Contributions
Abstract: S225
Type: Oral Presentation
Session title: Lymphoma - Translational research
Background
In Hodgkin lymphoma, treatment individualization is hindered by a lack of genomic characterization and technology for sensitive, molecular response assessment due to the scarcity of Hodgkin-Reed Sternberg (HRS) cells in the tumor tissue. Sequencing of cell-free (cf)DNA is a powerful strategy for genomic characterization and development of assays for sensitive disease monitoring.
Aims
The aim of this study is to introduce a targeted cfDNA sequencing platform that is then used to decipher the genomic landscape of HL and to facilitate minimal residual disease assessment (MRD).
Methods
We comprehensively genotyped and assessed minimal residual disease in 324 samples from 121 patients targeting approximately 3 Mb of the genome. All samples underwent pre-analytical processing, extraction of ctDNA and matching germline control DNA and sequencing at > 3000x average coverage post deduplication. An in-house, customised bioinformatics pipeline was developed for two-fold error reduction.
Results
First, we performed technical and clinical validation. We show that our approach can identify mutations with an allele frequency of 0.5% in cfDNA at 92.8% sensitivity and 100% specificity. 91.3% of mutations could be confirmed in tissue samples. For MRD detection, we identified a limit of detection of 0.00945%.
We identified 23 tumor suppressor genes (TSG) and 20 oncogenes (OG) in our HL cohort. Recurrent mutations were identified in several known oncogenic pathways in HL such as JAK-STAT signaling (SOCS1, STAT6, CSF2RB, IL4R), Class I MHC mediated antigen processing and presentation (B2M) and NF-kappa B signaling (TNFAIP3, IKBKB, NFKBIE). Novel findings with potential relevance to HL pathogenesis included recurrent mutations in EZH2, IRF4 and IRF8.
Six distinct COSMIC single base substitution (SBS) signatures were identified, the most common being SBS1 (aging), followed by SBS3 (defective homologous recombination (HR)-based DNA damage repair), SBS6 and SBS15 (defective DNA mismatch repair) SBS9 (somatic hypermutation in lymphoid cells) and finally SBS25 (so far only identified in HL cell lines). Interestingly, SBS3 - the defective HR signature - contributed majorly to mutations in genes crucial for HL oncogenesis, highlighting a potential novel role for defective HR in HL pathogenesis.
When linking HL genotypes to phenotypes, mutations in some genes, such as TP53 or NOTCH1, were associated with a high-risk phenotype and inferior treatment response, while other alterations, such as BRCA1 or MGMT loss were associated with a low-risk phenotype possibly mediated by increased sensitivity to commonly used cytotoxic drugs.
Furthermore, we identified a distinct genotype of large mediastinal mass HL, characterized by a higher frequency of B2M and IKBKB mutations, as well as a higher number of mutations in genes involved in Interferon signaling and JAK-STAT signaling. These alterations suggest remarkable biological similarity of large mediastinal mass HL to PMBCL.
Finally, we examined the relationship between MRD measured by following each patient’s unique mutational fingerprint at different timepoints, and interim PET-imaging after 2 cycles of treatment. MRD trajectories separated very early on, with clear differences already observable after 1 week. MRD after 1 week differentiated with 100% accuracy between Deauville 1-3 and Deauville 4-5 patients.
Conclusion
Here, we present the largest genomic landscape of HL to date and introduce our targeted ctDNA sequencing platform allowing for accurate genotyping using pre-treatment samples and highly sensitive minimal residual disease (MRD) detection.
Keyword(s): Genomics, Hodgkin's lymphoma, Minimal residual disease (MRD)
Abstract: S225
Type: Oral Presentation
Session title: Lymphoma - Translational research
Background
In Hodgkin lymphoma, treatment individualization is hindered by a lack of genomic characterization and technology for sensitive, molecular response assessment due to the scarcity of Hodgkin-Reed Sternberg (HRS) cells in the tumor tissue. Sequencing of cell-free (cf)DNA is a powerful strategy for genomic characterization and development of assays for sensitive disease monitoring.
Aims
The aim of this study is to introduce a targeted cfDNA sequencing platform that is then used to decipher the genomic landscape of HL and to facilitate minimal residual disease assessment (MRD).
Methods
We comprehensively genotyped and assessed minimal residual disease in 324 samples from 121 patients targeting approximately 3 Mb of the genome. All samples underwent pre-analytical processing, extraction of ctDNA and matching germline control DNA and sequencing at > 3000x average coverage post deduplication. An in-house, customised bioinformatics pipeline was developed for two-fold error reduction.
Results
First, we performed technical and clinical validation. We show that our approach can identify mutations with an allele frequency of 0.5% in cfDNA at 92.8% sensitivity and 100% specificity. 91.3% of mutations could be confirmed in tissue samples. For MRD detection, we identified a limit of detection of 0.00945%.
We identified 23 tumor suppressor genes (TSG) and 20 oncogenes (OG) in our HL cohort. Recurrent mutations were identified in several known oncogenic pathways in HL such as JAK-STAT signaling (SOCS1, STAT6, CSF2RB, IL4R), Class I MHC mediated antigen processing and presentation (B2M) and NF-kappa B signaling (TNFAIP3, IKBKB, NFKBIE). Novel findings with potential relevance to HL pathogenesis included recurrent mutations in EZH2, IRF4 and IRF8.
Six distinct COSMIC single base substitution (SBS) signatures were identified, the most common being SBS1 (aging), followed by SBS3 (defective homologous recombination (HR)-based DNA damage repair), SBS6 and SBS15 (defective DNA mismatch repair) SBS9 (somatic hypermutation in lymphoid cells) and finally SBS25 (so far only identified in HL cell lines). Interestingly, SBS3 - the defective HR signature - contributed majorly to mutations in genes crucial for HL oncogenesis, highlighting a potential novel role for defective HR in HL pathogenesis.
When linking HL genotypes to phenotypes, mutations in some genes, such as TP53 or NOTCH1, were associated with a high-risk phenotype and inferior treatment response, while other alterations, such as BRCA1 or MGMT loss were associated with a low-risk phenotype possibly mediated by increased sensitivity to commonly used cytotoxic drugs.
Furthermore, we identified a distinct genotype of large mediastinal mass HL, characterized by a higher frequency of B2M and IKBKB mutations, as well as a higher number of mutations in genes involved in Interferon signaling and JAK-STAT signaling. These alterations suggest remarkable biological similarity of large mediastinal mass HL to PMBCL.
Finally, we examined the relationship between MRD measured by following each patient’s unique mutational fingerprint at different timepoints, and interim PET-imaging after 2 cycles of treatment. MRD trajectories separated very early on, with clear differences already observable after 1 week. MRD after 1 week differentiated with 100% accuracy between Deauville 1-3 and Deauville 4-5 patients.
Conclusion
Here, we present the largest genomic landscape of HL to date and introduce our targeted ctDNA sequencing platform allowing for accurate genotyping using pre-treatment samples and highly sensitive minimal residual disease (MRD) detection.
Keyword(s): Genomics, Hodgkin's lymphoma, Minimal residual disease (MRD)