The impact of genetic heterogeneity on biomarker development in kidney cancer assessed by multiregional sampling.

Primary clear cell renal cell carcinoma (ccRCC) genetic heterogeneity may lead to an underestimation of the mutational burden detected from a single site evaluation. We sought to characterize the extent of clonal branching involving key tumor suppressor mutations in primary ccRCC and determine if genetic heterogeneity could limit the mutation profiling from a single region assessment. Ex vivo core needle biopsies were obtained from three to five different regions of resected renal tumors at a single institution from 2012 to 2013. DNA was extracted and targeted sequencing was performed on five genes associated with ccRCC (von-Hippel Lindau [VHL], PBRM1, SETD2, BAP1, and KDM5C). We constructed phylogenetic trees by inferring clonal evolution based on the mutations present within each core and estimated the predictive power of detecting a mutation for each successive tumor region sampled. We obtained 47 ex vivo biopsy cores from 14 primary ccRCC's (median tumor size 4.5 cm, IQR 4.0-5.9 cm). Branching patterns of various complexities were observed in tumors with three or more mutations. A VHL mutation was detected in nine tumors (64%), each time being present ubiquitously throughout the tumor. Other genes had various degrees of regional mutational variation. Based on the mutations' prevalence we estimated that three different tumor regions should be sampled to detect mutations in PBRM1, SETD2, BAP1, and/or KDM5C with 90% certainty. The mutational burden of renal tumors varies by region sampled. Single site assessment of key tumor suppressor mutations in primary ccRCC may not adequately capture the genetic predictors of tumor behavior.