Mutations and Copy Number Changes Predict Progression from Smoldering Myeloma to Symptomatic Myeloma in the Era of Novel IMWG Criteria

Introduction: Despite novel International Myeloma Working Group (IMWG) criteria, Smoldering Myeloma (SMM) remains a heterogeneous disease for which correctly identifying patients that will eventually progress to myeloma (MM) is essential. The genetic and molecular factors that underlie disease progression are not well elucidated, therefore, we examined samples from SMM patients in order to identify molecular determinants of progression.

Methods: CD138-sorted and control samples from 77 non-treated SMM patients according to IMWG 2014 underwent targeted sequencing and gene expression profiling (GEP). The median follow-up was 4.81 years (95% CI: 4.19-6.16). Targeted sequencing consisted of 140 genes and additional regions of interest for copy number, as well as tiling of the immunoglobulin and MYC loci for detection of translocations and was performed on a NextSeq500 using 75 bp paired end sequencing. Results were aligned to the hg19 genome and mutations, translocations and copy number were determined. Nonnegative matrix factorization (NMF) (NMF package in R) was used to identify mutation signatures. The median mean coverage was 365 (88-696) and 783 (161-1559) for translocations (Tx) and mutations respectively. We compared these samples to 199 newly diagnosed MM samples.

Results: Significant differences in the frequencies of mutated genes were seen, including fewer NRAS, KRAS, FAM46C, LRRK2 and TP53 mutations and more PCLO and MAFB mutations than expected in comparison to MM (p<0.05). Regarding structural changes, there was no difference in the incidence of Tx (including those involving MAF and MAFB) but significantly fewer del(1p), del(12p) and del(14q) cases in SMM (p<0.05). There was no difference in the incidence of MYC translocations (19% of cases) and MYC rearrangements (23%). The 4-year progression rate was 25 percent. The presence of KRAS mutations (n=9) and del(6q) (n=11) were statistically associated with shorter progression free survival (PFS) [median 49m (26-∞) vs 147m (67.6-∞) and 26m (9.6-∞) vs 147m (80-∞) for KRAS and del(6q) respectively] and treatment free survival (TFS) [median 6m (9.6-∞) vs 19m (9.6-∞) and 9 m (13.4-∞) vs 16m (13.4-∞) for KRAS and del(6q) respectively]. MYC alterations and NF-κB alterations (BIRC2 and BIRC3 loss, TRAF2 and TRAF3 mutation or gain, CYLD loss, MAP3K14 mutations) did not influence progression. There were no double-hit patients in this cohort defined by bi-allelic-TP53 or ISS III with amp(1q). Ten percent of patients were identified as high-risk according to GEP4 risk-score.

Pearson correlation was performed between patients that progressed (n=24) against those who did not (n=53) for genetic events with n≥6. Del(6q) [χ2=0.32, p=0.004], LRP1B [χ2=0.27, p=0.015] and KRAS mutations [χ2=0.28, p=0.01] were positively correlated to progression, but only del(6q) remained significant after Bonferroni adjustment.

Of particular interest, we did not identify the APOBEC mutational signature in the t(14;16) SMM samples, which is heavily associated with a poor prognosis in t(14;16) MM (4/11 in MM and 0/5 in SMM).

Discussion: As previously reported, copy number changes, Tx and mutations predate MM. The lower frequencies of copy number changes and mutations suggest an ongoing process whereby cells acquire successive events eventually leading to MM. KRAS and del(6q) were significant predictors of both PFS and TFS with hazard ratios of 2.8 and 3.71, respectively. We comprehensively analyzed both the NF-κB pathway mutations and copy number changes, that did not bear, unlike previous reports, any clear relationship to PFS. Although we are limited by the power of this analysis, this supports the idea that the NF-κB dependency preexists symptomatic myeloma and is present throughout disease stages. Further analysis of the NF-κB 11-gene signature expression are ongoing. This is the first broad analysis of both MYC rearrangements and Tx in SMM. Previous studies have focused on FISH analysis of IGH-MYC Tx that underestimate the extent of MYC rearrangements present. Finally, our data also shows that absence of an APOBEC signature in SMM may account for the rather indolent phenotype of MAF and MAFB Tx in comparison to MM.

Conclusion:KRAS mutations as well as del(6q) were associated with shorter PFS and TFS in this dataset. The absence of APOBEC signature may explain part of the indolent phenotype of the MAF and MAFB translocation SMM patients.