Fax: 011 +81-2-478-0884
Polymerase chain reaction-based diagnosis of bone marrow involvement in 170 cases of non-Hodgkin lymphoma
Article first published online: 12 JUN 2002
Copyright © 2002 American Cancer Society
Volume 94, Issue 12, pages 3073–3082, 15 June 2002
How to Cite
Kang, Y. H., Park, C. J., Seo, E. J., Huh, J., Kim, S. B., Kang, Y.-K. and Chi, H. S. (2002), Polymerase chain reaction-based diagnosis of bone marrow involvement in 170 cases of non-Hodgkin lymphoma. Cancer, 94: 3073–3082. doi: 10.1002/cncr.10584
- Issue published online: 12 JUN 2002
- Article first published online: 12 JUN 2002
- Manuscript Accepted: 8 JAN 2002
- Manuscript Revised: 4 JAN 2002
- Manuscript Received: 23 SEP 2001
- non-Hodgkin lymphoma;
- bone marrow involvement;
- polymerase chain reaction;
- immunoglobulin heavy chain gene;
- T cell receptor gamma gene
Up to the current time, diagnosis of bone marrow (BM) involvement in non-Hodgkin lymphoma (NHL) has been based on morphologic findings. Polymerase chain reaction (PCR) for antigen receptor gene rearrangements has the potential to increase the detection sensitivity of minimal degrees of BM involvement. The authors therefore assessed PCR-based clonalities of BM concurrently with morphology from 170 cases with NHL and evaluated the usefulness of comparative analysis of clonalities between bilateral BMs and the lymph node and the clinical significance of PCR based clonalities of BM.
Bilateral BM clot sections of 170 cases and 47 lymph nodes were tested for immunoglobulin heavy chain gene rearrangement or T-cell receptor gamma gene rearrangement according to the B- or T-lineage of the lymph node.
When compared with morphology, the results of PCR showed an unexpectedly low positive concordance rate of 61.0% for B-cell NHL and 57.1% for T-cell NHL. When the clonality of BM was compared with that of lymph nodes in B-cell NHL, bilateral clonalities of BM showed high concordance with the clonality of the lymph nodes. PCR-based clonality did not show significant impact on survival.
Morphology remains the gold standard in the evaluation of BM involvement by NHL. Although the comparative analysis of BM clonality and that of the lymph nodes is considered a valuable tool that increases the reliability of clonality, PCR-based clonality of BM does not significantly add to the sensitivity of diagnosing BM involvement by NHL. Cancer 2002;94:3073–82. © 2002 American Cancer Society.
The role of bone marrow (BM) examination is well established in the staging and posttherapy assessment of non-Hodgkin lymphoma (NHL). The documentation of BM involvement results in a Stage IV clinical classification for the patients. Staging of disease as determined by BM involvement has a clear implication for survival, and the current therapeutic protocols may be directed according to the stage of disease.1, 2 Bone marrow examination contributes significant information in patients with diffuse large cell type NHL, which is the most common type of NHL in Asia,3 because the documentation of BM involvement is linked to worse outcome.4
Diagnosis of BM involvement in NHL has traditionally been based on morphologic findings. However, the evaluation of BM involvement in NHL by morphologic features alone has limitations. First, BM involvement in NHL is frequently patchy, hence, there may be a 10— 50% discordance rate between unilateral and bilateral BM examination.5 To improve the detection rate, immunophenotypic analysis has been studied, but it has been proven to be cost ineffective.6, 7 Positron-emission tomography with 18-F-fluorodeoxyglucose was potentially more sensitive than morphology8 but was impractical in routine practice. In addition to the limitation of low sensitivity, it is often difficult to distinguish benign lymphoid aggregates composed of small lymphoid cells from BM involvement in low-grade NHL.9 This problem is especially common in BM in elderly patients, the age group with the highest incidence of both benign lymphoid aggregates and NHL.10 Immunohistochemic stains have not been shown to have diagnostic value,7, 11 though one study suggested L-26 (CD20) as a helpful adjunct in evaluating marrow lymphoid aggregates.12 Differences in nuclear characteristics between lymph node and bone marrow biopsies have been subtle and often not appreciable.13 In contrast, three-dimensional reconstruction of BM biopsies have provided clear, unambiguous information, but this technique needs representative biopsy samples and skillful image processing.14
Molecular analysis to detect clonally rearranged immunoglobulin genes and T-cell receptor genes could theoretically help make a more sensitive and specific diagnosis of BM involvement in NHL. Molecular testing initially employed Southern blot (SB) analysis to detect these rearrangements.15 Such analysis of BM aspiration cells showed good concordance with the morphology as well as increased sensitivity and specificity.16, 17 Patients with evidence of BM involvement by SB without morphologic evidence showed significantly shorter disease free survival.17 But several disadvantages, such as the amount of time and labor needed, the requirement of large quantities of DNA, the reliance on the radioactive probes, and the relative insensitivity, made SB analysis unfavorable for routine diagnostic testing.18, 19 After the extraction of DNA from fixed, paraffin-embedded tissues was made possible, polymerase chain reaction (PCR) has been widely used as a tool to detect the clonality. The development of consensus primers for the hypervariable region of the immunoglobulin heavy chain (IgH) gene and T-cell receptor gamma (TCRγ) gene permitted the routine use of gene rearrangement (GR) studies in NHL and other B- or T-cell lymphoid malignancies.20, 21 However, BM involvement in NHL has just recently been investigated with PCR-based GR techniques,22–26 and the clinical significance of PCR sensitivity in the diagnosis of BM involvement in NHL has still not been determined.
The purpose of the current study was to compare the morphologic involvement and cytogenetic abnormalities of BM with the PCR-based clonality of BM and to evaluate the usefulness of comparative analysis of the clonality between BM and lymph nodes or extranodal tissue. In addition, the study also aimed to assess the clinical significance of PCR-based clonality of BM. Accordingly, we performed retrospective PCR-based molecular studies with paraffin-embedded BM tissues in 170 cases of NHL.
PATIENTS AND METHODS
Patient Characteristics and Morphologic Studies
One hundred seventy patients with NHL who were newly diagnosed or previously treated between 1997 and 1999 at Asan Medical Center were included in the current study. Clinicopathologic characteristics of the patient population are outlined in Table 1. Stages were defined according to the Ann Arbor staging system. Histologic classification of NHL was based upon the revised European-American lymphoma classification of lymphoid malignancies (REAL classification).27 The patients ranged in age from 3 to 84 years (median age, 54). There were 101 men and 69 women. Bone marrow aspiration and trephine biopsies were performed at the initial staging procedure in 130 patients, at the re-staging procedure in 33 patients, and at the followup status after the chemotherapy in 7 patients. Bilateral BM aspirations, clot sections, and biopsies were thoroughly reviewed for the presence or absence of BM involvement of NHL, patterns of involvement, and other characteristics, including fibrosis, coagulation necrosis, or hemophagocytosis. The patterns of BM involvement of NHL were classified as one of four categories: FP, focal paratrabecular; NO, nodular and interstitial; DI, diffuse interstitial; PM, packed marrow pattern. Medical records were reviewed for the presence of B symptoms, overall survival months, and the parameters of international prognostic index including age, serum lactate dehydrogenase (LD), performance, stage, and extranodal involvement.28
|Histology by REAL classification||Ann Arbor staging||Total|
|Diffuse large B-cell L||29||12||18||32||91|
|Marginal zone B-cell L||4||2||1||2||9|
|Mantle cell L||1||1||6||8|
|Follicular center L||3||1||3||7|
|Small lymphocytic L||5||5|
|High grade B-cell L||2||1||2||5|
|Precursor B-cell L||1||1|
|Peripheral T-cell L||2||3||9||14|
|Precursor T-cell L||1||7||8|
|Anaplastic large cell L||1||2||1||4|
|Intestinal T-cell L||1||1|
|Hepatosplenic γ/δ L||1||1|
Eighty eight BM aspirations were analyzed. The BM aspirations were disaggregated immediately in RPMI 1640 medium and were re-suspended in RPMI 1640 supplemented with 20% fetal calf serum at a concentration of 106 cells/mL. Direct or short-term (48 hour) unstimulated culture methods were used. Metaphase cells were arrested by exposing them to colcemid at a final concentration of 0.02 μg/mL for the last two hours before harvest; after hypotonic treatment with 0.075 M KCl for 15 minutes at room temperature, the cells were fixed with methanol and glacial acetic acid (3:1). Metaphase preparations were made either by an air- or steam-dry method. After aging slides for four to seven days at 37 °C, the trypsin G-banding technique with Giemsa stain was used for the analysis of karyotypes. The karyotypes were defined according to the International System for Human Cytogenetic Nomenclature;29 the rearrangements were regarded as clonal if at least two cells carried the same translocation or showed gain or deletion of a chromosome.
DNA extraction and β-globin PCR
Bone marrow clot sections were chosen for the clonality test since, in our institute, the clot sections had been reported to have the highest diagnostic sensitivity.30 Three hundred forty BM clot sections and 49 lymph nodes or extranodal tissues were obtained. All PCR reactions were performed on formalin-fixed, paraffin-embedded tissues as previously described. To assess the appropriateness of DNA extraction, the human β-globin gene was chosen to be amplified, since this gene is found in all human cells.
PCR protocol for immunoglobulin heavy chain gene rearrangement
Polymerase chain reaction for immunoglobulin heavy chain gene rearrangement (IgH-GR) was performed in 260 BM clot section specimens and 35 lymph nodes or extranodal tissues with B-cell lineage lymphoma, using standard methods.31 Briefly, the primer pair consisted of a universal VH primer based on an oligomeric sequence located at the 3' end of the FRIII and a consensus JH primer to amplify complementarity determining region-III (Table 2). The PCR cycles consisted of denaturation at 93 °C for three minutes, annealing of the primers at 64 °C for one minute, and then extension of the DNA at 73 °C for one minute. A total of 40 cycles were performed for each reaction. Each PCR round was preceded by a three minute denaturation at 93 °C and was followed by a final extension of three minutes at 73 °C. The reaction products were then run on a 2% agarose gel and examined under short wavelength UV light after ethidium bromide staining. Clonal bands ranged from 70 to 150 base pairs (Fig. 1).
|Immunoglobulin heavy chain gene rearrangement PCR primers31|
|FRIII||5′ CTGTCGACACGGCCGTGTATTACTG 3′|
|JH||5′ AACTGCAGAGGAGACGGTGACC 3′|
|T-cell receptor gamma gene rearrangement PCR primers32|
|V2||5′ CTTCCTGCAGATGACTCCTACAACTCCAAGGTTG 3′|
|V3||5′ CTTCCTGCAGATGACGTCTCCACCGCAAGGGATG 3′|
|V4||5′ CTTCCTGCAGATGACTCCTACACCTCCACCAGCGTTG 3′|
|V5||5′ TTCC TGCAGATGACGTCTCCAACTCAAAGGAT G 3′|
|V9||5′ GG(A/G/C/T)ACTGCAGGAAAGGAATCTGGCATTCCG 3′|
|V10||5′ CACTGCAGGCTCAAGATTGCTCAGGTGGG 3′|
|V11||5′ ACTGCAGGCTCAAGATTGCTCAGGTGGG 3′|
|V12||5′ ACTCTGCAGCCTCTTGGGCACTGCTCTAAA 3′|
|JTG12||5′ AAGTGTTGTTCCACTGCCAAA 3′|
|JGT3||5′ AGTTACTATGAGC(T/C)TAGTCCC 3′|
|JGT4||5′ TGTAATGATAAGCTTTGTTCC 3′|
PCR protocol for tcr gamma gene rearrangement
Polymerase chain reaction of TCRγGR was performed in 80 clot section specimens and 14 lymph nodes or extranodal tissues of T-cell lineage NHL.32 Unlike IgH genes, the TCRγ gene V segments differ from each other throughout their length, so different primers were used, as previously described (Table 2). These primers were used as two mixes: mix I, which was used for initial testing of samples and contained TCRγ V 2,3,4,8, and 9; and mix II, which was used for negative samples at the initial testing with mix I and contained TCRγ V 5,10,11, and 12. For the J segments, three primers were included: JGT12 for J1.3 and J2.3, JGT3 for J1.1 and J1.2, and JGT4 for J1.2. The PCR cycles consisted of denaturation at 94 °C for 30 seconds, annealing of the primers at 55 °C for 30 seconds, and then extension of the DNA at 72 °C for 30 seconds. A total of 45 cycles were performed for each reaction. Each PCR round was preceded by a 3 minute denaturation at 94 °C and was followed by a final extension of 10 minutes at 72 °C. The reaction products were then run on a 2% agarose gel and examined under short wavelength UV light after ethidium bromide staining. Clonal bands ranged from 170 to 230 base pairs (Fig. 2).
Serial dilution of 1,000 ng of DNA from an REH cell line or an H9 cell line with genomic human DNA from paraffin-embedded normal BM was performed to assess the minimal amount of sample needed for a clearly detectable clonal signal. The detection limit was reached when the DNA mixture contained 500 pg DNA of clonal DNA (the equivalent of 5 × 10−4 of total cell population) of the REH cell line used as a positive control for IgH-GR and 10 pg DNA (the equivalent of 1 × 10−5 of total cell population) of the H9 cell line was used as a positive control for TCRγ-GR.
The difference of the proportions between the groups of comparison was evaluated by the chi-square test. Univariate analysis of each parameter to predict the survival was carried out using the Kaplan-Meier product method.
Morphologic Studies of the Bone Marrow
Fifty five cases (32.4%) among 170 cases of NHL were diagnosed as having BM involvement of NHL by morphology; this included 41 cases (31.5%) of 130 B-cell lineage lymphomas and 14 cases (35.0%) of 40 T-cell lineage lymphomas (Table 3). Thirty seven cases (67.3%) showed bilateral BM involvement, and 18 cases (32.7%) showed unilateral involvement. Bone marrow involvement was observed in 72.0% of aspirations, 82.5% of clot sections, and 89.5% of biopsies.
|Histology by REAL classification||Total||Pattern of bone marrow involvement||Positivity of morphology/PCR|
|B-cell lineage||130||6 (3)a||26 (14)||4 (2)||5 (3)||89 (53)||41 (31.5%)/78 (60.0%)|
|Diffuse large B-cell L||91||5 (3)||12 (7)||2 (1)||2 (1)||70 (42)||21 (23.1%)/54 (59.3%)|
|Marginal zone B-cell L||9||—||2 (0)||—||—||7 (2)||2 (22.2%)/2 (22.2%)|
|Mantle cell L||8||—||4 (2)||—||1 (1)||3 (3)||5 (62.5%)/7 (75.0%)|
|Follicular center L||7||1 (0)||2 (2)||—||—||4 (2)||3 (42.9%)/4 (57.1%)|
|Small lymphocytic L||5||—||4 (3)||1 (1)||—||0 (0)||5 (100%)/4 (80.0%)|
|High grade B-cell L||5||—||—||1 (0)||—||4 (3)||1 (20.0%)/3 (60.0%)|
|Burkitt L||4||—||1 (1)||—||2 (1)||1 (1)||3 (75.0%)/3 (75.0%)|
|Precursor B cell L||1||—||1 (1)||—||—||—||1 (100%)/1 (100%)|
|T-cell lineage||40||1 (1)||6 (3)||6 (3)||1 (1)||26 (4)||14 (35.0%)/12 (30.0%)|
|Peripheral T-cell L||14||—||2 (1)||6 (3)||—||6 (0)||8 (57.1%)/4 (28.6%)|
|Angiocentric L||11||—||—||—||—||11 (2)||0 (0%)/2 (18.2%)|
|Precursor T-cell L||8||1 (1)||3 (2)||—||1 (1)||3 (0)||5 (62.5%)/4 (50.0%)|
|Anaplastic large cell L||4||—||—||—||—||4 (2)||0 (0%)/2 (50.0%)|
|Intestinal T-cell L||1||—||—||—||—||1 (0)||0 (0%)/0 (0%)|
|Hepatosplenic γ/δ L||1||—||—||—||—||1 (0)||0 (0%)/0 (0%)|
|Angioimmunoblastic L||1||—||—||—||—||1 (0)||1 (100%)/0 (0%)|
|Total||170||7 (4)||32 (17)||10 (5)||6 (4)||115 (57)||55 (32.4%)/82 (48.2%)|
Frequencies and patterns of BM involvement of NHL by morphology according to the histologic classification are outlined in Table 3. Mantle cell lymphoma, small lymphocytic lymphoma, Burkitt lymphoma, peripheral T-cell lymphoma, and precursor T-cell lymphoblastic leukemia/lymphoma showed high frequencies (more than half) of BM involvement by morphology. The most frequent pattern was nodular, followed by a diffuse interstitial pattern.
Five patients showed nodular infiltration on biopsies alone, which included two cases of conjunctival marginal zone B-cell lymphoma and primary central nervous system (CNS) lymphoma in otherwise clinical Stage I (if BM involvement by morphology had been excluded). Six patients showed nodular infiltration on clot sections alone, which also included two cases of gastric marginal zone B-cell lymphoma and CNS lymphoma in otherwise clinical Stage I (if BM involvement by morphology had been excluded).
Discordant morphology of neoplastic lymphocytes between lymph nodes or extra-nodal sites and BM was seen in four patients with diffuse large B-cell lymphoma with small lymphocytic infiltration in BM.
Characteristic parameters of peripheral blood (PB) and BM, such as thrombocytopenia, leukoerythroblastic reaction, PB involvement, BM fibrosis, BM necrosis, or BM hemophagocytic histiocytosis, were exclusively seen in the group with BM involvement by morphology, with the exception of mild thrombocytopenia observed in two patients at recurrence of NHL without BM involvement by morphology. Diffuse extensive or localized extensive fibrosis was seen in four cases of BM involvement by morphology, including one case each of diffuse large B-cell lymphoma, mantle cell lymphoma, follicular center lymphoma, and Burkitt lymphoma. Coagulation necrosis was also seen in four cases of BM involvement by morphology, including two cases of diffuse large B-cell lymphoma, mantle cell lymphoma, and peripheral T-cell lymphoma. Hemophagocytic features were exclusively noted in five cases of peripheral T-cell lymphoma.
Among 88 cases tested for cytogenetic studies, 20 cases (22.7%) showed abnormal karyotypes.In 40 cases of BM involvement by morphology, 19 cases (47.5%) showed abnormal karyotypes. One case without BM involvement by morphology showed an abnormal karyotype involving 14q32 where the immunoglobulin heavy chain was located. This case had a diffuse large B-cell lymphoma at Stage III with a huge abdominal mass. The modal chromosome number of abnormal karyotypes was pseudo- or near-diploidy, including four cases (20%) of hypodiploidy and eight cases (40%) of hyperdiploidy. The structural abnormalities were very complex, with more than three abnormalities in most cases. The recurrent abnormality was rare, and only del(6)(q15) was observed in four cases. The most common breakpoints were 6q (seven cases), 1p (three cases), and 14q32 (three cases).
IgH GR or TCRγ GR of lymph nodes or extranodal tissues
A total of 47 lymph nodes or extranodal tissues were tested. Each case was tested for IgH GR or TCRγ GR according to the result of immunohistochemistry. Among 35 cases of B-cell NHL tested for IgH GR, 23 cases (65.7%) showed monoclonality, 4 cases (11.4%) showed oligoclonality (2 –cases biclonality, 2 –cases more than two), and 8 cases (22.9%) showed no evidence of clonality. Overall clonality was 77.1% (27 out of 35). Among 12 cases of T-cell NHL tested for TCRγ GR, 8 cases (66.7%) showed monoclonality. Oligoclonality was not observed in TCRγ GR.
IgH GR or TCRγ GR of the bone marrow
Frequencies of IgH GR and TCRγ GR according to the histology and pattern of morphologic BM involvement pattern are outlined in Table 3. Seventy eight cases (60.0%) of B-cell lineage NHL and 12 cases (30.0%) of T-cell NHL showed the clonality of IgH GR or TCRγ GR in BM. Among 78 cases of PCR positive clonality of B-cell NHL, 21 cases (26.9%) showed oligoclonality of BM. None of the T-cell NHLs showed oligoclonality of BM. Among B-cell NHL, mantle cell lymphomas, small lymphocytic lymphomas, and Burkitt lymphomas showed higher frequencies of IgH GR. Among T-cell NHL, precursor T-cell lymphoma and anaplastic large cell lymphoma showed higher frequencies of TCRγ GR.
A positive rate of IgH GR or TCRγ GR of BM did not show a significant correlation with the pattern of BM involvement by morphology (Table 3). Four cases with stages increased by BM involvement by morphology with nodular infiltration on biopsies or clot sections alone showed negative results for PCR-based clonality of BM. Among four cases with discordant morphology of neoplastic lymphocytes between lymph nodes or extra-nodal sites and BM, two cases showed positive clonalities and two cases negative clonalities.
The frequencies of the characteristic findings, such as peripheral blood thrombocytopenia, peripheral blood leukoerythroblastic reaction, peripheral blood involvement, BM fibrosis, BM necrosis, or BM hemophagocytosis, were similar in the two groups defined according to PCR-based clonality of BM.
In 20 cases of abnormal karyotypes, the frequencies of PCR-based clonalities were 58.8% (10 out of 17) in B-cell NHL and 66.7% (2 out o f 3) in T-cell NHL, which were similar to those in the cases of normal karyotypes. In one case that had an abnormal karyotype but no evidence of BM involvement by morphology, PCR-based clonality was negative.
Concordance between Morphology and PCR Based Clonality of the Bone Marrow
Neither B-cell nor T-cell NHL groups showed significant differences in the rate of PCR-based clonality according to BM involvement by morphology (Table 4), though T-cell NHL cases showed borderline statistical significance (P = 0.08). B-cell NHL cases showed an overall positive concordance rate of 61.0%, and T-cell NHL cases 57.1%. In B-cell NHL cases, diffuse large B-cell lymphoma showed the lowest positive concordance rate of 57.1%.
|Pathology by REAL classification||Morphology positive||Morphology negative|
|PCR positive||PCR negative||PCR positive||PCR negative|
|B-cell lineage||130||25/41 (61.0%)||16/41 (39.0%)||53/89 (59.6%)||36/89 (40.4%)|
|Diffuse large B-cell L||91||12/21 (57.1%)||9/21 (42.9%)||42/70 (60.0%)||28/70 (40.0%)|
|Marginal zone B-cell L||9||–/2 (–)||2/2 (100%)||2/7 (28.6%)||5/7 (71.4%)|
|Mantle cell L||8||3/5 (60.0%)||2/5 (40.0%)||3/3 (100%)||–/3 (–)|
|Follicular center L||7||2/3 (66.7%)||1/3 (33.3%)||2/4 (50.0%)||2/4 (50.0%)|
|Small lymphocytic L||5||5/5 (100%)||–/5 (–)||–/– (–)||–/– (–)|
|High grade B-cell L||5||–/1 (–)||1/1 (100%)||3/4 (75.0%)||1/4 (25.0%)|
|Burkitt L||4||2/3 (66.7%)||1/3 (33.3%)||1/1 (100%)||–/1 (–)|
|Precursor B cell L||1||1/1 (100%)||–/1 (–)||–/– (–)||–/– (–)|
|T-cell lineage||40||8/14 (57.1%)||6/14 (42.9%)||6/26 (23.1%)||20/26 (76.9%)|
|Peripheral T-cell L||14||4/8 (50.0%)||4/8 (50.0%)||2/6 (33.3%)||4/6 (66.7%)|
|Angiocentric L||11||–/– (–)||–/– (–)||2/11 (18.2%)||9/11 (81.8%)|
|Precursor T cell L||8||4/5 (80.0%)||1/5 (20.0%)||–/3 (–)||3/3 (100%)|
|Anaplastic large cell L||4||–/– (–)||–/– (–)||2/4 (50.0%)||2/4 (50.0%)|
|Intestinal T-cell L||1||–/– (–)||–/– (–)||–/1 (–)||1/1 (100%)|
|Hepatosplenic γ/δ L||1||–/– (–)||–/– (–)||–/1 (–)||1/1 (100%)|
|Angioimmunoblastic L||1||–/– (–)||1/1 (100%)||–/– (–)||–/– (–)|
|Total||170||33/55 (60.0%)||22/55 (40.0%)||59/115 (51.3%)||56/115 (48.7%)|
Comparative Analysis of the Clonality between Lymph Node and the Bone Marrow
Comparison of PCR clonality between lymph node or extranodal tissue and BM is outlined in Table 5. Fifteen cases (93.8%) among 16 cases with bilateral clonality of BM showed same clonality of the lymph node. Seven cases (87.5%) among eight cases with unilateral clonality of BM showed no evidence of BM involvement by morphology.
|Clonality of lymph node or tissue||Clonality of bone marrow||Morphology|
|B-cell lineage NHL (n = 35)|
|Positive (n = 27)||Bilateral positive||7||5|
|Negative (n = 8)||Bilateral positive||1||—|
|T-cell lineage NHL (n = 12)|
|Positive (n = 8)||Bilateral positive||3||—|
|Negative (n = 4)||Negative||3||1|
Among 12 cases of T-cell NHL, 7 cases (58.3%) showed concordant results between lymph node or extranodal tissue and BM. Three cases of positive concordance results showed the evidence of BM involvement by morphology.
Clinical Significance of PCR Based Clonality
Univariate analysis of clinical factors, including international prognostic indices (age, serum LD, performance, stage, extranodal involvement), gender, cell lineage, histology, BM involvement by morphology, and PCR-based clonality is outlined in Table 6. Serum LD, performance, stage, extranodal involvement, and BM involvement by morphology showed statistically significant impacts on overall survival (P < 0.05). Polymerase chain reaction based clonality did not show a significant impact on overall survival in B-cell NHL and T-cell NHL (P = 0.2, Table 6).
|Parameters||Total patients||Median survival (mo)||P|
|Serum LD (IU/L)|
|Bone marrow by morphology|
|Bone marrow by PCR|
Use of PCR for the detection of IgH or TCRγ GR is a relatively new method for the study of lymphoid neoplasms. Most studies have been reported on lymph node based specimens and presented relatively good results. A few studies have recently been investigated for BM involvement of NHL.22–26
In the current study, 59 cases (34.7%) among 170 cases with NHL showed no evidence of BM involvement by morphology but were positive for PCR-based clonality (Table 4). These cases may represent false negative cases by morphology, in other words, minimal BM involvement of NHL, or false positive results by PCR. A minimal degree of BM involvement could be explained by an extremely higher sensitivity of PCR techniques than in morphologic detection. False positive results by PCR might be caused with circulating, not proliferating, neoplastic lymphoid cells in BM or benign reactive lymphocytes, since detection of IgH GR or TCRγ GR is not specific for neoplastic DNA. Both benign and malignant lymphocytes undergo GR early in their maturation, which can be shown by PCR amplification.20, 21 In most lymphomas, there are variable numbers of benign lymphocytes in the bone marrow in addition to the malignant clone. Although these are not detected by Southern blot, they may produce background smearing on PCR based detection of GR. Polyclonal patterns and combinations of a dominant clone and a polyclonal background have been reported in reactive lymph nodes and in lymphomas with a significant benign lymphocytic background, respectively.31 Therefore, careful interpretation is essential. To improve the specificity of the clonality, denaturing gradient gel electrophoresis, heteroduplex analysis, and DNA sequencing may be used, although these methods are also limited in that they cannot distinguish between benign and malignant lymphoid cells.33–35 The current study suggests that bilateral studies of BM and comparative analysis between the clonality of BM and that of lymph node or extranodal tissue are valuable in increasing the reliability of BM clonality, especially in B-cell NHL (Table 5).
In the current study, 22 cases (40.0%) among 55 cases with BM involvement by morphology showed no evidence of PCR-based clonality (Table 4). These cases may represent false negative cases by PCR or overdiagnosis by morphology. Due to these cases, PCR-based clonality showed an unexpectedly low positive concordance rate (60.0%) with morphology (Table 4). Coad et al.22 reported the positive concordance was 57% in a total of 225 cases of B-cell NHL, including a large proportion of follicular center cell lymphoma (38.2%) and diffuse large B-cell lymphoma (37.8%). Crotty et al.23 reported that 9 cases (75.0%) showed positive clonality among 12 cases of BM involvement by morphology. Kim et al.24 reported that 13 cases (86.7%) showed positive clonality among 15 cases of BM involvement by morphology. Pittaluga et al.26 reported that 46 cases (85.2%) of 54 cases with BM involvement by morphology showed clonality among 104 cases of hairy cell leukemia and mantle cell lymphoma. The false negative results by PCR might be caused by insufficient DNA extraction, an inability of consensus V primers to recognize complementary DNA sequences in all of the V segments, and/or an inability of V and J primers to recognize genetic alterations such as partial rearrangements, chromosomal translocations, and somatic mutations involving the antigen receptor gene loci.21 In the current study, the unexpectedly low positive concordance rate of PCR-based clonality with BM involvement by morphology may be due to the overrepresentation of diffuse large B-cell lymphoma (53.5% among total NHL cases), in which the positive concordance rate between morphology and PCR-based clonality has been reported to be as low as 42%.22 This suggests that the detection of PCR based clonality may be less useful in Asia because diffuse large B-cell lymphoma is the most common type of NHL in Asia. Also, to overcome this low concordance rate, more complex primers are needed, such as VH-FRI;36 in addition, Southern blot analysis is needed when PCR-based clonality is negative.21, 37
Four cases of B-cell NHL interpreted as BM involvement for nodular B-cell infiltrates on either biopsy or clot section alone in otherwise clinical Stage I raise the possibility of overdiagnosis by morphology. Polymerase chain reaction studies failed to give evidence of clonally rearranged immunoglobulin genes in these cases. It cannot be ruled out that these nodular infiltrates might represent benign lymphoid nodules, which emphasizes the diagnostic difficulty of BM involvement by NHL.
The associated characteristic parameters of BM involvement by morphology by NHL, such as PB leukoerythroblastic reaction, PB involvement, BM fibrosis, and BM necrosis, were very specific, seen exclusively in cases with BM involvement by morphology, with the exception of mild thrombocytopenia seen in two cases of recurrent NHL without BM involvement by morphology, probably due to hypoplastic marrow for previous chemotherapy.
Compared with the chromosomal abnormality of BM, the evaluation of BM by morphology was a convincing method with high specificity to detect BM involvement of NHL. The only case with abnormal cytogenetic results was diagnosed as having no evidence of BM involvement of NHL. When this case was carefully reviewed, there was no evidence of BM involvement by morphology on the biopsy or clot section, but large lymphoid cells with high nuclear/cytoplasmic ratios and deep blue cytoplasm with or without cytoplasmic vacuoles were rarely seen.
There have been a few reports of minimal involvement of NHL identified by culturing BM cells from morphologically negative cases,38 PCR analysis for specific translocation such as bcl-2,39 or Southern blot analysis,40 associated with poorer clinical outcomes. But the clinical significance of PCR based clonality of BM in NHL is uncertain. To our knowledge, the current study is the first to analyze the clinical significance of PCR-based clonality of BM in NHL, and the results show that PCR based clonality has no impact on overall survival, although BM involvement by morphology does have a significant impact on survival.
In conclusion, morphology remains the gold standard in the evaluation of BM involvement of NHL. Although the bilateral analysis of BM clonality and comparative analysis with lymph nodes are valuable in increasing reliability, PCR-based clonality of BM does not seem to support the clue of BM involvement or the prognostic significance in NHL.
- 9Lesions simulating lymphoma and miscellaneous tumor-like lesions in the bone marrow. In: RosaiJ, editor. Atlas of tumor pathology. vol. 9, 3rd series. Washington, DC: Armed Forces Institute of Pathology, 1994: 409–414., .
- 10Lymphoid aggregates in bone marrow. In: RogersP, editor. Bone marrow pathology. Chicago: ASCP Press, 1995: 259–260..
- 24Detection of minimal lesion and identification of clonality in malignant lymphoma. Korean J Pathol. 1998; 32: 298–308., , , et al.
- 26How reliable is histologic examination of bone marrow trephine biopsy specimens for the staging of non-Hodgkin lymphoma? A study of hairy cell leukemia and mantle cell lymphoma involvement of the bone marrow trephine specimen by histologic, immunohistochemical, and polymerase chain reaction techniques. Am J Clin Pathol. 1999; 111: 179–184., , , , .
- 29ISCN: An International System for Human Cytogenetic Nomenclature. Basel: S. Karger, 1995..
- 30Bone marrow involvement by non-Hodgkin's lymphoma: incidence, histopathology and diagnostic efficacies. Korean J Clin Pathol. 1994; 14: 20–29., .