by Azra Raza
Over the twelve-year period from 2002 to 2014, seventy-two new anticancer drugs gained FDA approval; they prolonged survival by 2.1 months. Of eighty-six cancer therapies for solid tumors approved between 2006 and 2017, the median gain in overall survival was 2.45 months. Of the cancer drugs approved during the past two decades, 70 percent of them were at best useless, showing no measurable survival benefit. Between 30 and 70 percent of the drugs may actually be harmful to patients. A study published in the British Medical Journal showed that thirty-nine of sixty-eight cancer drugs approved by the European regulators between 2009 and 2013 showed no improvement in survival or quality of life over existing treatment, placebo, or in combinations with other agents. My own specialty, MDS, is a case in point. There are two approved strategies to treat MDS. One drug, lenalidomide (Revlimid), is restricted for a subset of patients, roughly 10 percent, whose MDS cells have a deletion in the long arm of chromosome 5. For the remaining 90 percent, one of two approved drugs, azacitidine (Vidaza) or decitabine (Dacogen), are recommended. With either drug, the chance of improving anemia in low-risk MDS, to the point at which transfusions would no longer be needed, is approximately 20 percent. There is currently no way to preselect the 20 percent of patients likely to respond. This means 80 percent will receive chemotherapy for five to seven days every month for a minimum of six months with little or no benefit but with all attendant toxicities and at prohibitive financial expense. In responding patients, the drug administration must continue for as long as there is no progression in the disease. Responders are not cured; the median duration of response is ten months, and an occasional patient continues in remission for years.
So what advice should an oncologist give to a patient faced with these options? In a larger sense, the choices we make for our patients are made by people we never meet. Even if I felt differently, I could not make a truly independent decision. Other experts have devised formal criteria for best practices, and any nonconformity could leave the deviant open to legal challenges. Driven by internal forces, we seek refuge in emulation. Responsibility is assumed collectively by a group of key opinion leaders, or KOLs, in the field. The group takes into account all the existing scientific literature and a fair summary of innumerable clinical trials to eventually distill the experience into a broad set of principles. The guidelines that emerge are at the heart of evidence-based medicine, and the wider community of oncologists subsequently uses these to classify, stage, and treat their cancer patients, evaluating the results of their treatment in a uniform and universally interpretable language.
This is a good thing. Indeed, evidence-based medicine is essential. But it is not sufficient by itself when caring for individual patients. No matter how large or statistically significant the data are from which the universal rules are derived, application of population-based insights to specific patients remains very challenging. The typical experimental trial with a 30 percent response rate is really telling us that if a hundred patients with similar clinical and biologic characteristics were treated with the drug, thirty will likely respond. For an individual patient today, we have no way of knowing whether they are one of the 30 percent who respond or the 70 percent who don’t. Besides, how meaningful is the response? If the median duration of response is, say, ten months, then of the thirty patients who responded, fifteen will lose the response before ten months, and fifteen will continue to benefit beyond that. A few of those will be long-term responders. The disease will come back. This rule applies to even the most successful targeted therapies today with only rare exceptions. They offer improvement in survival by a few months over supportive care. Nevertheless, when I’m faced with an elderly patient with lower-risk MDS without deletion of chromosome 5, receiving two units of blood every two to three weeks, the KOLs say I should give the FDA-approved treatment, despite its 20 percent chance of a limited-duration response. And when faced with a patient like Omar, should experimental trials be offered if the treatment is of no survival benefit? Again, the KOLs say yes.
Imagine now, with these data in hand, you are sitting across the desk from Omar. It is impossible to use the best evidence-based medicine derived from large populations to make decisions about him as an individual. There is embarrassingly little information to predict the most likely outcome for Omar. If he were one of the fortunate ones, then we hoped he would be that rare long-term responder. We had to give it a shot. Nothing ventured, nothing gained.
The oncologists believed that by treating Omar with experimental drugs and chemoradiation therapy, they at least offered him a chance of response, no matter what the odds. But the problem isn’t simply that the drugs he was given ultimately didn’t help him—the problem was with the advice he got, too. It’s possible the advice we gave him wasn’t realistic or explicit enough—perhaps we should have suggested he spend whatever time he had left enjoying life rather than vomiting his guts out after each round of chemotherapy and living on revolting, tasteless liquids because of the raw carbuncles that studded his throat. He could have spent at least a little time traveling with his new wife, visiting friends in England and his family in Pakistan and Bangladesh. Instead, Omar was a perpetual captive; either he was receiving one kind of therapy or another or suffering their side effects, which, beside the vomiting and the ravaged mouth, included very low blood counts and a highly suppressed immune system, landing him in the hospital regularly with bouts of infections.
Was it really the best solution to do nothing? If we had withheld treatment, the tumors would have grown rapidly and caused tremendous pain as well. Which would be less excruciating? Subjecting patients to painful toxicities of futile treatments with their enormous attendant physical, financial, emotional, and psychological burdens is challenging. Would palliation of the pain with local control of growing tumor masses have been any less painful? Did we ever give Omar the choice of no treatment at all? And should we have? The past is some guide. The toxicities of chemotherapy and radiation therapy are well recognized today, whereas it has become rare to see the ravages of unconstrained cancer. Stephen Hall, in his excellent book A Commotion in the Blood, describes the last stages of a highly malignant sarcoma in a young girl at the end of the nineteenth century:
The endgame in cancer is never pretty, less so in an era where doctors chased rather than managed the less ghastly symptoms. The breast tumors had become the size of goose eggs, the abdominal tumor even larger; the length of her body from head to toe was stippled by small tumors that Coley likened to buckshot or split peas. Last came the vomiting, several times a day, though she had no solid food; soon, she was regurgitating copious amounts of blood. “The attacks occurred almost hourly,” Coley noted, “and were very exhausting to the patient in her extremely weak condition.” Elizabeth Dashiell remained conscious of this horrific piracy of her eighteen-year-old body until very nearly the end, when finally, mercifully, she died at home in New Jersey at 7:00 a.m. on January 23, 1891.
Not only is such an uncontrolled death horrible, hopes are squandered on chasing cures that can’t be found. But then, unexpected benefit can also occur, even after ten years of repeated failures, if the right drug is given. The challenge is how to match the right drug to the right patient from the start.
One patient of mine, Philip Kolman, suffering from a lower-risk MDS, was essentially giving himself up for dead. In his own telling, “One day in early 2017, my Florida doctor told me that he had nothing left to give me. My transfusions were becoming very frequent, two or three units of blood a week. He said that I should contact everyone I knew to see if there was a [research] program available for me.” In stark contrast to Omar and his siblings, Kolman says, “I accepted the news with the understanding that I didn’t have much time left, and I started to make final arrangements.” Among them was to write to me. Although he was prepared to lie down, I was not; I told him to fly to New York for tests for a new research program. Once he was in, his need for transfusions quickly dropped from every week to every four to five weeks; hi
s condition worsened a bit before stabilizing at a transfusion every two to three weeks. “I’m now waiting for the next drug to come along with its promise of a new beginning and hope.”
WEEKS BEFORE HIS death, I visited Omar at home on his fortieth birthday. He was quite the dandy, and that evening, he had taken care to dress up. He wore a formal black jacket and beautifully fitting trousers. With an impossible innocence, he took me aside. He had something to show me: a rock-hard reddish growth that had appeared out of nowhere on his arm in the preceding forty-eight hours. With an indefatigable will to live, the exceptionally intelligent young man stared intently at his arm and asked me whether I thought it meant the return of the sarcoma. He hoped I would say no, that it was an infection. It was the one time during the course of my time with Omar that I felt physically ill—and I was not even his family. It wounded me to think of how Mursi and Kamal, Sara and Farid, and most of all Naheed would take the cancer’s resurgence. I could not bear to stay at the party. Despite Naheed’s remonstrations, I left within minutes, and before I could reach the subway, I was retching on the sidewalk.
My husband, Harvey Preisler, was directing the Rush University Cancer Center in Chicago when, at fifty-seven years of age, he received the diagnosis of cancer. He had personally supervised my training in oncology. One rule he emphasized was not to become too close to patients. I am not certain that I have followed his advice as faithfully as he wanted me to. He appalled me when he said, “You are going to take care of me.”
“But, Harvey,” I objected, “all my life, you are the one who insisted that I could no longer remain objective if my feelings clouded my clinical decisions.”
He simply said, “Sorry, I only trust your judgment.”
In the subsequent five years, we looked at countless blood reports, MRIs, and CAT scans together, staring at the growing masses in his abdomen, the persistent fungal infection spreading menacingly in the lungs. Harvey knew precisely what those images meant. He was not someone looking for false hope. He was not a man easily duped. Yet he would invariably turn to me and ask, “So what do you think, Az?” He needed to suspend his judgment and looked to me to decide how he should feel. I took infinite care never to break his spirit.
Julie Yip-Williams, who blogged about her colon cancer and died on March 19, 2018, at age forty-two, said, “Cancer crushes hope, leaving a wasteland of grief, depression, despair and a sense of unending futility. Hope is a funny thing, though. It seems to have a life and will of its own that I cannot control through the sheer force of my mind. It is irrepressible, its very existence inextricably tied to our very spirit, its flame, no matter how weak, not extinguishable.”
What were Omar’s choices? Succumb to hopelessness and despair, face the terrified looks of his wife and mother who followed his every move, or pin his hopes on the oncologists pushing the limits of modern medical offerings? With cancer, it is rarely a matter of either-or; there is seldom a choice between hope and despair. Patients face both simultaneously, or serially. Omar did, too, with a stoic’s sobriety combined with an unflagging optimism of will.
OMAR’S EXPERIENCES, AND Philip’s, point to some devastating concerns about the state of cancer research today.
A common semantic distortion relates to the description of an ineffective therapy as “the patient failed the drug” instead of the other way around. The drugs, not the patients, arrive at the bedside for clinical trials when confidence in their success is 5 percent at best. The preclinical lab data used to identify the potential benefits of a drug cannot predict what will actually work in a clinical setting. We were forced to use trial and error both in Omar’s case and Philip’s, instead of being able to identify sooner what could or could not work for each, at great financial and personal cost. What are we doing wrong? Why have we failed to translate the scientific advances of high-profile publications into improved outcome for our patients?
It is high time to question the current paradigm of research. There are bright spots—many subsets of patients, even with aggressive tumors, have been successfully treated with drugs developed using the present approaches: chronic myeloid leukemia, most childhood malignancies, and some forms of adult bone marrow and lymphoid cancers. We shall see why. But we shall also see that the exceptions exist among a litany of failures. These failures are systemic. The vast majority of researchers are studying diseases they never see, in animals who don’t get them spontaneously, or in test tubes where the “cancer” must be artificially created and maintained. Such contrived data bear little resemblance to the actual tumors, yet these “models” are the ones turned over to industry for further clinical development. This approach to drug development, the exceptions notwithstanding, has been stupendously unhelpful. How did we get here?
IN JANUARY 1912, Alexis Carrel, soon to be a Nobel laureate for work in surgery, removed cells from the heart of a chicken embryo, plated them on a dish in his laboratory, and, to the great surprise of the scientific community, kept them alive and growing robustly for the next three decades. The cells thrived as long as they were fed the right cocktail of nutrients, and Carrel’s miraculous culture led to the conclusion that living cells have the potential for immortality. Unfortunately, no one else could replicate Carrel’s results—in general, investigators could maintain cells in culture, but no one could demonstrate the continuity of survival for weeks, let alone decades—nor explain what enabled them to survive in Carrel’s flasks.
The question whether cells possess the potential for immortality remained unresolved until 1960, when Leonard Hayflick provided the answer. Through a complex series of experiments, Hayflick succeeded in routinely growing cells in culture for long periods, but not forever. Cells are not immortal. They age and they die. If external forces do not kill them first, Hayflick found that, after roughly forty-five divisions—known today as the Hayflick limit—cells follow one of two paths. Either they eventually dial down their activities to the bare minimum necessary for viability, curl up, and enter a period of senescence, or they commit suicide. Carrel, Hayflick argued, could not have been culturing his original cells all those years. Instead, the nutrient solution Carrel used daily to feed the cultures most likely contained viable embryonic stem cells, which seeded and grew on their own.
The Hayflick limit, accepted as a golden rule of biology, has proved to be true for normal cells ever since. Cancer cells, however, are different. One tumor took off in the laboratory and achieved immortality. On February 8, 1951, a cervical cancer was removed from Henrietta Lacks and brought to the laboratory of George Otto Gey. HeLa cells, labeled using the first two letters of the patient’s first and last names, began to thrive in culture, giving rise to the first human tissue culture “cell line.” Acting almost as if they were a monstrous superorganism, HeLa cells have steamrollered their way from test tubes to animals, gulping cocktails of nutritious chemicals, floating in flasks and cutting jagged paths across methylcellulose-coated petri dishes, climbing, creeping, fanning, and expanding perpetually for six decades. They metamorphosed; compared to the normal human cell’s chromosome number of 46, their chromosome number varies between 70 and 164. HeLa cells are unique in their ability to survive under the most challenging environmental conditions, carving out a space for themselves with unmatched velocity, be it in inorganic flasks or in mice.
To date, some forty thousand pounds of HeLa cells have been grown, studied, molecularly dissected, genetically reprogrammed, used as teaching tools for graduate students, formed the backbone of elaborate, major grant proposals, and otherwise spread throughout science. This orgy has led to an embarrassment of riches for the researchers, earning for them thousands of patents covering diseases ranging from polio to cancer. Ironically, this unexpected gift, an enormous boon for researchers, exchanged hands and laboratories, crossed oceans and continents, all without the knowledge or consent of Ms. Lacks, who died eight months after the original tumor was plucked from her pelvis. (Rebecca Skloot skillfully recounts the scandalous drama of He
La cells, involving interactions of race and research, greed, business, and bioethical issues, in her 2010 best-selling book, The Immortal Life of Henrietta Lacks.)
The consistent, predictable growth and behavior of HeLa cells provided researchers with an opportunity to experiment, including tests of the efficacy of a number of agents, on a reproducible in vitro model. The success with HeLa led to the broader discovery that, with practice, skill, and a little luck, malignant cells from a variety of tumors could be induced to grow continuously in the laboratory. This in turn gave birth to development of additional cell lines, and researchers flooded the field with a deluge of experiments conducted on all types of cancers.
Many such experiments examined the effects of potential anticancer agents on these tissue culture cell lines with the hopes of developing reliable methods to predict responsiveness. The question was, how faithful are cell lines to their ancestry? Partially so. The success of a tumor in a human (or any other animal) depends on many factors, including how well it has managed to subvert the tissues in which it exists to support its growth at the expense of normal cells surrounding it. Cell lines are created by removing tumor cells from this natural habitat, forcing them to adapt to a new, and hostile, environment. The journey from an organ to plastic containers results in the creation of almost a new species of cells that diverge wildly from their parents in morphology, genotype, phenotype, and biologic behavior. The artificially grown cells can only replicate some but not all the characteristics of the cells from which they originated. As a rule, for example, they don’t grow in perpetuity. To survive for any length of time, however, additional transformative changes occur, affecting not just the raw material of genome but also the expression of genes, so that before long, cells in vitro bear little resemblance to the parent from which they originated. For one thing, the doubling time of cultured cells is much faster. In fact, they are selected for long-term passages in the lab precisely because of their ability to divide rapidly and grow furiously. Cultured cancer cells also have a very different relationship with oxygen. In the body, cancer cells exist with low levels of oxygen, whereas those in the lab come to require significantly higher oxygen levels—up to ten times as high.