Learning From the Best: On Patients Who Respond Well to Cancer Treatment

CURESpring 2016
Volume 15
Issue 2

Scientists are studying exceptional responders, who benefit from cancer therapies when others don’t, to learn how to duplicate those results in broader groups of patients.

Randy Hillard by Hetler Photography

Randy Hillard by Hetler Photography

Randy Hillard by Hetler Photography

For a brief moment in medical school, Randy Hillard considered becoming an oncologist. Then he looked at the low success rates of chemotherapy at the time — he graduated in 1977 — and decided he probably wasn’t cut out for that brand of medicine.

So imagine his dismay in December of 2010 when he was diagnosed with stage 4 stomach cancer. The disease had spread to his peritoneum, the membrane that protects the abdominal organs.

He soon discovered that, while there had been significant advances in cancer treatments, gastric cancer was still a very difficult disease. Few with his diagnosis survived more than a year.

“I read everything I could and got increasingly depressed,” says Hillard, a professor of psychiatry at Michigan State University. “I even thought about not getting treated.”

Ultimately, he chose a hopeful path and had most of his stomach removed, along with 30 lymph nodes, parts of his liver and gall bladder, and most of the peritoneum. He began a course of oxaliplatin, along with two years of oral 5FU (capecitabine).

Added to that pair of chemotherapies was the drug now believed to be his silver bullet: Herceptin (trastuzumab). The targeted therapy, first used to treat breast cancer, had been approved for metastatic stomach cancer that overexpressed the protein HER2 in the weeks before his diagnosis because it increased life expectancy in eligible patients — barely — from 11 to 13.5 months. Hillard’s tumor was tested, and was found to overexpress the protein.

As he approached his one-year anniversary, Hillard’s oncologist commented on how well he seemed to be doing. “Once I hit three years, he and I both started thinking this was exceptional,” says Hillard, who is done with chemotherapy but remains on Herceptin.

For most of us, exceptional is what we’ve always wanted but rarely gotten. Exceptional is Carnegie Hall and blue chip. It’s when you hit the ball past the outfield, over the fence and into the second parking lot.

When it happens with a cancer therapy you’re called an exceptional responder, and it’s the ultimate jackpot. “The only recognition I’ve ever gotten that really mattered was being an exceptional responder,” says Hillard, now five years out. “I wake up every morning shocked at how non-dead I am.”

But only in the last few years have researchers been able to share in the winnings.


Exceptional responders have been part of the lore of cancer treatment for decades, but the cruel irony was that there wasn’t much a researcher could learn from them.

“All you could really do was be happy for them,” recalls Barbara A. Conley, a medical oncologist with the Division of Cancer Treatment and Diagnosis at the National Cancer Institute (NCI). In 2014, Conley and Louis Staudt, director of the NCI Center for Cancer Genomics, took the reins of the first systemic attempt to gather these tales of exceptional healing and forge them into clinical advances. The research was made possible decades ago when scientists first learned to decode the information in DNA and other genetic molecules. But it wasn’t realistic until the last few years, when continuing technological advances made it financially feasible to gather and interpret the data from individual patients.

This data on exceptional responders could eventually be used throughout cancer care to help match the right patients to the right medications, or to develop new medications that target biomarkers first identified in these outliers.

Understanding what makes patients exceptional responders is ridiculously complicated: A single tumor cell can contain thousands of mutations, most of which have little meaning. But somewhere in this mad scramble hides a clue — or a series of clues — which explains how the cancer runs amok. In the case of exceptional responders, there are additional clues that should explain why a particular drug works. Beyond looking at gene mutations and the problems they cause, scientists can consider changes in tumor proteins that are not associated with genetic alterations, or not easily mapped to a long list of mutations.

“Can we find a reason why these patients should be responding better than others?” asks Conley. “That’s the whole crux of what we’re doing.” The study of exceptional responders is new, but there are a handful of studies recruiting. Adera Labs is looking to better understand the response of patients with pancreatic cancer to standard chemotherapy by comparing the genomes of exceptional responders to those of patients who respond less robustly. And the Breast International Group is considering the mechanisms of either response or resistance to systemic therapy for metastatic breast cancer, with a search for biomarkers in exceptional responders part of the effort.

The NCI began its foray into exceptional responders by delving into its own archives. While there are many stories of people unexpectedly outliving their diagnoses, the scientists were looking for more than just a happy ending: They wanted to see a response to some kind of drug. “It has to be a complete response where we would expect to see a complete response less than 10 percent of the time,” Conley explains. “Or it has to be a partial response, where either a partial response lasting at least six months is seen less than 10 percent of the time, or where the response lasted three times as long as has been reported in the literature.”

That’s not eight out of 154, but eight so far. There is a lot of work yet to be done, followed by further study in the lab and in trials. How long before these discoveries begin to influence standard treatments? “It could be a while,” warns Conley.

Sure, something really big could happen, such as the discovery of driver mutations that override less consequential genetic changes. But the breakthroughs are more likely to be incremental. “Many of us think that the more that we do this, the more we can find,” she says. And it’s a very young science. “We’re not going to get good at it immediately.”


In January of 2010, a 73-year-old woman with advanced bladder cancer was enrolled in a clinical trial at Memorial Sloan Kettering Cancer Center (MSKCC). There was nothing remarkable about the trial; it was just one of several recently opened studies. When treatment options are exhausted, the doctors at Sloan encourage patients to consider investigational treatments. Ultimately, the trial came up negative. The drug being tested, Afinitor (everolimus), has since been approved in several cancer types, but not for the treatment of bladder cancer.

But for this one patient, Afinitor was a revelation: by six months, she'd had a complete response. More than six years later she’s still on the drug, still cancer-free.

David Solit, MSKCC’s director of the Kravis Center for Molecular Oncology, wanted to know why. To find out, he and his team sequenced the patient’s entire tumor genome, the first entire bladder cancer genome ever sequenced. It took months to analyze the resulting cache of information, which totaled 3 billion DNA molecules. Solit and his colleagues found 17,136 different mutations in their patient — and from this managed to ultimately distill two crucial ones, a deletion in the TSC1 gene and a different loss-of-function mutation in the NF2 gene. These mutations made the woman’s cancer particularly sensitive to treatment with a drug like Afinitor, which inhibits mTOR — a protein that regulates cell growth, proliferation and survival. The finding may help direct treatment for other patients.

Solit’s goal is not just to analyze exceptional responders, but to use the information gleaned to accelerate drug discovery. It’s possible to set that sort of a goal, in part, due to the ability to finally read the fine print of the cancer genome. Already, that technology has accelerated the delineation of cancers into smaller and smaller sub-groups of patients who have specific shared mutations, or biomarkers, regardless of where in their bodies their cancers first emerged. None of this would matter if not for the third leg of the stool: targeted therapies that use those genetic signatures to acquire a bullseye.

Developing drugs for smaller groups of people has been a tough sell to drug companies because they want something that will work broadly. “The holy grail in drug development is to develop a drug that works for everybody,” says Solit. “Unfortunately, individualcancers are very different. If you try to develop a drug that works for all cancer patients, often it works for no one.”

But sequencing technology has gotten cheap enough and powerful enough to justify a change in strategy: “It is now feasible to develop a drug for a much smaller population of patients because it is increasingly easy to find patients with rare mutational signatures,” he explains. And as we get better at matching the genetic jumble to the cancer, these drugs are likely to become even better targeted, resulting in treatments that are more effective and less toxic.

Sloan Kettering has embarked on a significant commitment to this approach. Every patient with metastatic disease — more than 10,000, so far — has his or her tumor screened for 410 known cancer-related genes. It doesn’t benefit everybody: The results direct only about a third of the patients toward standard-of-care treatment. For another third, doctors can see what they believe to be the key mutation, but don’t have a treatment option. For the remainder, they don’t yet understand what makes the cancer grow, or how to treat it.

These are critical steps on the road to precision medicine. “It will soon be the standard of care to perform tumor genetic testing on every single patient and then to decide what treatment they receive based on their personal tumor genetic profile,” Solit believes. “My hope is that we’re not going to need the extraordinary responders initiative in the future, because we’ll know the genetic drivers of every patient’s tumor prior to their enrollment in a clinical trial.”

This tumor profiling is leading to a new kind of drug trial called a basket trial. Whereas trials have long been organized by disease type, or the site where a cancer arose, a basket study groups together different cancers with similar mutations. And because Sloan Kettering profiles so many tumors, it’s enrolling patients into basket studies at much higher rates than many other institutions.

“We are seeing enough activity in these basket studies that it will lead to more rapid approval of some drugs,” Solit says. Already, it’s benefited research into several rare cancers, for which it was previously difficult to recruit enough patients for a trial. In one recent basket trial of Zelboraf (vemurafenib), which is prescribed for melanomas with BRAF mutations, the drug proved itself effective in some non-melanoma cancers, as well. “Every patient with histiocytosis, a rare cancer, benefited,” Solit says. “You never would have had a trial of this rare disease without this basket study design.”

These small victories eventually add up. In lung cancer, the alphabet soup of known mutations includes EGFR (15 to 20 percent), ALK (5 percent), MET (3 percent), HER2 (2 percent) and ROS1 (1 percent). That adds up to targeted treatment options for a third of patients with non-small cell lung cancer. “You can view it as glass half full or glass half empty. When I went into oncology, we did not have any of these drugs, and the median survival of lung cancer patients was less than a year,” says Solit. “Unfortunately, approximately 20 percent of lung cancer patients have KRAS mutations, and we do not yet know how to attack this mutation.”

“We still need to do better,” he says. “Every patient has to be considered as an individual. We need to determine what makes their tumor grow and then consider the opportunities to counteract the key cancer mutations found in their particular cancer. A single approach is not going to work. We need to be willing to develop drugs for small populations of patients.”


“This is a really exciting decade coming up,” says Funda Meric- Bernstam. As medical director of MD Anderson’s Khalifa Institute for Personalized Cancer Therapy, she is charged with harnessing the power of in-depth molecular characterization and turning it into significant patient benefits.

Like Solit, she’s very interested in what we can learn from exceptional responders. For example, it’s not uncommon for patients to have a mixed response to treatment. Because metastatic tumors can originate from cells with different mutations, they can react differently to a drug, even in a single patient, with one tumor growing while the other declines. Comparing the genetics of the two tumors could crack a few puzzles about how the drugs work and who is likely to have the best response.

Exceptional responders might also hold the key to an understanding of how tumors evolve, particularly in response to treatment. “Most of the time, the tumor figures out how to outsmart the drug,” says Meric-Bernstam. To outsmart the tumor, we need to figure out how it gets around the drug. That process may be most clear when a patient starts out having a stellar response to a drug and then becomes resistant to it.

Now that we have these molecular tools, the real challenge in keeping research going will move from the technical to the emotional. The problem crops up when doctors need to ask patients for biopsies to be used for research, sometimes at a particularly difficult moment.

“The drug was working and everything was so great. And now the drug has stopped working. And that’s a really awkward conversation to have,” Meric-Bernstam explains. Patients are disappointed and often not ready to stay another day to get a biopsy. Yet she emphasizes that this could be a crucial moment in the life of the tumor — and in research. “If we could now get a sample, we could understand why your tumor stopped responding,” she says, adding that the biopsy may help the patient by identifying another druggable target.

So far, that’s the unrealized promise of these exciting new targeted therapies: coordinating the hits into a calculated knockout punch. Hundreds of clinical trials are already underway in a brute-force effort to crack the combination to a solution that allows more people to benefit from these drugs. Exceptional responders could accelerate this process.

“I actually think most discovery is in that realm: understanding the mechanisms of acquired resistance in those patients that were fortunate enough to have a full-on response of clinical significance,” Meric-Bernstam says. “It’s how we can drive the targeted therapy’s success to the next level.”

“Whether a drug works or not, if we don’t have the tissue at hand, we will never know whether it was a resistant target or the drug was not good,” she points out. And scientists can even learn from negative trials. “Tissue sampling can really make or break our ability to design our next trial better,” she says.

Meric-Bernstam recognizes that patients may find it inconvenient or upsetting to offer these biopsies, which often don’t end up helping them personally. “We understand what a huge ask it is, especially for a research biopsy,” she says. But by participating in exceptional responder research, she and Conley agree, patients are making an important contribution to science.

“I think the patients that volunteer to have this done are amazing,” Conley says