Diagnostic Advances in Multiple Myeloma

With the exception of non-secretory MM, the identification and quantification of a monoclonal protein (M-protein) in the serum or urine is essential to the diagnosis and monitoring of MM. The standard serologic evaluation of a patient with suspected myeloma should include serum electrophoresis and immunofixation (SPEP/IFE) and serum free light chains (FREELITE, The Binding Site Ltd, Birmingham, U.K.; FLC) [4]. These tests can identify a monoclonal protein in the majority of patients; however, further testing is needed in certain situations. Although uncommon, IgD and IgE MM should be considered in patients suspected of MM with light chain only MM identified in those who have normal to elevated total protein levels or an M-protein on SPEP that is not able to be identified by routine immunofixation for IgG, IgA, or IgM. In some laboratories, such scenarios result in reflex testing for IgD and IgE.

The 24-h urine PEP/IFE is also part of the standard diagnostic evaluation for and monitoring of MM; however, this is a cumbersome test that is not always reliable, particularly, due to incomplete collection [5]. In a large retrospective analysis of 428 patients with plasma cell dyscrasias at the Mayo Clinic, they found that using SPEP/IFE and serum FLC identified the monoclonal paraprotein in 99.5 % of patients, and only in 0.5 % of patients was the 24-h urine sample required to make the diagnosis [6]. Similarly, in an analysis of 116 patients with advanced MM, using the combination of SPEP, IFE, and FLC, had 100 % sensitivity in identifying the M-protein, and values correlated with 24-h urine protein levels over time [7]. If confirmed in other studies, this suggests 24-h urine collections may not be routinely necessary, although it should still be done in all patients for evaluation of renal dysfunction or nephrotic syndrome in the setting of a monoclonal plasma cell disorder.

Another study evaluated the use of spot protein to creatinine ratio as a surrogate for 24-h urine total protein in patients with newly diagnosed MM, which is an accepted alternative in many other forms of chronic kidney disease. They found spot protein to creatinine ratios to be highly sensitive and specific with a receiver operating characteristic of 0.86 [8]. A shortcoming of this study is that it only examined total protein, not UPEP or IFE, and the electrophoretic pattern is important from distinguishing light chain cast nephropathy from amyloid and from other unrelated forms of proteinuria. An alternative approach could involve using a spot urine protein electrophoresis and creatinine. The spot UPEP would provide a urinary M-protein and other components, principally albumin, and the ratio of each component to spot urine creatinine could be used to estimate the 24-h M-protein, 24-h albumin, etc. Additional studies are required to validate this approach. Free light chain measurement in the urine using the FREELITE assay has not been well validated and should not be performed outside of the context of a clinical study [9].

Conventional serologic tests are reliable in most patients, but the M-protein, particularly IgA, does not always migrate in the γ region, and can be obscured by other proteins, particularly transferrin or complement C3 in the β region or the acute phase reactant haptoglobin in the α2 region [10]. This can result in an uninterpretable SPEP or even an SPEP that does not appear to have a monoclonal component. In 2009, a novel heavy/light chain nephelometric assay was developed (HevyLite, The Binding Site Ltd, Birmingham, U.K.; HLC), which use antibodies with epitopes that span the heavy chain/light chain junction and quantify each of the intact immunoglobulins and their ratios (IgAκ/IgAλ, IgGκ/IgGλ, IgMκ/IgMλ) [11].

Several studies have assessed the diagnostic performance of HLC in MM. Wolff et al. examined HLC in 68 patients with IgA monoclonal gammopathies, half of whom were unable to be quantified by SPEP or FLC due to migration outside of the γ region. In this selected population, the sensitivities of the quantitative SPEP, FLC, and HLC were 50, 49.3, and 95.6 %, respectively, when compared to the qualitative standard of immunofixation, and HLC was able to quantify patients unquantifiable by SPEP [12]. Other studies have found similar results in IgA and IgG MM with bands migrating outside of the γ region, including some patients with normal quantitative IgA but abnormal HLC ratio due to suppression of the uninvolved IgA (e.g., suppression of IgAκ in patients with IgAλ MM). [13]. Additionally, the HLC ratio may also be abnormal in patients with normal total IgA due to suppression of uninvolved IgA (e.g., suppression of IgAκ in patients with IgAλ MM) [13, 14•]. In these studies, HLC correlated with during the course of treatment; however, in a small group of patients, HLC took longer to normalize than SPEP when achieving CR, and HLC became abnormal prior to recurrence of M-protein at relapse suggesting that HLC may be more sensitive in monitoring disease than SPEP/IFE [14•].

To date, there have been no studies that have examined the utility of HLC specifically in plasma cell disorder (PCD) patients where the disease burden is difficult to quantitate by conventional testing (as defined by IMWG criteria). In this population, in addition to the HLC ratio, we are evaluating the use of the HLC difference (dHLC) of involved minus uninvolved isotypes (e.g., IgAκ minus IgAλ in patients with IgAκ MM), similar to the difference between the involved and uninvolved FLC (manuscript in process). Further studies are needed to determine the optimal use of HLC, including a reference for dHLC, as well as the use of HLC to characterize the depth of response.

As monoclonal antibodies are becoming available for MM, there will be new challenges in the quantification and monitoring of M-proteins. Daratumumab is a humanized IgG1κ monoclonal antibody against CD38 with promising activity in relapsed and refractory MM [15]. Daratumumab is given in concentrations that will create an M-protein on SPEP that can impair interpretation of the MM-related M-protein. Similarly, elotuzumab is an IgG1κ monoclonal antibody against SLAMF7 with promising activity in relapsed and refractory and ongoing studies in newly diagnosed patients, which runs on SPEP in the early γ region and confounds the interpretation of SPEP results [16]. Interference assays which use monoclonal antibodies against elotuzumab and daratumumab to alter its migration pattern on SPEP and confirm responses are in development (personal communication from manufacture re: elotuzumab) [17]. In clinical trials, daratumumab interference testing was performed when patients with IgGκ MM developed a response plateau with a persistent 0.1–0.2 g/dL M-protein or when patients with non-IgG MM developed a new 0.1–0.2 g/dL IgGκ M-protein to assess for clonal evolution. Alternative approaches to quantify disease in the presence of monoclonal antibodies include the use of HevyLite or IgG subclass testing; however, additional studies are required for both of these tests, and the baseline IgG subclass of the M-protein would need to be known in advance.

While osseous plasmacytomas with extension beyond bones can be diagnosed by imaging or biopsy, extramedullary disease at other sites can be more difficult to diagnose. There are not established criteria to diagnose central nervous system (CNS) or meningeal involvement, and the varying diagnostic criteria used in different case series make it difficult to determine the sensitivities and specificities of various diagnostic tests. MRI is frequently used to assess for brain parenchymal lesions or leptomeningeal enhancement, but is not always abnormal in patients with leptomeningeal disease that has been diagnosed on the basis of positive cerebrospinal fluid (CSF) cytology [18]. Similarly, in one of the larger case series of patients with CNS involvement of MM, only 70 % had positive CSF cytology [19]. Owing to the difficulty in diagnosing CNS involvement in MM, and the very poor prognosis it portends, there is a need for more sensitive diagnostic tests. In our series of 16 CSF specimens from nine patients with a clinical diagnosis of CNS MM, CSF free light chains were assessed. CSF FLC were highly concordant with a clinical diagnosis, and were more sensitive than both cytology and MRI in diagnosing CNS involvement with MM [20]. In all patients, the uninvolved FLC was undetectable in the CSF, suggesting the blood–brain barrier is intact even in patients with leptomeningeal disease and that there is not contamination from the serum. CSF FLC have been evaluated and found to be minimally present in normal controls, and they have also been validated in diseases such as multiple sclerosis, but, outside of this case series and other reports, they have not been well studied in MM [21].

Unlike cytology, flow cytometry, CSF immunofixation, or imaging findings, CSF FLC would offer a quantitative assessment of CSF involvement and may be able to be used to monitor quantitatively the response to treatment in the CSF, which may be discordant from the systemic response to treatment. While more studies are needed to validate CSF FLC and provide ranges of normal and abnormal values, the rarity of the condition would require collaboration of multiple myeloma centers of excellence, and these data support their use with careful interpretation. In practice, when there is concern for CNS involvement of MM, we perform an MRI of the brain or spine as clinically indicated and evaluated the CSF by cytology, flow cytometry, and FLC.

Pleural effusions are not rare in MM, but are frequently due to congestive heart failure, renal failure, or hypoalbuminemia; however, malignant pleural effusions due to MM involvement of the lung parenchyma or serosa are very rare and can be difficult to diagnose [22]. Cytology, flow cytometry, and immunofixation are also used malignant ascites in MM. The rarity of malignant effusions and ascites limits assessment of the sensitivity and specificity of cytology, flow cytometry, and immunofixation, and a limitation of cytology and immunofixation is that each of these tests are qualitative and cannot provide a quantitative assessment of tumor burden or disease monitoring [23, 24]. In a patient with very elevated serum M-protein, even a transudative effusion may have a monoclonal component identified by immunofixation.

We have retrospectively evaluated FLC in 11 patients with pleural effusions and found them useful in the diagnosis of malignant effusions. Unlike the CSF, in patients with elevated serum FLC and a non-malignant pleural effusion, there is an elevation in FLC in the effusion as well, as is the case with pleural effusions and other serum proteins such as albumin. We found the most useful method for identifying a malignant effusion was to use the principles of Light’s criteria and compare the pleural FLC to the serum FLC, which we performed using the ratio of the involved minus uninvolved FLC of the pleura to the serum (manuscript submitted). This approach could not be used with qualitative tests such as immunofixation. There is less published information regarding FLC in ascitic fluid, and we are exploring this in a similar approach to pleural effusions. FLC in ascitic or pleural fluid is a quantitative test that could potentially provide an assessment of extramedullary tumor burden at diagnosis and a quantitative marker of treatment response; however, more research is needed [25]. In our current practice, when there is a suspected malignant effusion or ascites, in addition to usual studies, we send flow cytometry, cytology, and FLC.

All patients with newly diagnosed MM will undergo bone marrow biopsy and aspiration. Immunophenotyping with flow cytometry or immunohistochemistry is an important technique in diagnosing MM. Owing to differential expression of surface proteins, the immunophenotype can help distinguish malignant plasma cells from normal or reactive plasma cells. Normal plasma cells express CD138, CD19, and CD45 and typically do not express CD38 or CD56. Plasma cells in MM express CD38, CD138, CD56, and CD117 but do not express CD27. The combination of these surface markers as well as the presence of light chain restriction on neoplastic plasma cells is usually able to distinguish normal plasma cells from MM plasma cells. Recent efforts have explored the use of other surface markers with differential expression in normal and malignant plasma cells. By examining newly diagnosed and relapsed patients with MM and normal controls, Muccio et al. found that CD150, CD 86, and CD200 could help further classify the plasma cells [26]. If this research is confirmed, larger flow cytometry panels may improve the specificity of flow cytometry in the diagnosis of MM. As monoclonal antibodies directed toward surface proteins continue to be developed and approved for treatment (e.g., daratumumab, elotuzumab), a more detailed immunophenotype of the malignant plasma cells at diagnosis and relapse may become important.

Flow cytometry is also being used for risk stratification of smoldering MM and also for detection of minimal residual disease and these topics are covered in articles in this issue. It is also important send some additional tests at diagnosis if they are intended to be used for monitoring in the future. If gene expression profiling is going to be used to risk stratify the patient prior to treatment, it should be sent at initial evaluation. Flow cytometry does not require any baseline evaluation for minimal residual disease monitoring, but if allele-specific polymerase chain reaction (PCR) with VDJ sequencing is going to be used to monitor minimal residual disease, then sequencing will need to be sent either at baseline or from stored formalin-fixed samples.

The traditional standard imaging technique for evaluation of bone disease in MM is the skeletal survey. To identify more sensitive tests to identify bone lesions earlier so treatment can be instituted to prevent complications, multiple other imaging techniques have been investigated. An initial study found whole body low-dose CT (WBLDCT) feasible in the diagnosis of lytic bone lesions with a rapid test time of less than 2 min and low radiation dose of 4.1 mSv, less than twice that of a conventional skeletal survey [27]. In multiple studies and a recent systematic review and meta-analysis, the sensitivity of WBLDCT has consistently been found to be >90 %. Up to 80 % more total lesions are found with WBLDCT than skeletal survey, and up to 30 % of patients with no bone lesions on skeletal survey would be newly diagnosed as having lytic bone lesions [28•, 29, 30]. Reliable testing is very important to obtaining consistent and accurate diagnosis, and the inter-observer agreement for WBLDCT is excellent with a κ statistic of 0.95, which could improve consistency across centers in the diagnosis of lytic bone lesions in MM [27]. Total testing time, including time interpreting by the radiologist, is significantly shorter with WBLDCT than skeletal survey, and the costs are similar [31].

¹⁸ F-Fluorodeoxyglucose-positron emission tomography-computed tomography (PET-CT) is able to evaluate for lytic bone lesions (CT portion) as well as extramedullary disease (PET to assess FDG avidity and CT for anatomic localization). Compared to the reference test of skeletal survey, PET-CT is also >90 % sensitive across multiple studies. PET-CT can diagnose 27–45 % more patients as having MM-associated bone disease than skeletal survey in meta-analysis. In addition to the improved detection rate of lytic bone lesions, PET-CT was also able to identify previously unrecognized extramedullary disease [28•, 32]. While PET-CT is able to identify more lytic lesions, some studies suggested that skeletal survey was more sensitive in identifying lesions in the ribs than PET-CT; however, this observation was not consistent across all studies [33, 34]. PET-CT may also provide a way to monitor response of patients with oligo- or non-secretory disease without requiring serial bone marrow biopsies during therapy. Importantly, PET-CT should include the entire skull and not just the skull base as it is frequently protocolled for other indications, and PET-CT should be performed prior to administration of dexamethasone, which may alter FDG uptake in MM lesions, and temporally separated from filgrastim administration, which may increase background marrow FDG avidity.

Magnetic resonance imaging (MRI) has frequently been used to evaluate for disease involvement in the spine and pelvis, but if MRI of the spine and pelvis is used in place of skeletal survey, up to 50 % of lesions would be missed [35]. This led to the use of whole body diffusion weighted MRI (WB-MRI). In comparison to skeletal survey, WB-MRI identifies significantly more lytic lesions, with the exception of the skull, and there is greater inter-observer reliability in identifying lesions [36, 37•, 38]. WB-MRI can also evaluate the bone marrow, and the marrow shows five different patterns of infiltration: normal appearance, focal involvement of lesions >5 mm, diffuse infiltration, diffuse infiltration, and focal lesions, or a variegated inhomogeneous pattern. The pattern and degree of marrow infiltration correlates with the degree of marrow plasmacytosis and other high-risk features [39]. WB-MRI is also capable of identifying extramedullary plasmacytomas, but the sensitivity of WB-MRI for this indication at different body sites is not well defined. Limitations of MRI are longer acquisition time of approximately 45 min and lack of availability specifically of WB-MRI at some institutions.

WBLDCT, PET-CT, and WB-MRI all have superior detection rates in diagnosing lytic bone lesions as compared to skeletal survey. They all also provide some degree of increased ability to diagnose extramedullary disease. A limited number of studies have directly compared these advanced imaging techniques. In a study of 41 patients with MM who underwent both WBLDCT and WB-MRI, 15 patients showed no involvement with either technique, four patients had identical involvement on both scans, 21 patients had involvement on both studies but more lesions on WB-MRI, and one patient had involvement on both studies but more lesions on WBLDCT. No patients were classified as not having lytic lesions on one technique and having lytic lesions on the other technique, but the increased detection rate on WB-MRI was statistically significant [40]. A similar study included 22 consecutive patients with newly diagnosed MM underwent PET-CT and WB-MRI at diagnosis and after treatment. PET-CT was positive in 18 patients and four patients had no lesions, but WB-MRI was positive in all 22 patients. However, in two of these patients with discordant results, the WB-MRI did not show evidence of lytic lesions, only diffuse marrow involvement [41].

The limited number of small studies available to compare these advanced imaging techniques precludes definitive conclusions regarding the superiority of any of these techniques. WBLDCT has advantages of both cost and time. PET-CT may be superior at diagnosing extramedullary disease and assessing residual disease posttreatment that can potentially impact prognosis [41]. WB-MRI is potentially more sensitive at diagnosing bone lesions and degree of marrow infiltration, but has long acquisition time and is more prone to motion artifact. Hybrid PET and MRI scanners are now available at some centers, and studies are underway to determine the potential utility of PET-MRI in a variety of malignancies [42, 43]. Due to the excellent sensitivity of MRI in the evaluation of bone lesions and bone marrow involvement, and the ability of functional imaging with PET to evaluate both extramedullary disease and whether residual sites of disease posttreatment are still active, PET-MRI is attractive option in MM, and we have ongoing studies evaluating the utility of PET-MRI.

If only a skeletal survey is performed at baseline and a more sensitive technique is performed later for other reasons (e.g., CT chest to evaluate for pulmonary embolism or pneumonia), any new findings may be misinterpreted as progression of disease, when they may have actually been present at diagnosis and not identified by the initial skeletal survey; thus, it is important to use more sensitive testing at diagnosis, particularly in patients with presumed SMM where radiologic findings would impact the management. PET/MRI may be most useful in patients who would benefit from both the advantages of PET in identifying extramedullary disease and the superiority of MRI in identifying spinal cord compression and candidacy for kyphoplasty. For example, a newly diagnosed patient with a high-risk gene expression profile with concern for extramedullary disease with severe lower back pain may be the ideal candidate for a PET-MRI. The IMWG currently recommends all patients with suspected MM or smoldering myeloma undergo evaluation with WBLDCT, PET-CT, WB-MRI, or spinal MRI [44].