have also been reported for primary and metastatic pancreatic carcinomas8,9. All of these examples (renal, breast and pancreatic cancers) were from advanced tumors with extensive metastatic disease. A very recent paper10 has reported similar clonal and mutational evolution in primary breast cancers, operated on
with curative intent and without any evidence of distant metastasis, by analyzing the data from 20,000×-depth sequencing of each mutant position. The results suggest that, for most early breast cancers, which are unifocal, deep sequencing will be more important than multiple biopsies to fully characterize
the cancer.
Altogether, the data from the three papers highlighted here confirm the predictions of the clonal-evolution model using sequencing data. So what are the practical implications of these findings? As Nowell presciently wrote: “...each patient’s cancer may require individual specific therapy, and even this may be thwarted by emergence of a genetically variant subline resistant to the treatment”1. Yet clonal evolution may not be an insurmountable
barrier to cancer therapy if treatments can be designed to target mutations present in all cancer cells. Today we are able to fully characterize the basic components and dynamics of the cancer evolutionary process using deep sequencing of primary tumors and metastases
as well as single-cell sequencing. These
methods should now be applied in the clinic to begin to unravel the molecular determinants of drug resistance by sequencing tumors before and after treatment, and by sequencing relapsed cancers after primary therapy. For now, such approaches should be regarded as a research enterprise, but the accumulated data will help drive strategies for overcoming drug resistance.
For most common cancers, we will soon know the full complement of driver mutations that occur at a ‘reasonable’ frequency (for example, in at least 1% of tumors), and all early indicators suggest that this number will not be more than 100–200 genes at most. Targeted sequencing of this limited number of genes will be relatively simple, and, moreover, can be done at great depth, which will enable the identification of all subclones, even minor ones. A greater challenge will be to determine how to sample multiple sites that might be very heterogeneous and how this will influence targeted therapies. If the ‘actionable mutation’ is present in the founding clone, then only one site will have to be biopsied, but mutations unique to particular sites will of course imply more biopsies. Finally, clonal heterogeneity
suggests that combinations of targeted therapies will be more effective, and we can anticipate