Circulating tumor DNA <i>NRAS</i> and <i>TP53</i> mutations to identify ultra-high-risk primary refractory myeloma patients.

Abstract

7557 Background: Approximately 5% of newly diagnosed multiple myeloma (MM) patients progress on bortezomib-containing first-line (1L) therapies and develop primary refractory disease (1REF). These patients have unacceptably poor outcomes. Underlying drivers of 1REF remain unclear. We hypothesized that circulating tumour DNA (ctDNA) profiling would identify the genetic drivers of 1REF. Methods: We analyzed plasma ctDNA, bone marrow (BM) MM cells from 89 patients, comprising 33 1REF and 56 contemporaneous 1L-responsive (RES) patients. ctDNA was available for all patients; BM was available for 63% (18 1REF and 38 RES). Ultra-deep targeted sequencing used a 21-gene panel enriched for MM drivers. High-confidence pathogenic variants were defined as variant allele frequency ≥1% with ≥15 supporting reads. Genomic enrichment analyses used Fisher's exact test. Survival outcomes (progression-free survival, PFS and overall survival, OS) were available for 27 1REF patients and analyzed using Kaplan–Meier methods. Results: Interrogation of BM showed more mutations in 1REF versus RES (67% vs 34%; p=0.04) with RAS/RAF predominating (67% vs 26%; p=0.007). Mutation patterns in 1REF versus (vs) RES comprised predominantly KRAS (44% vs 21%), followed by NRAS (22% vs 8%). BRAF was detected exclusively in 1REF (17% vs 0%) and TP53 showed no difference (6% vs 8%). ctDNA profiling revealed pathogenic variants in 18% of patients, enriched in 1REF vs RES (30% vs 11%; p=0.02), with RAS/RAF pathway mutations predominating (24% vs 2%; p=0.001), in line with BM observations. However, ctDNA RAS/RAF mutation patterns differed strikingly from BM. In 1REF vs RES, these comprised mainly NRAS (12% vs 0%; p=0.01) and BRAF (9% vs 0%; p=0.04), detected exclusively in 1REF, and KRAS (6% vs 2%). TP53 did not co-occur with RAS/RAF and were comparable across both cohorts (6% vs 5%). Mutation-survival associations in 1REF patients also revealed differences between BM and ctDNA. BM mutations lacked prognostic value while ctDNA mutations predicted survival, suggesting ctDNA captures treatment-resistant systemic disease. Patients with any ctDNA mutations had significantly shorter PFS and OS compared to those without (PFS: 8.5 months vs not reached [NR], p=0.0005; OS: 12.0 months vs NR, p=0.004). Patients with NRAS mutations showed consistently poor outcomes, while the impact of BRAF varied. The worst outcome was seen in an ultra-high-risk subset (15%, 4/27) with TP53 or NRAS , with all patients demonstrating an early fatal outcome (PFS 5.3 and OS 5.6 months; both p&lt;0.0001). Conclusions: ctDNA TP53 or NRAS identify a high-risk subset of 1REF patients with exceptionally poor outcomes, demonstrating that liquid biopsy, not BM evaluation, captures treatment-resistant systemic biology. These findings establish ctDNA as superior for risk stratification in 1REF MM and highlight an urgent need for alternative therapies in these patients.