Prognostic Molecular Signatures for Metastatic Potential in Clinically Low-Risk Stage I and II Clear Cell Renal Cell Carcinomas

• Published in: Frontiers in oncology Vol. 10; p. 1383
• Main Authors: Shih, Andrew J., Murphy, Neal, Kozel, Zachary, Shah, Paras, Yaskiv, Oksana, Khalili, Houman, Liew, Anthony, Kavoussi, Louis, Hall, Simon, Vira, Manish, Zhu, Xin-Hua, Lee, Annette T.
• Format: Journal Article
• Language: English
• Published: Frontiers Media S.A 11.08.2020
• Subjects: canonical correlation analysis; clear cell; gene expression; miRNA; molecular biomarker; Oncology; renal cell carcinoma
• ISSN: 2234-943X; 2234-943X
• DOI: 10.3389/fonc.2020.01383

Introduction: For patients with localized node-negative (Stage I and II) clear cell renal cell carcinomas (ccRCC), current clinicopathological staging has limited predictive capability because of their low risk. Analyzing molecular signatures at the time of nephrectomy can aid in understanding future metastatic potential. Objective: Develop a molecular signature that can stratify patients who have clinically low risk ccRCC, but have high risk genetic changes driving an aggressive metastatic phenotype. Patients, Materials, and Methods: Presented is the differential expression of mRNA and miRNA in 44 Stage I and Stage II patients, 21 who developed metastasis within 5 years of nephrectomy, compared to 23 patients who remained disease free for more than 5 years. Extracted RNA from nephrectomy specimens preserved in FFPE blocks was sequenced using RNAseq. MiRNA expression was performed using the TaqMan OpenArray qPCR protocol. Results: One hundred thirty one genes and 2 miRNA were differentially expressed between the two groups. Canonical correlation (CC) analysis was applied and four CCs (CC32, CC20, CC9, and CC7) have an AUC > 0.65 in our dataset with similar predictive power in the TCGA-KIRC dataset. Gene set enrichment showed CC9 as kidney development/adhesion, CC20 as oxidative phosphorylation pathway, CC32 as RNA binding/spindle and CC7 as immune response. In a multivariate Cox model, the four CCs were able to identify high/low risk groups for metastases in the TCGA-KIRC ( p < 0.05) with odds ratios of CC32 = 5.7, CC20 = 4.4, CC9 = 3.6, and CC7 = 2.7. Conclusion: These results identify molecular signatures for more aggressive tumors in clinically low risk ccRCC patients who have a higher potential of metastasis than would be expected.