In City of God, Saint Augustine tells the story of Innocentia of Carthage. Diagnosed with breast cancer, she was advised by a physician to follow the ancient recommendation of Hippocrates: “It is better and safer not to treat than to treat a hidden cancer : let a good diet be sufficient.” Later, in a dream, she was instructed to go to a church and ask a newly baptized stranger to make the sign of the cross on her breast. Upon waking, Innocentia did just that and was cured. Medicine derived from dreams and miracles continued at least until the 14th century, when French physician Jean de Tournemire, whose 18-year-old daughter was afflicted with breast cancer, prayed for intercession from Avignon cardinal Peter of Luxembourg. De Tournemire had diagnosed his daughter himself, palpating a nodule the size of a hazelnut with corrosio fortis, or the rough, ulcerated fissures that he recognized as the “unmistakable” signs of an advanced cancer. A colleague recommended surgery, only with “great fear and caution, after the surgeon is persuaded by many insistent requests and a very high fee.” Application of curative ointments, or insufficient surgery, might only expedite ulceration of the tumor, hastening death. De Tournemire, too, recalled Hippocrates’s aphorism. His daughter didn’t need an operation; she needed a miracle.
A 74-year-old man with metastatic lung cancer, Mr. R, was referred to radiation oncology for possible treatment for an asymptomatic tumor in his chest. A surveillance scan showed that the primary tumor, in the right lower lobe of his lung, was growing, while the few metastatic sites he was living with, in his sacrum and liver, appeared to remain controlled by the chemotherapy he’d been on for 10 months—about two months longer than the median survival for metastatic lung cancer.
At first, it was not clear why he’d even been referred to radiation oncology. As a rule, patients with metastatic cancer cannot be cured. Chemotherapy can add months, which is why it’s usually recommended, but radiation is generally used only to manage a specific problem, such as pain, bleeding, an obstructed airway, or neurological changes from spinal cord or brain metastases. Mr. R did not have any of these problems. The indication for radiation, said his medical oncologist, Dr. Charles, was merely to go after the growing tumor in his chest, a problem neither patient nor doctor would have known about but for frequent surveillance scans to assess Mr. R’s response to chemotherapy. Systemic treatments for metastatic cancer, like chemotherapy, offer diminishing returns—the chance of seeing a durable response, or any response at all, declines with each additional cycle of treatment. Moreover, once patients stop benefiting from the first approach, the odds they’ll see improvement from other lines of treatment dramatically decrease.
In the 1990s, the term oligometastatic was introduced to describe a subcategory of patients with metastatic cancer who may respond more favorably to treatment. Institutional series and case reports focused attention on patients with oligometastases (from the Greek “few” or “scanty,” describing a limited number of metastatic sites) who, after receiving surgery or radiation on those sites, appeared to live longer compared with patients treated with chemotherapy alone. These reports suggested a different way of looking at metastatic cancer: its spread was not an all-or-nothing phenomenon with malignant seeds exploding simultaneously throughout the body; instead, one metastasis evolved the capacity to seed an additional one, or several, in sequence. By this hypothesis, it made sense to treat the first few visible sites and, by extension, to treat even the primary tumor site with curative intent—in effect, to approach a metastatic patient with the same intensity of treatment as if the patient didn’t have metastases. Still, no medical intervention is without risk: sometimes, as de Tournemire’s colleague warned, more aggressive treatment can end up doing more harm than good. I called Dr. Charles to ask whether this approach was what he intended—to treat a metastatic patient using curative radiation chemotherapy to the chest, with all its harms and risks, even though the chances for cure were almost zero.
“That’s right,” he said. Dr. Charles had earned a reputation—and gained a measure of fame—for continuing to treat lung cancer patients long after they had any hope of survival. He did not sit down with patients to inquire about their beliefs and values. He knew that they sought him out not for his bedside manner but for his willingness to do everything possible to extend their lives, even if that meant sending them to me for risky radiation treatment that carried uncertain benefits. When I tried to raise my concerns with Dr. Charles, he cut me off. “It’s not investigational!” he said.
Only recently has cancer become a problem that can be “beat.” For most of history, it was seen as mostly untreatable. In antiquity, Hippocrates and Galen wrote of an excess of “black bile” as a possible cause of cancer. The fluid resisted natural elimination through the pores, and in the soft tissues hardened into a tumor. (The fleshy, dark ulceration, and the strong necrotic odor of an untreated breast cancer appeared to support the black bile hypothesis.) Fourteenth-century causes of black bile itself included bear meat, heavy wine, worries, and sadness. For centuries, diagnosis brought despair, and physicians and patients seemed willing to try anything, from frankincense, human milk, donkey milk, river crabs, and honey to plant-derived ingredients such as calamine, soot, swallowwort, and stinging nettle. Medical papyri surviving from ancient Egypt recommended uterus cancer treatment using a stone from the shore, broken with water, left overnight in dew and then poured into the vagina. Another approach recommended fumigation of everything the patient associated with the smell of roasted meat. Medieval prescriptions recommended the application of animal droppings, such as from pigeons or goats.
An 1821 publication described cancer’s characteristic trait as “the obstinacy with which it resisted” treatment, noting that surgery “had been practiced in some of these instances, but … with perfect failure of cure.” The Boston Medical and Surgical Journal recommended in 1893 that “a patient with cancer not amenable to surgical operation should at once be put upon small, continued doses of morphia or opium and made permanently subject to the morphia habit, purposely induced.” In 1971, when the Army biological warfare site at Fort Detrick, Maryland, was converted to a cancer research center, President Nixon declared a “war on cancer.” Fifty years later, one researcher wrote in Nature Medicine, “it’s been a wash.”
The week before I met Mr. R, he was admitted to the hospital with signs of septic shock—fevers, chills, and low blood pressure. He spent two days on a mechanical ventilator. The care team’s first impression was pneumonia with sepsis, but after blood and urine cultures, scans, x-rays, and even a lumbar puncture, no infectious source could be found. Instead, the team determined that Mr. R was having an idiosyncratic reaction to one of the chemotherapy drugs he’d been on for more than a year. It was odd to have such a bad reaction, given that he had tolerated the drug well until then, but in cancer medicine, the risks of treatment can be unpredictable. A note on Mr. R’s ICU chart read, “Discussed with family poor prognosis—metastatic cancer, septic shock, multi-organ failure.”
But he survived.
When I met him for the first time in the radiation oncology clinic, he wore thick glasses, and he had loose skin under the eyes, sagging cheeks, patches of uneven skin tone, accentuated wrinkles: the signs of a smoker. He lived in the suburbs, had a daughter in a faraway city, was accompanied to appointments by his wife. Ten months earlier, he had visited his internist with pain in the tailbone. X-rays didn’t find anything, but a CAT scan and MRI showed what looked like cancer metastases in the sacrum, which prompted more scans, a bone biopsy, and a bronchoscopy to get an additional biopsy specimen from the lung tumor, for a diagnosis of metastatic lung cancer. “I quit smoking the day I found out,” he said.
None of that mattered now. In the exam room, I introduced myself, turned to the computer on the desk between us, and pulled up the PET scan he’d already seen. “This is why you’ve been sent to us,” I said, and pointed to the cancer in his right lung.
“That’s my understanding,” he said, and smiled. It was my first time to see it, his wonderful, easy, broad smile, disarming in its welcome.
“I’m not sure how Dr. Charles framed it,” I said. “Here’s what the scan shows, and here’s how he described it to me: you’ve been living with metastatic lung cancer for 10 months. The PET scan shows that the tumor in your chest—the primary site—appears to be growing, while the metastases in your liver, in your bones, are stable. Dr. Charles has asked us to consider giving radiation to your chest while he continues to give chemo. It’s the same approach we’d use if you did not have metastases—curative treatment.” I stopped to make sure he was still with me.
“I don’t mind telling you,” I said, “I have not done this before. It’s one thing to ask a patient with lung cancer limited to the chest to take on the risks of treatment—it’s a small chance, but at least it’s a chance, that he might be cured. For you, we already know it’s in your liver.” I scrolled the PET scan images to the liver and showed him the metastases. “Some clinicians are convinced that there’s a subgroup of metastatic patients out there who don’t fit the textbook, who may have a more favorable tumor biology, a part of the science not well understood. For these patients, maybe aggressive treatment can give them more time.”
“Dr. Charles explained all that,” Mr. R said. “He told me he’s hopeful this will work.”
I had mixed feelings going into my meeting with Mr. R, but sitting with him in the exam room, I reminded myself that as his doctor, I was not there to tell him what to do. My job was to help him, not only by managing his radiation treatment but also by being his advocate: I was there for guidance, as well as support, to see him through, one way or another. Never take hope from the patient, oncologists learn in training, and indeed, there are patients who outlive their predicted survival by years.
I have had several patients like this. One was a 75-year-old man with advanced lung cancer whose estimated survival was 17 months. Four years after treatment, he developed a single brain metastasis, which was treated by surgery followed by radiation. Two years later, he developed a recurrent metastasis at the same site in the brain, and radiation was used again. Eight years after the patient was first diagnosed, his chest and brain showed “clean” scans. “See you next year,” I told him. Another 75-year-old man, with terrible lung function, reliant on full-time supplemental oxygen running from a tank he carried that connected plastic tubes to his nostrils, was diagnosed with advanced, inoperable lung cancer. Fourteen radiation treatments through the course of a planned 37, he had already been hospitalized eight times for a life-threatening complication called a pneumothorax, or air trapped between the two linings of the lungs. In a somber discussion with him and his family, we said we’d decided to stop treatment, even knowing we’d reached only one-third of the intended dose. The risk of hurting him, we’d seen, was higher than the chance we could help him. Even with the full dose, the chance for cure was less than 20 percent; the chance of dying from a complication was two to four percent, a number that was starting to seem more likely. “I understand,” he said. Three years later, his scans remained “stable,” with the cancer in his chest controlled, and no sign of recurrence or metastases.
Recalling these patients and others, I feel transported to Carthage in the fifth century, to Saint Augustine’s story of Innocentia. “It’s a bad prognosis,” I can hear myself telling a patient newly diagnosed with advanced lung cancer, something I’ve said many times. “Even so, some patients are cured. I have met them. They return every year for follow-up. And until we know different, you’re going to be one of them.” The man several years out from treatment for brain metastases from lung cancer asked whether he had to continue to get scans, eight years after he was first treated.
“For what?” he said. “I’m 83 years old.”
It was a good question. “At this point,” I said, “it’s partly for me.” Never take the hope from the doctor, the aphorism should read.
Stories like these seem like modern miracles, but the explanation is likely more rudimentary. These patients, or their tumors, have some distinctive but as yet dimly understood nuance in molecular biology. Whether a scientist can break it down or not, positive stories—proofs—give hope to patients, their families, and their doctors. When I met Mr. R, less than a week after he’d almost died from the treatment he was receiving, I started to see the world like Dr. Charles: maybe we should keep pushing.
In the exam room with Mr. R, I asked if he had any questions.
“What did you say the difference was?” he asked, “How patients do with the radiation versus without it?”
“I didn’t say exactly,” I said. “Part of the answer is the nature of the evidence, or the science supporting one recommendation or another. One kind of evidence would be experience reported by an institution—a big cancer hospital, say. Another kind is a clinical trial. It turns out, a few trials have tried to answer your question; they’re called phase 2—preliminary trials that test efficacy among small numbers of patients. One trial found a big difference in survival for the patients treated with chest radiation, compared with the patients treated merely with chemotherapy—the ‘old’ way.”
“Researchers use the term hypothesis generating,” I continued. “Although the results are promising, the study wasn’t subjected to the more rigorous statistics required in higher-level scientific trials.” I waited a minute, hoped this would turn over in him. “I’m always hopeful,” I said. “Still, I think it’s fair to call it controversial.”
He took off his glasses and rubbed the bridge of his nose. “Wow,” he said. “It’s a lot to think about.” He shook his head, smiled again. “What are the risks?”
“It’s the same as if we were treating you for advanced lung cancer without metastases,” I said. “The risk of death from the treatment, or complications from the treatment, is small but real. That’s bad enough, but other risks can hurt patients so badly that they might prefer death. Stroke, sepsis, esophageal or airway perforation or stricture—those are some of the more serious risks. We could injure your lungs, such that you needed oxygen, or steroids.”
“For how long?” his wife asked.
“For the rest of his life,” I said, “if he didn’t die from it,” which had happened to more than one patient I’d treated.
Mr. R was quiet for a moment. “Can I think about it?”
“Of course,” I said. “We want you to think about it. Meanwhile, we need a brain MRI, to complete the workup. We don’t think anything’s wrong. It’s called restaging—if the cancer progressed in one place we can see, your chest, we want to rule out the chance it also progressed somewhere we haven’t looked recently, your brain.”
“Of course,” he said.
In cancer care, decisions are based on the experience of previously treated patients and the outcomes observed. Sometimes this experience includes a randomized trial, or better yet, a series of randomized trials, where all the statistics point to the same result—so-called level 1A evidence, considered the most robust scientific support. At the other end of the continuum is level 5 evidence, based on expert opinion without any supporting medical research. In the middle, a range of scientific support might back up a given treatment approach: case series research or reporting on a group of patients treated a similar way; cohort studies, looking at a population given similar treatment; or first principles, intuitive observations about how biology works.
The treatment recommended for Mr. R. was supported by cohort experiences, expert opinion, and most prominently, two small investigational randomized trials, published just a few months before he was diagnosed. One of the trials had 49 patients; the other, 99 patients. In one of the trials, almost five percent of the patients in the test arm died, possibly from the treatment itself. All the same, Mr. R might reasonably ask, What are the chances that the results from the phase 2 trials are correct?
The goal of a phase 2 trial is to establish safety and efficacy of a new treatment. If the results show promise, a phase 3 trial will compare the new treatment against the existing approach. In 2017, the Food and Drug Administration published 22 cases where phase 2 and phase 3 trials had divergent results. The paper looked at different medical situations, different kinds of patients, in different subspecialties—devices, vaccines, cancer medications, in patients treated by psychiatrists, cardiologists, pulmonologists, pediatricians—and found many examples of phase 2 results overturned by phase 3 trials. In another example, in the 1980s and ’90s, high-dose chemotherapy with bone marrow transplant gained broad acceptance as treatment for women with advanced or metastatic breast cancer. Though the risk of death from the treatment itself was reported as high as 20 percent, results from preliminary investigations seemed promising: trial patients appeared to have a substantial improvement in survival, compared with women treated using the old standard, chemotherapy. The new procedure was used for thousands of women, for more than 10 years—until phase 3 trials showed that women who received this treatment had a higher risk of death.
In small-cell lung cancer, a tumor with a slightly different biology from Mr. R’s non-small-cell lung cancer, a phase 2 trial that tested treatment directed to the primary tumor and multiple metastatic sites—the program proposed for Mr. R—closed early because of excess toxicity. The risk in following some unorthodox or undertested program was unknown: for Mr. R, perhaps the biology of metastases conferred some extra, unknown risk, or perhaps the year of chemotherapy given before chest radiation predicted that he would not be able to tolerate the treatment as well as a de novo patient—one who had yet to receive any treatment at all. Although those scenarios were unlikely, what if he did have a serious problem? There was no high-level evidence to rely on.
Not least, perhaps there were biological reasons that the subgroup of oligometastatic patients appeared to do well. A 1980 report from Sweden described 82 patients with lung metastases. Seventy had surgery to remove the metastases, 12 did not, and at five years, there was no statistically significant difference in survival. The authors hypothesized that the patients well enough to undergo the operation, and exhibiting fewer metastases, might have a tumor-host relationship with a favorable biology—these patients were going to live, or not live, whether the metastases were treated or not.
It remained unknown whether oligometastatic patients benefited from radiation administered to the primary tumor site or other sites. Possibly it was harmful.
Before Mr. R returned, he got a brain MRI. When the scan showed metastases, the members of the care team were surprised. But we shouldn’t have been: the chance of developing brain metastases from non-small-cell lung cancer is almost 50 percent. As with the cancer that had spread to other parts of his body, Mr. R was completely asymptomatic; he didn’t have headaches, weakness of one body part or another, altered gait, or any other sign something might be wrong. The metastasis was limited to one small area in the cerebellum—“not an eloquent region,” the neurosurgeon wrote in his note, recommending radiation. “Brain metastases might give us pause,” I told Mr. R and his wife. “Chest radiation was already a mixed recommendation—high risk, for a small chance we could help you. Some clinicians might say that brain metastases show the direction this cancer is going, which is out of control. Other clinicians would say the brain metastasis is small, asymptomatic, and easily treated—so the cancer is just a step or two away from being under control.”
Mr. R shook his head, as if to say no. Was he bewildered, frustrated, both, something else? “What do you think?” he asked.
“I’m not sure what to tell you,” I said. “I can understand both ways of looking at it.”
He paused, took a deep breath. “Ten months ago, I was told to make plans,” he said. “Now I feel fantastic. I played golf yesterday. What do I have to lose?”
The following week he started radiation, and the next day he was admitted to the hospital with septic shock and put on a ventilator for a day. This time, it was a usual suspect: pneumonia. The hospitalization and ICU stay were signs of how challenging treatment can be for lung cancer patients, for metastatic patients: they’re already compromised by the cancer, knocked down further still by the physiological stress of the treatment. Pneumonia hits them harder.
Almost a month later, when Mr. R returned to restart chest radiation, having almost died a second time, I asked him how he was feeling.
“Tired,” he said, and smiled. “Tired.” Despite his hospital admission, a near-death experience, the brain metastases, and his cancer’s progression, the plan remained intact: treat the chest, using a curative radiation dose together with the complicated and high-risk addition of chemotherapy, even though everybody knew there was no chance for cure.
Only a decade or so ago, patients with brain metastases would be advised to settle their affairs. The best survival, with treatment, was in the range of three to six months. But doctors encouraged Mr. R to undergo stereotactic radiosurgery, an approach supported by level 2 or 3 evidence. It comes with a cost. The financial component is breathtaking: about $80,000 for brain radiosurgery, which takes about half of one morning to complete. But there’s another cost: time. Leaving aside the time required to travel to the hospital and receive treatment, Mr. R also had to account for existential time: coming to terms with the possibility of dying sometime in the next several months. He needed a coach, I thought, an advocate—not another doctor recommending more treatment.
Yet if I were his coach, I’d be obligated to tell him that brain radiosurgery, while supported only by second-line evidence, rarely led to complications. Scans beget scans, and treatment begets treatment. People want to live. Each small additional step doesn’t seem like much, after the previous thousand small steps.
Mr. R struggled to complete chest radiation, and as I sat with him week after week in the exam room, the computer monitor between us, I felt the strange distance of the radiation oncologist. Here we were, deciding on treatments that would be challenging for him—physically, mentally, emotionally. Risks and complications didn’t unfold for me, sitting at my desk, as for a surgeon hovering over a patient, making physical medical choices in the moment. Nevertheless, radiation could hurt Mr. R just as badly as an operation.
As he got closer to the end of the course, Mr. R picked up a kind of rhythm. I saw in him what I’d seen in other patients struggling through combined radiation chemotherapy for lung cancer: he may have been surprised, at first, at how tough it was to get through, but eventually he began to see he was going to make it. The week he finished, I reviewed with him the follow-up plan: in three months, he’d get a CAT scan of his chest, an MRI of his brain, before a recommendation for the next step. “Any questions?”
He looked to the ceiling for a moment, then returned his gaze to me. “It’s hard to know what to ask,” he said. “I feel like I’ve been through so much. Anything I should be doing or not doing?”
“Avoid salted meats,” I wanted to say. “Omit sadness, worries, and bear meat.” Instead, I said, “Right,” and nodded—in acknowledgment, I suppose, that nobody knew.
We want to believe that this time, things are different. Stories of patients who have beaten the odds, apparent miracles, hold out hope for the rest of us. With his daughter dying of breast cancer, de Tournemire obtained a thread from the belt of Peter of Luxembourg and asked his wife to rub it onto the afflicted breast morning and evening. In time, the tumor’s hardness started to give way, and the breast was no longer corroded by the fissure—findings, he wrote, “contra naturam cancri,” or against the nature of cancer. He concluded that with Peter’s intercession, a miracle had occurred.
Mr. R died two months after completing chest radiation with chemotherapy. Just a week before, he participated as a patient representative at a lung cancer conference, in a discussion called “Your Cancer Journey.” When he died, he was not in a hospital, not on a ventilator, and not in any pain. He merely fell asleep one night and didn’t wake up. It’s one way—a common one—that patients die from liver or brain metastases. He had both.
In a strange sense, his illness remained apparent only on scans, almost invisible to me—arguably, the doctor giving him the riskiest, least scientifically validated treatment. I had read about his hospitalizations and his stay in the ICU but found out that he had died only after he failed to appear at a follow-up appointment. Did chest radiation “expedite” his death, as Hippocrates suggested? Although that was unlikely, this very question had not been put to the test of a phase 3 trial. The real answer was that nobody knew. Mr. R would not be added to my list of miracles.
I thought back on our interactions. Mr. R had only months to live, but nobody told him, “You don’t have to do any of this. Go for a walk, look at the river, gaze into the upper reaches of the trees, see if you can hear the butterfly’s wings. Or go to a bar, be with friends and strangers on a Sunday afternoon for your favorite team’s baseball game.” Instead, we talked about statistics from clinical trials that might be obsolete in 20 years, if not sooner. I wondered whether our communication was limited by trying to use mere words to translate the weight of illness and death. It’s always the challenge: words, even long ones such as oligometastases, reach for something possibly untranslatable.
A year later, I saw a patient whom I had treated two years previously. He was dying from lung cancer—the treatment had proved too toxic, and he was too sick to take any more. “I wish it were different,” I said. “Radiation to asymptomatic brain metastases detected on a scan will not make you live longer, or feel better.”
“But I’m a fighter,” he said. His face was turned away from me.
“The brain metastases are not the problem. You can’t live with it in your liver.”
He turned back to me, then looked at me for a long time.
I’d already talked with his sister, who wanted to bring him home. “It’s great she’s there for you,” I said. A few minutes later, as I walked out, I could hear him crying.
The same day, I saw a different follow-up patient, the man with advanced lung cancer who remained alive years after his radiation treatment ended early because of repeated pneumothoraces. “How do you feel?” I said, after one more yearly surveillance scan showed “no concerning findings.” He required continuous oxygen because of bad lung disease—circumstances no different from before he received a subtherapeutic dose of radiation.
“Great,” he said, and stood up to give me a fist-bump.
“You look it,” I said, meeting his fist with mine. “See you next year.”
Where was the truth?
My mind traveled back to the first time I met Mr. R, when I’d tried to break down statistics, describe different phase trials, his scan results, and radiation toxicity. He just wanted something he could grasp. “One more time,” he said. “What did you say the difference was in patients who received radiation versus patients who didn’t?”
“Survival was 41 months for the treated patients, versus 17 months for the patients who received usual care.”
“Forty-one months versus 17 months?” Again he flashed that trademark smile—his way of keeping things light, of helping all of us get through difficult conversations.
“It’s one trial,” I said. After a moment of quiet between us, I thought of something. “I can imagine that sitting where you’re sitting, it’s tough to appreciate the risk, if you just reduce it to those numbers.”
“Forty-one months?” he said again. “I’m starting to see why people do this.”
Permission required for reprinting, reproducing, or other uses.