Background: Multiple myeloma (MM) demonstrates significant disease heterogeneity both intratumorally with different clonal subtypes and spatially with focal involvement throughout the body. The presence of large focal lesions (FL) is a strong independent prognostic factor in MM. However, the genetic composition of FLs and the impact of treatment on them have not been entirely characterized. The aim of our study is to prospectively compare the genetic profile of myeloma samples between standard bone marrow (BM) biopsy versus systematically selected image guided FL biopsy.

Method: Newly diagnosed MM patients were enrolled in our institution's prospective global response assessment study with incorporation of advanced imaging (NCT03375567) and received paired biopsies of the iliac crest for standard BM biopsy evaluation along with CT-guided biopsy of a FL before and after induction therapy with 4 cycles carfilzomib, lenalidomide, and dexamethasone (KRd). FL were identified for biopsy following diagnostic radiology review for the most aggressive appearing lesion. Plasma cells were enriched using anti-CD138 bead selection in aspirate samples with sufficient cellularity. Whole exome sequencing (WES) using Illumina Next Generation Sequencing with a depth of 150x was performed. Peripheral blood of patients following induction was used as normal control. Variants were identified with an in-house developed pipeline. Exonic variants were filtered using the FATHMM-MLK score of > 0.7 for nonsynonymous SNVs and for those identified in the COSMIC90 database.

Results: 13 patients with paired BM and FL biopsies at diagnosis and/or after induction were included in the analysis. Plasma cells (PC) were able to be enriched from 3 diagnosis BM samples. Mutational profile of whole BM samples shared on average 76.7% of mutations (range: 51.1-89.6%) with PC enriched samples. Overall number of mutations identified in BM and FL at initial diagnosis and end of treatment (EOT) were similar. The average percent of unique mutations found in the FL compared to BM biopsy samples was 18.9% at diagnosis and 10.9% at EOT (p = 0.28, Figure). Presence of a high-risk mutation (HRM), including TP53, MAPK, KRAS, NRAS, BRAF, CKS1B, FAF1, CDKN2C, DIS3, MAF, and MMSET, was found in 30% and 22.2% of diagnosis BM and FL biopsies, respectively. No HRM were detected in EOT BM biopsies, but 20% of EOT FL biopsies had a HRM. In the patients with longitudinal biopsies and a HRM identified at initial diagnosis, those HRMs were not detected in their EOT BM and FL biopsies. Additionally, one patient with longitudinal paired biopsies demonstrated the development of a HRM in the EOT FL biopsy despite no detectable HRM at initial diagnosis. TP53 mutations represented 28.6% of all HRM mutations detected and were exclusively found in EOT FL biopsies. In the 10 patients who had WES performed in their EOT paired samples, 30% achieved a CR, 50% achieved a VGPR, and 20% achieved PR by IMWG response criteria.

Conclusion: Using whole exome sequencing, we demonstrate spatial genetic heterogeneity between BM and FL biopsies in newly diagnosed MM patients prospectively. The degree of mutational heterogeneity found in FL was not significantly different between initial diagnosis and following induction therapy with KRd. HRMs were seen in 20-30% of BM and FL samples. Evaluation of longitudinal samples show patterns of HRM loss in the BM and HRM gain in FL following treatment. Overall mutational burden persisted despite clinical response.

Dhodapkar:Lava Therapeutics, Sanofi, Janssen: Membership on an entity's Board of Directors or advisory committees. Halene:Forma Therapeutics: Consultancy. Xu:Pure Marrow: Consultancy; Seattle Genetics: Membership on an entity's Board of Directors or advisory committees; Blueprint Medicines: Consultancy; Seattle Genetics: Consultancy. Gorshein:Janssen: Speakers Bureau; AbbVie: Membership on an entity's Board of Directors or advisory committees. Haims:Pfizer: Consultancy. Wei:Takeda: Research Funding. Neparidze:Janssen: Research Funding; GSK: Research Funding.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution