Immunohistological analysis in diagnosis of plasma cell myeloma based on cytoplasmic kappa/lambda ratio of CD38-positive plasma cells

Abstract

The accurate determination of cytoplasmic immunoglobulin (cIg) light chain (LC) expression is important to differentiate reactive plasmacytosis from a clonal plasma cell neoplasm such as plasma cell myeloma (PCM). Through retrospective analysis, we studied the cytoplasmic kappa/lambda ratio of CD38-positive plasma cells in the bone marrow from 19 PCM patients and 19 controls. To demonstrate cIg LC expression, the bone marrow was immunostained for IgA, IgG, IgM, kappa, and lambda. The kappa/lambda ratio was defined as the ratio of the kappa-positive cell to the lambda-positive cell in plasma cells. PCM cells were distinguished from normal plasma cells by cut-off levels between 0.59 and 4.0, a sensitivity of 94.7%, and a specificity of 94.7%. The detection of the cytoplasmic kappa/lambda ratio of CD38-positive plasma cells may be a useful tool in the diagnosis of PCM and the correct diagnosis of PCM may be achieved more simply.

Keywords:

• Plasma cell myeloma
• Kappa/lambda ratio
• Immunohistology
• CD38

Introduction

Plasma cell myeloma (PCM) is a neoplastic disorder of the B-lymphocyte lineage associated with the bone marrow infiltrating plasma cells, producing monoclonal immunoglobulin. The diagnosis of PCM is made in combination with other laboratory and clinical findings, which comprises a variety of marrow abnormalities. To diagnose PCM, the criteria of World Health Organization (WHO) are most commonly used.¹ However, several laboratory parameters are required for a proper diagnosis. The diagnosis involves evaluation of the clinical burden of plasma cell infiltration, analysis of radiologically detectable bone lesions, electrophoretic determination of the monoclonal immunoglobulin, and assessment of plasma cells in the bone marrow or extramedullary tissue. Recent technological advances in the detection of plasma cell pathology reshapes the way we define plasma cell disorders as their clinical relevance becomes clearer.^2–4 However, morphological analysis of the bone marrow remains the ‘gold standard’ for quantifying the volume of medullary plasma cell infiltration and assessing the degree of plasma cell dysplasia. Through retrospective analysis, we investigated the diagnostic value of an abnormal cytoplasmic kappa/lambda ratio of CD38-positive plasma cells based on immunohistopathological assessment. We found the diagnostic range for the kappa/lambda ratio to maximize the diagnostic sensitivity and specificity, and to minimize the false-positive and false-negative results.

Materials and methods

Case and control samples

We selected the patients with PCM. We retrospectively analyzed records of the 19 patients, and the collected bone marrow samples had CD38-positive plasma cells, which were demonstrated on immunohistological analysis. All patients were diagnosed with PCM according to the WHO criteria between 2000 and 2010.¹ There were 9 males and 10 females; their ages ranged from 45 to 85 years (mean ± SD, 65.05 ± 10.69 years). The monoclonal component was IgA-type (n = 6), IgG-type (n = 11), and Bence-Jones-type (n = 2) myeloma. The monoclonal component Ig chain was kappa type (n = 12) and lambda type (n = 7) in the study cases.

In the control group, the bone marrow samples were obtained from 19 patients who were either healing from inflammatory disorders or were healthy individuals who had initially been suspected of having a hematological disease. We retrospectively analyzed records of the patients and collected the bone marrow samples with CD38-positive plasma cells demonstrated by immunohistological analysis. There were 12 males and 7 females; their ages ranged from 32.0 to 87.0 years (mean ± SD, 59.21 ± 18.48 years). Their serum and urine monoclonal component was undetected by immunofixation electrophoresis.

Histopathology

Tissue was fixed in 10% formalin, embedded in paraffin sections (5 µm thick), and stained with hematoxylin and eosin (HE), periodic Acid-Schiff, Giemza, and Gomori's silver impregnation. The histological sections were examined by two independent pathologists.

Immunostaining

Plasma cell infiltrates in all biopsies and smears were morphologically assessed with reference to the work of Bartl et al.⁵ The morphological features of plasma cells were assessed with the help of immunostaining. Immunohistological analyses were performed on formalin-fixed, paraffin-embedded tissue sections mounted on glass slides using antibodies directed against the following: CD38 (1:100; SPC32, monoclonal mouse anti-human CD38; Novocastra Laboratories, Ltd, Newcastle upon Tyne, UK), IgG (1:40 000; polyclonal rabbit anti-human IgG; Dako, Glostrup, Denmark), IgA (1:10; polyclonal rabbit anti-human IgA; Nichirei, Tokyo, Japan), IgM (1:4; monoclonal mouse anti-human IgM; Nichirei), kappa (1:40 000, polyclonal rabbit anti-human kappa; Dako), and lambda (1:80 000; polyclonal rabbit anti-human lambda; Dako). The tissue sections (5 µm) were deparaffinized in xylene, rehydrated in a graded series of ethanol solutions, and washed with a solution of Tris-buffered saline (TBS). Endogenous peroxidase activity was quenched by immersion in a 1:4 solution of 3% hydrogen peroxide in methanol for 20 minutes, followed by several rinses in TBS. The slides were placed in Coplin jars with 10 mM citrate buffer (pH 6) and heated in a water bath filled with distilled water at 95°C for 40 minutes to unmask the antigen. After heating, the jars were allowed to cool for 20 minutes. The slides were incubated with the primary antibody for 30 minutes at room temperature. These slides were rinsed in TBS (three times for 10 minutes) and incubated with the secondary antibody (EnVision Detection Kit; Dako), according to the manufacturer's instructions. The slides were rinsed in TBS (three times for 10 minutes) and stained with 3,3′-diaminobenzidine tetrahydrochloride included in the kit. The sections were counterstained with Mayer's hematoxylin solution. Negative controls were run in parallel, replacing the primary antibody with the antibody diluent. These tissue sections were evaluated using light microscopy to determine the positivity. Positive cells were counted in five low-power (×20) fields for each tissue section. The number of immunoreactive cells was counted. The positive cell ratios of IgA, IgG, IgM, kappa, and lambda to the plasma cell population were analyzed. The kappa/lambda ratio was defined as the ratio of kappa-positive cells to lambda-positive cells in plasma cells. More than 20% positivity of the plasma cells was judged as indicating positivity for the purposes of the present study.^6,7

Statistical analysis

The number of observations, mean, standard deviation, median, minimum, and maximum for continuous variables, and the number, percentage, sensitivity, and specificity for categorical data were calculated. All analyses were performed using SPSS software (ver. 18 for Windows; SPSS Japan Inc., Tokyo, Japan) software. In our data consisting of the kappa/lambda ratios of 19 patients and 19 healthy individuals, the kappa/lambda ratios of the healthy individuals were limited to a narrow range from 0.0106 to 4.00, while those of the patients were distributed over a wider range from 0.0005 to 975.8. Thus, by choosing two levels, l and u, with l < u, to distinguish patients from controls, we could create a frequency table with two rows and three columns (Table 1) and define (n_b + n_d + n_f)/38 as the ‘error rate’ for disease diagnosis using the kappa/lambda ratio. We decided that the values for both the lower and upper cut-off levels would be those with the lowest error rate among the error rates of 4000 frequency tables created by moving l (0–0.99 by 0.01) and u (1.1–5.0 by 0.1), independently. We also computed the sensitivity and specificity of these cut-off values. All procedures were performed using R statistical software (http://www.Rproject.org).

Table 1. Result of frequency table in staistical analysis

	x≦l	l < x≦u	u < x	Total
Multiple myeloma	na	nb	nc	19
Control	nd	ne	nf	19

Results

In the patients with PCM, the positive cell ratios with cytoplasmic immunoglobulin (cIg) were: 22.00–70.00 (mean ± SD, 37.33 ± 17.62%) for IgA in cases of IgA type (n = 6), and 90.00−99.00% (mean ± SD, 97.36 ± 3.64%) for IgG in cases of IgG types (n = 11). The kappa/lambda ratio was 69.857−975.800 (mean ± SD, 325.819 ± 311.626) for kappa in cases of kappa type (n = 12), and 0.0005−1.1972 (mean ± SD, 0.1822 ± 0.4479) for lambda in cases of lambda type (n = 7). In controls, the positive cell ratios with cIg (IgA, IgG, IgM, kappa, and lambda) were: 0.01−22.00% (mean ± SD, 7.68 ± 7.04%) for IgA, 0.01−68.00% (mean ± SD, 9.47 ± 17.55%) for IgG, 0.01−37.0% (mean ± SD, 6.26 ± 8.56%) for IgM, 1.00−93.00% (mean ± SD, 19.11 ± 22.00%) for kappa, and 2.00−94.0% (mean ± SD, 18.00 ± 27.04%) for lambda. The kappa/lambda ratio was 0.0106−4.00 (mean ± SD, 1.99 ± 1.17).

In the 19 patients diagnosed with PCM, the samples of 18 patients exhibited kappa/lambda ratios of >4.0 and ≤0.59, respectively. The kappa/lambda ratio was >0.59 and ≤4.0 for the remaining patients with PCM. This patient was diagnosed with IgG PCM according to the WHO criteria. In the 19 controls, the samples of 18 controls exhibited kappa/lambda ratios of >0.59 and ≤4.0. The kappa/lambda ratio was <0.59 for the remaining control (Tables 2 and 3). PCM cells were distinguished from normal plasma cells by cut-off levels between 0.59 and 4.0, a sensitivity of 94.7%, and a specificity of 94.7%.

Table 2. Result of frequency table in statistical analysis

	x ≦ 0.59	0.59 < x ≦ 4.0	4.0 < x	Total
Multiple myeloma	6	1	12	19
Control	1	18	0	19

Table 3. Results of frequency table evaluation

	Control	Multiple myeloma
Kappa/lambda ≦ 0.59 or kappa/lambda > 4.0	1	18
0.59 < kappa/lambda ≦ 4.0	18	1

Discussion

The quantitative and qualitative morphological changes in the bone marrow are of great value in differentiating benign plasmacytosis from malignant plasmacytosis.⁵ However, reactive conditions with high plasma cell counts (up to 50%) and presence of immature cells have been reported. The morphological criteria used for the diagnosis of PCM are not always reliable since there is great variation in the appearance of individual marrow particles.⁸ Moreover, focalization or slight infiltration of the malignant process may be present in the early cases of PCM. PCM cells also show wide range of morphological appearances, resembling other hematocytes such as lymphoma cells and other blast cells.^9–11 Therefore, PCM cells cannot be differentiated from other hematocytes based only on morphology.

The detection of small plasma cell collections on HE-stained sections can be challenging, even for the experienced observer. This is particularly so when the quality of the HE sections is suboptimal. Immunohistological analysis ‘lights up’ interspersed and small collections of plasma cells, allowing a more straightforward assessment of plasma cell infiltration. The bone marrow aspirate count may underestimate the degree of bone marrow plasmacytosis when compared with immunohistological analysis.¹² It is well known that non-secretory myeloma can be detected by immunohistological analysis,¹³ and immunohistological detection of pathological plasma cell aggregates, atypia, and cytoplasmic light chain (LC) restriction may be fundamental in diagnosing.¹⁴

On their differentiation pathway to plasma cells, normal B lymphocytes cease to express characteristic B-cell surface antigens such as surface immunoglobulin (sIg), CD10, CD19, CD20, or HLA-DR. They acquire plasma cell markers: cIg and the CD38 surface antigen.¹⁵ CD38 is the most typical plasma cell marker. Neoplastic cells retain some of the phenotypic characteristics of normal plasma cells, such as CD38 expression.¹⁶ Thus, the analysis of CD38-positive plasma cells could prove to be efficient in the diagnosis of the malignant plasma cell disorder PCM.

In the diagnostic process for detecting PCM, it is sufficient to quantify serum-free LCs (SFLCs) and kappa/lambda ratios in patients suspected of plasma cell dysplasia from clinical, biochemical, or serum protein electrophoresis (SPE).¹⁷ In particular, in some patients with non-secretory myeloma, SFLC assays have allowed the detection of monoclonal proteins that were previously undetectable.¹⁸ Commonly agreed-upon decision limits of <0.26 and >1.65 for abnormal kappa/lambda ratios have been used.¹⁹ However, measurement and interpretation of SFLCs is not without problems, including deciding which reference range or decision limits to use.^19,20 In addition, a recent study using commonly agreed-upon decision limits reported a considerable number of false-positive kappa/lambda ratios when all sera sent for SPE analysis were screened. An abnormal kappa/lambda ratio was a sensitive, but not very specific, marker of monoclonal gammopathy. Polyclonal immunoresponse and renal impairment was identified as the main causes of false-positive kappa/lambda chain. A recent study has reported the positive predictive value as 62% for an abnormal kappa/lambda ratio.¹⁷

The accurate determination of cIg LC expression is important to differentiate reactive plasmacytosis from a clonal plasma cell dysplasia. The former is caused by polyclonal proliferation of plasma cells, while the latter results from clonal proliferation of plasma cells showing LC restriction. Currently, several methods are available for measuring cIg LC, such as immunohistochemical stain, flow cytometry, and immunofluorescent stain.²¹ The LC restriction was defined as a kappa/lambda ratio >4.0 or <0.5 based on commonly used criteria.^22–24 However, these studies showed cIg LC restriction in patients with B-cell lymphoma and B-cell chronic lymphocytic leukemia, not in PCM. Only one report has proposed the threshold values for the ratio of cytoplasmic kappa to lambda LCs in plasma cells for detecting monoclonality, in addition to the criteria for discriminating between PCM and benign reactive plasmacytosis using immunohistological analysis.²⁵ This study determined the cytoplasmic kappa/lambda ratio in plasma cells to be 0.4−3.5 in reactive plasmacytosis, and <0.2 or >11.1 in PCM. However, this study did not show the decision limits of an abnormal kappa/lambda ratio in the diagnostic process. Moreover, the plasma cell infiltrate in all biopsies and smears was only morphologically assessed with HE staining, and therefore the detection of plasma cells may be ambiguous.

To the best of our knowledge, this is the first published report on the cytoplasmic kappa/lambda ratio in CD38-positive plasma cells by an immunohistological analysis, showing a diagnostic rate of ≤0.59 and >4.0 with a sensitivity of 94.7% and a specificity of 94.7%. Pathological analysis is the cornerstone for quantifying the volume of medullary plasma cell infiltration and assessing the degree of plasma cell dysplasia in establishing the diagnosis of PCM. Furthermore, the immunohistological analysis was easily performed. Therefore, this method may improve the accuracy of the diagnostic kappa/lambda ratio directly shown in plasma cells, and could be a useful tool in the diagnosis of PCM.