Abstract T026

Circulating tumor DNA sequencing facilitates biological classification and individualized risk stratification in patients with Hodgkin lymphoma

Introduction: The development of biomarkers identifying high-risk Hodgkin lymphoma (HL) patients based on biological risk factors available before treatment initiation remains a high unmet medical need. We previously presented a biological classification of HL consisting of three subtypes based on plasma-derived circulating tumor (ct)DNA sequencing: Inflammatory immune escape HL is characterized by frequent copy number variations including immune escape variants such as high-level amplifications of the PD-L1 locus and an inflammatory tumor microenvironment. Virally-driven HL shows strong association with Epstein-Barr virus (EBV) and/or Human herpesvirus (HHV)6 as well as a tumor microenvironment with increased presence of cytotoxic T-cells and NK-cells. Oncogene-driven HL is defined by a high tumor mutational burden including recurrent mutations in common oncogenic drivers known in HL.

Methods: To assess clinical applicability and prognostic relevance of our classification, we performed a blinded clinical validation in an event-enriched cohort consisting of 72 patients from the GHSG HD21 trial. To increase clinical feasibility, we used a novel, validated assay in this study (LymphoVista HL, validation data presented in a separate abstract at this meeting).

Results: 64/72 (88.9%) patients were successfully assigned to one of the three subtypes. We weighted the outcome analysis to reflect the HD21 trial population. Despite the use of highly efficient treatment regimen in the HD21 trial (eBEACOPP and BrECADD), we were able to detect clinically meaningful differences in progression-free survival (PFS) between Inflammatory immune escape HL (3-year PFS 86.4%), Virally-driven HL (3-year PFS 92.7%), and Oncogene-driven HL (3-year PFS 97.1%) (Figure 1). When additionally assessing minimal residual disease using ctDNA, we were able to identify patients at very high risk of relapse within the subtypes.

Conclusion: We propose a clinically feasible, noninvasive method for upfront individualized risk stratification in patients with HL based on ctDNA sequencing. MRD assessment during treatment using the same assay further refines risk assessment.