Cutting-edge cancer drug hobbled by diagnostic test confusion

Physicians struggle to identify which patients are likely to respond to a recently approved therapy.

Search for this author in:

Coloured scanning electron micrograph of a cancerous colon

Cancerous cell growth (orange) in the colon.Credit: Steve Gschmeissner/SPL

A landmark cancer drug approved last year seemed to herald a long-anticipated change in the treatment of some tumours: with medicines selected on the basis of molecular markers, rather than the tissue in which the cancer first took root.

But clinicians and researchers are struggling to put that theory into practice. Although the drug itself works well in a variety of tumour types, some of the tests used to identify the molecular markers, it turns out, do not.

On 15 April at the American Association for Cancer Research annual meeting in Chicago, Illinois, researchers and representatives from the US Food and Drug Administration (FDA) will discuss how best to tackle the situation. “If you get a false negative result, you’re not going to give that patient the therapy, which is terrible,” says Zsofia Stadler, an oncologist at the Memorial Sloan Kettering Cancer Center in New York City. “That’s why there’s such a debate.”

The drug in question, pembrolizumab (Keytruda), works by firing up the body’s immune responses against tumours. First approved by the FDA in 2014 to treat melanoma, it has since been given the go-ahead to treat a handful of other cancers, including lung cancer .

But last year, researchers reported that patients whose tumours had a disabled DNA-repair system also responded to the drug, regardless of where the tumour originated1. Damaged DNA can yield mutant proteins, which the immune system could target as potential invaders. Scientists think this increases the chances that immune cells unleashed by pembrolizumab will find and attack the tumour.

Mixed results

In May 2017, the FDA allowed pharmaceutical giant Merck of Kenilworth, New Jersey, to market pembrolizumab to people with advanced-stage cancer who had any solid tumour with that particular DNA-repair defect. “This is absolutely a breakthrough approval,” says Razelle Kurzrock, an oncologist at the University of California, San Diego. “We have seen some dramatic responses in our patients.”

But the three kinds of tests commonly used to look for the DNA damage that arises from that defect can produce conflicting results, says Heather Hampel, a genetic counsellor at Ohio State University in Columbus. One relies on PCR, a process that amplifies specific regions of the genome; a second looks for certain proteins; and a third relies on DNA sequencing. “Which is the best? Is any positive on any test sufficient?” Hampel says. “Does that mean you should try them all? No one wants to miss a patient who might benefit from pembrolizumab.”

And there are signs that some of the tests might work better in certain tissues than others, says Shridar Ganesan, a physician and cancer researcher at the Rutgers Cancer Institute of New Jersey in New Brunswick. PCR assays, for example, look for changes in certain regions of DNA called microsatellites. Particular microsatellites might be more prone to damage in some tissues than others, he says.

Stadler notes that the degree to which the DNA changes may also vary from tissue to tissue: colon cancers tend to accumulate many mutations, whereas tumours in the adrenocortex generally have fewer. That can lead to a false negative result in tissues with fewer mutations, she says.

Complicated future

Similar complications might arise for some future tissue-agnostic drug approvals, particularly those based on DNA-repair defects. This could include drugs called PARP inhibitors, which are currently approved in the United States for breast and ovarian cancers caused by mutations in either of two genes involved in DNA repair: BRCA1 or BRCA2. Researchers are looking at whether PARP inhibitors might also work in any solid tumour that carries similar DNA-repair defects, even if they aren’t caused by BRCA1 or BRCA2 mutations. There are multiple tests available to identify the patterns of DNA damage in such tumours, says Hampel.

Evidence has also been building that the overall number of mutations in a tumour could indicate how likely it is to respond to immunotherapies such as pembrolizumab. Tests for this might also be complex, notes Stadler.

Eventually, some of these issues will be ironed out, says Michael Overman, an oncologist with the University of Texas MD Anderson Cancer Center in Houston, as researchers gather data on which tests work best in which cancers. But the FDA was wise to move forward with the approval rather than wait for more evidence to sort out the issues with the molecular marker tests, he says. “There are still a lot of open questions, but the therapy works exceptionally well,” he says. “It was the right thing to do.”

Nature 556, 161-162 (2018)

doi: 10.1038/d41586-018-03862-6
Nature Briefing

Sign up for the daily Nature Briefing email newsletter

Stay up to date with what matters in science and why, handpicked from Nature and other publications worldwide.

Sign Up


  1. 1.

    Le, D. T. et al. Science 357, 409–413 (2017).

Download references