Circulating tumor DNA of classical Hodgkin lymphoma.

Abstract

7516 Background: Classical Hodgkin lymphoma (cHL) is the most common lymphoma in young adults. The morphologic peculiarity that makes cHL unique are the few tumor cells surrounded by a polytypic inflammatory microenvironment. It is also the cancer with the highest sensitivity to immune checkpoint inhibitors. But why is it so immunogenic and how it escapes the immune control at the same time have only been partially elucidated. Although cure rates of cHL have constantly increased over the years there is still a high need for identification of biomarkers for more precise treatment decisions. The need for enriching tumor cells before assaying has limited cHL biological studies. Patients with cHL have significant amounts of circulating tumor DNA (ctDNA), suggesting an alternative approach for studying this enigmatic tumor. Methods: IOSI-EMA003 (NCT03280394) is a prospective, observational, multi-center study of adult patients with previously untreated cHL aiming at: i) identifying subgroups of patients with phenotype- and outcome-associated molecular signatures; ii) testing and validating baseline ctDNA load as a prognostic biomarker; iii) testing if ctDNA can be used for the early identification of chemoresistance. Blood samples were collected at staging and disease response assessment. PET scans were centrally and blindly reviewed. ctDNA was analyzed by using deep targeted and shallow whole genome sequencing to measure ctDNA load, capture mutations and profile ctDNA fragmentation patterns. The composition of the tumor microenvironment of cHL was probed by single cell RNA-seq. Results: A total of 215 patients were enrolled. Based on ctDNA fragmentation patterns reflecting chromatin accessibility in the regulatory region of germinal center (GC) B-cell genes, we segregated cHL into two subgroups that we named SNCD (for sub-nucleosomal cfDNA) and NCD (for nucleosomal cfDNA). We hypothesized that SNCD cHL stems from a cell that is closer to the GC B-cell differentiation stage than NCD cHL. SNCD cHL and NCD cHL differed in many aspects, including activation induced cytidine deaminase (AID)-hypermutation profile, neoantigen load, immune editing mechanism and response to chemotherapy and checkpoint inhibitors. Clinically, SNCD cHL displayed less sensitivity to both chemotherapy and anti-PD1 antibodies. Immune editing of SNCD cHL points loss of MHC-I, recruitment of Treg and upregulation of LAG3 as prominent immune checkpoint. High load of pre-treatment ctDNA nominated high-risk patients more accurately than clinico-radiological features. In addition, persistence of residual 18FDG avid lesions and measurable ctDNA was a better proof of chemoresistance than the sole persistence of residual 18FDG avid lesions. Conclusions: Our results provide a roadmap for cHL subtyping and molecular classification. We also propose ctDNA as a quantifiable and radiation-free biomarker to be used in clinical trials testing a more personalized treatment approach. Clinical trial information: NCT03280394 .