Kent Haffer remembers when his oncologist approached him with an unusual idea. It was 2013, and Haffer, a 55-year-old computer programmer living in St. Peters, Missouri, had been receiving treatment for advanced melanoma for eight years. His oncologist proposed that Haffer be part of a clinical trial, but not one involving hundreds of people. Instead, this study would have only one person in it: Haffer.
A one-person study is often called an
N-of-1 trial (the N stands for “number”). Cancer researchers are increasingly
looking at this investigational strategy to capitalize on the molecular information provided by genomic sequencing, in which cancer cell mutations are identified by analyzing a patient’s tumor. The idea is that the patient receives an experimental therapy targeted to those mutations and is monitored for disease progression and side effects while researchers collect data.
In Haffer’s case, that experimental therapy involved three doses of a cancer vaccine that he would receive every six weeks. His oncologist’s goal was twofold: One, he wanted the vaccine to help bolster Haffer’s immune system to prevent a recurrence of melanoma, which had been treated successfully in a traditional clinical trial. And two, he wanted to study Haffer’s response to the vaccine to see if it would be a valid approach in other patients.
A Rough Path To N-of-1
Haffer had already had a bumpy journey. He found a suspicious spot on his left calf and was later diagnosed with stage III melanoma in 2005. After the initial biopsy and surgery to excise the first tumor and lymph nodes in his left leg, he self-administered interferon injections to his abdomen three times a week for six months. Despite these efforts, the melanoma reappeared and spread. Additional cancerous spots showed up on his leg, and CT scans showed new tumors growing in his pelvis. Over the next several years, he had nine more surgeries to remove tumors, “sometimes seven or eight of them at a time,” he recalls. He also underwent isolated limb perfusion—in which high doses of chemotherapy are administered only to one limb and removed from the bloodstream before they can reach and damage internal organs. Still, scans showed that the disease had progressed to stage IV, spreading further into his pelvis. By 2008, his oncologist declared his disease unresectable.
In that same year, Haffer’s oncologist, Gerald Linette at the Washington University School of Medicine in
St. Louis, enrolled him in a traditional phase III clinical trial of Yervoy (ipilimumab), a then-experimental treatment that helps the body’s immune system attack cancer cells. Haffer took 26 doses of the drug, stopping when his immune system became overstimulated and caused mild inflammation in his lungs. But he experienced a complete response: All detectable signs of tumors disappeared, and his scans remain clean nearly eight years later.
Linette says few patients respond so completely to immunotherapy, and he wanted to know what was different about Haffer. So in 2013, Linette approached him with another experimental treatment. Linette and his colleagues had developed a method of making a personalized cancer vaccine—Haffer calls it a “custom cocktail”—based on mutations in a patient’s own tumor cells and designed to give the T cells a boost. T cells are white blood cells, part of the body’s immune system, that help protect the body from infections and can be used to fight cancer. Unlike preventive vaccines for diseases like measles or shingles, which can protect healthy people from infection, therapeutic cancer vaccines aim to fortify the immune systems of people who already have cancer. Linette wanted to use the vaccine to probe and even strengthen the molecular mechanism that made Haffer’s immune system respond so well.
Many patients develop resistance to immunotherapy, and Linette wanted to counter that. “Based on our understanding of the human immune system, we thought that [Haffer’s] immunity against the tumor would also diminish over time,” says Linette, who believed the vaccine might help keep recurrence at bay and assist researchers to better understand how the immune system works against cancer.
Haffer didn’t hesitate. He told Linette, “Let’s do it.”
Pros and Cons Of N-of-1 Studies
Large phase III clinical trials provide evidence that the U.S. Food and Drug Administration (FDA) relies on in deciding whether to approve a drug for a specific patient population. But most trials report average responses, meaning, in theory, that no one in the trial may have experienced the reported outcome. Some patients taking the drug may survive longer; others may die sooner. It’s impossible for a large-scale clinical trial to predict how each individual cancer patient will respond to a given drug. Furthermore, researchers may stop studying a drug if not enough patients respond to it during a clinical trial—even though some might have benefited.
But in papers, panels and presentations, experts argue that individualized studies involving only one or a small number of patients may be a more insightful way to determine treatment—especially in some cancers, like those that are rare or treatment-resistant. A group of physicians and epidemiologists in Canada described the N-of-1 approach in a 1986 paper in the New England Journal of Medicine, outlining a design for studies on individuals. But Laurence Collette, a biostatistician with the European Organization for Research and Treatment of Cancer in Brussels, Belgium, says researchers still have a long road ahead of them to make N-of-1 studies robust and clinically applicable to a wide patient population. “It’s very experimental and not regarded as a statistically valid approach,” she cautions. She and other experts note that in order to be useful to a larger community, data from N-of-1 studies must be collected in a standardized fashion, anonymized and stored with other data.
In April 2015, biologist Nicholas Schork from the
J. Craig Venter Institute in La Jolla, California, wrote in Nature that in many cases, N-of-1 studies are exactly the right tool for developing personalized therapy, in which a person’s treatment matches his disease. Schork noted that if N-of-1 studies are done properly—sticking close to guidelines like the ones proposed in 1986—their findings about a patient’s response to intervention could be as statistically valid as those from large, classical clinical trials.
Cancer researchers are increasingly calling attention to the need to formalize the N-of-1 trial. Done correctly, the approach has the potential to help a wider population than just the person at the center of the trial. Analyzing and sharing individuals’ data—about genetic mutations, diagnosis, treatment, response and survival—can help researchers better understand tumor biology and prescribe treatment based on the particular mutations that drive a person’s disease.
Explaining Exceptional Responders
In 2009, researchers at Memorial Sloan Kettering launched a phase II trial of Afinitor (everolimus) in metastatic bladder cancer patients. At the end of the trial, nearly all of the study participants showed either no obvious change in their cancer or disease progression. By most measures, the trial was unsuccessful.
Except for one patient.
One woman participating in the trial saw her disease disappear entirely, and six years later she still shows no signs of cancer, despite having a diagnosis with a five-year survival rate of about 5 percent. She is an example of what researchers call an exceptional responder—a patient whose unique reaction to treatment is strikingly dissimilar to that of other patients with the same disease undergoing the same treatment.
“Patients who have an exceptional response can teach us about the biology of these tumors,” says computational biologist Barry Taylor at Memorial Sloan Kettering Cancer Center in New York City. He and his colleagues sequenced the genome of the exceptional responder’s tumor, which means they mapped its DNA and identified genetic mutations. Like detectives looking for evidence, the researchers sought mutations that would help them understand the patient’s atypically successful response.
They first identified 17,136 potentially involved mutations in the patient’s tumor genes, then winnowed that list to 140, and finally to two. They searched for those two mutations in 96 bladder cancer patients and finally narrowed their focus to a mutation in the TSC1 gene. That analysis eventually led them to hypothesize that patients like the exceptional responder, who had mutations in the TSC1 gene, may respond well to Afinitor. They pursued that hypothesis in yet another group of 13 bladder cancer patients who had been part of the Afinitor trial and found that patients with the TSC1 mutation did have a longer time before a recurrence—though they didn’t have the same dramatic recovery. The findings were published in Science in 2012.
Earlier this year, two separate research groups reported in the same issue of the Journal of the National Cancer Institute how genetic sequencing was used to develop successful treatments for two individual patients. Each had similar genetic mutations but different types of cancers. In a commentary published in 2013 in the Journal of Clinical Investigation, researcher A. Rose Brannon and oncologist Charles Sawyers, from Memorial Sloan Kettering Cancer Center in New York City, describe how researchers have used an individual’s tumor to pinpoint genetic mutations associated with a lethal prostate cancer. Sawyers is a past president of the American Association for Cancer Research (AACR).
Last summer at the AACR precision medicine conference in Salt Lake City, genomics researchers outlined a way to use the approach in pediatric patients with rare or recurrent disease. And in a study published in November 2014 in ecancer-medicalscience, medical oncologist Arturo Loaiza-Bonilla at the Abramson Cancer Center of the University of Pennsylvania in Philadelphia and his colleagues reported on a woman with stage IV liver cancer—with metastases to her lungs and bone—who responded initially to a dual treatment of targeted drugs. The researchers devised this treatment strategy by matching the genomic profile of the patient’s tumor to the actions of drugs that targeted one of the tumor’s mutations. The patient died, says Loaiza-Bonilla, but he thinks the treatment was beneficial. “She gained more than a year and a half through this approach to be with her family. It was a remarkable response,” he says.
Right now, patients with rare cancers are most likely to benefit from an N-of-1 approach, but other projects are expanding the reach of individualized treatments. In 2014, the National Cancer Institute (NCI) announced the Exceptional Responders Initiative, which will collect tumor tissue from patients who, for example, responded to a drug that was largely ineffective in early-phase clinical trials. This research into exceptional responders is now a part of a wider NCI effort called the Precision Medicine Initiative, launched in 2015, which aims to use genetic information to diagnose and treat cancer.
Another project, called the NCI-MATCH (Molecular Analysis for Therapy CHoice) trial, uses genetic sequencing to match patients to other treatments after their disease has progressed following standard therapy. NCI-MATCH offers a new avenue of treatment for patients with rare or treatment-resistant cancers. It began enrolling cancer patients in August 2015 and uses an approach called a “basket” trial. Patients who enroll in the trial have their tumors sequenced and are assigned to a treatment group, or basket, based on their tumor’s genetic mutations. The treatments include those approved by the FDA, some of which may be used off-label, and investigational drugs that have not been approved but for which a dose is known.
Obstacles To Wider Use
Even with projects like these and recent advances in genetic sequencing of tumors, many hurdles remain before N-of-1 studies or other small-scale precision efforts will become common.
Elaine Mardis, a cancer researcher at Washington University in St. Louis and part of the team that helped design the vaccine for Haffer, says many physicians today aren’t even aware of recent advances or these types of opportunities for patients.
“Many oncologists, especially if they trained and were certified in their profession 10 or more years ago, really aren’t familiar with what genomics can and can’t do in the clinical setting,” she says.
In addition, the benefits of tumor sequencing have been shown only in small studies—and large studies are needed in order to demonstrate that the approach can extend a person’s life. So for now, genomic sequencing, which can be costly, generally is not covered by most insurance companies or Medicare, although that might be changing. In January, the Independence Blue Cross health insurance company, based in Philadelphia, announced it would cover the cost of genomic sequencing for some cancer patients. Many insurance companies already offer coverage for genetic testing in people who have a family history of cancer, as in the case of carriers of mutations in the BRCA1 or BRCA2 genes.
If insurance doesn’t cover the testing, either a patient pays out of pocket or the institution covers the cost. At Memorial Sloan Kettering, the majority of patients with recurrent disease have their tumorDNA-sequenced. The cancer center funds much of this work through philanthropic donations and institutional funds. At Washington University, Haffer didn’t have to pay for his treatment because it was covered by philanthropic donations to the medical center.
For his part, Haffer says he feels fortunate to have landed in a customized study. In May 2015, Linette, Mardis and their colleagues reported in Science that Haffer and two other melanoma patients who had received personalized vaccines showed signs of a strong immune response. The other two had some disease but remained stable.
Haffer says that when he was first diagnosed with melanoma, he had limited treatment options. Now, he says, “there are so many things you can do. You can try a little of this and a little of that. They’re making a vaccine specifically for me. And you can’t get any better than that.”