Researchers at Memorial Sloan Kettering Cancer Center in New York have discovered that cancer cells with high numbers of faults in their DNA are more likely to respond to immune checkpoint inhibitors (ICI), a major class of immunotherapy drugs, which includes Keytruda.

The study, published today in Nature Genetics adds important pieces to the puzzle as to why some cancer patients respond to immunotherapy whereas others do not. The researchers measured ‘tumor mutation burden (TMB)’, essentially counting how many DNA faults a tumor contains by looking for errors in the DNA sequence.

“People assume that TMB is important in predicting response to immunotherapy in all cancers, but up until now, all we’ve had is data from small studies and clinical trials on mostly lung cancers and melanoma,” said Luc Morris, MD, surgical oncologist at Memorial Sloan Kettering Cancer Center and one of the lead authors of the paper.

The researchers studied the DNA of 1,662 patients with advanced cancer (classified as stage IV or metastatic disease) treated with one or more of several FDA-approved ICI drugs and DNA from 5,371 patients with advanced cancer who had not had ICI. They used a tool called MSK-IMPACT, which looks at just 3% of the coding-regions in DNA, but is correlated to the number of mutations in the genome.

“Is TMB associated with likelihood that immunotherapy has benefit? Is this true in all cancers? We wanted to find out whether TMB had broad applicability,” said Morris.

The researchers found that if they took the 20% of cancers in their data with the most mutations, these people responded better to ICI than those with lower numbers of mutations in their tumors.

However, this correlation did not hold true for all tumors, for example, in people with a type of brain cancer called glioma, those with TMB in the top 20% did no better on ICI than those with lower TMB. Also in breast cancer there was no conclusive evidence that a higher TMB predicted response to ICI, although the study included relatively few breast cancer patients as ICI is not currently widely used for the disease.

Researchers don’t exactly know why high numbers of mutations make cancers more susceptible to immunotherapy, but they do have a very plausible theory. They think that the more mutated a cell is, the more likely it is to produce incorrect, mangled proteins. These displayed on the cell surface are called neoantigens and they are so far from what would be considered normal, the immune system identifies them as foreign and attacks the cells.

This is not a unique study in concept, with previous research on a smaller number of cancers of specific types, notably lung and melanoma, indicating that TMB is likely predictive of immunotherapy response. However, it is the largest and most comprehensive study to date, providing the most persuasive evidence that this is true for a greater number of cancer types.

“Only in lung cancer is TMB being used in a clinical trial. Hopefully this data will give us permission to include it in future clinical trials on other cancer types,” said Morris.

However, some patients with high TMB don’t respond to ICI at all, so there is still work to be done to figure out why high TMB is not a universal predictor of response to ICI.

“TMB by itself is not going to give you a high confidence in predicting whether a patient is going to respond to immunotherapy or not. It is one biomarker for response, but a number of other factors are important. I would not suggest you take the data from this paper and apply it to a patient in the clinic,” said Morris.