That cancer is a heterogeneous disease has long been recognized, but until recently therapeutic interventions have almost exclusively been stratified based on the anatomical location of the primary site. Although tumour site carries prognostic information and informs treatment options, treatment outcome is not adequately captured by site alone. In the last few decades, it has become clear that cancer is a genetic disease driven by damage to cellular DNA, in particular by mutations that activate oncogenes and inactivate tumour suppressor genes. To understand precision medicine, it is important to have an understanding of the mutational processes that shape the genome, both before and after cancers arise.
Mutational processes in this context may involve single base changes (point mutations); genomic rearrangements including gain or loss of whole segments of DNA; or translocations, where different chromosomal regions can combine, potentially creating fusion oncogenes. DNA damage is continuously at work in the cells of every living organism, and mutations may occur through myriad exogenous and endogenous carcinogenic processes, causing different types of genomic instability. As the size of the human genome is vast (about 3.2 billion base pairs per haploid genome), there is considerable space for the acquisition of somatic mutations over the course of a lifetime. When tumours arise, they will carry a combination of the somatic mutations accumulated in the precancerous cell prior to carcinogenesis all the way back in time to the fertilized egg, as well as mutations selected during carcinogenesis and evolution of the tumour.1 Recently, large-scale sequencing efforts have shown that cancer genomes at the exome level may harbour anything from just a few mutations to several thousand mutations, chromosomal aberrations and whole genome doublings.2,3
Since different types of genomic instability can leave a certain footprint on cellular DNA, it is possible to determine some of the mutational processes that have shaped different cancer genomes by analysing the nucleotide context of somatic mutations in tumours.4 These analyses have revealed that, as the cancer evolves through time and space, different mutational processes may become active due to carcinogenic exposure (e.g. tobacco smoke), oncogenic loss of DNA repair efficacy, disruption of DNA replication polymerases, or activation of genes that induce DNA damage directly, which may considerably increase the level of genomic instability.5,6 Altogether, this may lead to an elevated level of somatic aberrations upon which selection can act in a Darwinian manner for cells with increased malignant potential. This in turn is one of the major sources of heterogeneity in cancer, both between patients within the same cancer type (inter-tumour heterogeneity) and between subclones within individual tumours (intra-tumour heterogeneity).
Even within the same cancer type, significant inter-tumour heterogeneity exists between patients. For example, two lung cancer patients with cancer originating from the same cell type may have a different subtype, due to different combinations of point mutations, chromosomal aberrations ...