Integrated Genomic and Transcriptomic Profiling Reveals Distinct Biological Programs and Therapeutic Opportunities in Cancer of Unknown Primary
En palabras de los autores
Background: Cancer of unknown primary (CUP) is a metastatic, heterogeneous malignancy accounting for 2-5% of cancer diagnoses globally. Failure to identify the primary site precludes tumor-type-specific therapy and, together with an aggressive clinical course, contributes to poor survival. The biological mechanisms driving these behaviors remain poorly understood. A deeper understanding of CUP biology is therefore essential to improve diagnostic precision and therapeutic strategies. To address this knowledge gap, we comprehensively characterized the multi-omics landscape of CUP and identified biological and potentially therapeutic targets beyond tissue-of-origin (ToO) assignment. Methods: Whole-genome sequencing (WGS) data from 283 CUP patients were analyzed for ToO predictions, mutational signatures, driver alterations, genetic immune escape (GIE), and clinically actionable alterations. Matched RNA-sequencing (RNA-seq) data from 79 cases were used to characterize functional gene expression signatures and define transcriptomic subtypes. Clinical, WGS, and RNA-seq data were integrated via weighted gene co-expression network analysis (WGCNA). Results: High-confidence ToO predictions (likelihood >80%) were obtained for 58% of CUPs, most commonly predicting lung, gastrointestinal, and biliary tract origins. Mutational signatures clustering revealed four subgroups dominated by tobacco smoking, defective DNA mismatch repair (dMMR), UV exposure, and APOBEC activity, while a fifth group, comprising most tumors (n=169; 60%), showed mixed mutational processes. The cohort also frequently exhibited mutations in established cancer driver genes, including TP53, CDKN2A, KRAS, and epigenetic modifiers ARID1A, SMARCA4, and ARID2. GIE alterations occurred in 35% of tumors, mostly involving HLA-I genes. Transcriptomic profiling defined five subtypes with different degrees of immune infiltration, fibrosis, and tumor proliferation. These subtypes were not significantly associated with clinical or genomic features, indicating that the tumor microenvironment (TME) represents another layer of CUP heterogeneity. WGCNA further linked proliferative programs to cell-cycle, DNA-replication, and neuroendocrine features. WGS data identified potentially targetable alterations in 64% of the CUP patients. Conclusions: CUP comprises biologically heterogeneous tumors with diverse predicted origins, mutational histories, and TME subtypes. Transcriptomic profiling and integrated genomics suggest that immune escape and rapid proliferation contribute to the clinical aggressive behavior. Together, these findings provide a molecular framework for understanding CUP heterogeneity and support therapeutic stratification by both ToO predictions and tumor-specific molecular alterations.
Apareció: miércoles, 18 de noviembre. medRxiv. Preprint, todavía sin revisión por pares.