Distinct Molecular Aberrations of Classic Hodgkin Lymphoma in Older Adults Identified By Comprehensive Genomic Profiling

Introduction

Despite curative advances in classic Hodgkin lymphoma (cHL), patients (pts) >60 years have worse outcomes, especially those with EBV-related disease. Older pts are more likely to be ineligible for intensive treatment. However, the underlying biological mechanisms are also not well understood. To address this, we employed a comprehensive approach using circulating tumor DNA (ctDNA) for non-invasive genotyping, gene expression inference, and viral characterization; along with tumor transcriptome & microenvironment profiling.

Methods

We comprehensively profiled 311 newly diagnosed cHL pts participating in prospective LYSA clinical trials (PVAB, AHL2011) and real-world cohorts (REALYSA, BIO-LYMPH). We defined older age as >60y. ctDNA was genotyped using CAPP-Seq to identify SNVs and CNAs (Alig, Nature 2024). EBV status and virome profiling were determined with VirCAPP-Seq (Garofalo, ASH2022), plasma T-cell repertoire (TCR) profiling using SABER (Sworder, Cancer Cell 2023) and transcriptional profiling by EPIC-Seq (Esfahani, Nat. Biotech. 2022). Tumor bulk RNA-Seq was used for differential expression and Ecotypes (Steen, Cancer Cell 2021), and cytokines assessed by ELLA immunoassay.

Results

Tumor mutational burden was markedly lower in older pts (median 90 vs 162.5 mutations/MB, p=0.01). Mutations in younger pts had significantly higher contribution of specific AID signatures (SBS84, SBS85), as determined by Single Base Substitution spectra. Conversely, older pts had more frequent somatic deletions (7 vs 2, p<0.01). Older pts harbored significantly more BCL2 and TET2 mutations and chr15p deletions, but fewer CIITA, XPO1, B2M, STAT6, GNA13, ITPKB, and NFKBIE mutations, and fewer chr2p and chr17q amplifications. Comparing these findings with inferred transcription by EPIC-Seq, genes in the JAK-STAT and TARC pathways showed significantly higher inferred expression (p=0.005) in younger pts. Cytokine assays confirmed lower TARC protein levels in older pts (p=0.018).

We identified a subset of 21 pts with multiple noncoding BCL2 mutations, who were exclusively >45y with the majority harboring concurrent t(14;18) BCL2-IGH in ctDNA. In contrast with the prior literature, we confirmed these fusions to reside in tumor Hodgkin-Reed-Sternberg (HRS) cells by BCL2 FISH. Fragmentomic patterns, allelic levels, and ctDNA dynamics also supported these BCL2 lesions as originating from tumor HRS cells, with no evidence for composite FL. BCL2-mutated pts were almost invariably EBV-neg by EBER ISH. Clinically, BCL2-mutated pts had more advanced disease (91% vs 69%), poorer outcomes (4-years PFS 58% vs 81%), and a higher relapse rate compared to BCL2wt/EBV-neg older pts (HR=2.6, p=0.03).

We used VirCAPP-Seq to enable clinical stratification of viral EBV burden. The ratio of EBV to HL ctDNA accurately distinguished EBV+ pts (ratio=21 in EBER+ vs 0.06 in EBER-; p<0.001). Elevated polyomavirus and anellovirus DNA in older pts served as a surrogate for impaired immune function.

Expression profiling by EPIC-Seq and tumor RNA-Seq both revealed stronger expression of IFNG signaling and macrophage activation genes in older pts (p=0.026). By EcoTyping tumors, we identified distinct macrophage states by age, revealing enrichment of key immune escape genes in older cHL pts. Additionally, older pts had lower lymphocyte counts (p=0.006). Unlike in younger pts, EBV status in older pts had no effect on plasma TCR clone count (p=0.58) or TCR entropy (p=0.79).

Clinically, older pts had less frequent bulky disease (19% vs 35%, p=0.003) and a higher prevalence of EBV+ (61% vs. 26%, p<0.001), despite having similar IPS, and PET/CT metrics. With a median follow-up of 4.3 years, relapse rates were significantly higher in older pts (HR=3.4, p<0.001). By combining the risk factors (RF) of ctDNA burden, EBV status and BCL2 status, older pts with >= 2 RF had significantly worse PFS than those with 1 or less. Thus, 3 highly distinct risk groups of older pts could be stratified (p<0.001).

Conclusions

Older cHL pts harbor distinct genomic, transcriptomic, and immunological characteristics. Notably, our identification of a previously unrecognized high-risk subgroup with BCL2 mutations has significant clinical and therapeutic implications. These findings could guide newly personalized treatment strategies for older cHL pts.