Sequencing-Based Detection of Circulating Tumor DNA in the Autologous Stem Cell Grafts of Patients with Diffuse Large B-Cell Lymphoma Undergoing Hematopoietic Stem Cell Transplantation

Introduction: High dose chemotherapy followed by autologous stem cell transplantation (ASCT) cures a subset of patients with chemosensitive relapsed or refractory (rel/ref) diffuse large B-cell lymphoma (DLBCL). Several factors associated with post-ASCT outcome have been identified, including pre-ASCT PET status, but better biomarkers are needed in order to optimally select candidates for the procedure. In other lymphoma subtypes with defining chromosomal translocations, PCR detection of pre- and post-ASCT minimal residual disease (MRD) in peripheral blood and of tumor contamination in the stem cell product is associated with inferior outcome. Until recently, MRD detection in DLBCL was limited by the rarity of detectable circulating disease using conventional techniques. The immunosequencing platform (Adaptive Biotechnologies, Corp.) is a next-generation-sequencing (NGS)-based MRD assay that detects small amounts of circulating tumor DNA (CTD) in patients with lymphoid malignancies. The assay detects CTD at diagnosis in most DLBCL patients and CTD levels track with response to induction therapy (Armand, 2013). Persistence of CTD or recurrence of CTD after completion of therapy is highly associated with DLBCL relapse (Kurtz, 2015; Roschewski, 2015). We evaluated whether CTD in autologous stem cell grafts was predictive of outcome in patients with rel/ref DLBCL undergoing ASCT.

Methods: We retrospectively studied patients with rel/ref DLBCL, including transformed indolent lymphoma (TIL), who had paired archival tumor and autologous stem cell specimens and underwent ASCT at Brigham and Women's Hospital/Dana-Farber Cancer Institute from 2003-2013. Genomic tumor DNA was extracted from archival formalin-fixed paraffin-embedded (FFPE) tissue and analyzed using the NGS-based MRD assay. PCR amplification of IGH-VDJ, IGH-DJ, and IGK regions using universal consensus primers was performed followed by NGS to determine the tumor clonotype(s), defined as having a frequency > 5% in the tumor specimen. DNA from all available autologous peripheral blood or bone marrow stem cell specimens from each patient was amplified using universal consensus primers and sequenced to determine the level of CTD, defined as the number of lymphoma molecules per diploid genome.

Results: We identified 98 eligible patients with rel/ref DLBCL/TIL. The median age was 60 (range 22-77) years; 63% were male; 65% had DLBCL, 29% had TIL, and 5% had primary mediastinal DLBCL; the median number of prior lines of therapy was 2 (range 2-5); all had received prior rituximab; 38% had primary refractory disease; 60% were in complete remission at ASCT; 96% received CBV conditioning. Median follow-up was 56 (range 19-123) months. The 4y progression-free survival (PFS) and overall survival (OS) in the entire cohort were 46% and 64%, respectively.

Among 83 patients (85%) with sufficient DNA for clonotype determination, a clonotype was identified in 59 (71%). CTD data was complete in 53 patients (52 received peripheral blood stem cells (PBSC) and 1 received bone marrow). Eight patients (15%) had detectable CTD (CTD+) in the stem cell autografts (all PBSC) and 6/8 relapsed after ASCT. One CTD+ patient had early non-relapse mortality less than 1 month after ASCT and was never restaged. Seven of 8 CTD+ patients had TIL histology, 5 of whom relapsed (4 with aggressive lymphoma). The 4y PFS and OS in CTD+ v CTD- patients were 13% v 48% (p=0.01), and 38% v 67% (p=0.013), respectively [Figure 1]. In multivariable models including CTD status and pre-ASCT characteristics, CTD+ was the only factor associated with OS (HR 3.1, p=0.018), but was not significantly associated with PFS.

Discussion: In patients with rel/ref DLBCL undergoing ASCT, the presence of CTD in the autologous stem cell graft is associated with inferior survival. CTD detection in the autograft may be more common in patients with TIL. In studies evaluating CTD detection in DLBCL, the plasma compartment has been more sensitive for detecting CTD than mononuclear cells. The use of concentrated cell specimens in this study may have decreased the sensitivity of the assay.  Nevertheless, if the present findings are confirmed in a larger population, CTD detection may permit the identification of a subgroup of patients with a particularly poor outcome after ASCT, for whom alternative approaches could be considered.

Figure 1 Overall (A) and Progression-Free (B) Survival in CTD+ vs CTD- Patients