OAB-059: Towards a comprehensive multimodal minimal residual disease assessment in multiple myeloma: the role of circulating cell-free DNA to define the extent of disease spreading

Background Multiple Myeloma (MM) is a plasma cell (PC) disorder characterized by the presence of multiple lytic lesions at the time of diagnosis. Recently, cell-free DNA (cfDNA) has proven to resume the heterogeneity of spatially distributed clones. However, the potential of cfDNA to track the evolutionary dynamics and the heterogeneity of MM, possibly anticipating the emergence of therapy resistant residual cells, remains to be confirmed. Aim of this study is to evaluate cfDNA at diagnosis and during follow-up in order to compare this approach with both bone marrow (BM) molecular and whole-body imaging residual disease assessment. Methods A total of 97 patients (pts) were screened at baseline and during follow-up with 18F-FDG PET/CT and NMR, and molecularly assessed by Ultra Low Pass-Whole Genome Sequencing (ULP-WGS). ULP-WGS was used to characterize both the neoplastic PC clone from BM (gDNA) and the cfDNA from peripheral blood. Data were analysed by ichorCNA and Clonality R packages. Results Overall, cfDNA tumour fraction (TF) at diagnosis was significantly lower as compared to gDNA TF [median (M) TF: 3.0%vs.74.4%, respectively]. Nevertheless, high cfDNA TF levels (49/97 pts = 50%, M cfDNA TF: 10%, range : 3.0-40.6%) correlated with high gDNA TF levels (M gDNA TF: 84%, range: 5.9-95.2%). Similarly, a significant correlation was shown between cfDNA TF and the BM plasma cells clone throughout disease progression, highlighting a parallel dynamics of tumour burden from SMM to MM to +3m post-induction (cfDNA TF vs. % CD138+/CD38+ BM cells: 1.5 vs 1.0; 3.0 vs. 2.3; 1.2 vs. 0.01, respectively, p: 0.008). Pts with high cfDNA TF were likely to have more sites of extramedullary disease (EMD) at PET-CT, a higher number of PET lesions and higher tumour metabolic activity, as compared to pts with low TF (EMD 8/49, 17% vs. 2/49, 3.4%; M n. PET lesions: 8 vs. 2; SUVmax: 12.9 vs. 3.9). Similarly, bone involvement as detected by MRI, was more evident in pts with high cfDNA TF (M n. focal lesions: 6 vs. 1). More interestingly, after 3 cycles of induction therapy, imaging and cfDNA TF dynamics were concordant: indeed, in those cases where FDG activity and focal lesions still persisted (4/10=40%: SUVmax 5.8; >2 PET lesions), cfDNA TF as well was still detectable (>1 TF).Finally, the comparison between cfDNA and BM genomic profiles showed an overall concordance of the copy number alterations (CNAs) landscape in most patients (83/97; 86.4%), whereas three patients (10/97; 10,3%) displayed a different genomic profile in cfDNA, as compared to BM. Conclusions although only few cases showed genomic evidence of spatial heterogeneity, in pts with high cfDNA TF, imaging data overall suggested a propensity to a metastatic spread of the disease. Future studies will be addressed to exploit the use of cfDNA in disease monitoring. Acknowledgements AIRC IG2018-22059.