Screening and Diagnosis of Monoclonal Gammopathies: An International Survey of Laboratory Practice

In April 2016, a voluntary supplemental questionnaire was distributed to 923 laboratories as part of the 2016 A mailing of a College of American Pathologists proficiency testing survey for electrophoresis (survey ELP). Survey questions were designed as multiple-choice responses, permitting one or more selections by the participant depending on the question. Open-ended choices (eg, “Other, specify”) were included in some questions. These qualitative questions were not recategorized into quantitative tabulations of multiple-choice question responses. They are included in the Results section, as applicable, with subtabulation of thematic content across responses. Questions that allowed participants to select multiple responses were tabulated according to total number of responses, as well as the percentage of laboratories that selected individual response items. Percentage totals were rounded to 1 decimal place. Differences between US and non-US responses were assessed using the χ² test with a significance level of P < .05. All summaries and analyses were performed with SAS 9.3 (SAS Institute, Cary, North Carolina).

A total of 774 laboratories from 38 countries and regions completed the questionnaire (84% response rate). Most respondent laboratories (85.7%) were located in either the United States (n = 614) or Canada (n = 49). Survey respondents who provided a facility type description (n = 760) were from voluntary nonprofit hospitals (33.9%; n = 258), university hospitals (16.4%; n = 125), regional/local independent laboratories (except clinic or group practice and not owned by national corporation[s]; 13.2%; n = 100), city/county/state hospitals (12.9%; n = 98), national/corporate laboratories (owned by national corporation[s]; 8.9%; n = 68), veterans' hospitals (6.4%; n = 49), proprietary hospitals (5.9%; n = 45), Army/Air Force/Navy hospitals (1.2%; n = 9), and public health nonhospital facilities (1.1%; n = 8). The distribution of institution types (academic versus nonacademic) did not differ between US and non-US participant laboratories (P = .16).

Questions directed toward understanding methods used by participant laboratories are included in Table 1. A total of 68.6% of respondents (520 of 758) reported using AGE methods for SPE, whereas 31.4% (238 of 758) reported using CZE methods (Table 1, question A). Use of SPE methods differed between 600 US laboratories (AGE 72.5% [435]; CZE 27.5% [165]) and 158 non-US laboratories (AGE 53.8% [85]; CZE 46.2% [73]) (P < .001). A total of 71.7% of respondents (536 of 748) used IFE for clonal characterization, 7.5% (56 of 748) used IS-CE, and 17.8% (133 of 748) did not reflexively use either method when a monoclonal band was observed by AGE or CZE (Table 1, question B).

When an M protein was identified by either AGE or CZE methods, the vast majority of respondents (77.3%; 521 of 674) used a perpendicular-drop method of M-spike quantification from the electropherogram (Table 1, question C; see Figure, A and B, for illustration). A much smaller number of participants used a tangent-skimming approach for quantification (9.1%; 61 of 674). Method of quantification differed slightly between US (perpendicular drop 80.2% [425 of 530]; tangent skimming 7.4% [39 of 530]) and non-US (perpendicular drop 66.7% [96 of 144]; tangent skimming 15.3% [22 of 144]) laboratories (P = .007). Distribution of 28 responses in the “other” category for this question included primarily different combinations of the multiple-choice items listed. A total of 65 participants reported that quantitation was not performed.

More than half (55.0%) of respondents (410 of 745) provided or suggested additional testing for possible IgD or IgE when reactivity to only κ or λ light chains was observed on routine clonal characterization, and 45.0% of respondents (335 of 745) did not test for IgD or IgE in that situation (Table 1, question D). This practice of providing (or suggesting) testing for IgD or IgE reactivity differed slightly between US (53.0%; 312 of 589) and non-US (62.8%; 98 of 156) laboratories (P = .03). Few respondent laboratories (4.0%; 30 of 745) offered any tests that could detect the presence of monoclonal therapeutics (Table 1, question E).

Questions directed toward understanding testing approaches are included in Tables 2, 3, 4, and 5. The most common strategy used at participant institutions in the initial screening for the presence of a monoclonal immunoglobulin was SPE (AGE or CZE) with reflex to IFE or IS-CE (39.3% [299 of 760]; Table 2, question F). Serum protein electrophoresis was included in answer choices selected by nearly all participants (95.8%; 728 of 760), with variation directed toward the content of additional reflexive or panel-based testing (Table 2, question F). Practice differences were observed between US and non-US laboratories. The order of the 2 most frequent responses was reversed between US (SPE with reflex to IFE/ISE 42.8% [257 of 600]; SPE only 16.3% [98 of 600]) and non-US (SPE with reflex to IFE/ISE 26.3% [42 of 160]; SPE only 29.4% [47 of 160]) laboratories (P < .001). Practice patterns were also analyzed for participants who responded to question F and also provided a facility type designation in our survey. Serum protein electrophoresis with reflex to IFE/ISE was the most frequent response chosen by respondents at nearly all facility types (Table 3). Although SPE only was the most frequent response chosen by participants at national/corporate laboratories (owned by national corporation[s]), SPE with reflex to IFE/ISE was still the second most frequent response in that setting (20.6%; 13 of 63; data not shown).

The majority of participant laboratories (60.2%; 462 of 767) did not provide panels of these tests for clinician ordering convenience (Table 2, question G). As the combination of test choices listed in question F may not fully represent laboratory practice across all facilities, a “fill all that apply” version of this question (“Which testing does your laboratory perform for new cases of multiple myeloma to detect serum monoclonal proteins”) was also included in the survey (Table 2, question H). This allowed additional data analyses (Table 2, question H) based on the number of responses and the percentage of participant laboratories. Additionally, a matrix that displays the most common combination of tests selected (Table 4) was created using response data from Table 2, question H. Serum protein electrophoresis (88.1%; 656), IFE (79.1%; 589), and sFLC (34.4%; 256) were used in the highest percentage of 745 laboratories (Table 2, question H). The combination of SPE and IFE was chosen by participants most frequently (36.9% [275 of 745]; Table 4), followed by SPE, IFE, and sFLC (26.6%; 198 of 745), and SPE alone (11.9%; 89 of 745). Of the 47 “other” responses to question H, 19 mentioned that quantitative immunoglobulins (eg, IgG, IgA, IgM) were also commonly ordered for new patients.

A forced-choice question (Table 2, question I) was included to determine which test was primarily used by laboratories to follow patients with monoclonal gammopathy. Most of the 724 participants chose SPE (55.7%; 403), followed by IFE (18.9%; 137) and sFLC (11.7%; 85). However, 84 respondents selected “other.” Of this subset, 50 listed a combination of multiple tests, most of which (44) also included SPE.

Participants were also asked which methods were most frequently ordered for monoclonal free light-chain testing and/or monitoring (Table 2, question J). Serum free light-chain (48.4%; 355), urine IFE on 24-hour urine (47.1%; 345), and urine IFE on an early-morning void (43.5%; 319) were used in the highest percentage of 733 laboratories. Of the 70 “other” responses, 19 noted that testing was performed on a “random” urine specimen, whereas 25 noted that free light-chain testing was not performed in house. The “fill in all that apply” format of this question also permitted creation of a matrix displaying the most frequently selected combination of methods used for free light-chain testing (Table 5).

Interpretation of test results (eg, SPE, IFE, and IS-CE) at most participant laboratories is conducted by pathologists who have MD degrees (72.3% [540 of 747]; Table 6, question K). A smaller number of the 747 participant laboratories noted that interpretations were conducted by medical laboratory scientists (11.9%; 89) or scientists with PhD degrees (8.2%; 61). Open-ended text in the 44 “other” responses made thematic or specific reference to interpretation by those with MDs (n = 30), those with PhDs (n = 19), and medical laboratory scientists and/or technologists (n = 19), often in collaboration (eg, “medical laboratory scientist does interpretation, pathologist reviews”). Practice differences in who performed interpretations existed between US and non-US laboratories. For example, MD pathologists were far more likely to perform interpretations in US laboratories (80.7%; 486 of 602) than in non-US laboratories (37.2%; 54 of 145) (P < .001).

A majority of participant laboratories (66.6%; 499 of 749) kept a file on all known monoclonal gammopathy patients to assist in review (Table 6, question L). Most laboratories (76.9%; 576 of 749) did not cancel IFE orders on previously characterized specimens with M proteins (Table 6, question M). This was true even if the M protein had the same migration as the originally described M protein (78.2% [577 of 738] did not cancel; Table 6, question N). Trends in the 73 “other” responses to question M included 13 responses noting that intervals of time were used in deciding upon cancellation (eg, “if IFE has been done in last 30 days”), 10 responses noting that a change in results might be used in deciding whether an IFE is required (eg, “if change in electrophoresis pattern”), 20 responses noting that IFE was either nonorderable by clinicians or done reflexively, 7 responses noting that IFE testing might be performed based on ordering clinician specialty (eg, “hematology”), patient location, or specific clinician request, and 4 responses noting that IFE testing was performed at the discretion of the pathologist. Immunofixation electrophoresis order cancellation on specimens with previously characterized M proteins was less common in US (18.6%; 109 of 585) than in non-US (34.0%; 52 of 153) laboratories (P < .001). More than half of participant laboratories (59.6%; 442 of 742) did not comment on whether an M protein in a follow-up had “increased, decreased, or not changed” from a previous quantitation (Table 6, question O).

Agarose gel electrophoresis remains the most common method of SPE used by participant laboratories, although CZE methods are gaining widespread adoption, particularly in non-US laboratories. It should be noted that a limitation of the survey design is that laboratory size and focus (eg, inpatient versus outpatient) was not assessed and therefore could not be used to assess practice variations in facilities of different sizes or practice types. The distribution of methods reported by respondents in our questionnaire (Table 1, question A) is concordant with analytical methods reported in the 2016 ELP-A proficiency testing challenge, the survey that contained our questionnaire. For example, the proficiency testing event included ∼66% AGE and ∼34% CZE methods for the total γ globulin challenge.¹⁴  This provides supportive evidence that the voluntary questionnaire responses accurately reflect current laboratory practices by participants. CZE methods are gaining a more widespread adoption by clinical laboratories, as studies have shown that they provide an analytically acceptable alternative to AGE for most routine analysis.¹⁵  Additionally, CZE methods and instrumentation are more automated (eg, specimen/bar code traceable) than traditional gel-based techniques.

The survey also showed that laboratories are conducting clonal characterization most commonly by IFE as opposed to IS-CE (Table 1, question B). This distribution is also roughly concordant with analytical methods reported in the 2016 ELP-A proficiency testing challenge (∼81% IFE, ∼3% immunoelectrophoresis, and ∼16% IS-CE).¹⁴  Slightly higher percentages in the challenge likely reflect the fact that respondents who answered no to survey question B would not be participating in that particular component of the proficiency testing event, as they were presumably not offering IFE, immunoelectrophoresis, and/or IS-CE methods in house. IS-CE methods provide a more automated alternative for clonal characterization than IFE, although some performance differences between methods exist, particularly in the detection and characterization of small or atypical M proteins.^16,17 

The vast majority of participants reported using a perpendicular-drop approach to M-protein quantitation, and a smaller percentage use tangent skimming (Table 1, question C). Perpendicular drop includes the entire area under the M-protein gate on the electropherogram, whereas tangent skimming provides one way to exclude potential underlying polyclonal immunoglobulin from the overall M-protein quantitation (Figure 1).^7,18  Much smaller percentages of participants reported using quantitative immunoglobulins, sFLC ratios, or heavy/light chains as methods of M-protein quantitation. Not offering quantitation was noted by a minority of laboratories (8.8% of respondents; 65 of 739), although the survey did not assess whether quantitation was available in these settings as a send-out test. M-protein identification without quantitation does not provide adequate information to clinicians in establishing diagnoses or evaluating the outcome of therapeutic intervention. Because of the differences in actual measurement with these techniques, one should use the same technique when following patient M proteins.

There are several scenarios in which IFE and/or IS-CE results may be misleading or inaccurate. One such scenario is IgD and IgE myeloma,¹⁹  as routine IFE/IS-CE testing does not include antisera to detect δ or ɛ heavy chains. Only 55.0% of participant laboratories provide or suggest additional testing to exclude these rare disorders when only free light chains are observed by routine testing (Table 1, question D). This suggests that IgD and/or IgE M proteins may go mischaracterized as free light chain disease in a subset of patients.

Growing use of monoclonal antibody therapeutics (MATs) for clinical care has led to increased detection of these drugs as interferents on AGE, CZE, IFE, and IS-CE methods.^20–22  Monoclonal antibody therapeutics such as daratumumab (human anti-CD38 IgG₁-κ)²³  and elotuzumab (humanized anti-SLAMF7 IgG₁-κ)²⁴  are now being used for treatment of multiple myeloma and may be detected as false-positive bands on electrophoresis.²⁵  Although a method to detect one MAT (daratumumab-specific immunofixation electrophoresis reflex assay) has recently been developed,^26,27  few respondents reported that they offer any methods for the detection of MATs. Laboratories should consider the effect that MATs may have on test interpretation and downstream patient care. Medication history and clinical notes may also serve to identify possible MATs in facilities where that information is available (eg, via the electronic health record) to those responsible for interpreting test patterns and results.

Serum protein electrophoresis is used by virtually all participant institutions, and most often as a reflex to IFE or IS-CE. It is interesting to note that most facilities do not provide panels for ordering convenience, although this may be intentional to prevent overuse that can occur with panel-based testing. Use of sFLC testing for screening and monitoring is also quite common. Urine studies are still widely used in the detection and assessment of free light chain disease, although recent studies support the value of sFLC assays over urine studies in many circumstances.²⁸  Light-chain testing is another area of laboratory medicine that is evolving, and laboratories should work together with ordering providers to perform the most appropriate testing for their patients.

The present survey demonstrated that a variety of test combinations are used for screening and follow-up of patients with monoclonal disorders (Tables 2, 3, 4, and 5). When asked about the most common approach used in screening for the presence of an M protein in an initial evaluation (Table 2, question F), 6.1% mentioned the combination of SPE, serum IFE, and sFLC for screening. A larger percentage (11.8%) included quantitative immunoglobulins (IgA, IgG, and IgM) in addition to the testing above. In a separate “fill in all that apply” question to select which testing was used to detect serum M proteins (see Table 2, question H), the matrix analysis (Table 4) revealed a larger percentage of respondents (26.6%) who chose the test combination of SPE, serum IFE, and sFLC, although it should be noted that quantitative immunoglobulins were not included as a choice in question H, a limitation of the present questionnaire. An additional 2.1% chose the 3-test combination but replaced IFE with IS-CE, and 2.0% chose the 3-test combination with the addition of an “other” test. A limitation of Table 2, question H, and Table 4 that cannot be excluded in our analysis is the possibility that some respondents may have interpreted the question as asking which tests were available on their respective test menus, and not which tests were commonly performed, which was the intended goal of the question.

Another limitation of the current survey is that not all combinations of panel and/or reflexive testing were listed in individual question options, and not all laboratories performed all components of test panels or options. Different-sized facilities (particularly smaller laboratories) may not have the capacity or staffing to support an extensive menu of specialized testing. Respondents also may not have expertise or knowledge of send-out tests (eg, Table 2, question J; “other” included 25 respondents who noted that they did not perform free light-chain testing in house), and respondents may have not incorporated send-out tests into the responses for questions in Tables 2, 3, 4, and 5. These factors may influence responses to questions related to test practices in these areas.

Our survey demonstrated that most tests were interpreted by MD pathologists (Table 6, question K), although medical laboratory scientists, PhD scientists, and nonpathologist physicians were often also involved. Analysis of open-ended responses suggested that interpretation may also have been conducted as a collaborative effort, with initial interpretation by a medical laboratory scientist and final review and interpretation conducted by an MD. The predominance of MD interpretation in US facilities may be influenced by requirements associated with Medicare Part B professional component billing.²⁹  Differences in MD versus PhD interpretations may also reflect differences in the number and/or distribution of physicians and pathologists versus clinical chemists in practice (or assigned to these tasks), although this is difficult to assess with available data.^30–32 

Most laboratories kept a file on all known monoclonal gammopathy patients to assist with interpretation. The survey did not specifically address other methods of accessing prior test information (eg, databases of results or images associated with electrophoresis instrumentation, electronic health record lookup); thus, the percentage of laboratories with access to historical patient information may actually be higher than reported. Regardless, fewer reported actually using that information to comment on whether an M protein had increased, decreased, or not changed since a previous sample. This would require clinicians to trend results in the patient chart or electronic health record to make that determination for the purpose of clinical care. Harmonization of reporting structure and interpretative comments, as well as interfacing of results to the laboratory information system, could provide better clarity and more consistent information to clinicians.^33,34  Approximately one-third of laboratories reported not keeping such files of known monoclonal gammopathy patients. Without the ability to refer to prior information about the migration and shape of the M protein, it would be difficult to detect changes in migration that could suggest a second M protein, monoclonal antibody therapy, or even a mislabeled sample.

Regarding potential cancellation of testing, most laboratories did not cancel follow-up IFEs even if the M protein migrated to the same position on SPE as a previous sample. The ability to cancel testing may be limited by compliance and billing considerations.³⁵  In many settings, the laboratory may be considered a “service” that must follow all order requests based on a fee-for-service approach. In a more practical sense, lack of access to clinical information may also hinder the ability to integrate meaningful utilization strategies. Additionally, patient enrollment in clinical trials (with corresponding test requirements) may also conflict with optimization strategies that are otherwise intended to limit unnecessary testing. Test cancellation practices in the clinical laboratory were the focus of a recent College of American Pathologists Q-probes study, which emphasized the importance of having defined policies and procedures.³⁶ 

Extraordinary advances have been made in the detection and treatment of monoclonal disorders such as multiple myeloma, and these advances are linked to appropriate diagnostic testing. All cases of multiple myeloma are preceded for many years by a monoclonal gammopathy of undetermined significance demonstrated by SPE, IFE, and/or sFLC testing.³⁷  Treating high-risk cases of smoldering multiple myeloma in patients who are clinically asymptomatic improves both progression-free and overall survival of these patients.³⁸  Availability of current novel therapies in additional to stem cell transplantation has added years of survival for patients with these disorders.³⁹  Because early detection and monitoring of monoclonal gammopathy of undetermined significance, smoldering multiple myeloma, and multiple myeloma are heavily dependent on optimal use of the clinical laboratory,^6,40,41  we hope that describing baseline differences in current practices, as evident in our survey, may serve as an impetus for clinical laboratories and providers to consider and/or adopt evidence-based guidelines for monoclonal gammopathy testing as appropriate for their setting and practice characteristics.

A limitation of the present survey, however, is that questions were designed to primarily focus on the use of electrophoretic techniques for the diagnosis and management of monoclonal disorders. Serum protein electrophoresis and urine protein electrophoresis have also been used for decades to aid clinicians in identification of a wide variety of other conditions that alter electrophoretic patterns. For example, SPE has been used to report acute-phase reactions, cirrhosis, nephrotic syndrome, α₁-antitrypsin deficiency, and other clinically relevant findings.⁴²  In many of these areas, more sensitive and specific assays are now available, but may not be considered in the absence of recognizing an unusual electrophoretic pattern. Further studies may be required to assess the ongoing use of electrophoretic testing for such purposes.

In conclusion, the present survey results provide an overview of current laboratory and clinical practice in the context of monoclonal gammopathy ordering and testing. These results suggest significant variation across facilities. Further efforts at harmonization may help to streamline and simplify ordering and testing practices and create greater clarity for clinicians who use this information in patient care.

The authors would like to thank Mu Shan, MS, for statistical assistance in the initial survey analysis. Drs Genzen, Murray, Abel, Meng, Baltaro, Rhoads, Delgado, Bashleben, and Ansari are members of the College of American Pathologists Diagnostic Immunology Resource Committee.