Diagnostic Impact of Molecular Lineage Analysis on Paraffin-Embedded Tissue in Hematolymphoid Neoplasia Reclassified by Current WHO Criteria

BACKGROUND AND OBJECTIVE

By current WHO criteria, most - though not all - cases of hematolymphoid neoplasm can be diagnosed immunomorphologically, diminishing the role of molecular tests for lymphoid antigen receptor clonality in lymphoma diagnosis. Hence, our objective was to glean immunomorphological and molecular correlates from hematolymphoid neoplasms that had remained unresolvable without diagnostic molecular input.

METHODS

Thirty-five such cases were reviewed histologically and with standard immunoperoxidases. In situ hybridization for Epstein-Barr virus (EBV)-encoded RNAs (EBER) was performed on selected cases. PCR amplification of genes encoding T-cell receptors (TcR) and immunoglobulin heavy chains (IgH) [TR and IGH genes, respectively] was performed on whole tissue in all cases, and on microdissected cells in two cases.

RESULTS

Twenty-five cases (71%) requiring diagnostic molecular genotyping had some form of peripheral T-cell lymphoma (PTCL). Twenty (80%) of these were complicated by a proliferation of B-lineage cells, either within the same tissue ('syntopic') as large B cells (LBC) or Reed-Sternberg (RS)-like cells (17 cases), florid lymphoid hyperplasia (two cases, one also with syntopic LBC) or monotypic plasma cells (one case), or at a separate ('metatopic') site as a B-cell lymphoma (two cases, one of which also had syntopic LBC) or Hodgkin lymphoma (HL; one case, also showing syntopic LBC). Fifteen (75%) of these 20 PTCLs with B-lineage proliferation yielded monoclonal TR gene rearrangements, and only two (10%) showed IGH monoclonality, which was transient in one case. Three (18%) of the PTCLs with LBC had originally been misinterpreted as some form of HL. Conversely, of the remaining cases, three of four (75%) that had been diagnosed initially as some form of large cell non-HL (NHL), including two of three that were called 'anaplastic', had to be revised to grade II/syncytial nodular sclerosing (NS) HL, yielding polyclonal TcRgamma gene (TRG) rearrangements, with one case, in addition, disclosing a biallelic clonal IGH gene rearrangement that excluded anaplastic large cell lymphoma.

DISCUSSION/CONCLUSION

Paradoxically, monoclonality of TR rather than IGH gene rearrangement may more often be detectable in a predominantly dispersed ('hodgkinoid'), large B-lineage cell proliferation, consistent with release from immune regulation in the milieu of impaired immunosurveillance within a PTCL. This is compounded by the difficulty in ascertaining clonal IGH gene rearrangements resulting from (1) poor consensus primer hybridization due to somatic hypermutations, and (2) 'dilution' in a T-cell-rich milieu. These same difficulties also account for the long-elusive identification of the RS cell lineage. Conversely, anaplastic lymphoma, which is of non-B lineage, may be mimicked by NSHL, which is of B lineage.

[Optimized protocols for interphase FISH analysis of imprints and sections using split signal probes]

Fluorescent in situ hybridization (FISH) analysis is a molecular technique allowing the detection of recurrent translocations in cancer. Several hybridization protocols were assayed in order to evaluate their performances for interphase FISH analysis of histological sections and imprints using split probes. Adult and foetal lymphoid tissues were selected. Touch imprints of fresh (EF) or frozen (EC) tissues, sections (CF) and isolated nuclei (NI) of formol-fixed paraffin-embedded tissues were performed. The cut-off values of the IGH, IGlambda, BCL-2, BCL-6, CCND1 and MYC DNA FISH split signal probes were calculated for adult reactive lymph nodes on the different histological preparations (EC, CF, CC, NI) and on several tissues for the IGH and BCL-6 probes. In reactive lymph nodes, the cut-off values of the probes were between 3 and 13% and found independent of the preparation type. Conversely, slight but significant variations of the cut-off level were observed when different foetal control tissues were assayed with the same probe set. Finally, this study provided optimized-protocols for FISH analysis of either fresh/frozen imprints or formalin-fixed paraffin-embedded sections using split signal DNA probes.