Stefanie Joho found out that cancer ran in her family in her early teens, when her 44-year-old mother got sick with colon cancer. A genetic counselor at the Fox Chase Cancer Center in Philadelphia explained that Stefanie’s mother had inherited a faulty gene from her father’s side of the family. That single copy of a single gene left her cells unable to repair spelling mistakes that inevitably occur when DNA gets copied in cell division. As a result, Stefanie’s mom, like all people with the inherited condition known as Lynch syndrome, was at an increased risk for many cancers, and especially colorectal cancer. Stefanie and her two sisters each had a 50 percent chance they carried the mutant gene, too.
Stefanie remembers her parents sitting them down to explain that when they got older, they would have to decide whether to get tested for that fateful genetic mutation. If they decided not to test, her parents’ advice — or perhaps “plea” is a better word — was for the girls to act as if they had Lynch syndrome by getting frequent colonoscopies and other recommended cancer screenings.
Stefanie made a firm decision: She’d get tested when she turned 25. As she understood it at the time, age 25 was the earliest a colonoscopy would have been recommended for a person with Lynch syndrome anyway. She didn’t see any point in knowing before then. But the massive tumor that nearly obstructed Stefanie’s colon would soon force her to make other plans.
Running On Empty
When Stefanie went to see her general practitioner, she hadn’t thought anything was seriously wrong. She felt fatigued, but that was probably just stress, she thought. When a blood test showed severe anemia, her doctor had the clinical sense to order a colonoscopy. At the age of 22, Stefanie learned she did indeed have Lynch syndrome and an aggressive case of colon cancer, too.
Surgeons removed the mass from her transverse colon. Her lymph nodes came back clear and no further treatment was recommended. That sounded like good news, but Stefanie’s cancer would soon return. She had another surgery, with a recovery even more brutal than the first, and her doctors told her that the cancer had advanced. If it came back again, surgery might not be an option. She suff-ered through several rounds of chemotherapy and had another scan. The cancer wasn’t backing down.
Doctors told Stefanie there was little chance, based on the proximity of her tumor to a major artery, that she’d survive the operation if they went in again. She continued with chemo in hopes of shrinking the cancer enough to make surgery possible, but to no avail. Her tumor kept growing.
With the chemo causing only further aggravation, Stefanie’s tumor seemed unstoppable. Meanwhile, she’d become dependent on regular doses of prescription painkillers — oxycodone and fentanyl — in an attempt to keep her excruciating and debilitating back pain under some kind of control. Her weight had dropped to nearly 100 pounds, and she couldn’t remember when she’d last felt anything like hunger. Stefanie remembers feeling so depleted that she didn’t even feel human.
“Everyone was telling me I was running out of options,” she says.
Back in her New York apartment, Stefanie remembers lying in her room in the pitch dark. “I didn’t want to see the light,” she says.
In the next room, Stefanie’s sister Jess Joho, whom Stefanie refers to as her “other half,” was scouring the Internet for any glimmer of hope. She plugged in buzzwords she’d heard Stefanie’s doctors throw around: “microsatellite instability” (MSI), and “immunotherapy.” Her search turned up a clinical trial, “Phase 2 Study of MK-3475 in Patients With Micro-satellite Unstable (MSI) Tumors.”
The study was underway at a few centers, including the Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins in Baltimore. The clinic there was enrolling patients, age 18 and older, with heavily mutated cancers that had already failed at least two prior treatments. One arm of the study was devoted to people with mismatch repair-deficient colorectal cancers, including those who’d inherited Lynch syndrome from birth. It sounded perfect for Stefanie.
Jess knocked on Stephanie’s bedroom door. “I really think I found the trial that could save your life,” she said to her sister.
Stefanie sent an email to her first oncologist in Philadelphia, who put her and Jess in touch with Luis Diaz Jr., an oncologist then at Hopkins. Shortly thereafter, Diaz called Stefanie personally. Stefanie likes to say his words saved her life right there on the phone.
“Get over here as fast as you can. We’re having tremendous success with patients like you.”
Follow the Evidence
The investigational drug referred to as MK-3475, which Stefanie would soon begin receiving via infusions every two weeks, is now known as Keytruda (pembrolizumab). The drug is an antibody, designed to block the function of a protein called PD-1 at the surface of T cells in the immune system. Blocking the PD-1 protein releases “brakes” on the immune system. Those brakes normally keep T cells from going to battle against the body’s own cells, including cancerous ones. With PD-1 disabled, the hope is that the immune system can be freed to fight cancer.
It hadn’t always looked as though PD-1 inhibitors would work to treat colorectal cancers. A 2012 safety study of another very similar PD-1 inhibitor showed the general approach to cancer treatment could hold promise for some patients with non-small cell lung cancer, melanoma, and kidney cancer — but for colo-rectal cancer patients, not so much. Of 19 colo-rectal cancer patients enrolled in that early clinical trial, only one showed any response.
That might have been the end of the story. It turned out, however, that the one elderly colorectal cancer patient who had responded had done so in a remarkable way. He’d been in treatment since 2003, undergoing surgery and multiple chemotherapy regimens that seemed to work, but only temporarily. In July 2007, with his cancer progressing, he enrolled in the early clinical trial and got his first dose of the experimental PD-1 inhibitor now marketed as Opdivo (nivolumab). Six months later, CT and PET scans showed no sign of cancer.
His seemingly anomalous response piqued the interest of a team including Diaz and Dung Le, both investigators on the trial Stefanie would later sign up for at Hopkins. “What was different about this single patient?” they wanted to know.
The team suspected that his cancer must have carried a large number of mutations due to defects in DNA mismatch repair. Mismatch repair is a system within cells for recognizing and repairing typos that arise as DNA is duplicated in preparation for one cell to divide into two. When that system breaks down, cells lose the essential ability to spell check their work. Without quality control, the mistakes can add up. Colorectal cancers with a defective mismatch repair system carry 10 or even 100 times as many mutations in comparison to other cancers, and each mutation might be translated into novel proteins not found elsewhere in the body. Diaz and Le’s theory was that those novel proteins might help immunotherapy along by making it more likely for a patient’s own immune system to recognize a tumor as a foreign invader and launch an attack against it. It turned out they were on to something: the single responsive patient did indeed have a heavily mutated, mismatch repair-deficient tumor.
To gather more evidence, the Hopkins team could have launched a trial focused only on colorectal cancer. But they realized a more powerful way to test their theory was a “basket trial,” one that included patients with multiple cancer types that all carried the same genetic defect. Their trial design ultimately included three groups of patients: those, like Stefanie, who had mismatch repair-deficient colorectal cancers, those with colorectal cancers that didn’t have the repair deficiency, and those with mismatch repair-deficient cancers in other parts of the body.
The trial opened in September 2013. Stefanie would make her way from her presumptive deathbed in New York to Baltimore to receive her first Keytruda infusion almost a year later. The next month, the Food and Drug Administration (FDA) granted Keytruda accelerated approval for the treatment of patients with advanced and inoperable melanoma, making the drug the first approved PD-1 inhibitor.
By that time, Stefanie was already getting better. The pain in her back subsided within three days of her first infusion. In fact, without even realizing it, Stefanie spontaneously cut back on her pain medications. She didn’t understand what was happening to her, but that sudden drop in narcotic pain meds landed her in the emergency room with her heart racing at 230 beats per minute. She remembers someone asking if she had a family history of heart failure. She assumed it was a response to the Keytruda, but Stefanie was really suffering from opioid withdrawal. She no longer needed a large dose of narcotics to dull the pain, but it took a little time for her body to get the message.
The Keytruda treatment brought other challenges, including joint pain, an overactive thyroid, and fatigue, all common side effects of immunotherapy. Le, her oncologist on the trial, encouraged Stefanie to view the side effects as potentially encouraging signs that the treatment was revving up her immune system in the hoped-for ways. Within three months, her tumor had shrunk by 65 percent. By 18 months, “no one could find a tumor in my body,” Stefanie says.
Stefanie wasn’t the only one to show such a positive response. In a 2015 New England Journal of Medicine study, Le, Diaz, and colleagues reported on the first 41 patients, ahead of schedule. After 20 weeks of treatment, four of 10 patients in the trial with mismatch repair-deficient colorectal cancer had seen their tumors shrink; in seven patients, the cancer hadn’t progressed. In contrast, patients whose colorectal cancers lacked the genetic glitch rarely if ever responded to the treatment. Only two of 18 in the non-hypermutation group avoided cancer progression; none saw their tumors shrink.
And there was more encouraging news in support of the theory: Patients with mismatch repair defects in other cancer types responded similarly well. The numbers were still small, but it looked as though a positive response to Keytruda could be predicted based on the presence of mismatch repair defects, no matter where in the body a tumor was found. That kind of predictive ability is especially valuable in the case of Keytruda, because the drug costs about $150,000 a year. (A recent analysis suggested it can cost more for patients who need higher doses of Keytruda as a first-line treatment for lung cancer.)
In June 2017, Le, Diaz, and colleagues reported the latest trial results, this time from 86 patients with 12 different cancer types. A little over half of those patients who’d begun Keytruda infusions only after they’d run out of other options had responded to the treatment. One in five of those trial participants — Stefanie among them — had shown a complete response. For the time being at least, their bolstered immune systems had won the battle.
In the beginning, “it was just a simple hypothesis,” says Diaz, who recently moved from Hopkins to Memorial Sloan Kettering Cancer Center in New York. To see it pan out so well, he says, has been nothing short of unbelievable. “These patients are not only having responses in the tumor, but in about 20 percent of cases the tumors disappear,” he says, and that’s “in patients who were already signed up for hospice.”
An Important First
The FDA didn’t wait for those latest results to come out in the journal Science before making its decision. On May 23, 2017, the agency announced that Keytruda had been granted accelerated approval for treating any solid tumor with a defect in mismatch repair. To find out how many patients might fall into that category, the Hopkins team analyzed tumors from more than 12,000 people with 32 cancer types and found mismatch repair problems in about 5 percent of patients with 11 different cancers. So, while most tumors don’t carry those types of genetic abnormalities, the expanded approval could mean the difference between life and death for about 60,000 U.S. cancer patients every year.
The findings and the FDA’s decision mark an important turning point for cancer treatment by showing something cancer genomics researchers have long suspected: Sometimes, when it comes to making treatment decisions, a cancer’s tissue of origin won’t be as important as its molecular features. In a press release, Richard Pazdur, the director of the FDA’s Oncology Center of Excellence, called the expanded Keytruda approval “an important first for the cancer community.” In the past, he explained, the FDA had always limited the approval of cancer drugs to particular cancer types defined according to where in the body a cancer first arose — lung, breast, or pancreas, for example. In an email, the FDA’s Steven Lemery, an associate division director in the Center for Drug Evaluation and Research, says it was an indication that “a new understanding of some forms of cancer is possible.”
Several factors tipped the balance in favor of an unprecedented “tissue-agnostic” approval, Lemery says. First, the trial data showed “consistent evidence” of drug response in multiple tumor types. The underlying science was clearly demonstrated, and it helped that patients with mismatch repair-deficient tumors lack other approved treatments. On top of all that, he says, Keytruda had already been approved for multiple other cancer indications. Melanoma was the first, but the drug has since been approved for head and neck cancer, non-small cell lung cancer, and bladder cancers. “We had a good understanding of the safety profile of this drug and a good understanding of the efficacy of this class of checkpoint inhibitors,” Lemery says.
Keith Flaherty, director for the Center for Targeted Therapy at Massachusetts General Hospital, sees the decision as “revolutionary” for drug development and approval. Most of the new trial evidence pertained to patients with mismatch repair-deficient colorectal cancer, he says. It’s not surprising, based on the evidence, that the FDA would grant approval for colorectal cancer patients fitting that description.
“But the fact that they were willing to extend the approval to any cancer type based on a relatively small amount of clinical evidence demonstrated in endometrial and gastric cancer is quite unprecedented,” he says. “This represents the high water mark for FDA following the biology, and not only the clinical data, with regard to oncology approvals.”
Flaherty and Lemery both are careful to add that similar approaches surely won’t work in all cancers or with all drug targets. There are already examples to the contrary. They note that targeted drugs have been effective in treating melanoma and lung cancers with mutations in the BRAF gene (which affects cell division and differentiation and is not a mismatch repair gene). But, as single agents at least, those drugs haven’t worked so well in treating colorectal cancers carrying the same genetic mutations.
Flaherty says it’s still hard to say whether the underlying biology will support the development and approval of many other drugs that might work reliably across cancer types. There are other candidates working their way through the pipeline, but for the vast majority it’s too soon to know. There is buzz about an experimental drug developed by Loxo Oncology, where Flaherty is a director and chair of the company’s scientific advisory board. At the 2017 American Society of Clinical Oncology meeting in Chicago, Loxo announced that its drug larotrec-tinib showed a 76 percent response rate across tumors of different types carrying an abnormal gene fusion.
There have been setbacks and failures, and those will surely continue. For example, the FDA recently put three trials testing Keytruda in combination with other drugs for treating multiple myeloma on hold based on evidence that more patients receiving the drug combination died than in the control group. And in July, Merck announced that, in a follow-up study, Keytruda had failed to extend survival in previously treated patients with advanced head and neck cancer any more than standard combination therapy did.
Progress rarely goes in a straight line in cancer treatment. But the molecular approach remains promising. As the examples and evidence add up, Flaherty says, we’re moving toward a future when it may become valuable or even necessary “to analyze the genetic and molecular composition of every patient’s cancer at the time that we are making decisions regarding systemic treatment options” — if not at the initial diagnosis, then perhaps at least for those with cancers that have reached a more advanced stage. (See “Make Me a Match,” page 40.)
The Gift of Time
For Le, it’s hard to put into words what it feels like to see many of her patients, once without hope, doing so well. She offered an anecdote from one older gentleman she’d recently seen for a routine follow-up appointment. “He said to me that when he came to me, both his children were not married and he had no grandchildren. Now, he’s watched both of his children get married, and he has four grandchildren. When you think about the timespan it would take to marry off two children and have multiple grandchildren … it’s totally amazing to be able to give somebody his life back. A lot of times we give people time, but not this kind of time.”
Stefanie has given up the stressful New York career in digital marketing she’d pursued before her diagnosis. She’s back in the Philadelphia area where she grew up and looking forward to working as an advocate for cancer patients. She feels strongly that all cancer patients receive MSI testing. “I now understand a lot of the science, and I’m trying to learn [more] about it,” Stefanie says. “I understand it’s not a miracle. I understand it was 30 years of basic science research that went into this. But, on the receiving end, it certainly feels like a miracle.”
Make Me a Match
In the National Cancer Institute’s Molecular Analysis for Therapy Choice Trial (NCI-MATCH), qualifying patients with advanced cancer of any type are assigned to one of 19 treatment arms based on a genetic analysis of their tumor. The goal is to find out whether it is effective to assign cancer treatment based on genetic changes.
In June, the ECOG-ACRIN Cancer Research Group leading the trial announced that it had reached its 6,000 tumor-sequencing goal two years ahead of schedule. The trial is still recruiting patients who carry certain rare genetic variants discovered through genomic analyses performed as part of standard oncology practice.
Barbara Conley, an oncologist and an associate director in the National Cancer Institute’s Cancer Diagnosis Program, says the rapid accrual shows unexpectedly strong interest from patients and doctors in incorporating genetic information into treatment decisions.
For now, she says, it’s “still an unproven idea that sequencing everybody is clinically effective or has clinical utility.” The trial will help to determine when a genetic analysis should be recommended and when it shouldn’t. Conley says some early results from the trial could be available later this year for some of the treatment arms that have already closed.
Whichever way the findings go, says Keith Flaherty, who serves as the ECOG-ACRIN study chair, the NCI-MATCH trial promises to generate evidence for dozens of potential therapy-biomarker pairings.
“Everyone was telling me I was running out of options.”
Seal of Approval
Since the Keytruda approval in May, the Food and Drug Administration (FDA) has made a flurry of other approvals in targeted cancer therapies. Some of those approvals include the following:
• On May 26, the FDA granted regular approval to Zykadia (ceritinib) for patients with metastatic non-small cell lung cancer whose tumors are anaplastic lymphoma kinase (ALK)-positive.
• On June 22, regular approval was granted for Tafinlar (dabrafenib) and Mekinist (trametinib) administered in combination for patients with non-small cell lung cancer with the BRAF V600E mutation.
• On July 11, Blincyto (blinatumomab) was approved for the treatment of relapsed or refractory B-cell precursor acute lymphoblastic leukemia.
• On July 17, the FDA approved Nerlynx (neratinib) for extended adjuvant treatment of adults with early stage HER2-positive breast cancer.
• On July 31, accelerated approval was granted for Opdivo (nivolumab) in patients with mismatch repair-deficient and microsatellite instability-high (MSI-H) metastatic colorectal cancer that progressed after prior treatment.
• On August 1, the FDA granted regular approval to Idhifa (enasidenib) for relapsed or refractory acute myeloid leukemia with an IDH2 mutation.
• On August 17, Besponsa (inotuzumab ozogamicin) was approved for the treatment of adults with relapsed or refractory B-cell precursor acute lymphoblastic leukemia.
• On August 17, the FDA approved Lynparza (olaparib) tablets for maintenance treatment of adults with recurrent epithelial ovarian, fallopian tube, or primary peritoneal cancers, who have a complete or partial response to platinum-based chemotherapies.
“These patients are not only having responses in the tumor, but in about 20 percent of cases the tumors disappear,” he says, and that’s “in patients who were already signed up for hospice.”