The basic biology that differentiates a primary cell from a transformed cell has been elucidated over the past several decades. Mutations in a limited number of oncogenes and tumor suppressors free the cell from growth-inhibitory checkpoints and allow proliferation in the absence of external stimuli, leading to tumor formation. Yet, primary tumors themselves are rarely deadly: greater than 90% of cancer mortality results from metastasis. What drives metastasis, and, more broadly, what distinguishes a lethal cancer from one with a favorable prognosis, is poorly understood.

To address this question, we have performed a comprehensive meta-analysis on genomic data from primary tumors that are linked to patients’ clinical outcomes. Utiliizing data from >20,000 patients, we have identified protein-coding genes, lncRNAs, miRNAs, methylation sites, and CNVs in primary tumors that are significantly associated with patient prognosis across cancer types. Multivariate analysis reveals a hierarchy of survival determinants: the strongest mortality-associated factors are enriched for components of the mitotic cell cycle, while secondary clusters of genes are involved in extracellular matrix remodeling, cell motility, and angiogenesis. Survival-associated genomic features are indicative of immune infiltration into primary tumors and oxidative phosphorylation activity.

In vitro and in vivo analyses reveal that mortality-associated genes rarely promote the direct transformation of primary cells. Instead, CRISPR mutagenesis reveals that these genes largely represent tumor dependencies, and in their absence cancer cells fail to proliferate. Through this analysis, we have identified new genetic dependencies common across cancer cells, including C16ORF59 and C5ORF46. In total, our results represent the largest assessment of genomic features linked to cancer prognosis completed to date, and offer several lines of insight into the biological differences between fatal and benign cancers.

Note: This abstract was not presented at the meeting.

Citation Format: Joan Smith, Ann Lin, Chris Giuliano, Jason M. Sheltzer. Identification and characterization of survival-associated genomic features across tumor types [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 390. doi:10.1158/1538-7445.AM2017-390