That’s not the important conclusion of the primary article. Highly mutated cancers of certain types respond better to checkpoint inhibition and so forth. Melanoma, lung adeno, colorectal with MMR defects, for example. Others (glioma, high-grade serous ovarian, germ cell for example) don’t seem to get “hot” immunologically in proportion to their mutation burdens. (And there are other tumors, serous ovarian hypercalcemic for example, that have a vanishingly low mutational burden and respond to checkpoint inhibitors anyways).

The initial results suggested that more mutations == more immune response. That’s clearly not universal enough to rely upon clinically, although in some tumor types it’s suitably reliable to stratify clinical trials. In other tumor types, it looks like we are stuck with trial and error to find good enough biomarkers for response, as with most drugs.

So while TMB can be useful, it has limitations, and this study (the primary results, published in Nature Genetics, a related but different journal) helped clarify where those limitations are. The news & views piece just has pull quotes.