Within weeks of the cancer’s detection, a tumor had grown in her small intestine, making it difficult to eat, and her peritoneal cavity, the space surrounding the organs in the abdomen, began to fill with fluid — the latter a sign she had just months, perhaps only weeks, to live. “I began writing goodbye letters to my children,” said the 57-year-old McKeown of Valley Center, Calif.
But then her luck changed. Through a clinical trial at Moores Cancer Center at UC San Diego Health, a test designed to reveal the genetic abnormalities locked inside tumors found a large number of them in McKeown’s cancer. The results raised the possibility that Opdivo, a drug that triggers the immune system to attack cancer and which was approved to treat melanoma, might work against the metastatic breast cancer cells ravaging McKeown’s body.
McKeown started on the drug in early 2017. After two infusions, her tumor markers, substances in the blood that reflect the amount of underlying disease, dropped dramatically. Within eight weeks, imaging scans showed no sign of cancer anywhere in McKeown’s body. Today, she remains cancer-free.
McKeown and other patients like her are rare but powerful examples of why sequencing the genome of a tumor to uncover the mutant genetics housed in its DNA — a practice known as comprehensive tumor profiling — is fast becoming a cornerstone of cancer treatment. The approach enables doctors to match individuals with cancer with drugs that target specific genetic mutations or other abnormalities in their tumor, often leading to unexpected recommendations, such as using a melanoma therapy for breast cancer.
But whether such stunning outcomes should be the exception rather than the rule has become a contested issue among oncologists. As these tests become more routine — helped largely by Medicare, which now covers them — patients and cancer physicians alike are asking whether comprehensive tumor profiling heralds a new era for cancer care or simply a higher price tag.
Every cancer is unique. Individual combinations of genetic changes, or mutations, distinguish one tumor from another, even when they begin in the same part of the body. Two people with lung cancer, for example, could have tumors with notably different genetic profiles. And these mutations can shape how a tumor behaves, including whether it responds to treatment.
For this reason, doctors are increasingly interested in tumor profiling for their patients. “A tumor might have a mutation that suggests they might respond to a certain targeted therapy, including treatments that would not have been considered had the mutation not been found,” says Razelle Kurzrock, who leads the Center for Personalized Cancer Therapy at Moores.
The bevy of treatments aimed against genetic mutations — called targeted therapies — means oncologists have a range of options to sort through for every patient. The Food and Drug Administration has approved 31 targeted therapies, including several immunotherapies, potent new drugs that have brought patients meaningful improvements in survival. Many more are under investigation.
The sheer number of options, says David Hyman, who specializes in gynecologic cancers at Memorial Sloan Kettering Cancer Center in New York, means profiling tumors makes sense.
“Many of these drugs work in a wider range of patients than traditional cancer drugs, but we won’t know who stands to benefit unless we look,” he says.
Hyman cites the example of Keytruda, which the FDA approved for advanced melanoma treatment in 2014. Last year, approval was expanded to include patients whose tumors carry a genetic mutation called microsatellite instability-high or MSI-H, making it the first drug approved based on genetic characteristics rather than on the cancer’s location in the body. This abnormality impairs the cells’ ability to fix damaged DNA, regardless of where in the body those tumors are located.
Just 4 percent of cancer patients’ tumors are MSI-H, most commonly colorectal and endometrial cancers. But Hyman says any patient for whom standard treatment is not working, or not available, should be tested for the abnormality. If the search for MSI-H tumors is confined to the two cancers where it’s usually found, he says, “we’d miss out on a lot of patients who could be helped.”
Likewise, tumor profiling can also find clinical trials for patients, particularly those, like McKeown, who are out of options. Increasingly, experimental studies enroll patients based on the genetic alterations in their tumors, not on the organ in which their tumor initially grew, says Timothy Yap, who heads a clinical trials research program at University of Texas MD Anderson Cancer Center in Houston. Sequencing tumor DNA, Yap says, is “immensely useful” for finding clinical trials for these patients.
Looking at tumor DNA for treatment clues isn’t new.
For years, oncologists have routinely checked for specific mutations in a few tumor types — such as BRAF mutations in melanoma and ROS1, ALKr, EGFR in lung cancer — using tests that scan for a single gene or a select set of genes. But comprehensive tumor profiling is different. This testing identifies mutations in large swaths of the tumor at once, providing “a far more detailed look” into the tumor’s genetic abnormalities, Yap says.
Several academic medical centers — including Sloan Kettering, MD Anderson and the University of California at San Francisco’s Helen Diller Family Comprehensive Cancer Center — already offer this testing as part of the standard treatment for patients with rare or advanced cancers. It’s also part of many clinical trial protocols.
Patients treated elsewhere or outside of clinical trials have not always found it easy to have their tumor comprehensively profiled because most private insurers don’t cover the expense. Many doctors have been hesitant to order the test — which can reach several thousand dollars — for fear of patients getting stuck with the bill, says Eli Van Allen, a prostate cancer researcher at Dana-Farber Cancer Institute in Boston. “Medicare’s decision to cover the test has made it a lot easier to just go for it,” he says.
This year, Medicare agreed to cover the test for beneficiaries with advanced solid tumors, regardless of the organ or tissue of origin. One of the tests now covered — FoundationOne, or F1CDx, made by Foundation Medicine — can identify cancer-linked mutations in 324 genes in any type of solid tumor. F1CDx can also determine whether a tumor is MSI-H and how many mutations it has overall, both of which, Kurzrock says, are “strong predictors” of how likely a person is to respond to cancer immunotherapies.
Many cancer centers have developed their own profiling test similar to or even more comprehensive than F1CDx. Sloan Kettering, for example, has examined the genetic abnormalities lurking in the tumors of more than 25,000 patients using its MSK-IMPACT test, which scans for mutated versions of 468 genes. “Findings from tests are increasingly used to guide treatment selections,” Hyman says. He estimates that about 15 percent of patients screened with MSK-IMPACT can be matched with an FDA-approved drug, and another 10 percent with a drug in clinical trials.
But not all cancer practitioners are on board with comprehensive tumor profiling as a routine part of care. “The fact is most patients don’t benefit from it,” says Jack West, medical director of the Thoracic Oncology Program at the Swedish Cancer Institute in Seattle. Often, he says, tumor profiling reveals cancer-causing mutations for which there are no known treatments.
West says he worries that simply having the tumor profile in hand could lead some doctors to try treatments that have not been shown to work in the patient’s particular cancer. Prescribing “off-label” treatments — using a drug for a use other than its FDA-approved one — is common in oncology, but it brings financial and logistical hurdles because insurers are usually unwilling to pay for them.
And even if patients do manage to secure off-label treatments, the drugs may not work the way they expect or hope.
“Anecdotes of someone who has benefited from a drug,” says West, “should not be mistaken for proof that it will help other people with that condition.”
Other experts warn that the evidence supporting routine tumor profiling outside a clinical trial is insufficient. Anthony Letai, a cancer researcher at Dana-Farber, says he believes the FDA should require companies to conduct clinical trials comparing the survival time among patients whose treatment was based on their genetic tests with those whose treatment was not.
Only one randomized clinical trial has investigated whether treatment decisions influenced by genetic analyses were better for patients than an oncologist’s best educated guess. The study found no difference in progression-free survival — the length of time that a patient’s disease does not worsen — between the two strategies. “Without evidence,” Letai says, “there is risk of doing harm.” He cites the history of prostate cancer screening: Years after screening became a routine part of health care, studies found that this screening injured more men than it helped, and guidelines were changed to advise against it. (Recently, those guidelines were softened.)
But what about patients such as McKeown? West likens their remarkable response to “winning the lottery,” rather than evidence in support of widespread comprehensive tumor sequencing.
“What you don’t see reflected in those stories,” he says, “are the thousands of people who have their tumor sequenced and learn nothing to help them.” Still, for McKeown, winning the lottery seems to have meant nothing short of staying alive.