Background: Individualization of treatment in Hodgkin's lymphoma is necessary to improve cure rates and reduce treatment side effects. Currently, it is hindered by a lack of genomic characterization and sensitive molecular response assessment. Sequencing of cell-free DNA is a powerful strategy to understand the cancer genome and can be used for extremely sensitive disease monitoring. In Hodgkin's lymphoma, a high proportion of cell-free DNA is tumor-derived, whereas traditional tumor biopsies only contain a little tumor-derived DNA. Methods: We comprehensively genotype and assess minimal residual disease in 121 patients with baseline plasma as well as 77 follow-up samples from a subset of patients with our targeted cell-free DNA sequencing platform. Findings: We present an integrated landscape of mutations and copy number variations in Hodgkin's lymphoma. In addition, we perform a deep analysis of mutational processes driving Hodgkin's lymphoma, investigate the clonal structure of Hodgkin's lymphoma, and link several genotypes to Hodgkin's lymphoma phenotypes and outcome. Finally, we show that minimal residual disease assessment by repeat cell-free DNA sequencing, as early as a week after treatment initiation, predicts treatment response and progression-free survival, allowing highly improved treatment guidance and relapse prediction. Conclusions: Our targeted cell-free DNA sequencing platform reveals the genomic landscape of Hodgkin's lymphoma and facilitates ultrasensitive detection of minimal residual disease.

Individualizing treatment is key to improve outcome and reduce long-term side-effects in any cancer. In Hodgkin lymphoma (HL), individualization of treatment is hindered by a lack of genomic characterization and technology for sensitive, molecular response assessment. Sequencing of cell-free (cf)DNA is a powerful strategy to understand an individual cancer genome and can be used to develop assays for extremely sensitive disease monitoring. In HL, a high proportion of cfDNA is tumor-derived making it a highly relevant disease model to study the role of cfDNA sequencing in cancer. Here, we introduce our targeted cfDNA sequencing platform and present the largest genomic landscape of HL to date, which was entirely derived by cfDNA sequencing. We comprehensively genotype and assess minimal residual disease in 324 samples from 121 patients, presenting an integrated landscape of mutations and copy number variations in HL. In addition, we perform a deep analysis of mutational processes driving HL, investigate the clonal structure of HL and link several genotypes to HL phenotypes and outcome. Finally, we show that minimal residual disease assessment by repeat cfDNA sequencing as early as a week after treatment initiation is feasible and predicts overall treatment response allowing highly improved treatment guidance and relapse prediction. Our study also serves as a blueprint showcasing the utility of our platform for other cancers with similar therapeutic challenges.

No abstract