Abstract LB-275: 5-miRNA signature to predict clear cell renal cell carcinoma metastasis and prognosis

Abstract

Abstract Introduction: Clear cell renal cell carcinoma (ccRCC) represents the most common renal cancer histology. In the setting of metastatic disease, few patients achieve a durable remission with currently available therapies. Early determination of metastatic potential may help guide therapy and improve clinical outcome. MicroRNA (miRNA) is a group of small non-coding RNAs that regulate gene expression during development and differentiation. miRNA expression is altered in malignant tissue, and signatures based on miRNA expression can aid in diagnosis and prognostication. In this study, we have characterized ccRCC miRNA expression in a 28-sample training cohort using microarray technology. From this, we have developed a 5-miRNA expression signature to predict the risk of metastasis and overall prognosis. This signature has been further validated by an independent 34-sample testing cohort. Study Design: Training and testing cohorts were established, comprised of 28 and 34 ccRCC frozen tissue specimens, respectively. The training cohort included specimens from patients characterized as stage I (T1; n=14) and stage IV (M1; n=14). The testing cohort included specimens from patients with (n=20) and without (n=14) metastatic disease. All cases used for the testing cohort had been followed for at least 5 years if there was no tumor metastasis reported. Total RNAs of these samples were analyzed using Agilent miRNA microarray (probes for 723 human miRNAs, Sanger miRBase 10.1). Results: (1) Differentially expressed miRNAs between metastatic ccRCCs and their non-metastatic counterparts in the training cohort were identified, using the criteria of fold change >1.5 (p value <0.05 by ANOVA). The 28 samples in the cohort were found to form distinct groups with different miRNA expression profiles using hierarchical clustering method. (2) A 5-miRNA ccRCC metastasis-specific signature has been identified using a customized computational method, which was based on logistic regression and Linear Discriminant Analysis classification, by comparing miRNA expression between metastatic and non-metastatic samples. (3) The signature has been successfully validated in the 34-sample testing cohort. With the clinical follow-up information (> 5 years if no metastasis reported), the signature had very high sensitivity (77%) and specificity (100%) when it was used to predict the tumor status of metastasis and metastatic potential. (4) This 5-miRNA signature may supplement widely used prognostic tools in RCC, such as the UCLA Integrated Staging System (UISS). The 5-miRNA signature stratified outcome within groups classified as high, intermediate or low risk by this schema. Conclusions: We have developed a 5-miRNA expression signature to determine ccRCC metastasis and prognosis. With further validation in larger cohorts, the signature may be applied towards early prediction of metastatic potential, and may augment currently available risk stratification tools for RCC. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr LB-275.