Recognition of cancer's heterogeneity is not new. But time and again when looking at clinical trial designs and even preclinical modeling of novel agents, for reasons of practicality but also perhaps tunnel vision this dynamic diversity of the tumor is mostly disregarded. 

And yet the heterogeneity is implicit in the recognition of the need for combination therapy. One might argue that the omics view has led in some cases to a reductionist approach, thinking the right targeted agent even to a single driver mutation will have significant benefit. It may, but it is usually not durable.

The advent of immuno-oncology (IO) approaches has somewhat reinvigorated the more holistic thinking as these approaches, for the most part and for the moment, are applied broadly, like chemo regimens of yore, though clearly in most cancers and most patients the responses and the outcomes remain suboptimal.

Ultimately, as we continue to grapple with the plasticity and diversity of tumors, we will be better able to exploit our omics learnings, but I suspect we will also rely more and more on the inherent diversity of the immune system as the best opponent to that so wily of foes that is cancer.