Monoclonal gammopathy of undetermined significance, new diagnosis. After selective gating to include viable cells and exclude doublets, plasma cells are gated based on their bright CD38 and dimmer CD45 (panel A). There is an increased proportion of plasma cells present (normally plasma cells comprise just a few percent of bone marrow leukocytes). Panels B–D highlight the other immunophenotypic changes of this atypical population. The cells also express CD138 (panel B), with dim to absent CD19 (panel C, gated in black), and predominantly lacking CD56 (panel C). The neoplastic plasma cells are color-gated in black and demonstrate cytoplasmic kappa immunoglobulin light chain restriction (panel D). Nonneoplastic plasma cells are depicted in red (kappa expressing) and green (lambda expressing)
Plasma cell myeloma, new diagnosis. After selective gating to include viable cells and exclude doublets, plasma cells are gated based on their bright CD38 and dimmer CD45 (panel A). There is an increased proportion of plasma cells present (normally plasma cells comprise just a few percent of bone marrow leukocytes). Panels B–D highlight the other immunophenotypic changes of this atypical population. The cells also express CD138 (panel B), with dim-to-absent CD19 (panel C), bright aberrant CD56 (panel C), and cytoplasmic lambda immunoglobulin light chain restriction (panel D)
Light chain expressing plasma cells in a patient with nonsecretory plasma cell myeloma. Clinically, this patient lacked an M-protein and was deemed to have nonsecretory myeloma. Though 3 % of myeloma patients are nonsecretory, over 85 % of the non-secretors have detectable immunoglobulin protein that has impaired secretion. This patient has an increase in plasma cells based on the CD38 versus CD45 plot, and a good proportion of these cells have dim-to-absent CD45 (panel A). The cells also express partial CD138 (panel B), with dim-to-absent CD19 and lack of CD56 (panel C). The plasma cells uniformly have cytoplasmic kappa immunoglobulin light chain restriction (panel D)
Immunoglobulin-negative (non-producing) plasma cells in a patient with plasma cell myeloma. This patient has typical immunophenotypic features of plasma cell myeloma, including bright CD38, dim CD45, partial CD138, and dim CD19 (panels A–C). However, the patient’s plasma cells predominantly lacked CD56, completely lacked cytoplasmic light chains, and the patient’s serum did not have an M-spike. When this patient presented to our laboratory, we repeated the cytoplasmic immunoglobulin stains multiple times, using different combinations of anti-kappa and lambda immunoglobulin antibodies (polyclonal and monoclonal, in different fluorochromes) and confirmed that the cells lacked light chain expression.
Plasma cell leukemia in a young man. This patient presented with circulating blast-like cells. The flow study presented contained numerous blast-like cells by morphology, and by flow cytometry, they fall into the traditional “blast gate” based on dim CD45 and low side scatter (SSC) (panel A). In the initial workup by flow cytometry, the cells lacked CD19, as well as cytoplasmic CD3, cytoplasmic CD79a, and cytoplasmic MPO (not shown). The cells lacked CD117 and CD15 (panel B), but expressed CD13 and partial CD33 (panel C). Though the cells had some myeloid antigens, including CD13 and CD33, it was not a perfect fit for a myeloid leukemia. Due to the bright CD38 staining, additional markers were obtained to evaluate for plasma cell neoplasm. The cells express bright CD38 and CD138 (panels D and E), but lacked CD56 (panel E), and were restricted for cytoplasmic lambda immunoglobulin light chains.
Flow-cytometric immunophenotyping usually performed on bone marrow specimens obtained for the evaluation of PCN is one of the most important laboratory studies that can provide semiquantitative and phenotypic assessment of plasma cells that has diagnostic, prognostic, and predictive utility. In addition, flow cytometry is sensitive enough to detect clonal circulating plasma cells (cPC) in blood also to define high-risk disease. A recent study by Mayo clinic , suggests that presence of clonal cPC in blood is a marker of high-risk disease in all stages of newly diagnosed monoclonal gammopathy. The cPCs were discriminated from polyclonal/normal PCs based on the differential CD19 and CD45 expression. The cPCs detected were reported as the number of clonal events/150,000 collected total events. For those samples where less than 150,000 events were gated or examined, the number of final clonal events was adjusted to 150,000 events. The lower limit of cPC detection by this method is 20 cells/150,000 (0.013 %). Thus, according to this study, 400 cPC was considered as the optimal cutoff for defining high-risk disease and also associated with higher plasma cell proliferation and adverse cytogenetics .
Predictive and prognostic implications
The number of nonneoplastic plasma cells as a percentage of total plasma cells at diagnosis is an important flow cytometric assessment in patients with symptomatic plasma cell myeloma, since patients with more than 5 % normal plasma cells at diagnosis have a significantly lower frequency of high-risk cytogenetic abnormalities, a higher rate of response to treatment and an overall favorable baseline clinical prognosis . A study by Perez-Persona et al.  found that at 5 years, the risk of progression of MGUS and asymptomatic plasma cell myeloma to symptomatic plasma cell myeloma was 25 % and 64 %, respectively, in cases where 95 % of all bone marrow plasma cells had an aberrant phenotype on flow cytometry, at diagnosis . As previously suggested, these counts may be affected by multiple factors, including cell loss during processing. Recent studies, however, have shown that despite a generally lower yield of plasma cell counts on flow cytometry, there remains a significant positive correlation between the counts obtained by morphology and flow cytometry, and the count of bone marrow myeloma cells obtained by flow cytometry is an independent prognostic factor for overall survival .
In addition to their diagnostic utility, flow cytometric expression of CD45, CD117, CD28, CD20, and/or CD56 has been found to have prognostic and predictive implications. Studies have found higher percentages of CD45 + myeloma cells in MGUS and AMM compared to symptomatic and relapsed cases, and a lower percentage of CD45 + in patients with bone lesions and high-grade angiogenesis . CD45 positivity in MM, therefore, correlates with better overall survival (CD45 +, 39 months vs. CD45 −, 18 months) . CD117 is expressed with equal intensity in both MGUS and MM and is associated with better progression-free and overall survival in MM  . Conversely, CD28, when expressed, is bright in MM compared to MGUS, and its expression correlates with the presence of t(11;14), t(4;14), deletion of 17p and 13q, and non-hyperdiploid karyotype, and is considered an aggressive phenotype . Cases with expression of both CD28 and CD117 on multivariate analysis have bad prognosis comparable to CD28 + cases, when compared to favorable prognosis associated with CD28− CD117 + cases . Similarly, though CD19 is expressed in 5 % of myelomas , in a study by Mateo et al., its presence has been shown to have lower progression-free and overall survival in univariate but not in multivariate analysis . About 10–30 % of myelomas express CD20 and tend to have small, mature plasma cell morphology and with the presence of t(11;14) and overexpression of cyclinD1 protein . The loss of CD20 during the disease course in cases that were previously positive at diagnosis correlated with poor survival , but the impact of cyclin D1 protein overexpression is unclear. While few studies show that cyclin D1-positive cases have prolonged survival, others have not been able to show any significant correlation [26, 27]. An interesting example of CD20-positive myeloma that showed t(11;14) is shown in Fig. 6. Although lack of CD56 expression in MM has been shown to correlate with fewer lytic bone lesions  and CD56 positivity is associated with presence of neoplastic plasma cells in circulation, conflicting studies  regarding flow cytometric expression of CD56 and prognosis in MM undermine its utility as a prognostic marker .
CD20-positive plasma cells in a patient with t(11;14) associated plasma cell myeloma. This patient has an increased proportion of plasma cells with increased CD38 and dim to absent CD45 (panel A). The plasma cells have dim CD138, but a majority has bright CD20 expression (panel B). The plasma cells have dim-to-absent CD19 (panel C), supporting that they are neoplastic plasma cells (rather than B cells with plasmacytic differentiation). They have partial dim CD56 expression (panel C) and are kappa monotypic (panel D)
Recent observations has shed some light about the prospective identification of patients at risk of showing unsustained CR (those cases with high-risk cytogenetics and persistent minimal residual disease) when compared to patients failing to achieve CR but with an indolent disease course because the latter patient group showed MGUS-like signature at baseline by gene expression profiling and are associated with 10-year survival rates after HDT/autologous stem cell transplantation (ASCT) despite displaying a lower incidence of CR [29, 30]. The prospective identification of this signature may contribute to discriminate patients with a suboptimal response that require additional treatment from a residual “MGUS-like component” that may remain stable without further treatment; therefore, investigation of new biomarkers and easy-to-perform assays for routine diagnostic laboratories could potentially contribute to identify such MM patients. In a recent study  , the potential utility of a MFC immunophenotyping-computerized algorithm based on the simultaneous assessment of the tumor burden and the degree of clonality of the bone marrow plasma cell compartment to identify symptomatic MM patients with an MGUS-like profile (8 % of total MM patients, showed a lower frequency of CD81-positive clonal PC) at baseline are associated with long-term disease control, irrespective of the depth of remission achieved after treatment.
Utility in monitoring residual disease
Another use of flow cytometry is in the detection of minimal residual disease (MRD) after the institution of chemotherapy or transplantation (an example is illustrated in Fig. 7). The final achievable goal of current plasma cell myeloma therapy is to control disease , making MRD testing an important tool in management of MM because achievement of MRD negativity is considered a powerful predictor of favorable outcome . The rates of MRD negativity predict the efficacy of anti-myeloma drugs . Recently, the benefit of patients with high-risk cytogenetics requires the addition of bortezomib, a proteasome inhibitor . Bortezomib has been shown to overcome, at least in part, the negative impact of cytogenetic factors, whereas results with lenalidomide have been more mixed. However, the extended use of bortezomib may be limited by peripheral neuropathy . In a recent study, carfilzomib, a selective proteasome inhibitor, with less peripheral neuropathy which was studied as a single agent in relapsed and/or refractory myeloma showed no evidence of cumulative toxicity after extended treatment and with high rates of deep remission and MRD negativity (assessed by 10-color MFC assay) among these patients .
Low-level residual disease in a polyclonal background. The neoplastic plasma cells are highlighted in black, and the polyclonal plasma cells are highlighted in red (kappa) and green (lambda). The neoplastic plasma cells in this case have dimmer CD38; therefore, they are harder to identify on CD38 vs. CD45 plots. Thus, in panel A, we use both CD38 and CD138 to select the plasma cell populations for further interrogation. In panel B, the neoplastic and nonneoplastic plasma cells express CD138, but the neoplastic plasma cells have slightly dimmer CD38 and form a more cohesive group. CD19 and CD56 can easily distinguish the neoplastic from nonneoplastic plasma cells (panel C). The neoplastic plasma cells have dim-to-absent CD19 and bright aberrant CD56, and are kappa-restricted (panel D). These gating strategies can help identify and enumerate both the clonal and non-clonal plasma cells present in the marrow sample
Studies have depended on flow cytometry-based detection of tumor cells to define MRD, but other studies have used allele specific oligonucleotide PCR (ASO-PCR) based approaches for MRD detection with majority showing equivalent ability of the two techniques to detect residual disease, but practical differences in terms of feasibility [36, 37]. ASO-PCR has the inability to obtain successful primers in up to a third of the patients, and the absolute requirement for a baseline sample hampers universal adaptation of this technique . These, along with the required expertise and the more universal access to flow cytometry in most of hematology laboratories, makes flow cytometry the method of choice today, and increasing ability to interrogate millions of cells in a short time and increasing number of markers that can be accommodated at a time has led to increasing sensitivity, specificity, and ease of use of MFC technique. Though the typical recommendation for a discrete population of cells comprising MRD in flow cytometry immunophenotyping assay is 20 neoplastic cells, the European Myeloma Network-2008 recommends that at least 100 neoplastic cells should be acquired for accurate enumeration to reduce the coefficient of variation. Thus, to achieve a sensitivity of 0.01 % abnormal events, at least 1 × 106 total events must be evaluated, distributed in single or several tubes . Newer studies suggest collecting 5 × 106 events (after bulk lysis) to achieve a sensitivity of 0.0005 %. Assessment of MRD by flow cytometry is becoming particularly important, since patients without MRD at day100 following autologous stem cell transplant have significantly better progression-free and overall survival than those with MRD  . Analysis of the number of neoplastic plasma cells as a percentage of total plasma cell count rather than as a percentage of total leukocyte count is, therefore, recommended by some, since hemodilution would likely cause a proportionate reduction in the numbers of both the normal and the clonal plasma cells . A study by Gupta et al. has shown that a neoplastic plasma cell index (percentage of neoplastic/nonneoplastic plasma cells) of less than 30 on flow cytometry might further be of potential use in differentiating a complete treatment response (immunofixation negative) from a partial response (immunofixation positive)  .
One of the challenges in assessing MRD by flow cytometry is “antigen shifts” in neoplastic cells following treatment . These changes include gain or loss of immunoreactivity for a previously unexpressed antigen or changes in the intensity of an antigen expressed by the clone pre-treatment. The shifts are most often seen in CD45 in terms of gain of expression and brightness of intensity followed by gain of CD19 and CD20 . Interestingly, Bataille et al.  demonstrated a cellular model for myeloma cell growth and maturation based on an intraclonal CD45 hierarchy by in vitro studies using normal and myeloma plasma cells to indicate that bright CD45 expression is seen on the most actively proliferating myeloma cells but these differences in expression change with therapy and have not been confirmed by clinical FC. As shown by Morice et al. , myelomas contain CD45-positive and CD45-negative subsets and that the CD45-positive subset has proliferative activity, while the CD45-negative compartment is resistant to apoptosis; thus, modulation of CD45 expression is seen in treated myelomas. Any panel including all these markers, as suggested previously, might therefore be able to account for changes in the immunophenotype. Finally, expression of cytoplasmic κ and λ alone is not adequate for the assessment of MRD, especially when the number of monoclonal plasma cells is low and can be obscured by a background population of normal polytypic plasma cells, or when the disorders are bi-clonal with subsets positive for κ- and λ-Ig light chain, respectively  .
Thus, to summarize the major difficulty in defining MRD is that there remains a considerable heterogeneity in methodology (4C vs. 6C vs. 8–10C FC or ASO-PCR or sequencing), major differences in the antibody panel used to distinguish abnormal plasma cells from normal plasma cells, a major discordance in the number of events (varied from 100,000 to 4,000,000 events collected), the cutoffs for MRD positivity (varied from 20 to 50 cells comprising a population) and even when to begin monitoring . On the other hand, MRD cutoffs are not easy to define clinically, as they are dependent on the disease phenotype, molecular/cytogenetic characteristics of the disease, and newer and better advancements in the therapeutic agents. Therefore, certain methods of MRD measurement may be applicable to some patients with myeloma, but not to others, thus making it difficult to propose uniform guidelines using the current methodologies . In an earlier study, Paiva et al.  compared MRD by immunofixation analysis (IFX), serum-free light chain ratio (sFLC), and immunophenotyping by four-color MFC in 102 patients enrolled in the GEM05.65 year trial. Three 4-color combinations (CD38/CD56/CD19/CD45, CD38/CD27/CD45/CD28, and ß2 microglobulin/CD81/CD38/CD117) were used to detect phenotypic aberrancies in PCs. Seven percent of patients with no MRD by MFC (sensitivity ≤ 10−4–10−5) after induction therapy (i.e., immunophenotypic response) remained IFX positive initially, although all subsequently became IFX negative. Discrepant results were common among all methods tested . Those patients with immunophenotypic response had significantly increased PFS and time to progression compared to those with complete response (CR being IFX negative and < 5 % PCs on BM biopsy) or stringent complete response (sCR defined as CR and normalization of the sFLC ratio); however, no OS benefit was noted. In another study by Rawstron et al. , the prognostic value of MRD assessment using MFC in patients with MM treated in the MRC (Medical Research Council) Myeloma IX trial was investigated after induction therapy (n = 378) and at day 100 after autologous stem cell transplantation (ASCT; n = 397) in intensive-pathway patients and at the end of induction therapy in non-intensive-pathway patients (n = 245). In intensive-pathway patients, the absence of MRD at day 100 after ASCT was highly predictive of a favorable outcome. This outcome advantage was demonstrable in patients with favorable and adverse cytogenetics and in patients achieving immunofixation-negative complete response. MRD assessment after induction therapy in the non-intensive-pathway patients did not seem to be predictive of outcome . Thus, MRD assessment by MFC was predictive of overall outcome in patients with myeloma undergoing ASCT. In a recent paper by Martinez-Lopez et al. , the investigators demonstrate that detectable disease at a much lower level of detection (10−6) can be achieved by deep sequencing using high-throughput approach (the Lympho SIGHT™ platform, which relies on amplification and sequencing of immunoglobulin gene segments using consensus primers and high-throughput IGH sequencing) for MRD testing than that achieved by ASO-PCR or MFC in which the limit of detection can be achieved close to 10−5. Thus, sequencing can measure depth of response at least 1 log lower than MFC. Additionally, the use of sequencing technology for MRD detection adds an additional advantage by providing a window into the biology of the remaining tumor cells as one can identify what the potential mutations are present in the residual disease and use this information to target specific “add on” strategies to further drive down or cure the disease burden .