U.S. study will test whether drug combinations can wipe out tumors

The U.S. government is launching a wide-ranging effort to study whether combining two drugs matched to molecular weak spots in a patient's tumor will work better to wipe out the cancer than a single such drug.

Although cancer drug combinations are nothing new, the National Cancer Institute's (NCI's) ComboMATCH study formally announced today is demanding unusually rigorous evidence—in the form of data showing the drug pairs shrink tumors in mice—before letting them move into clinical testing. The effort will also take advantage of NCI's sprawling clinical networks to find patients whose tumors have rare mutations that could be targeted.

This framework "will allow us to explore many exciting novel drug combinations," says oncologist Roisin O’Cearbhaill of Memorial Sloan Kettering Cancer Center, one of the leaders of ComboMATCH.

The project follows NCI-MATCH, an effort launched in 2015 amid much enthusiasm for the idea of sequencing a cancer patient's tumor to detect key mutations that might be driving unchecked cell replication—such as a mutated gene coding for an overactive growth protein. The patient could then be matched with a drug blocking that problem protein.

Before that effort, a few such "targeted" drugs had been approved for specific cancers such as leukemia and breast tumors. NCI-MATCH hoped to add more by testing compounds matched not to a patient's broad cancer type, but to their tumor's individual genetic profile and a resulting protein target. Tumor sequencing has since become a routine part of care for many cancer patients.

In one success, NCI-MATCH data helped lead to U.S. regulatory approval last year for a treatment (as it happens, a drug combination) for solid tumors carrying a specific mutation that alters a protein called BRAF that signals cells to grow.

But such successes were rare, and the results from NCI-MATCH overall have been discouraging. Only about 18% of 6000 participants screened had a mutation that could be matched to an existing drug, and in only six of 27 trials of mostly single drugs did cancer patients improve enough or live longer than expected to meet the project's bar for success.

That's largely because tumors that shrink when treated with a single targeted therapy eventually start growing again, as a few preexisting or newly evolved drug-resistant cancer cells, or "clones," expand and take over a tumor. "The one tumor-one gene-one drug approach oversimplified cancer biology and did not take into account tumor heterogeneity, clonal evolution, or other mechanisms of resistance," ComboMATCH's leaders wrote in an April paper.

Since NCI-MATCH started, drugs that instead harness the immune system to wipe out tumors have garnered more medical success and public attention, sometimes keeping cancers at bay for years. But despite skepticism from some oncologists, NCI isn't giving up on targeted therapies.

For ComboMATCH, hundreds of investigators from five NCI clinical trial networks have pooled data from lab experiments to find drug combinations that shrank tumors in mice implanted with lab-grown human cancer cells or bits of actual human tumors. "The rigor of the preclinical evidence here, I think, is unique," said Jeffrey Moscow of NCI's Division of Cancer Treatment and Diagnosis at a press briefing.

The project is starting with 12 small trials, each enrolling about 30 to 200 advanced cancer patients for whom standard therapies have failed. Each trial will test a pair of targeted drugs donated by companies, or in some cases standard chemotherapy and a targeted drug. For example, a trial co-led by O’Cearbhaill for a specific type of breast cancer is combining a targeted drug to block a hormone that helps an overabundant protein drive the cancer's growth and a second drug blocking a growth pathway involving an altered protein from a gene called NF1.

If researchers see hints of significant improvement in how long tumor growth is stopped compared with just one of the drugs, the combos will then be tested in larger studies.

Unlike most industry-sponsored trials, NCI's ComboMATCH, which could eventually enroll 2900 patients, will combine drugs from different companies. The study's 300-and-counting sites will also make it easier to find patients with unusual cancer types or mutations, Moscow says. "It really requires looking across the whole country."