OAB-057: Temporal-weight estimation of the copy number alterations of of 1384 Multiple Myeloma patients defines an ancestrality index impacting patients survival

Background MM is a hematological malignancy always evolving from pre-malignant stages, with progressive increase of genomic complexity. MM is characterized by a large abundance of copy number alterations (CNA); many of them, regarded as “driver”, stack up progressively from early tumor stages, causing biological changes that give rise to tumor hallmarks and malignant phenotypes. The combined application of whole genome analysis and mathematical models allows to deeply describe these alterations and to infer their order of acquisition during oncogenesis from their clonality levels, assuming that clonal ones are more ancestral than subclonal. Aims: (1) To define the temporal order of acquisition of CNA, leading to the onset of symptomatic MM and (2) to define a scoring model able to stratify patients (pts) according to the ancestrality of the alterations observed in their genomic landscape. Methods Genomic data collected from a total of 1384 newly diagnosed MM pts were included in the study: SNPs array data were collected from 514 pts of our Institution (BO dataset); in 870 pts, WES data were downloaded from CoMMpass study. CN calls and clonality levels were harmonized by an analysis pipeline including ASCAT, GISTIC v2 and custom R scripts. Timing estimates were obtained with BradleyTerry2 package. Survival analysis were performed on R. Results A full call-set of CNAs was obtained by harmonizing BO and CoMMpass datasets. The clonality information was first extrapolated from the whole call-set, to define the temporal order of acquisition of non-primary CNAs. CNAs were then accurately ranked, by using the obtained timing estimates, characterized by a quite narrow confidence interval. Of interest, chr 1q gains and chr 13q losses were frequently clonal and ranked as ancestral events, whereas chr 17p losses were late occurring events. By weighting the CNAs carried by any given pts at diagnosis with their relative timing estimate in a combinatorial process, an Ancestrality Index (AI) was defined for each pts (median AI=3.4, IQR=1.7-6.0). The AI was found to be significantly associated with progression free (PFS) and overall survival (OS) (p3.4 (i.e. with a more “ancestral” profile) had a worse outcome as compared to the rest of pts (OS 40% vs 58%, PFS 42% vs 56%, at a median follow up of 92m and 34m, p<0.001).The risk attributed to this “ancestral” category was independent from other high-risk cytogenetic features (i.e. del17p, t(4;14), t(14;20), t(14;20)). Conclusions By means of whole genome analysis and dataset harmonizing, the temporal order of acquisition of MM CNAs has been confidently described. A score reflecting the disease ancestrality of MM pts at diagnosis was generated and associated to survival outcomes. Overall, these findings support the evidence that MM pts at diagnosis carrying an excess of ancestral alterations, expected to likely be drivers, are prone to have a dismal prognosis.