OA – 27: Para-medullary (PMD) and extra-medullary (EMD) myeloma demonstrate increased copy number aberration, mutational burden, structural variants and genomic complexity compared to marrow-based myeloma

Introduction: Growth beyond the bone marrow (BM) represents clinically aggressive myeloma and is associated with inferior therapy response. We describe the largest PMD/EMD cohort to date, with comprehensive genomic characterization from combining SNP-array, extended targeted sequencing (MSK-IMPACT-Heme) and whole genome sequencing (WGS) data.

Methods: We examined all available in-house PMD/EMD samples, comparing with paired BM where available, with our published WGS from NDMM BM (Maura et al. Nature Cancer 2023) and with WGS from PMD/EMD collected during autopsy (Landau et al. Nat Comm 2020). WGS data were processed via Isabl, MSK-IMPACT-Heme via cBioPortal, SNP-array by clinical pathology, all using in-house validated pipelines.

Results: 148 plasmacytoma samples were assessed; 99 PMD, 49 EMD, with 53 pre-therapy and 95 post-therapy. 48 patients had paired BM samples (28 coincident and 23 distant in time), with 11 patients having > 1 PMD/EMD biopsy. Including our previously published data, the cohort comprised 258 samples from 176 patients: 105 WGS, 133 MSK-IMPACT-Heme, and 20 SNP-array.

Within both the whole cohort and samples with paired BM, PMD/EMD had considerably more copy number aberration (CNA) than BM. GISTIC analysis for significant CNA detailed broad gains affecting 1q, 8q (MYC), 20p and 20q, with focal gains at 1q21, 6p25 and 18q21. Broad deletions affected 4p, 8p and 17p (TP53), and numerous focal deletions were detailed. Notably, gain/amp1q was observed in 78% of EMD overall, and in 83% of post-therapy EMD samples.

Prevalent mutations in PMD/EMD affected KRAS, NRAS, and TP53, with KRAS being enriched in PMD (36%), while EMD was enriched for NRAS (38%) and TP53 (31%). Dndscv analysis revealed multiple genes to be positively selected as driver events, including epigenetic modifiers (KMT2A, KMT2C, TET2, SMARCA4), immune response (PRMD1, LTB), proliferation (STAT3, NF1, BTG1), adhesion (FAT1), and DNA-damage response (TP53, ATM, BRCA2).

Paired BM-PMD/EMD showed 2 patterns; some had increased mutational burden, while others had retained mutations but increased CNA. Within those with > 1 PMD/EMD sample, diverging clones were defined by a variety of mutations (including KRAS, TP53, STAT3, IGF1R, DNMT3A), CNA affecting multiple chromosomes (including 8q; MYC, 17p; TP53), noncanonical translocations, and CNA consistent with chromothripsis.

In the WGS data, we noted spatial divergence in mutational signature contribution and the emergence of complex structural variants. Several melphalan-exposed patients had SBS99 evident in the phylogenetic tree trunk of multiple biopsies, consistent with single cells surviving transplant and subsequently seeding in multiple sites.

Conclusions: PMD and EMD demonstrate multiple features of genomic complexity when compared with BM-based myeloma, including emerging copy number aberration, mutational burden, and complex structural variants. Ongoing studies include expanding the WGS dataset, and correlation of genomic features with clinical response to therapy.