Purpose: The clinical management of small renal masses (SRMs) is challenging since the current methods for distinguishing between benign masses and malignant renal cell carcinomas (RCCs) are frequently inaccurate or inconclusive. In addition, renal cancer subtypes also have different treatments and outcomes. High false negative rates increase the risk of cancer progression and indeterminate diagnoses result in unnecessary and potentially morbid surgical procedures.
Experimental design: We built a predictive classification model for kidney tumors using 697 DNA methylation profiles from six different subgroups: clear cell, papillary and chromophobe RCC, benign angiomylolipomas, oncocytomas, and normal kidney tissues. Furthermore, the DNA methylation-dependent classifier has been validated in 272 ex vivo needle biopsy samples from 100 renal masses (71% SRMs).
Results: In general, the results were highly reproducible (89%, n=70) in predicting identical malignant subtypes from biopsies. Overall, 98% of adjacent-normals (n=102) were correctly classified as normal, while 92% of tumors (n=71) were correctly classified malignant and 86% of benign (n=29) were correctly classified benign by this classification model.
Conclusions: Overall, this study provides molecular-based support for using routine needle biopsies to determine tumor classification of SRMs and support the clinical decision-making.
Keywords: DNA methylation; kidney cancer; small renal mass; tumor classification.