Diffuse large B-cell lymphoma (DLBCL) is the most common lymphoid neoplasm. Over the past years many studies have shown that DLBCL is not a homogenous entity. The most recent edition of the World Health Organization (WHO) Classification of Tumours reflects this new approach, having subdivided DLBCL into morphological variants, and immunophenotypical and molecular subgroups, thus forming new, distinct entities as well as recognising the existence of so-called ‘borderline cases’.1 Among these borderline cases, the category of ‘B-cell lymphoma, unclassifiable, with features intermediate between Burkitt lymphoma and DLBCL’ (I-BL/DLBCL) has been created; MYC being rearranged in approximately 50% of such cases.1 Less is known about the molecular epidemiology of MYC breaks in DLBCL cases, which do not fit the category of I-BL/DLBCL. Importantly, the clinical course of MYC rearranged DLBCL differs from other DLBCL and therefore application of Burkitt lymphoma (BL)-like treatment regimens might be advantageous.2^–4

In order to study the molecular epidemiology of MYC aberrations in DLBCL and to determine whether their presence can be predicted by conventional methods such as morphology and immunohistochemistry, we examined 333 cases on a standardised tissue microarray (TMA) platform.

METHODS

Patients

A total of 333 DLBCL cases diagnosed between 1988 and 2001 and reclassified according to the WHO criteria were included in this study1; DLBCL cases with features suggestive of BL (ie, medium-sized blasts with a diffuse, monotonous growth pattern, pronounced starry-sky pattern, Ki-67 >95%, and involvement of the terminal ileum, ovaries and breasts) were intentionally excluded. Clinical and follow-up data were obtained by reviewing patients’ charts.

Construction of tissue microarrays

TMA were constructed as described previously.5 Four different TMA were constructed containing tumour samples from different histopathological institutions (University Hospitals of Basel, Bologna and Innsbruck, Triemli Hospital in Zurich).

Immunohistochemistry

Sections (4 μm) of the TMA blocks were cut to adhesive-coated slides and stained using standard procedures. The following primary antibodies were used: Bcl2 (clone 124, DAKO M0887, final dilution 1:10), Bcl6 (clone PG-B6p, DAKO M7211, final dilution 1:10), CD5 (clone SP19, Cell Marquee, final dilution as supplied by company), CD10 (clone NLL-CD10-270 56C6, Novocastra, final dilution 1:10), Ki-67 (clone MIB-1, DAKO M7240, final dilution 1:100), and Mum1 (clone Mum-1P, DAKO M7259, final dilution 1:50). Twenty per cent of cases were re-evaluated by a second observer to assess inter-observer agreement. At least 200 cells were assessed in each tumour core and the percentage of positive cells (ie, cells with a distinct staining) was calculated. Cut-off levels were determined as described elsewhere.6

Fluorescence in situ hybridisation (FISH)

The MYC, BCL2, BCL6 and cyclin D1 gene status was assessed by FISH utilising dual-colour, break-apart probes (all probes from Vysis/Abott, Downers Grove, Illinois, USA: MYC, order no. 05J91-001; BCL2, order no. 07J75-001; BCL6, order no. 05J68-001; cyclin D1, order no. 05J72-001) on paraffin-embedded tissue sections according to the protocol published by Vysis (www.abbottmolecular.com). Slides were counterstained with 125 ng/ml 4′,6-diamino-2-phenylindole in antifade solution. FISH signals were scored with a Zeiss fluorescence microscope equipped with double-band pass filters for simultaneous visualisation of Spectrum Green and Spectrum Orange signals. Cases on the TMA were considered evaluable for FISH if at least 200 tumour cell nuclei per core displayed positive signals. Splits were recorded as percentage of cells bearing an abnormality out of all analysed cells. The cut-off score to consider a case rearranged was the mean plus three standard deviations of split nuclei in the reference cases (ie, 4% for MYC, 3% for BCL2, 1.5% for BCL6 and 7% for cyclin D1). The minimal distance of split signals was defined as at lest three single signal widths. Gains, ie amplifications, polysomies and/or polyploidies, were assessed as well. For controls, tonsil sections of five patients were used. Results gained by two observers (MC and AT) were compared in 80 cases to assess inter-observer agreement.

Statistical analysis

Statistical analysis was done using SPSS V.16.0 for Windows (SPSS, Chicago, Illinois, USA). Inter-observer agreement for FISH and immunohistochemical analyses was assessed using κ statistics, a κ-value of ⩾0.75 implying excellent agreement. Survival was analysed by the Kaplan–Meier method and compared by the log-rank test. Statistical significance was defined as p<0.05; two-sided tests were used throughout.

RESULTS

FISH analysis was successful in 220/333 (66%) cases. Failures were due either to unsuccessful hybridisation (n = 63) or to problems related to the array technology (n = 40). The inter-observer correlation of FISH results for MYC was excellent (κ = 0.95). Nine of the 220 cases (4%) showed a break in the MYC gene. None of the nine cases was associated with an Epstein–Barr virus infection; one patient was HIV positive. Gains of MYC were observed in 19 DLBCL cases. Of these, 15 showed gains of either one or the other analysed genes; thus the observed MYC signal gains likely reflected a general polyploidy of the respective cases. High level MYC amplifications were not observed. Three cases showed an additional BCL2 (case 3), BCL6 (case 4) or cyclin D1 break (case 2) (table 1).

Morphology and immunophenotype of the rearranged cases were carefully reevaluated (fig 1, table 1). None of the cases expressed CD5. In retrospect, only case 6 may have fulfilled morphological and phenotypic criteria to be categorised as an I-BL/DLBCL, but its proliferation rate was 85%. In our group, the proliferation rate, as assessed by expression of Ki-67 (mean 73% (SD 16%), range 45–90%), never exceeded 90%. This was supported also by the proliferation rate assessed on conventional full tissue sections from the respective cases (mean 69% (SD 25%), range 30–90%). Therefore we conclude that Ki-67 expression is not likely to be useful in identifying these MYC rearranged cases; the same is applicable to all other immunophenotypic markers studied.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

(A–I) Morphology of cases of diffuse large B-cell lymphoma with myc breaks; H&E stain, ×200. The consecutive figure letters correspond to the consecutive case numbers from table 1. Of note are cases with anaplastic (C) and centroblastic (D and H) morphology as well as a high number of apoptotic bodies only in case 4 (D) and predominantly small cell morphology only in case 6 (F). None of the latter two cases shows a starry sky pattern. (J) Fluorescence in situ hybridisation analysis with break-apart MYC gene probes: cells with clearly split signals (free red and green) can be easily detected; DAPI counterstaining, ×640.

Patients with lymphomas harbouring a MYC break had a mean and median survival time of 48 months (95% CI 0 to 96) and 9 months (95% CI 0 to 21), respectively, compared to patients without MYC breaks (mean 101, 95% CI 78 to 125 months; median 80, 95% CI 21 to 139 months). Cases with MYC gains also had shorter mean and median survival time (53 and 12 months, respectively), compared to cases without gains (105 and 84 months, respectively).

DISCUSSION

Similarly to other studies, we found a striking survival difference for DLBCL cases with MYC rearrangements as opposed to cases without this aberration.3 4 7^–9 The relative frequency of MYC breaks in our study group was comparable or slightly lower than in these previous studies (5–15%). This may be explained by our exclusion of cases with features suggestive of BL/I-BL/DLBCL, the main aim of our study being to determine whether MYC rearranged DLBCL cases without morphological and phenotypic features of BL/I-BL/DLBCL could be identified by conventional methods alone. Our data clearly show that in unselected DLBCL collectives, cases with MYC rearrangements could not be reliably identified by morphological and immunohistochemical methods, and thus routine FISH work-up for MYC aberrations is needed. In line with these considerations, two recent studies, which identified aggressive B-cell lymphomas with a morphology typical of DLBCL and a molecular gene expression signature of BL, stress the impossibility of recognising such cases by routine histological and immunohistochemical evaluation alone.7 10 From the practical point of view, in contrast to gene expression profiling, our approach using commercially available FISH probes has a number of advantages: applicability on routinely processed tissue, rapid assessment, and cost-efficiency. Based on our data and cumulative evidence form the literature,7 10 we suggest FISH work-up for genetic aberrations of MYC in all DLBCL. These patients should be treated in prospective randomised clinical trials in order to establish the best possible management.