Article Text

Download PDFPDF

Correlation between molecular and histopathological diagnoses of B cell lymphomas in bone marrow biopsy and aspirates
  1. R B Ilgenfritz1,
  2. K Kayasut1,
  3. A Le Tourneau1,
  4. O A Calendini2,
  5. L Ouafi1,
  6. C Marzac2,
  7. J Diebold1,
  8. F Devez1,
  9. V Ducruit1,
  10. P E Bouchet1,
  11. J Audouin1,
  12. T J Molina1
  1. 1AP-HP, Hotel Dieu, Service d’Anatomie et de Cytologie Pathologiques, Université Paris Descartes, Paris, France
  2. 2AP-HP, Hotel Dieu, Service d’Hematologie Biologique, Université Paris Descartes, Paris, France
  1. Raquel B Ilgenfritz, Hôtel-Dieu, Service d’Anatomie et de Cytologie Pathologiques, 1 place du Parvis Notre-Dame, 75004 Paris, France; raquel.fritz{at}


Aims: PCR has been shown previously to be the most sensitive technique to detect a clonal population in marrow aspirates (MAs), and the clinical standard for evaluation of bone marrow lymphoma involvement today is bone marrow trephine biopsy (BMTB). The goal of this study was to compare morphological evaluation of B cell neoplasm in BMTB (histology and immunohistochemistry) and PCR analysis in MA, with both specimens obtained at the same time, in patients with a known molecular marker of the disease.

Methods: This was a retrospective evaluation of 98 consecutive BMTB specimens from 60 patients with a known B-cell neoplasm and a previous PCR marker of the disease (BCL2 and/or IGH).

Results: Considering the IGH PCR cases alone, a B cell clone was detected in 85% and 39% of the morphology (M) positive and negative groups, respectively. Five M(+), IGH(−) cases were found, including two cases of follicular lymphoma (FL), one case of diffuse large B cell lymphoma, and two cases of mantle cell lymphoma. The FLs had about 20% and 50% of BMTB involvement each. All other cases had minimal lymphoma localisation. The two FLs were also BCL2-MBR(+). Use of BCL2-MBR detected all M(+) cases and 66% of M(−) cases whenever it was an initial marker of disease.

Conclusions: IGH PCR alone is not good enough for BMTB assessment, especially in FL. On the other hand, the PCR study for BCL2 is more sensitive than morphology, without any false negative results in this series, suggesting that BCL2-MBR PCR on MA can be used as an alternative and more sensitive examination for disease evaluation, providing that there is careful analysis of data, adequate knowledge of PCR pitfalls and absence of other haematological disorders.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Bone marrow trephine biopsy (BMTB) is routinely performed for evaluation of lymphoma involvement in staging, diagnosis and disease evolution. The diagnosis of lymphoma in BMTB is difficult in cases of minimal, focal random or interstitial infiltration of low-grade B cell lymphoma even when using the appropriate criteria.1 2 In recent years, PCR analysis of antigen receptor gene and oncogene rearrangements, including immunoglobulin heavy chain (IGH) and BCL2 genes, has increasingly been used to identify tumour clonality. This method can detect a clonal population below the morphological threshold, as well as one neoplastic cell in 105–106 normal cells.3 However, the sensitivity of PCR depends on the primer set, lymphoma subtype, type of specimen, tissue fixation and degree of competitive amplification between the DNA of monoclonal and polyclonal lymphocytes.4 5 The sensitivity of the PCR method for detecting IGH clonality in BMTBs and marrow aspirates (MAs) has been reported, ranging from 80% to as low as 40%.510 The aim of our study was to correlate PCR with histopathological and immunohistochemical evaluation of BMTB using IGH PCR and BCL2 markers in concurrently obtained MAs, and to further define the role of PCR for the follow-up of patients with consecutive BMTB specimens.


Case selection

From the files of the Department of Pathology, Hôtel-Dieu, Paris, France, we retrospectively evaluated 98 consecutive BMTBs from 60 patients with a known B cell neoplasm, in a 7-year period. The initial lymphoma was classified according to the World Health Organization (WHO) 2008 (table 1).2

Table 1 Number of patients according to initial diagnosis and initial positive molecular markers

Case selection was based on the presence of at least one clonal marker of the disease (IGH or BCL2) demonstrated by PCR in a previous reference sample. (Lymph node: 28 formalin-fixed and seven frozen samples; fresh peripheral blood: nine samples; frozen MA: nine samples; digestive biopsy: three formalin-fixed and two frozen samples; skin and lung biopsy: one sample from each site, both formalin-fixed material.) All cases had a BMTB for morphological analysis, and an MA that had been obtained at the same time for PCR study. BMTB was performed at the posterior superior iliac crest, and MA for molecular analysis was obtained through the same incision site, immediately after the biopsy. Twenty-seven patients had more than one BMTB/MA pair during the course of the disease (range from 2 to 4).

BMTB analysis

Biopsies were fixed in Bouin, decalcified with RDO rapid bone decalcifier and embedded in paraffin. Multiple 3 μm sections were analysed with H&E, Giemsa, Perls’, reticulin and periodic acid–Schiff stains. At least three additional H&E levels were analysed in negative cases (range from 3 to 6). Immunohistochemistry study (IHC) was performed using the streptavidin-biotin-peroxidase complex method, with the following antibodies: CD20 (L26; Dako, Glostrup, Denmark), CD79a (JCB117; Dako), CD3 (F7.2.38; Dako), CD5 (4C7; Tebu), CD23 (1B12; Tebu), CD10 (56C6; Tebu), and IgA, G, M, D, κ and λ (polyclonal; Dako). Slides were analysed without the knowledge of the PCR data, and cases were classified into two groups after morphology and IHC analysis (M groups). Group M(+) included cases with morphological and/or IHC evidence of lymphoma involvement, and group M(−) included cases without morphological and/or IHC evidence of lymphoma involvement. BMTBs showing discordance with PCR results were later reviewed by at least two pathologists, confirming in all cases the initial histopathological diagnosis.

Molecular studies

Fresh MA was collected in EDTA. DNA was extracted using the Nucleon kit (Nucleon BACC2, RPN 8502; Amersham Biosciences, Little Chalfont, UK) and was kept at −80°C until ready for use. Amplification was performed using a thermocycler (Gene Amp PCR System 2400; PerkinElmer, Waltham, Massachusetts, USA). BCL2-MBR rearrangement and IGH-PCR (FR1, FR2 and FR3 loci) were detected as previously described.1113 PCR band size analysis was evaluated on high-resolution migration gel analysis, testing in parallel the initial tumour, MA specimen and different size controls.


Results are shown in table 2. Forty-five (95%) of the 47 M(+) cases were positive for at least one PCR marker, while only 20 of the 51 M(−) cases were negative for all markers. Two M(+) cases were not detected by PCR; but they had only an IGH band as initial marker. When considering IGH PCR, 34 of the 62 cases were M(+). Of these 34 cases, 29 (85%) were positive, but five were negative, for IGH PCR on MA (table 3). The five cases included two follicular lymphomas (FLs) with about 20–50% of BMTB involvement in a paratrabecular and interstitial pattern (fig 1), a diffuse large B cell lymphoma (DLBCL) and two mantle cell lymphomas. The latter three cases had minimal (<10%) marrow infiltration. When considering only BCL2-MBR, PCR was positive in all of the 24 M(+) cases, and in 21 of the 32 M(−) cases. Not a single M(+) case was negative for this marker. PCR was also positive for the two IGH-negative FLs that had BCL2-MBR as an initial marker (table 3). The concordance rates (M(+), PCR(+); and M(−), PCR(−)) were 74% and 66% considering IGH alone and IGH with BCL2, respectively. The discordance rates (M(+), PCR(−); and M(−), PCR(+)) in the same groups were 25% and 34%, respectively. Furthermore, we identified three cases (two DLBCL and one FL) having a clonal rearrangement at the IGH level that was not present as a marker in the initial tumour. These bands, observed among patients undergoing chemotherapy, were interpreted as non-significant and in all cases were considered as negative PCR results. The FL case was BCL2-MBR-positive in the tumour but negative in the MA. There was a subsequent BMTB with PCR study 4 months later, and this showed neither IGH or BCL2-MBR PCR positivity, nor tumour infiltration (fig 2).

Figure 1

Bone marrow trephine biopsy of case no. 5, which was negative for IGH PCR but had an important lymphoma localisation (about 50% of the marrow population) in a paratrabecular pattern (arrows). H&E, original magnification: (A) ×10, (B) ×40.

Figure 2

Post-chemotherapy follicular lymphoma (lane 2) with an IGH PCR clonal band at the FR3 locus (white arrow). This band was not present as a marker in the initial tumour, and a second sample was negative 4 months later. Lanes: 5, protein size ladder; 6, positive control from initial tumour; 1, 3 and 4, other cases.

Table 2 Comparison of marrow aspirate PCR (IGH and/or BCL2) results and concomitant bone marrow biopsy morphological (M(+) or M(−)) evaluation
Table 3 Cases with morphologically detected disease (BMTB) and negative IGH markers in MA, where IGH was an initial marker


Our positive concordance rate for IGH cases is high (85%), close to the study of Braunschweig et al (81%), due to the utilisation of the same selection criteria.14 The requirement of a previous molecular marker of the disease to assess marrow involvement allows us to better evaluate the help of molecular tools in the follow-up of patients with B cell lymphoma having consecutive BMTBs. However, this high concordance rate requires molecular analysis of initial tumour; this is often lacking in daily practice, but is crucial to allow efficient analysis of PCR products. Indeed, in the M(−) group, a different clonal rearrangement of IGH was identified in two DLBCL and one FL compared with the initial tumour. These were post-chemotherapy cases and had a very few number of small B lymphocytes in the BMTB. The presence of a low number of normal B lymphocytes has been described as being able to produce a monoclonal band using IGH-PCR analysis.15 16 Our study therefore emphasised how careful one should be to interpret a clone detected by PCR on a marrow sample without knowledge of the molecular profile of the initial tumour. Such a clone may not represent markers of the disease and should be considered as non-significant. Although sensitive IGH PCR detected clonality linked to the initial tumour in 11 out of 28 M(−) cases, five M(+) cases were negative for IGH PCR. The tumour load in the BMTB was surprisingly high in two of these cases, ranging from 20% to 50%, and both were FL. Previous studies have shown that FL appears to be one of the more difficult lymphoma subtypes to evaluate with the standard PCR assays, probably due to the relatively high somatic mutation rate and ongoing mutation in the variable region.4 1719 The difficulty can also be explained by sampling errors, which are caused by heterogeneous or focal BMTB involvement by lymphoma, or the presence of fibrosis and dilution by co-aspirated blood, resulting in a low number of neoplastic cells in MA. Therefore, insufficient clonal tissue and sample variations between MA and BMTB can be relevant explanations to these discordant results. On the other hand, many recent studies confirmed the usefulness of BCL2 rearrangement analysis in peripheral blood and MA for detection of minimal residual disease by PCR, and of serial PCR analysis to determine the molecular response and for predicting the survival and relapse rate.2025 It is of note that low level BCL2 rearrangements have been detected in normal subjects and patients with cancer, varying from 12% to 23% of the subjects.2628 However, Ladetto et al reported a rare incidence (2.3%) of this translocation in chemotherapy-treated patients, and provides evidence for the value of persistent PCR positivity in the molecular follow up of post-chemotherapy patients with FL.26

Take-home messages

  • BCL2-MBR PCR on marrow aspirates may represent a tool for the marrow follow-up of B cell lymphomas when it is a marker of the disease.

  • IGH PCR on marrow aspirate is not sensitive enough for the follow-up of B cell lymphoma in bone marrow and should be interpreted with caution and always compared with the initial tumour PCR product, to avoid false-positive results.

  • Material for PCR analysis of the initial tumour should be collected on a routine basis, to allow comparison with subsequent samples.

Interestingly, in our cases, there was no single M(+) case associated with a negative BCL2-MBR PCR, and we could even detect minimal residual disease in 66% of the M(−) cases. Therefore, our study demonstrates the utility of BCL2-MBR PCR in MA for evaluation of lymphoma involvement in staging and follow-up of patients with B cell neoplasms whenever this marker is present on initial disease (the MBR breakpoint is present in only 50–60% of the FL). Our study suggests that BCL2-MBR PCR on MA may represent an additional tool for disease evaluation, providing that there is a careful analysis of data, adequate knowledge of PCR pitfalls, and absence of other haematological disorders.



  • Competing interests: None.

  • Ethics approval: Obtained.