Assessing the contribution of tumor mutational phenotypes to cancer progression risk

Author summary Cancer results from mutations in previously healthy cell populations that typically accumulate over a period of years before tumor growth is apparent. This process is governed by the laws of evolution, by which random mutations arising in cells will occasionally lead to selection for particular cell populations with growth advantages. How this process unfolds in any given tumor is highly random and idiosyncratic, however, making tumor growth difficult to predict. One reason tumor evolution is so random is that tumors frequently contain mutations that bias the cancer cells to generate new mutations in different ways, yet variations among these mechanisms of mutability (which we call mutational phenotypes) themselves provide predictive power for a cancer's future progression. However, little is known about the degree to which the risk of cancer progressing is characterized by these variations in mutation mechanism as opposed to other sources. In this work, we examine the question of how much of the risk of cancer progression is explained by these differences patient-to-patient in mutability preferences. We find that approximately a third of the risk of future cancer progression is accounted for by variations in mutational phenotypes. Furthermore, these mutational phenotypes are complementary to and only partially redundant with other sources of predictive information, a finding confirmed by showing that machine learning models using such information can enhance our ability to predict which cancers progress and recur beyond what can be accomplished from more traditional genomic and clinical data sources alone. Cancer occurs via an accumulation of somatic genomic alterations in a process of clonal evolution. There has been intensive study of potential causal mutations driving cancer development and progression. However, much recent evidence suggests that tumor evolution is normally driven by a variety of mechanisms of somatic hypermutability, which act in different combinations or degrees in different cancers. These variations in mutability phenotypes are predictive of progression outcomes independent of the specific mutations they have produced to date. Here we explore the question of how and to what degree these differences in mutational phenotypes act in a cancer to predict its future progression. We develop a computational paradigm using evolutionary tree inference (tumor phylogeny) algorithms to derive features quantifying single-tumor mutational phenotypes, followed by a machine learning framework to identify key features predictive of progression. Analyses of breast invasive carcinoma and lung carcinoma demonstrate that a large fraction of the risk of future clinical outcomes of cancer progression-overall survival and disease-free survival-can be explained solely from mutational phenotype features derived from the phylogenetic analysis. We further show that mutational phenotypes have additional predictive power even after accounting for traditional clinical and driver gene-centric genomic predictors of progression. These results confirm the importance of mutational phenotypes in contributing to cancer progression risk and suggest strategies for enhancing the predictive power of conventional clinical data or driver-centric biomarkers.

Automated summary

This article discusses the impact of differences in mutational phenotypes among different patients on the risk of cancer progression, finding that nearly a third of the risk can be explained by these differences. Mutational phenotypes serve as important parameters for predicting cancer progression, even after considering other predictive information, showcasing their independent predictive power.