DiscussionIn this biological validation study of NGS-based IG clonality testing, we show that NGS-based IG clonality and conventional EuroClonality/BIOMED-2 IG clonality analysis employing GeneScan provide highly similar results; the two techniques arrived at the same overall conclusion in 98% of the analyses. In addition, NGS-based IG clonality showed a very high interlaboratory concordance of 99% for the overall conclusion of the analysis. These results were obtained despite the enrichment of the samples for FFPE material and (post-) germinal center B-cell lymphomas. Our multicenter validation study thus shows that NGS-based IG clonality can reliably be applied as an alternative for conventional GeneScan clonality analysis in molecular lymphoma diagnostics. Of note, plasma cell neoplasms were not included in this study. Also, it has to be noted that the NGS assay as used in this study, does not utilize IGHV FR1/FR2 forward primer sets analogous to those in BIOMED-2/GeneScan analysis, and may yield polyclonal results in a small number of cases that would have been clonal by the latter assay (if the DNA-quality of these cases allows amplification of the larger amplicons). The three samples with a discordant result between NGS-based clonality analysis and GeneScan clonality analysis were histologically all lymphomas. In one of these, GeneScan but not NGS-based analysis was able to detect a clonal B-cell proliferation. In the other two, NGS-based analysis but not GeneScan demonstrated the clonal B-cell population. These discordances are probably a result of the different primer design for NGS- and GeneScan-based clonality analysis; a new set of primers was designed for NGS-based clonality detection to generate smaller amplicons which makes the technique more suitable for analysis of FFPE tissue. Indeed, although the overall results were highly similar, when comparing the results from GeneScan and NGS-based clonality assessment for the different targets (i.e., IGHV-IGHJ, IGHD-IGHJ, IGKV-IGKJ, IGKV-KDE/IntronRSS-KDE), NGS-based clonality was more often interpretable and was more often able to demonstrate a clonal product in B-cell neoplasia (Table 2). In addition to the smaller amplicon size that facilitates the detection of rearrangements in suboptimal DNAs from FFPE tissues, primer design of the NGS-based assay was performed based on different parameters than for the first BIOMED-2 primers. The new set of primers was designed to be gene-specific5Bruggemann M. Kotrova M. Knecht H. Bartram J. Boudjogrha M. Bystry V. et al.Standardized next-generation sequencing of immunoglobulin and T-cell receptor gene recombinations for MRD marker identification in acute lymphoblastic leukaemia; a EuroClonality-NGS validation study., while the BIOMED-2 primers were designed as gene-family-specific primers1van Dongen J.J. Langerak A.W. Bruggemann M. Evans P.A. Hummel M. Lavender F.L. Delabesse E. Davi F. Schuuring E. Garcia-Sanz R. van Krieken J.H. Droese J. Gonzalez D. Bastard C. White H.E. Spaargaren M. Gonzalez M. Parreira A. Smith J.L. Morgan G.J. Kneba M. Macintyre E.A. Design and standardization of PCR primers and protocols for detection of clonal immunoglobulin and T-cell receptor gene recombinations in suspect lymphoproliferations: report of the BIOMED-2 Concerted Action BMH4-CT98-3936.. Clearly, this new primer design with more optimal primer binding sites for NGS-based assays resulted in a higher detection rate of the different targets in both pre- and post-germinal center B-cell lymphomas. When comparing our current NGS-based approach to the results obtained in the original BIOMED-2 GeneScan clonality analysis studies, the BIOMED-2 assay shows a higher detection rate of clonality in B-cell lymphoma (99% vs. 96%). One could argue that in the BIOMED-2 study10Evans P.A. Pott C. Groenen P.J. Salles G. Davi F. Berger F. Garcia J.F. van Krieken J.H. Pals S. Kluin P. Schuuring E. Spaargaren M. Boone E. Gonzalez D. Martinez B. Villuendas R. Gameiro P. Diss T.C. Mills K. Morgan G.J. Carter G.I. Milner B.J. Pearson D. Hummel M. Jung W. Ott M. Canioni D. Beldjord K. Bastard C. Delfau-Larue M.H. van Dongen J.J. Molina T.J. Cabecadas J. Significantly improved PCR-based clonality testing in B-cell malignancies by use of multiple immunoglobulin gene targets. Report of the BIOMED-2 Concerted Action BHM4-CT98-3936., fresh frozen (FF) material was used, whereas in the current study the majority of DNA was derived from FFPE material with suboptimal DNA-integrity. Nevertheless, in the B-cell neoplasms in which NGS clonality could not detect a clonal B-cell proliferation (n=6), the DNA was derived from fresh frozen tissue/ peripheral blood (n=4) or from FFPE with sufficient DNA quality (n=2) (Supplementary Table S2). Moreover, comparison of the detection rate of a clonal result in lymphoma showed a better performance using DNA derived from FFPE material than from FF material (98 vs. 89%, Supplementary Table S5). A more likely explanation therefore seems to be that the BIOMED-2 study included multiple IGHV-IGHJ targets (i.e., FR1, FR2 and FR3) whereas only one (IGH FR3) was analyzed in this study. Indeed, additional analyses of IGH FR1 and FR2 PCRs gave a clonal result in 4 out of the 6 cases of B-cell neoplasia in which analysis with only IGH FR3 could not detect a clonal rearrangement (Supplementary Table S2). In addition, the BIOMED-2 study contained a larger number of lymphomas with lower levels of somatic hypermutation (i.e., chronic lymphocytic leukemia, mantle cell lymphoma).Analysis of reactive lymphoproliferations showed that the most dominant clonotype in that setting is often present in less than 5% of the reads although it can be present in as much as 14% of reads. Nevertheless, also most of the targets with a high percentage of dominant clonotypes were still scored as polyclonal in the light of the entire clonotype pattern, thus stressing the importance of evaluation of the pattern and not only the percentage of the most dominant clonotype. In the majority of B-cell lymphomas, the most dominant clonotype was represented by a much higher percentage of the reads, albeit that this is of course primarily dependent on the tumor cell percentage. In 81/82 reactive cases the NGS-based clonality assay was very helpful in confirming the polyclonal character of the reactive lesions. In one of the 82 reactive cases the final molecular result was reproducibly scored as clonal using the NGS-based clonality assay based on dominant clonotypes in all four different gene targets. Detection of small clones in reactive lesions has been described previously11Langerak A.W. Molina T.J. Lavender F.L. Pearson D. Flohr T. Sambade C. Schuuring E. Al Saati T. van Dongen J.J. van Krieken J.H. Polymerase chain reaction-based clonality testing in tissue samples with reactive lymphoproliferations: usefulness and pitfalls. A report of the BIOMED-2 Concerted Action BMH4-CT98-3936. and should result in a more detailed pathological review that might explain the presence of a small B-cell clone.As the NGS-based approach for clonality analysis allows quantitative assessment of the results, we made a first effort to find cut-off values to distinguish between a clonal and a polyclonal result. Indeed, using the percentages of the different clonotypes, the score of a particular target (clonal vs. non-clonal) could be predicted with high specificity and sensitivity. Although this is only a first and preliminary attempt to quantitatively assess clonality results, the results suggest that quantitative cut-off points could theoretically be useful to help interpret NGS-based clonality analysis. It has to be stressed that a quantitative analysis by strict cut-off values can never replace a careful assessment of the clonality analysis. This is because the DNA input, the DNA-quality of the sample and the PCR target are important parameters, as summarized in Table 3, which influence the number of reads and the clonotype pattern so that these not necessarily reflect the actual distribution of the different clonotypes in the tissue. Also, there are minor differences in efficiency of the different primer sets. In combination with the knowledge that immunoglobulin V, D, and J genes are utilized with unequal frequencies 15Pascual V. Wilson P. Liu Y.J. Banchereau J. Capra J.D. Biased VH4 gene segment repertoire in the human tonsil., 16Brezinschek H.P. Foster S.J. Brezinschek R.I. Dorner T. Domiati-Saad R. Lipsky P.E. Analysis of the human VH gene repertoire. Differential effects of selection and somatic hypermutation on human peripheral CD5(+)/IgM+ and CD5(-)/IgM+ B cells., 17de Wildt R.M. Hoet R.M. van Venrooij W.J. Tomlinson I.M. Winter G. Analysis of heavy and light chain pairings indicates that receptor editing shapes the human antibody repertoire., 18Arnaout R. Lee W. Cahill P. Honan T. Sparrow T. Weiand M. Nusbaum C. Rajewsky K. Koralov S.B. High-resolution description of antibody heavy-chain repertoires in humans. an entirely equal, unbiased, detection of V, D and J genes is not possible. Moreover, histomorphology, the suspected tumor cell percentage as well as immunological aspects 11Langerak A.W. Molina T.J. Lavender F.L. Pearson D. Flohr T. Sambade C. Schuuring E. Al Saati T. van Dongen J.J. van Krieken J.H. Polymerase chain reaction-based clonality testing in tissue samples with reactive lymphoproliferations: usefulness and pitfalls. A report of the BIOMED-2 Concerted Action BMH4-CT98-3936. should also be taken into consideration upon evaluation of clonality results, as summarized in Table 3. Implementation of NGS-based clonality analysis in the diagnostic laboratory has multiple potential advantages. First of all, the technique is highly sensitive, allowing the detection of small clones in a polyclonal background. The limit of detection via NGS-based clonality of 2.5% is not possible with the conventional assays because the clonal products will disappear in the polyclonal background. Secondly, the NGS-based clonality assay has been designed especially for FFPE specimens with suboptimal DNA-quality in contrast to the conventional assays (with size windows between 100 nt and 420 nt for detection of clonal rearrangements), via which potential clonal rearrangements in suboptimal DNAs will not be detected. Thirdly, a more accurate comparison of clonal B-cell populations is possible because the exact sequences can be compared rather than comparing peak sizes in GeneScan analysis. This is extremely valuable in determining clonal relatedness in patients with multiple lymphomas. Fourthly, some targets are more easily scored with the NGS-based approach. Especially IG kappa (IGK) can be difficult to score with GeneScan analysis due to the narrow Gaussian curve. With the NGS-based approach, each clonotype is visualized individually, allowing easier interpretation. The large amount of data that is generated with NGS-based clonality analysis opens up new possibilities. It carries the potential for a quantitative approach to the scoring of clonality analysis rather than scoring by pattern recognition only. Also, new applications become possible on a larger scale, for example assessment of the immunoglobulin repertoire.

Table 3Critical parameters for interpretation of NGS-amplicon-based clonality analysis of immunoglobulin genes

The introduction of NGS-based clonality testing needs an adapted workflow in a laboratory. Currently, in a diagnostic (pathology) laboratory many tests are NGS-based, therefore switching to NGS for clonality assessment can simplify the workflow from many different technologies to a few NGS-based workflows. Also, maintaining different technologies (NGS, GeneScan and/or heteroduplex analysis) in a laboratory requires validation of the workflows and the different assays under the accreditation standards.

The workflow for NGS-based clonality testing will probably affect the turnaround time, although this is highly dependent on the local situation. In general, it can be expected that the workflow for the conventional method of clonality analysis involving the EuroClonality/BIOMED-2 clonality assay with GeneScan has a turnaround time of 1-1.5 days excluding analysis time, whereas NGS-based clonality has an estimated turnaround time of 2.5 days for the wet-lab part and sequencing time, followed by the bioinformatics analysis.

Depending on the local situation, the ‘clonality sequencing samples’ (with amplicons of around 200bp) might be combined with other sequencing samples with similarly sized amplicons, which will save sequencing costs and reduce turnaround time. This will of course need local testing and validation. The sequencing costs will be highly dependent on the instrument platform, the number of samples and the maximum acceptable turnaround time.

In summary, in this study we have investigated the clonality profiles in 124 WHO-defined mature B-cell neoplasms, the majority of which concerned FFPE tissue. We showed that the NGS-based IG clonality assay, based on the combination of IGHV-IGHD-IGHJ FR3, IGKV-IGKJ as well as the unmutated targets IGHD-IGHJ and IGKV-KDE/intronRSS-KDE performs extremely well in mature B-cell neoplasms using good, but also suboptimal quality DNA. Even though we were able to formulate more objective parameters for interpretation, we want to stress that it remains important to evaluate clonality testing in the context of clinical, histological and immunophenotypic information.

Supplemental Data

Supplemental Figure S1: Results from cases with discordant interlaboratory results for NGS-based IG clonality analysis. A: Case EC-105 was a follicular lymphoma grade 3B with transformation to diffuse large B-cell lymphoma, diagnosed in a lymph node. In one laboratory (lab 01), analysis of IGKV-IGKJ and IGKV-KDE/IntronRSS-KDE gave a clonal result, but in the other lab (lab 02) the result was polyclonal. A repeat analysis gave a polyclonal result, resulting in a final score or polyclonality. B: For case EC-207, peripheral blood from a patient with Sjögren’s syndrome was analysed. In one laboratory (lab 01), analysis of IGHD-IGHJ gave a clonal result, but in the other lab (lab 02), the result was polyclonal. A repeat analysis gave a polyclonal result, resulting in a final score of polyclonality.

Supplemental Figure S2: NGS-based IG clonality analysis IGHV-IGHJ results of case EC-053 (reactive lymphoproliferation, not otherwise specified). This analysis shows a dominant clonotype that is present in a high percentage (green column, comprising 13.14% of reads) but due to the background which shows several additional dominant clonotypes, it was scored as polyclonal. The duplicate analysis by the other laboratory also gave a polyclonal result.

Supplemental Figure S3: Receiver operating characteristic curves for different ratios of the #1 clonotype % and the background clonotypes for the different targets of NGS-based IG clonality analysis.

Article InfoPublication History

Accepted: June 1, 2021

Received in revised form: May 17, 2021

Received: January 23, 2021

Publication stageIn Press Journal Pre-ProofFootnotes

Funding: this study was supported by the Dutch Health Insurers’ Innovation Fund (Project no. 17-179) and the revenues of the previously obtained patent (PCT/NL2003/000690), which is collectively owned by the EuroClonality/BIOMED-2 Consortium and licensed to Invivoscribe,

Conflict of interest disclosure: The EuroClonality-NGS Working Group is part of the EuroClonality Consortium, which is an independent scientific consortium that aims at innovation, standardization and education in the field of diagnostic clonality analysis. The revenues of the previously obtained patent (PCT/NL2003/000690), which is collectively owned by the EuroClonality/BIOMED-2 Consortium and licensed to Invivoscribe, are exclusively used for EuroClonality consortium activities, such as for covering costs of the consortium meetings, collective experiments, the External Quality Assessment schemes, and the EuroClonality Educational Workshops.

Identification

DOI: https://doi.org/10.1016/j.jmoldx.2021.06.005

Copyright

© 2021 Published by Elsevier Inc. on behalf of the Association for Molecular Pathology and American Society for Investigative Pathology.

User License Creative Commons Attribution (CC BY 4.0) | ScienceDirectAccess this article on ScienceDirect