceded by cMBL, but fewer than 1% of cMBL cases progress to CLL.⁶  In 1 study,⁵²  the median level of bone marrow lymphocytes in cMBL was found to be 20%.

In the 11 patients from our cohort for whom peripheral blood findings were available, only 2 patients had abnormal B-cell lymphocytosis at a level sufficient to fulfill the WHO definition of CLL (cases 8 and 13). In addition to case 13, three other cases involved patients who were earlier diagnosed with CLL (cases 12, 14, and 15). In 2 patients (cases 1 and 7), the level of CD5⁺ monoclonal B cells in the marrow was 30% or greater, which is significantly higher than the median value of 20% monotypic B cells noted in cMBL, and these cases also likely represent CLL rather than MBL. Thus, 7 patients in our cohort appear to have CLL (cases 1, 7, 8, 12, 13, 14, and 15). The remaining 8 cases are best considered as MBL. However, the actual distinction between MBL and early-stage CLL may not be clinically relevant as these 2 conditions are a continuum rather than discrete entities.⁴⁸  Admittedly, the incidence of very-low-level MBL in patients with PCM cannot be inferred from our study because high-sensitivity flow cytometric detection of monoclonal B cells in the blood is not performed routinely in every patient with PCM. Small numbers of clonal B cells are detected incidentally in the bone marrow at a low frequency. In a single center retrospective study,⁸  fewer than 1% of flow cytometry analyses of bone marrow specimens from patients without a diagnosis of overt lymphoma contained 0.05% to 4.5% (median, 1.28%) clonal B cells. Of these, fewer than a third had the CLL phenotype and about a third of the total cohort subsequently developed non-Hodgkin lymphoma. This rate of progression to overt lymphoma is much higher than the approximately 1% rate of progression from cMBL to CLL.⁴⁸  This may suggest that subclinical/undetectable lymphoma in extramedullary locations may exist in a higher proportion of patients showing low-level clonal B cells in the bone marrow than in those showing low-level clonal B cells in the blood. Notably, none of the patients in this study, in whom clonal B cells were incidentally detected in the marrow, showed plasma cell myeloma in that or subsequent marrow.⁸ 

While it is probably clinically not very significant to differentiate cMBL from CLL in patients with concomitant PCM, it is critical that this rare co-occurrence is correctly differentiated from other causes of a lymphoplasmacytic infiltration in the bone marrow. Most patients in published studies, as well as our cohort, required treatment for PCM, while CLL/MBL was often managed without specific treatment (Tables 1 and 4). On the other hand, hematologic malignancies with a lymphoplasmacytic infiltrate may need lymphoma-specific treatments. The differential diagnoses of a lymphoplasmacytic or lymphoplasmacytoid infiltration in the bone marrow include at least 3 neoplastic conditions: lymphoplasmacytic lymphoma (LPL) (also called immunocytoma or Waldenstrom macroglobulinemia), marginal zone lymphoma (MZL) with plasmacytic differentiation, and CLL with plasmacytic differentiation. The characteristic clinical, morphologic, and immunophenotypic features useful in this differential diagnosis are summarized in Table 5.

Deviation from the classic immunophenotype of the lymphoid neoplasms included in this differential diagnosis may confound the correct interpretation. While characteristically the small lymphocytes in LPL are CD5⁻,^53,54  from 5% to 43% of LPL cases may express CD5, at least partially.^55–57  Asplund et al⁵⁸  reported 1 case of LPL among 26 CD5⁺ B-cell neoplasms with ambiguous immunophenotype. Some of these CD5⁺ LPL cases may have partial expression of CD23 but other immunophenotypic features of CLL, such as dim CD20 and dim surface immunoglobulin,^56,57  are absent and cyclin D1 is not expressed.⁵⁸  The morphologic finding of plasma cell collections completely separated from the lymphoid collections can occasionally occur in LPL,⁵⁷  which mimics the pattern seen in concurrent marrow involvement by CLL and PCM (Figure 3, A), but the plasma cells in LPL do not express cyclin D1 or CD56,⁵⁵  in contrast to PCM cells, which expressed these markers in 53.3% and 42.8% of our cases, respectively. The presence of an intimate mixture of lymphocytes and plasma cells, reported in a prior case report of concurrent CLL and PCM,⁵⁹  more closely resembles LPL and needs more careful workup (Figure 1). The paraprotein is typically (but not invariably) IgM in lymphoplasmacytic lymphoma.⁵⁶  The paraproteins produced by plasma cells in our patients were either IgG or IgA, but among the published reports, several cases include IgM-producing PCM,^26,36,44,45  and patients with CLL can produce IgM paraprotein at levels overlapping those of LPL.⁶⁰  Thus, the isotype of the paraprotein is not a reliable discriminator between LPL and concomitant PCM and CLL.

Of the 3 major types of MZLs (splenic, nodal, and mucosa-associated lymphoid tissue type), only splenic marginal zone lymphoma involves the bone marrow with any regularity,^61,62  but a significant proportion (21%–74%) may show plasmacytic differentiation.^63,64 Up to 20% of splenic MZLs can express CD5,⁶⁵  leading to difficulty in distinguishing these cases from CLL; however, splenic MZL typically shows strong expression of CD20 and surface immunoglobulin, which aids in the distinction between these 2 entities. Splenomegaly can occur in CLL, as well as splenic MZL, and cannot be used to distinguish the two. Splenic MZLs lack expression of cyclin D1,⁶³  which is helpful to separate them from cyclin D1+ plasma cell neoplasms occurring concurrently with CLL. Rarely, a monotypic plasma cell proliferation can occur in marrow involved by splenic MZL, and molecular study may be necessary to identify the 2 monotypic populations if the light-chain restriction is the same for both monotypic processes.⁶⁴  Chronic lymphocytic leukemia/small lymphocytic lymphoma itself can show lymphoplasmacytoid differentiation, but these cases lack true plasma cells and may appear as “atypical” CLL. True plasmacytic differentiation in CLL appears to be very rare and may only be seen in the lymph node.⁶⁶  Clinical and radiographic findings, such as lytic bone lesions, hypercalcemia, non-IgM paraprotein, support a diagnosis of PCM because these features are very rare in the lymphoid malignancies discussed above (LPL, splenic MZL, and CLL). An erroneous diagnosis of mantle cell lymphoma is a critical diagnostic pitfall if it is not realized that the CD5⁺ monoclonal B cells are separate from the cyclin D1+ plasma cells (Figure 3, D and F). Indeed, 1 of our patients (case 2) was referred for treatment after an outside diagnosis of mantle cell lymphoma. Conversely, the plasma cells may express CD20 in about 10% of cases⁶⁷  and may have a lymphocyte-like “small cell” morphology in about 3% to 4% of cases.⁶⁸  When small plasma cells express CD20 and CD45, along with cyclin D1 (which is seen in 75% of “small cell” myelomas), distinction from mantle cell lymphoma is facilitated by demonstrating the CD138⁺ and CD5^– phenotype of the plasma cells.⁶⁸ 

Are the CD5⁺ abnormal B cells and monotypic plasma cells occurring concomitantly in the marrow biologically related and what are the clinical implications of this rare event? It is generally accepted that normal plasma cells are terminally differentiated, specialized cells arising from B cells,⁶⁹  and it is theoretically possible that the monotypic plasma cells are further differentiated or transformed CLL cells. But the 2 processes may arise independently from the same stem cell or arise from different B cells purely coincidentally. Interestingly, only 3 of 51 published reports,^17,44,45  and none of our cases, show PCM preceding CLL (Table 4). Putative transformation of CLL cells to plasma cells may be inferred only by showing clonal identity between the 2 processes. Such clonal identity can be established definitively by sequencing the rearranged immunoglobulin heavy-chain genes from pure populations of each of the 2 abnormal cell types, but these data are not available in most cases. Lacking molecular data, we can use the well-known property of allelic exclusion in B cells⁷⁰  to estimate clonal relatedness. Caution must be exercised when interpreting such data, because nonidentical light chains may occasionally occur in clonal evolution.²⁴  In our cohort, information for light-chain restriction of CLL and PCM cells was available for 13 cases: restriction was the same in 7 and different in 6 cases. In the published reports, excluding the 3 cases first presenting with PCM,^17,44,45  information for light-chain restriction of CLL and PCM cells was available for 28 cases and the light-chain restriction was the same in 15 cases and different in 13. A Kaplan-Meier survival analysis (data not shown) in which the published cases and our cases were divided on the basis of same or different light-chain restriction showed statistically similar disease-specific median survival rates of 22 months and 28 months, respectively. Notably, these survival rates are shorter than the 36- to 48-month survival seen in de novo plasma cell myeloma.^2,10  Whether this apparent shortening of survival reflects a true change in the biology of PCM occurring on a background of CLL/MBL needs further study.

In summary, we have presented the clinicopathologic findings in 15 cases in which the bone marrow showed simultaneous presence of CD5⁺ monoclonal B cells and monotypic plasma cells. We have compared our results with the 51 cases of concurrent CLL and PCM previously reported in 30 articles. The distinction between CLL and the newly described precursor entity “monoclonal B-cell lymphocytosis” is discussed. Differentiating cases with concomitant presence of PCM and CLL/MBL from other causes of lymphoplasmacytic or lymphoplasmacytoid marrow infiltration is crucial for correct treatment. A careful review of immunophenotype of the B cells and immunohistochemical staining for CD56 and/or cyclin D1 to identify abnormal plasma cells will allow separation from conditions such as LPL, splenic MZL, and CLL with plasmacytic differentiation. Notably, despite the presence of the 2 abnormal populations, almost all of our patients, and many patients from published reports, required treatment for myeloma rather than for CLL.

We thank Steven Conlon, AAS, Photopath Division, Department of Pathology, Duke University Medical Center, for expert help with the Figures.