Joana Heinzelmann1,2, Madeleine Arndt1, Ramona Pleyers1, Tobias Fehlmann3, Sebastian Hoelters1,4, Philip Zeuschner1, Alexander Vogt1, Alexey Pryalukhin5,6, Elke Schaeffeler7,8, Rainer M Bohle5, Mieczyslaw Gajda9, Martin Janssen1, Michael Stoeckle1, Kerstin Junker10,11. 1. Department of Urology and Pediatric Urology, Saarland University, Homburg, Saar, Germany. 2. Department of Ophthalmology, Martin-Luther University Halle-Wittenberg, University Hospital Halle (Saale), Halle (Saale), Germany. 3. Department of Clinical Bioinformatics, Saarland University, Saarbruecken, Germany. 4. SERVA Electrophoresis GmbH, Heidelberg, Germany. 5. Institute of Pathology, Saarland University, Homburg, Saar, Germany. 6. Institute of Pathology, Bonn University Medical School, Bonn, Germany. 7. Dr. Margarete Fischer-Bosch Institute of Clinical Pharmacology, Stuttgart, Germany. 8. University of Tuebingen, Tuebingen, Germany. 9. Institute of Pathology, Jena University Hospital, Jena, Germany. 10. Department of Urology and Pediatric Urology, Saarland University, Homburg, Saar, Germany. kerstin.junker@uks.eu. 11. Department of Urology, Jena University Hospital, Jena, Germany. kerstin.junker@uks.eu.
Abstract
BACKGROUND: In order to improve individual prognostication as well as stratification for adjuvant therapy in patients with clinically localized clear cell renal cell carcinoma (ccRCC), reliable prognostic biomarkers are urgently needed. In this study, microRNAs (miRNAs) have emerged as promising candidates. We investigated whether a combination of differently expressed miRNAs in primary tumors can predict the individual metastatic risk. METHODS: Using two prospectively collected biobanks of academic centers, 108 ccRCCs were selected, including 57 from patients with metastatic disease at diagnosis or during follow-up and 51 without evidence of metastases. Fourteen previously identified candidate miRNAs were tested in 20 representative formalin-fixed and paraffin embedded samples in order to select the best discriminators between metastatic and nonmetastatic ccRCC. These miRNAs were approved in 108 tumor samples. We evaluated the association of altered miRNA expression with the metastatic potential of tumors using quantitative polymerase chain reaction. A prognostic 4-miRNA model has been established using a random forest classifier. Cox regression analyses were performed for correlation of the miRNA model and clinicopathological parameters to metastasis-free and overall survival. RESULTS: Nine miRNAs indicated significant expression alterations in the small cohort. These miRNAs were validated in the whole cohort. The established 4-miRNA score (miR-30a-3p/-30c-5p/-139-5p/-144-5p) has been identified as a superior predictor for metastasis-free survival (hazard ratio 12.402; p = 7.0E-05) and overall survival (p = 1.1E-04) compared with clinicopathological parameters, and likewise in the Leibovich score subgroups. CONCLUSIONS: We identified a 4-miRNA model that was found to be superior to clinicopathological parameters in accurately predicting individual metastatic risk and can support patient selection for risk-stratified follow-up and adjuvant therapy studies.
BACKGROUND: In order to improve individual prognostication as well as stratification for adjuvant therapy in patients with clinically localized clear cell renal cell carcinoma (ccRCC), reliable prognostic biomarkers are urgently needed. In this study, microRNAs (miRNAs) have emerged as promising candidates. We investigated whether a combination of differently expressed miRNAs in primary tumors can predict the individual metastatic risk. METHODS: Using two prospectively collected biobanks of academic centers, 108 ccRCCs were selected, including 57 from patients with metastatic disease at diagnosis or during follow-up and 51 without evidence of metastases. Fourteen previously identified candidate miRNAs were tested in 20 representative formalin-fixed and paraffin embedded samples in order to select the best discriminators between metastatic and nonmetastatic ccRCC. These miRNAs were approved in 108 tumor samples. We evaluated the association of altered miRNA expression with the metastatic potential of tumors using quantitative polymerase chain reaction. A prognostic 4-miRNA model has been established using a random forest classifier. Cox regression analyses were performed for correlation of the miRNA model and clinicopathological parameters to metastasis-free and overall survival. RESULTS: Nine miRNAs indicated significant expression alterations in the small cohort. These miRNAs were validated in the whole cohort. The established 4-miRNA score (miR-30a-3p/-30c-5p/-139-5p/-144-5p) has been identified as a superior predictor for metastasis-free survival (hazard ratio 12.402; p = 7.0E-05) and overall survival (p = 1.1E-04) compared with clinicopathological parameters, and likewise in the Leibovich score subgroups. CONCLUSIONS: We identified a 4-miRNA model that was found to be superior to clinicopathological parameters in accurately predicting individual metastatic risk and can support patient selection for risk-stratified follow-up and adjuvant therapy studies.