Deep visual multi-omics profiling links morphology and molecular programs in clear cell renal cell carcinoma

ABSTRACT: Clear cell renal cell carcinoma (ccRCC) exhibits striking intra-tumoral heterogeneity (ITH) at both morphological and genetic levels, complicating treatment and contributing to disease progression. CcRCCs with focal rhabdoid differentiation stand out as highly aggressive tumors distinguished by a subset of cells with unique morphological features. However, the correlation between distinct morphologies, specific molecular alterations, and their influence on tumor behavior rremains largely unknown. Here, we present Deep Visual Multi-Omics, that integrates advanced AI-based image analysis with morphology-guided single-cell isolation and ultra-sensitive multi-omics profiling to dissect the link between clinically relevant morphological and molecular features. Across five tumours, we profiled ~40,000 morphology-defined epithelial cells selected by AI-based classification and expert curation. By doing so, we map patterns of progressive molecular dysregulation across tumor grades and identify molecular features associated with rhabdoid ccRCC cells. These include upregulation of FOXM1-driven proliferation, disrupted cell-matrix interactions, and enhanced immune evasion. Despite dense T-cell infiltration, rhabdoid cells exhibit a cell-intrinsic immunosuppressive network involving IFN-beta, integrin signalling, expression of PD-L1, and novel immunomodulators such as CD38 and ITGB2 potentially driving T-cell exhaustion. Our findings offer mechanistic insight into the aggressiveness of rhabdoid ccRCC and support new avenues for precision immunotherapy, emphasizing the power of Deep Visual Multi-omics to decode cancer complexity and target high-risk subpopulations.

SUBMITTER: Doro Rutishauser 

PROVIDER: S-BIAD3059 | bioimages |

REPOSITORIES: bioimages