Abstract
Cancers are a mosaic of clones of varying population sizes. Any single cancer sample encodes a tumor-metagenome, since it represents the aggregate genomes of diverse clones that coexist within the sample. We quantified genomic instability as the fraction of the tumor-metagenome affected by copy number variations (CNVs) and leveraged two tumor mixture separation algorithms, EXPANDS and PyClone, to quantify genetic intra-tumor heterogeneity (ITH) from single cancer samples. We tested the potential of measures of genomic instability and ITH as prognostic biomarkers across 1,165 exome sequenced primary tumors from 12 cancer types at TCGA. Our results suggest that a tradeoff between the costs and adaptive benefits of genomic instability governs differential radiotherapy sensitivity.
Between 1 and 18 clones were estimated to coexist per tumor sample at >10% cell frequency (median = 4). Clone number varied considerably within and between cancer types, with melanomas representing the most heterogeneous group. 86% all analyzed tumor samples contained at least 2 clones. Across cancer types, the presence of >2 clones was associated with worse overall survival as compared to tumors in which < = 2 clones were detected (Log-rank test: P = 8.6E-4, HR = 1.49). An exceptionally favorable outcome was observed when >75% genomic instability was shared among < = 2 clones. The highest risk was observed among individuals with an intermediate number of 4 clones - additional diversification beyond 4 clones did not impart further risk. The highest risk was also observed among individuals with 50-75% genomic instability, in both the original exome sequencing dataset and an independent SNP array dataset. Genomic instability <25% or >75% predicted reduced risk (HR = 0.15, 95% CI: 0.08-0.29).
We analyzed the relation between radiotherapy intensity and overall survival among 242 individuals (21%) treated with radiotherapy and found that not all individuals did benefit equally from therapy. In order to achieve the same benefit from therapy, individuals with 25-50% genomic instability required higher therapy intensity (regression slope = 1.83; P = 0.009) than individuals with 50-75% genomic instability (slope = 2.09; P = 0.005). In contrast, individuals with <25% genomic instability did not benefit from increasing therapy intensity (slope = 0.71; P = 0.311).
Radiotherapy may be particularly effective against tumors with intermediate CNV burden, by pushing them past the limit of ‘tolerable’ genomic instability. Our results from two independent pan-cancer cohorts suggest that this limit is exceeded when >75% of a tumor's metagenome is affected by CNVs. This upper limit of tolerable genomic instability may be responsible for the non-linear association we observed between genetic ITH and survival. Leveraging a clone's distance to the upper limit of tolerable genomic instability may represent a new strategy to optimize therapy intensity.
Citation Format: Noemi Andor, Trevor A. Graham, Marnix Jansen, Li C. Xia, Athena Aktipis, Claudia Petritsch, Hanlee P. Ji, Carlo C. Maley. Pan-cancer analysis of clonal evolution reveals the costs and adaptive benefits of genomic instability. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 2387.