Screening studies using serum free light chain analysis

From Wikilite

23.1. Introduction

Whether or not to measure serum free light chains (sFLCs) as part of an initial diagnostic request for “possible myeloma - please investigate” is an important issue. Traditionally, serum protein electrophoresis (SPE) and/or serum immunofixation electrophoresis (sIFE) tests have been performed first, sometimes alongside immunoglobulin (Ig) G, IgA and IgM measurements. SPE can detect intact immunoglobulin monoclonal proteins in the 1-5g/L range (Chapter 6) which is more than adequate to identify all intact immunoglobulin multiple myeloma (IIMM) patients. Some centres screen with IFE as it is ten times more sensitive. However, occasional monoclonal FLCs are missed, and IFE is non-quantitative. Furthermore, many low-level monoclonal gammopathy of undetermined significance (MGUS) proteins are detected that are unlikely to progress to malignancy, provided sFLC levels are also normal ^[1] (Chapter 19). A few laboratories measure total serum κ and λ in the initial screen, but this is clinically inadequate (Chapter 6).

Because of these deficiencies, urine protein electrophoresis (UPE) has been performed alongside serum tests, but many patients with nonsecretory multiple myeloma (NSMM), AL amyloidosis and other FLC-associated disorders are still missed. Furthermore, only a minority of patients have accompanying urine samples (typically 5-25% in the UK). It is therefore logical to test for sFLCs on receipt of the first blood sample. The results will also provide baseline values for subsequent disease monitoring.

Hence, a strategy of using SPE/sIFE combined with sFLC analysis “up-front” allows identification of all clinically significant monoclonal gammopathies, facilitates risk assessment for disease progression in MM, MGUS, smouldering (asymptomatic) multiple myeloma (SMM), plasmacytomas etc., and supports appropriate clinical decisions for monitoring patients.

This chapter discusses the current and potential screening options for identifying monoclonal proteins. As a general rule, intact monoclonal immunoglobulins can be identified using serum electrophoretic tests while monoclonal light chain diseases should be identified using sFLC assays. The combination of the 2 procedures produces good diagnostic accuracy and urine FLC analysis is not required at initial clinical presentation (Chapter 24).

23.2. Diagnostic protocols for monoclonal gammopathies

Table 23.1. Approximate diagnostic sensitivity of tests for monoclonal gammopathies. (See relevant clinical chapters for details).

The accuracy of seven different diagnostic protocols used for identifying monoclonal gammopathies is shown in Table 23.1. The rationale for each protocol is as follows:

1. SPE identifies all patients with IIMM. By definition, these patients have at least 10g/L of monoclonal protein, which is much greater than the sensitivity of the assays (1-5g/L). The test fails to identify approximately 30% of light chain multiple myeloma (LCMM) patients (the other 70% have sFLC bands or hypogammaglobulinaemia) and all patients with NSMM (Chapter 9). Approximately 50% of AL amyloidosis patients are abnormal by SPE (Chapter 15) . Many intact immunoglobulin MGUS individuals are also identified.

2. The combination of SPE and sIFE is more sensitive but fails to identify a few patients with LCMM and all patients with NSMM. Although 70% of patients with AL amyloidosis are identified, those producing FLCs only are frequently missed. Many more MGUS individuals are identified but monoclonal proteins of less than 1-5g/L are of little clinical consequence if sFLCs are normal (Chapter 19).

3+4. The addition of UPE/urine IFE (uIFE) to either of the above protocols identifies the remaining 30% of patients with LCMM and many patients with AL amyloidosis. Few additional MGUS patients are identified, as they rarely produce sufficient monoclonal FLCs to exceed the renal threshold and enter the urine (Chapter 19). Also, since serum tests for FLCs are clinically more useful than UPE tests, these protocols are illogical (Chapter 24).

5. sFLC analysis identifies FLC monoclonal gammopathies but cannot be used alone. It is, by definition, a test for FLCs and not for intact monoclonal immunoglobulins. 5% of patients with IIMM, 10% with SMM and 30-60% with MGUS do not have excess FLC production (see Chapters 11, 14 and 19).

6. This sensitive protocol includes SPE to identify all the intact monoclonal immunoglobulins, and sFLC immunoassays to identify all the monoclonal FLCs. Approximately 20% of NSMM patients will have no detectable monoclonal protein, using this strategy or by any other serum or urine tests (Chapter 9). Such patients are nonproducers or non-excretors and may have mutations in the light chain DNA causing protein shape distortion. Nearly all patients with AL amyloidosis and light chain deposition disease (LCDD) are identified (Chapters 15 and 17). Some additional MGUS individuals who produce monoclonal sFLCs are also identified (Chapter 19), as are patients with renal impairment (Chapter 20).

7. The addition of sIFE to protocol 6 is not likely to identify any more patients with MM. It identifies approximately 2-10% additional AL amyloidosis patients (Chapter 15), and also identifies some MGUS individuals with minor monoclonal intact immunoglobulins, who are unlikely to progress to overt disease if there is no accompanying abnormal FLC κ/λ ratio (Chapter 19). The extra cost, the inconvenience of performing the test and problems interpreting the clinical relevance of minor MGUS bands may not be justified in all clinical laboratories. A recent study by Katzmann et al. ^[2] (see below, screening study 12) concludes that the use of SPE plus FLC provides a simple and efficient initial diagnostic screen for the high tumour burden monoclonal gammopathies, and that urine studies and urine IFE can be ordered more selectively.

In the unlikely event that SPE/sIFE and sFLC tests are normal but the clinical picture still suggests a plasma cell disorder, urinalysis for FLCs is required. However, caution should be exercised when considering the clinical importance of minor urine FLC bands when sFLC tests are normal (see below and Chapter 24). After a monoclonal protein has been identified, IFE is required to characterise the heavy chain type. The FLC type can be identified by serum κ/λ ratios or from the IFE gels.

Table 23.2. Clinical and laboratory data in 16 patients (confirmed monoclonal lymphoproliferative diseases) that were normal by CZE but abnormal by sFLCs ^[3]. NI: Not indicated, B-CLL: B-cell chronic lymphocytic leukaemia, SLL: B-cell small lymphocytic lymphoma, LPL: lymphoplasmacytic lymphoma, MZL: marginal zone lymphoma. Case numbers refer to the patients also shown in Figure 23.1.

23.3. Screening studies using sFLC analysis

Figure 23.1 κ/λ ratios in 55 abnormal sera from a CZE and FLC screening study of 1,003 patients ^[3]. Sera with intact monoclonal immunoglobulins by CZE are shown in black or blue and samples containing only abnormal FLC κ/λ ratios are shown in yellow. Numbers against large symbols refer to additional patients with confirmed monoclonal gammopathy identified by FLC tests (Table 23.2).

Figure 23.2 sFLCs in 925 sera screened at Derby, UK. Normal results are the FLC levels in samples with normal SPE. (Courtesy of P Hill).

Figure 23.3 Test results on 370 urines screened for monoclonal proteins at a UK Hospital. Of 925 sera sent for screening, 370 (40%) had accompanying urine samples of which 126 were suspicious by UPE but were negative by uIFE. (Courtesy of P Hill).

Figure 23.4 Test results on 971 sera screened for monoclonal proteins at Nuneaton, UK ^[4]. Patient samples in the normal range have been excluded.

Twelve screening studies evaluating the use of sFLC assays, alongside other tests, as part of the initial diagnostic screen for monoclonal gammopathies, are described below:

1. Bakshi et al. ^[3] used a combination of CZE and sFLC analysis, studying 1,003 consecutive unknown samples. Of these 39 contained monoclonal proteins by CZE and 33 had abnormal κ/λ ratios, resulting in a total of 55 abnormal sera.

Sixteen samples (11 κ and 5 λ) were abnormal for monoclonal proteins only by sFLC analysis (Figure 23.1 and Table 23.2). Subsequent IFE showed that 5 of the 16 samples were abnormal, although some had barely visible bands, and 11 samples were clearly only abnormal by FLC assays. Of the 39 samples that were positive by CZE, 17 were associated with abnormal FLC κ/λ ratios. In every case, the FLC test results concurred with the light chain type of the intact monoclonal immunoglobulin.

Nine of the extra monoclonal gammopathy patients identified from abnormal κ/λ ratios were eventually shown to have plasma cell dyscrasias: 3 with MM, 4 with B cell chronic lymphocytic leukaemia (B-CLL)/small cell lymphoma, 1 with aplastic anaemia and 1 with lymphoma. Seven samples were not associated with significant diseases and were classified as FLC MGUS (Chapter 19).

Addition of the FLC analysis to the CZE test in this screening trial increased the yield of lymphocyte and plasma cell proliferative diseases by 56%. The authors noted that the study emphasised the utility of sFLC testing for patients with monoclonal proteins when CZE was essentially non-diagnostic.

2. Hill et al. ^[5] studied 923 consecutive serum samples referred to a UK District General Hospital, 71 of which had abnormal sFLC ratios (Figure 23.2). Eight new B-cell dyscrasias were detected among 43 patients with negative SPE, but abnormal sFLC κ/λ ratios. Two patients, with κ/λ ratios of 50 and 193, had κ FLC LCMM; 5 patients had FLC MGUS, and one a malignant lymphoma (but with no monoclonal band). Thirty-five patients with negative SPE had abnormal κ/λ ratios but no monoclonal diseases. This false positive rate for a ratio of >1.65 was associated with polyclonal increases in immunoglobulins and renal impairment and was higher than previously noted. Such minor abnormalities need to be considered in the context of reference ranges for hospitalised patients and renal diseases (Chapter 5 and Chapter 20).

In this study, the utility of serum and urine tests for screening for monoclonal FLCs was also compared. A total of 370 (40%) of the serum samples had been accompanied by urine samples. Of these, 141 (38%) had suspicious UPE tests and were further analysed by uIFE. Only 15 of these urine samples (4% of the total tested) had confirmed monoclonal FLC, indicating that 126 tests were performed unnecessarily (Figure 23.3). In 11 of the samples there was a corresponding sFLC abnormality; but the remaining 4 were normal. One of these 4 samples was associated with a serum IgGλ of 7g/L that was identified by IFE but had been missed on the initial SPE. The remaining 3 samples had urine FLC concentrations of approximately 50mg/L. One result normalised as the patient recovered from illness, another was subsequently interpreted as polyclonal urine FLCs and the third sample belonged to a patient who had persisting FLC proteinuria but was asymptomatic and had no evidence of a B-cell disorder from bone marrow biopsy. Hence, none of the “urine only” monoclonal FLCs was associated with clinically active disease.

On the basis of these results the hospital practice was changed to a policy of not performing urine tests when screening for monoclonal gammopathies. This allowed all patients to have proper assessments for monoclonal FLCs and improved operational efficiencies, at an increased cost of less than £5 per patient. The cost increase is largely due to testing all patients for FLCs, whereas previously urine samples were available in only 40% of cases.

3. Augustson et al. ^[6] studied 217 consecutive samples referred to a UK district general hospital. They found an extra 8 monoclonal gammopathies by abnormal sFLC κ/λ ratios. Three of these patients had LCMM that were missed by the routine SPE tests and resulted in serious diagnostic delays. Subsequently, the study was increased to 971 patients ^[4] and results were similar to those observed in other screening studies (Figure 23.4).

4. Abadie et al. ^[7] reported results on 312 consecutive samples from Seattle hospitals in the USA. sFLC tests identified a higher percentage of true positive and true negative samples and a lower percentage of false positive and false negative samples than SPE. The two assays in combination offered the most reliable results (Figure 23.5).

Figure 23.5 Combined assay performance of SPE and sFLC assays for identifying monoclonal gammopathies in a screening study ^[7]. PPV: Positive predictive value, NPV: Negative predictive value.

5. Foray and Chapuis-Cellier ^[8] studied 75 patients with FLC monoclonal gammopathies. They observed 6 patients who were normal by urine tests but had abnormal sFLC results, and concluded that the assays should be used whenever monoclonal FLC diseases were suspected.

6. Katzmann et al. ^[9] in a large retrospective study, made a direct comparison of the relationship between the diagnostic sensitivity of sFLC analysis and urine studies. A total of 428 patients with monoclonal urine proteins was studied, a cohort 10 times larger than in any other study (for details see Chapter 24) and clinical diagnoses had been established in all patients. The authors concluded that by adding sFLC analysis to sIFE, urine screening tests were no longer necessary. These results addressed the uncertainties that had been discussed by Katzmann in an earlier editorial in the Journal of Clinical Chemistry ^[10]. Furthermore, the data supported the diagnostic sensitivity of sFLC assays that Katzmann ^[11] had previously reported for assessing patients with known monoclonal gammopathies (see audit below).

7. Beetham et al. ^[12] in the UK, prospectively investigated 932 consecutive patients by SPE and sFLCs plus serum or urine IFE where appropriate. Of 449 patients who had serum studies only (because no urine samples were available), 53 had monoclonal proteins, but importantly, no significant diseases were missed. Three new cases of serum monoclonal FLCs only were detected but unfortunately no information was provided regarding their clinical relevance. The authors indicated that the results supported other published studies, in that SPE, sIFE and sFLC analysis could replace urine studies for detecting monoclonal gammopathies. The other 483 samples were investigated in the context of a paired urine study (Chapter 24.6).

8. Holding et al. ^[13] assessed impact on disease detection from 753 patient sera tested by SPE and FLC and 128 patient-matched associated urine samples. Sensitivity of FLC for BJP was 98%. Use of FLC in routine testing increased the number of monoclonal gammopathies detected by 7%.

9. Piehler et al. ^[14] evaluated, in a general hospital population, the combination of SPE and sFLC analysis, where additional clinical/biochemical information suggested a myeloma related disorder. Of the 332 patients investigated, 85 were diagnosed with a monoclonal gammopathy (MGUS, plasmacytoma, MM, WM and AL amyloidosis). The inclusion of the sFLC measurement detected 8 additional monoclonal gammopathies: 5 LCMM, 1 NSMM, 1 AL amyloidosis patient and 1 MGUS. One NSMM, one plasma cell leukemia and two plasmacytomas were not detected by the combined approach. The authors concluded that the combined approach is capable of detecting nearly all patients with clinically relevant monoclonal gammopathies. They also highlighted the number of borderline κ/λ ratios and thus the importance of considering additional clinical and laboratory parameters when interpreting sFLC results.

10. Robson et al. ^[15] examined the additional value of sFLC analysis on all 653 SPE requests over a 2-month period. The sFLC test was positive in 43 individuals, 17 of whom were also positive by SPE/IFE. Of the remaining 26 sFLC positive patients, 5 had previously been identified as CLL and 21 were being screened for M-proteins; one was an IgD myeloma patient, 1 an IgA MGUS and 1 had CLL. A further 14 had no other indication of haematological malignancy and ratios close to normal (0.2 – 0.25 or 1.7 - 2.0) and were not investigated further. The remaining 4 patients had more extreme ratios and were classified as FLC MGUS and annual follow up was recommended. Notably, the poor provision of urine samples (4.6%) meant that sFLC analysis provided the most effective means of determining monoclonal FLC production.

11. Vermeersch et al. ^[16] examined combinations of SPE, sFLC, sIFE and uIFE analysis on 833 consecutive patients being investigated for monoclonal gammopathy. Of these, 28 were diagnosed with malignant plasma cell disorders, 25 with B cell NHL and 156 with MGUS. sFLC was abnormal in 24 of the patients with malignant plasma cell disorders (15 IIMM, 2 LCMM, 3 AL amyloidosis, 2 WM, 1 plasma cell leukaemia, 1 osteosclerotic myeloma). Nine of the 25 B-NHL patients and 44 of the 156 MGUS patients had abnormal sFLC results. SPE and FLC (with follow up IFE) had a slightly higher sensitivity than SPE and uIFE (with follow up IFE) (81.8% vs 82.3%). The authors noted their observed specificity for sFLC (96.8%) to be similar to that reported by Hill et al., and highlighted the lower sensitivity of sFLC analysis for MGUS. This is of limited concern as risk stratification would suggest a lower risk of progression to a malignant plasma cell disorder for these individuals ^[1] (Chapter 19).

12. To date, the most extensive study evaluating screening strategies for the detection of monoclonal gammopathies was undertaken at the Mayo Clinic, with 1877 samples tested across 5 assay formats (SPE, UPE, sIFE, uIFE and sFLC), all within 30 days of initial diagnosis ^[2]. The cohort comprised 467 MM, 191 SMM, 524 MGUS, 29 plasmacytoma, 26 WM, 581 AL amyloidosis, 18 LCDD and 31 POEMS patients. The sensitivities for each testing combination across the separate plasma cell disorders are shown in Table 23.3. When comparing the combination of SPE and sFLC with SPE, sFLC and sIFE, 58 patients were missed (44 MGUS, 7 POEMS, 5 AL amyloidosis, 1 plasmacytoma and 1 SMM). However, no MM, WM or LCDD patients were missed. Furthermore the addition of the sFLC assay identified 23 AL amyloidosis patients, 6 MM and 1 LCDD that were not detected by the traditional panel of serum and urine tests. The omission of urine analysis from the testing panel missed 23 patients (15 MGUS, 1 extramedullary myeloma, 1 LCDD and 6 AL amyloidosis patients). As part of their conclusion, Katzmann et al. ^[2] stated that due to the small incremental sensitivity provided by urine studies and sIFE, the use of SPE with sFLC analysis provides a simple and efficient initial diagnostic screen for the high tumour burden monoclonal gammopathies such as MM, WM and SMM. They observed that urine studies and sIFE may be ordered more selectively.

Table 23.3. Sensitivity of monoclonal gammopathy screening panels (as shown by screening study 12 above ^[2]).

The results from the 12 studies described above are consistent. Adding sFLC assays to the SPE/IFE tests for routine screening (in the absence of urine studies), represents a sensitive combination and increased the tumour pick-up rate by approximately one patient per 100 samples tested. Some of the additional patients identified had significant changes to therapy resulting from their earlier diagnoses. It is of interest that many of the patients identified with this strategy had B-cell non-Hodgkin lymphoma or CLL (Chapter 18).

The diagnostic value of sFLC testing at the intial screening stage has recently been reviewed ^{[17][18][19][20]}. Furthermore, the International Myeloma Working Group (IMWG) has recently published specific guidelines on the use of sFLC analysis in diagnosis/screening, monitoring and prognosis (Chapter 25). These state that in the context of screening for the presence of MM or related disorders, the sFLC assay in combination with SPE and sIFE yields high sensitivity and negates the requirement for 24h-urine studies for diagnosis other than for AL amyloidosis ^[21].

23.4. Audit of sFLC usage

Katzmann et al. ^[11] performed an audit of sFLC analysis for the year 2003, on patients attending the Mayo Clinic (Table 23.4). Of the 1,020 patients, 88% had plasma cell disorders. All 120 patients without a plasma cell disorder had normal κ/λ ratios, despite evidence of a complex variety of different diseases. Thus, there were no false positive results in this study. The diagnostic sensitivity for AL amyloidosis by serum κ/λ ratios was 91% compared with 69% for sIFE and 83% for uIFE. Serum κ/λ ratios plus sIFE produced a sensitivity of 99% which was not improved by adding uIFE (Table 15.1). Of the 20 patients with NSMM, 14 had abnormal sFLC κ/λ ratios, including all 5 patients who had sFLCs measured at clinical presentation. The 6 NSMM patients with normal ratios had all received a peripheral blood stem cell transplant (PBSCT), and 5/6 had also achieved a complete bone marrow response. The conclusion of the study was that the performance of the FLC assays, in a prospective analysis, matched the results from published retrospective validation studies.

Table 23.4 Audit of sFLC requests at the Mayo Clinic for the year 2003. Sensitivity for detection by FLC analysis was available for some of the diseases.

References