by Azra Raza
It is hard to believe that since 1970, no better strategy has emerged. I remember hearing a talk in 1982 at Roswell Park when a visiting professor said, “We know very well that our children will look at us in disbelief and say, ‘You gave what to your cancer patients? Chemotherapy? Were you out of your mind?’” Thirty years and many, many children, grandchildren, and great-grandchildren later, we are still doing the same. Just having to repeat the same conversation, the same statistics, and the same list of side effects to hundreds of patients a year for the past forty years is embarrassing and deeply depressing.
Had Lady N. lived to 2019, she might have responded to a new drug called luspatercept. Luspatercept was not developed to treat MDS patients. Like most successful cancer therapies, it arrived by their bedside purely by serendipity, developed for one indication, found to be useful for another. Luspatercept traps molecules that would otherwise bind to receptors on the surface of cells; when something is bound to those receptors, they initiate a signal that is important during the formation of bone. Overactivity of this pathway can lead to loss of bone. In patients with multiple myeloma, bone can be eaten away, resulting in creation of holes known as lytic lesions. This class of drugs was developed with the hope of blocking that signal and thereby reducing the number of lytic lesions in bones. When these were tried in healthy volunteers and in multiple myeloma patients, the researchers running the studies noted that the recipients’ hemoglobin sharply increased, often to dangerous levels, ratcheting up so high in some cases that the patients had to undergo bloodletting. The researchers shifted course, and the focus was redirected to treating anemia instead. Enter MDS.
A phase 2 trial of luspatercept was conducted in Europe with encouraging results, particularly showing improvement in anemia of patients whose bone marrow contained ring sideroblasts, or early red blood cells where the nucleus gets surrounded by a ring of iron particles. It is an example of poverty in the midst of plenty. There is an abundance of iron (heme), but the red cell precursors are unable to combine this heme with globin to make hemoglobin. The ring sideroblasts cannot proceed without hemoglobin to become fully mature red blood cells. They drop dead, leading to anemia.
At Columbia University, we opened a multicenter phase 3 clinical trial of this agent in 2016 where 70 percent of patients would receive the drug and 30 percent would receive placebo. One such patient was Mrs. Fern Priestly; she had the ring sideroblastic anemia and was transfusion dependent, which were the two major eligibility criteria for the trial. Lady N.’s marrow also contained ring sideroblasts. The study was blinded, so we didn’t know whether Fern was receiving placebo or the drug when we began, but the side effects she experienced, especially early on in the trial, helped her put two and two together fairly quickly. “I know I’m not on placebo,” she told me, “because of how tired I feel after the shot every three weeks.” You don’t have to be a weatherman to know it’s raining outside. “But it is slowly getting better each time. I think my body is finally getting adjusted to it.” The study was for her, after nineteen years since being diagnosed with MDS, “the high point of the entire saga.” She was so excited about the stabilization of her hemoglobin count with one shot of luspatercept every few weeks. “I have literally gotten my life back.”
In one of the saddest, most tragic twists, dear Fern and her husband, Eldon Priestly, were involved in a deadly car accident. Fern died instantly on Sunday, August 12, 2018, and Eldon passed away from the mortal injuries he sustained, four months later.
How weary, stale, flat, and unprofitable seemed all the uses of this world.
—HAMLET, ACT 1, SCENE 2
WHILE A PATIENT is participating in a study, every new complaint must be eyed suspiciously for possible relationship to the drug. Ultimately, the symptom may or may not be directly related, but extreme caution is necessary when new signs emerge under experimental therapies, as short- and long-term toxicities are inadequately known at the experimental stage. This is the price of novel drug trials. It became an issue with one of my patients on the luspatercept protocol.
In 2011, I met a patient who became my ideal within a few meetings. Gerson Lesser, a tall, handsome, fiercely intelligent, generous, well-read, thoughtful New Yorker of Jewish descent, is a fellow doctor, teacher, and researcher. He came to see me with his lovely, equally intelligent wife, Debbie. We became close as we settled into a routine of regular clinic visits. He had been politically active for most of his life, beginning with marching for the Spanish Republic during the Spanish Civil War. He spent hours at Zuccotti Park, the center of the Occupy Wall Street protests in 2011. A photo of Gerson at the protest went viral, “putting the lie,” in Gerson’s words, “to Rush Limbaugh’s ugly remarks about the movement”: “These protestors who are actually few in number, have contributed nothing. They are pure, genuine parasites. Many of them are bored, trust-fund kids, obsessed with being something, being somebody. Meaningless lives they want to matter.” Gerson, at ninety-plus years of age, was out there with his walker every afternoon.
Once the medical part of his visits was over, we would spend twice that much time catching up on personal details, discussing politics and literature, science and music. He often brought me books he had just read or thought I would enjoy, I invited them to my apartment for book readings, and we went out to many dinners in lovely Manhattan restaurants. Gerson and Debbie have acquired a place in my heart reserved for only a special few. I feel incredibly fortunate every day for the opportunity my profession provides—bearing witness to some extraordinary lives, enjoying unparalleled, intimate glimpses into the most noble aspects of humanity. In the presence of such grace, all one can do is to be grateful.
By the time I first met him, he had been suffering from a chronic, slowly progressive anemia for eight years. Eventually, I put him on the luspatercept trial also. He, too, had a spectacular response, and his hemoglobin jumped by several grams to reach almost a normal level for the first time in over a decade, but he simultaneously developed shortness of breath on exertion. On the off chance that it was due to the medication, we withdrew Gerson from the trial. He promptly became transfusion dependent again and remains so to this day.
Even as both Fern and Gerson experienced dramatic responses to the trial drug, one died of a freak accident and the other could not continue, highlighting the uncertainties involved in the human condition. Nevertheless, luspatercept will be a welcome addition to the parched field of MDS therapeutics when it is approved by the FDA. The problem is that even when it is given to a hundred patients with ring sideroblast type of MDS, only thirty-eight will respond by becoming completely transfusion independent, while sixty-two will not, and based on experience so far, no one will be cured. It is discouraging to see that clinical trials today are designed in much the same way they were thirty to forty years ago. For example, it is clear that using ring sideroblasts as a marker to select patients for treatment was not good enough because not all patients responded. No serious attempt was made in the luspatercept phase 3 trial to understand why 62 percent of patients failed to respond and what is unique about those who did. We could have saved the pretherapy blood and marrow samples on the trial subjects and, once the outcome was known, compared the samples of responders and nonresponders by using the latest molecular tools. This comparison could have provided us with clues to preselect future potential responders. Equally disheartening is the attitude of the regulating agencies because they fail to demand more rigor from sponsors of the trial. What has the agency done to protect sixty-two of one hundred future patients who will have little response to luspatercept but will suffer the side effects of the drug and have to bear the exorbitant cost of therapy after the drug is approved? Nothing at all, unfortunately. The drug makers, on the other hand, expect a neat multibillion-dollar annual market for luspatercept between Europe and the United States once the agent is FDA approved. If I were younger, I would have concentrated more on the positive results for the 38 percent of patients on the trial rather th
an stress over the 62 percent of failures. Now that I am older, I cannot ignore the toxicities, nor the physical and financial tolls, that experimental medications take on patients. Even if Lady N. were alive today and tried luspatercept, there is no guarantee she would have responded and certainly no way of knowing for how long and at what cost of side effects. And once she stopped responding, her disease would still have progressed and killed her, either through transformation to acute leukemia or through increasing profundity of her cytopenias, causing the blood counts to plunge into irretrievable lows.
Such demoralizing news is not restricted to MDS and acute myeloid leukemia treatment. Vinay Prasad, a young hematologist-oncologist at Oregon Health & Sciences University, is a major critic of how the United States spends $700 billion on health care, identifying drug costs, conflicts of interest, poorly designed clinical trials for cancer drugs and diagnostics, and the fact that “more than half of all practiced medicine is based on scant evidence—and possibly ineffectual” as the major issues in the field. Prasad published an analysis of fifty-four cancer drugs approved by the FDA between 2008 and 2012. Of those fifty-four drugs, thirty-six, or 67 percent, were approved based on so-called surrogate end points—that is, on the basis of something other than a known effect on the tumor leading to improved survival. Indeed, follow-up over the next several years showed that thirty-one of those thirty-six approved drugs yielded no demonstrable gains in survival. What are we doing wrong? Perhaps the one-size-fits-all approach is the problem? Can we improve these grim numbers by custom-designing therapy to suit individual patient needs? Precision medicine.
THE IDEA OF individualized therapy is attractive and logical on the surface. Take, for example, the drug Vidaza, which Lady N. received for six months to no avail. Other patients with MDS respond quite well to Vidaza, with the drug being effective enough that treatments—which can be debilitating—become less frequent with time. Mark De Noble, at eighty, was able to drive across the country with his wife after going on the drug in 2015: “It is February of 2019 now, and I continue to receive Vidaza for five days every six weeks and regularly visit Dr. Raza for my periodic bone marrow biopsies. My wife and I travel several times a year, mostly by car, exploring new places and visiting family and friends. At home, we enjoy hosting friends and family. Now that we’re retired, we volunteer once a month at a residential facility for fifteen troubled teenagers. As we all prepare a three-course dinner, we teach them how to prepare foods, handle kitchen tools, set a table, serve food, etc. Then we enjoy a delicious meal together.”
Mr. De Noble had an extraordinary response to Vidaza while Lady N.’s counts did not budge on the same drug despite half a year’s treatment. Under the microscope, their disease looked similar. In fact, Mr. De Noble experienced such complete and durable benefit from the drug that he earned a special moniker reserved for exceptional responders: unicorn. Traditionally, clinical trials of experimental agents are statistically powered to deliver response in a predetermined minimum percentage of patients. If the number fails to meet the end point, the drug is thrown out like the baby with the bathwater. This changed in 2012 as a result of a trial in which the drug everolimus given to patients with urothelial cancers produced overall dismal results in forty-four treated subjects but one of them showed a truly outstanding response. A deeper investigation into the reason for such exquisite sensitivity revealed the presence of unexpected mutations not previously associated with that type of bladder cancer, demonstrating once again the profound biologic variability within morphologically identical tumors. This one case led to the initiation of a pilot study, funded by the NCI, directed at identifying molecular features associated with exceptional responses. In the study referenced above, everolimus was the perfect drug for the exceptional responder, but was it worth having forty-four others suffer only the drug’s toxicities without any appreciable benefits?
The ideal situation would be to administer the drug only to preselected potential responders. Identifying predictive markers that allow for individualizing therapy by matching drugs to patients remains the treasured yet elusive holy grail of oncology. To what extent is this strategy being pursued? More than 90 percent of trials ongoing around the country make almost zero attempt to save tumor samples for post hoc examination to identify predictive biomarkers. Even in the NCI-funded study of exceptional responders previously mentioned, only genetic mutations were investigated as the single potential predictive marker. What if the reason for response was not a mutated gene but abnormal expression of the gene at the RNA level, or that it resided entirely outside the tumor cell, related to the microenvironment of the tumor? Why are we not making the required efforts in as comprehensive a manner as needed? Who is pushing this short-term agenda driven by the singular goal of getting a drug approved with alacrity as long as it meets the bar of improving survival by mere weeks in a few patients?
A patient with MDS that I became extremely close to over the years was Barbara Freehill. She had a lower-risk MDS that evolved to an overlap myelodysplastic-myeloproliferative neoplasm (MDS/MPN). I saw her every two to three weeks as she was steadily transfusion dependent. Her poise, dignified personality, gorgeous looks, and her incredible wisdom combined to make her one of the most amazing people I have been fortunate to take care of. We could talk about anything under the sun. I treated her for a long time with Dacogen and then Revlimid. She was under my care, seeing me two to three times a month for several years when one day she showed up without an appointment. My nurse came and told me Barbara wanted an urgent word with me. I went out to see her, and she could hardly breathe, so anxious was she. Her youngest daughter, thirty-nine-year-old Kendra Seth, was in the ICU. I will let Kendra tell you this shocking story:
I was admitted through the ER for a pain I was experiencing on my right side (I thought maybe I had pulled a muscle running). That pain quickly went from mild to excruciating. After a long night, a CAT scan revealed that I had a massive clot in my portal vein. The clot was virtually strangling all of the blood flow to my major organs. In order for me to survive, the clot had to be cleared as quickly as possible. After 3 failed surgeries & little hope for a plausible next step, my mom suggested that her doctor stop in for a “visit.”
I begged my mom not to… after the failure of my “best chance” surgery I was at an all-time low both mentally & physically. My body seemed to go into complete revolt, although I hadn’t eaten in weeks, I gained over 30 lbs in water weight virtually overnight. I couldn’t bend my fingers or toes, or even roll over—I was literally a prisoner in my own bed.
Mentally I just could not wrap my head around how a year ago I had successfully climbed Mt Kilimanjaro—how my life had changed so dramatically in a number of months. The last thing I wanted was another team of doctors that only asked the same questions and provided no answers.
And then I met Dr. Raza… at the time she walked through the door we were discussing my “best” option which was a 5 organ transplant. All I could think about was getting home to my husband & 4 young children so I was all for it—shows you how desperate I felt. When Dr. Raza quietly came in, she didn’t ask me all the usual questions, she spoke to me as a person, not a case—her humanity was immediately apparent. Dr. Raza suggested I be tested* for a mutation in the gene Jak2 and when that came back positive about 10 days later it was our first breadcrumb to getting me healthy.
(*Note: My colleague Joe Jurcic, who had been consulted as the hematologist on service, had noticed the high platelet count also, and had preempted me in ordering the test.)
The most dramatic feature of Kendra’s story is that she went from being considered for a five-organ transplant as she lay in the ICU, deathly ill, to being managed by aspirin alone. This happened because it was her underlying bone marrow overlap syndrome (MPN/MDS) causing her to have high platelets, which, in turn caused blood to form clots in large and small vessels. Aspirin reduces the clumping action of platelets, preventing clot formation. The life of a beautiful thirty-nine-yea
r-old mother of four was saved because the right drug was matched to the right patient.
KENDRA’S CASE SUGGESTS a great plan: find a mutation by sequencing the DNA in the patient’s cancer cells, match a drug with activity against the mutated protein, and administer it irrespective of which organ bears the tumor. This approach combines the best of available technology and preselection of patients likely to respond, resulting in therapy tailor-made to suit the need of an individual patient. Precision medicine. Customized health care. Targeted therapy. Predictive modeling. Optimized strategy.
All sound terrific. The wave of the future. The fashionable thing to do. Mostly, it does not work. Here is what happened. Two types of trials were conceived. In one, called umbrella trials, tumors affecting the same organ but presenting with different genetic mutations could be matched with targeted therapies. For example, one lung cancer has an EGFR mutation—and the best treatment for that patient would be an EGFR inhibitor like erlotinib—while another patient’s lung cancer has a mutation in the HER2 gene, for whom Herceptin would be the right match. A second type of trials, called basket trials, pursues the same mutation as it appears in tumors in various organs; the idea is that the targeted therapy should work for all. For example, a mutation in the EGFR gene in a patient with pancreatic cancer should respond to erlotinib as well as a lung cancer patient with the same mutation. Many cancer programs have stepped forward to promote the idea of precision medicine because it seems the right thing to do.
