P-232: Comprehensive genomic sequencing predicts sub-optimal response to bispecific antibodies (BsAb) and chimeric antigen receptor T cells (CAR-T) in multiple myeloma (MM)

Introduction: BsAb and CAR-T provide impressive response rates in relapsed MM, however, some patients are either primary refractory or have short remission durations. Small cohorts have demonstrated target antigen loss and immune dysregulation contributing to resistance, requiring assays not routinely available in clinical practice. Response prediction in newly diagnosed MM is improved using extended targeted sequencing, with estimation of the structural variant chromothripsis using copy number variant (CNV) signatures (Maclachlan et al. Blood 2023).

Methods: We assessed genomic risk factors (RF) and treatment response in patients having received commercially available BsAb or CAR-T at MSKCC; combining FISH, SNP-array, and a clinical sequencing panel (MSK-IMPACT-Heme). If patients received >1 agent, each therapy response was evaluated. Survival was assessed by Kaplan-Meier, and GISTIC analysis examined significantly aberrant CNV.

Results: 221 patients were assessed; 126 received CAR-T, 133 BsAb, with 32 receiving >1 therapy. Median follow-up time was 194d (IQR 92-369), with 108 progression events. Genomic data was available in 191 (86%); FISH in 191, SNP array in 168 and MSK-IMPACT-Heme in 119. Samples were collected pre-therapy in the majority, post-therapy in 20 and at both timepoints in 15.

While t(14;16) predicted shorter PFS with both CAR-T (median 83d, p=0.007) and BsAb (median 24d, p=0.03), none of t(4;14), gain1q, amp1q, del1p or MYC aberration were significant in univariate analysis. In BsAb patients only, TP53mut predicted adverse PFS (median 66d, p=0.0004), as did del17p (median 58d, p=0.0003) and biallelic TP53 aberration (median 34d, p< 0.0001). Mutations in KRAS, NRAS or DNA-damage-response genes were not associated with short PFS.

GISTIC analysis in those who had progression or death within 3m showed an enrichment of focal gains at 7p12.1, 8q22.2, 16q22.1 and 17q22. Complex genomics, defined by CNV suggesting chromothripsis, showed a trend to significance, which when combined with biallelic TP53 aberration, predicted a median PFS of 21d. Considering the effect of multiple RF; for those reaching progression or death within 3m, 46/68 (68%) had ≥2 standard RF, which increased to 51 (75%) with the inclusion of TP53mut, MYC aberration and complex genomics.

Broad monoallelic CN loss at BCMA/TNFRSF17 (del16p13) or GPRC5D (del12p13) was detected in 22% and 24% pre-therapy and did not predict resistance to the relevant targeted agent. Post-therapy genomic evolution was observed in paired samples, with emerging TP53 and KRAS mutations, gain8q (MYC), focal deletion of GPRC5D and CNV suggesting chromothripsis affecting chr16.

Conclusions: BsAb and CAR-T may overcome individual standard genomic RF, however harboring multiple genomic RF remains adverse. Clinically available extended genomic characterization predicts response better than FISH alone in this large cohort treated with BsAb/CAR-T. This work will extend into pre-/post-therapy whole genome sequencing for more granular data analysis.