27O - Master gene regulation activity and methylation data reveals common patterns in well differentiated neuroendocrine tumors (NETs) from different tumor sites

Background

NENs are a heterogeneous family of tumors of wide anatomical distribution that share common clinical and biological traits. Although some studies have characterized the molecular landscape of specific primary locations, there is no cross-site molecular classification of NENs. The global purpose of our project is to improve our understanding of the biological basis of NETs and its impact on clinical outcome. We performed a multiomics factor integrative analysis (MOFA) with data from NETs of multiple anatomic sites and assessed the clinical relevance of inferred latent factors representing the underlying main axes of molecular heterogeneity across samples.

Methods

Transcriptomic and methylation data, were obtained from FFPE tumor samples of 194 patients with G1-G3 NETs from GEP, lung or other origins. Transcriptomic data was transformed into a master regulator activity (MRA) score matrix using the ARACNE and Viper, deeming as master regulators transcription factors, co-factors, cell surface receptors and main regulators of signalling pathways. MRA scores were combined with methylation information performing a MOFA modelling stratified by primary tumor site. All factors were associated to different molecular and clinical data, included tumor grade, site, functionality, and overall survival.

Results

MOFA identified 10 factors: 1, 2, 7, 8 and 10 were exclusively explained by MRA, 3 and 6 by methylation data, and 4, 5, 6 and 9 by both omics. All factors, except for factor 6, explained the variance by primary tumor site. Factor 6 elucidate differences between G1 and G2 tumors only in NETs from lung origin. Several factors were associated with grade (3,4,6 and 9), gender (1,6) and functionality (3,8 and 9), and factor 6 was also associated with a poor prognosis. Finally, we identified genes and pathways related to 10 factors included FGFR4 in factor 2 or several genes related to NETs in factor 4.

Conclusions

To our knowledge this is the largest series to date of comprehensively characterized NETs from a clinical and molecular perspective, and shall lay the ground to build a cross-site classification of NENs to identify unique and shared targetable vulnerabilities of potential clinical use.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

Rocio Garcia-Carbonero.

Funding

Pfizer.

Disclosure

R. Garcia-Carbonero: Financial Interests, Personal, Advisory Board: AAA, Advanz Pharma, Amgen, Bayer, BMS, HMP, Ipsen, Merck, Midatech, MSD, Novartis, PharmaMar, Pierre Fabre, Servier; Financial Interests, Institutional, Research Grant: BMS, MSD, Pfizer; Non-Financial Interests, Personal, Leadership Role, Global PI of investigator-initiated clinical trials (AXINET, NICENEC, PEMBROLA): BMS, MSD, Pfizer; Other, Personal, Other, Honoraria received by spouse for advisory board or invited speaker roles: Abbie, AstraZeneca, Bayer, Boehringer Ingelheim, BMS, Genomica, Lilly, MSD, Merck, Novartis, Pfizer, PharmaMar, Roche, Sanofi, Servier, Takeda. All other authors have declared no conflicts of interest.