Recognizing the Continuous Nature of Expression Heterogeneity and Clinical Outcomes in Clear Cell Renal Cell Carcinoma

PURPOSEEvaluation of 12 ccRCC publicly-available ccRCC gene expression datasets showed that previously proposed discrete molecular subtypes are unstable. To reflect the continuous nature of gene expression observed, we developed a quantitative score (Continuous Linear Enhanced Assessment of Renal cell carcinoma, or CLEAR) using expression analysis founded on pathologic parameters.MATERIALS AND METHODS265 ccRCC gene expression profiles were used to develop the CLEAR score, representing a genetic correlate of the continuum of morphological tumor grade. A signature derivation method based on correlation of CLEAR score with gene expression ranking was used to derive an 18-transcript signature. External validation was conducted in multiple public expression datasets.ResultsAs a measure of intertumoral gene expression heterogeneity, the CLEAR score demonstrated both superior prognostic estimates (94% vs 83% adequacy index in TCGA dataset) and inverse correlation with anti-angiogenic tyrosine-kinase inhibition (65% vs 55% adequacy index) in comparison to previously proposed discrete subtyping classifications. Inverse correlation with high-dose interleukin-2 outcomes was also observed for the CLEAR score (p=0.05). Multiple somatic mutations (VHL, PBRM1, SETD2, KDM5C, TP53, BAP1, PTEN, MTOR) were associated with the CLEAR score. Application of the CLEAR score to independent expression profiling of intratumoral ccRCC regions demonstrated its ability to reflect intratumoral expression heterogeneity and further analysis showed average intertumoral heterogeneity exceeded intratumoral heterogeneity.ConclusionsThe CLEAR score, a gene expression signature developed on histopathology, outperformed discrete subtype-classification in prognostic estimates and correlated better with treatment outcomes. Recognizing cancer as a continuum has important implications for laboratory and clinical research.