But the last part of the answer lies, surely, in how we imagine cancer and screening. We are a visual species. Seeing is believing, and to see cancer in its early, incipient form, we believe, must be the best way to prevent it. As the writer Malcolm Gladwell once described it, “This is a textbook example of how the battle against cancer is supposed to work. Use a powerful camera. Take a detailed picture. Spot the tumor as early as possible. Treat it immediately and aggressively. . . . The danger posed by a tumor is represented visually. Large is bad; small is better.”

  But powerful as the camera might be, cancer confounds this simple rule. Since metastasis is what kills patients with breast cancer, it is, of course, generally true that the ability to detect and remove premetastatic tumors saves women’s lives. But it is also true that just because a tumor is small does not mean that it is premetastatic. Even relatively small tumors barely detectable by mammography can carry genetic programs that make them vastly more likely to metastasize early. Conversely, large tumors may inherently be genetically benign—unlikely to invade and metastasize. Size matters, in other words—but only to a point. The difference in the behavior of tumors is not just a consequence of quantitative growth, but of qualitative growth.

  A static picture cannot capture this qualitative growth. Seeing a “small” tumor and extracting it from the body does not guarantee our freedom from cancer—a fact that we still struggle to believe. In the end, a mammogram or a Pap smear is a portrait of cancer in its infancy. Like any portrait, it is drawn in the hopes that it might capture something essential about the subject—its psyche, its inner being, its future, its behavior. “All photographs are accurate,” the artist Richard Avedon liked to say, “[but] none of them is the truth.”

  But if the “truth” of every cancer is imprinted in its behavior, then how might one capture this mysterious quality? How could scientists make that crucial transition between simply visualizing cancer and knowing its malignant potential, its vulnerabilities, its patterns of spread—its future?

  By the late 1980s, the entire discipline of cancer prevention appeared to have stalled at this critical juncture. The missing element in the puzzle was a deeper understanding of carcinogenesis—a mechanistic understanding that would explain the means by which normal cells become cancer cells. Chronic inflammation with hepatitis B virus and H. pylori initiated the march of carcinogenesis, but by what route? The Ames test proved that mutagenicity was linked to carcinogenicity, but mutations in which genes, and by what mechanism?

  And if such mutations were known, could they be used to launch more intelligent efforts to prevent cancer? Instead of running larger trials of mammography, for instance, could one run smarter trials of mammography—by risk-stratifying women (identifying those with predisposing mutations for breast cancer) such that high-risk women received higher levels of surveillance? Would that strategy, coupled with better technology, capture the identity of cancer more accurately than a simple, static portrait?

  Cancer therapeutics, too, had seemingly arrived at the same bottleneck. Huggins and Walpole had shown that knowing the inner machinery of the cancer cell could reveal unique vulnerabilities. But the discovery had to come from the bottom up—from the cancer cell to its therapy. “As the decade ended,” Bruce Chabner, former director of the NCI’s Division of Cancer Treatment, recalled, “it was as if the whole discipline of oncology, both prevention and cure, had bumped up against a fundamental limitation of knowledge. We were trying to combat cancer without understanding the cancer cell, which was like launching rockets without understanding the internal combustion engine.”

  But others disagreed. With screening tests still faltering, with carcinogens still at large, and with the mechanistic understanding of cancer in its infancy, the impatience to deploy a large-scale therapeutic attack on cancer grew to its bristling tipping point. A chemotherapeutic poison was a poison was a poison, and one did not need to understand a cancer cell to poison it. So, just as a generation of radical surgeons had once shuttered the blinds around itself and pushed the discipline to its terrifying limits, so, too, did a generation of radical chemotherapists. If every dividing cell in the body needed to be obliterated to rid it of cancer, then so be it. It was a conviction that would draw oncology into its darkest hour.

  * In addition to mammography, women also received a breast exam, typically performed by a surgeon.

  STAMP

  Then did I beat them as small as the dust of the earth, I did stamp them as the mire of the street, and did spread them abroad.

  —Samuel 22:43

  Cancer therapy is like beating the dog with a stick to get rid of his fleas.

  —Anna Deavere Smith, Let Me Down Easy

  February was my cruelest month. The second month of 2004 arrived with a salvo of deaths and relapses, each marked with the astonishing, punctuated clarity of a gunshot in winter. Steve Harmon, thirty-six, had esophageal cancer growing at the inlet of his stomach. For six months, he had soldiered through chemotherapy as if caught in a mythical punishment cycle devised by the Greeks. He was debilitated by perhaps the severest forms of nausea that I had ever encountered in a patient, but he had to keep eating to avoid losing weight. As the tumor whittled him down week by week, he became fixated, absurdly, on the measurement of his weight down to a fraction of an ounce, as if gripped by the fear that he might vanish altogether by reaching zero.

  Meanwhile, a growing retinue of family members accompanied him to his clinic visits: three children who came with games and books and watched, unbearably, as their father shook with chills one morning; a brother who hovered suspiciously, then accusingly, as we shuffled and reshuffled medicines to keep Steve from throwing up; a wife who bravely shepherded the entire retinue through the whole affair as if it were a family trip gone horribly wrong.

  One morning, finding Steve alone on one of the reclining chairs of the infusion room, I asked him whether he would rather have the chemotherapy alone, in a private room. Was it, perhaps, too much for his family—for his children?

  He looked away with a flicker of irritation. “I know what the statistics are.” His voice was strained, as if tightening against a harness. “Left to myself, I would not even try. I’m doing this because of the kids.”

  “If a man die,” William Carlos Williams once wrote, “it is because death / has first possessed his imagination.” Death possessed the imagination of my patients that month, and my task was to repossess imagination from death. It is a task almost impossibly difficult to describe, an operation far more delicate and complex than the administration of a medicine or the performance of surgery. It was easy to repossess imagination with false promises; much harder to do so with nuanced truths. It demanded an act of exquisite measuring and remeasuring, filling and unfilling a psychological respirator with oxygen. Too much “repossession” and imagination might bloat into delusion. Too little and it might asphyxiate hope altogether.

  In his poignant memoir of his mother’s illness, Susan Sontag’s son, David Rieff, describes a meeting between Sontag and a prominent doctor in New York. Sontag, having survived uterine and breast cancer, had been diagnosed with myelodysplasia, a precancerous disease that often sours into full-blown leukemia. (Sontag’s myelodysplasia was caused by the high-dose chemotherapy that she had received for the other cancers.) The doctor—Rieff calls him Dr. A.—was totally pessimistic. There was no hope, he told her flatly. And not just that; there was nothing to do but wait for cancer to explode out of the bone marrow. All options were closed. His word—the Word—was final, immutable, static. “Like so many doctors,” Rieff recalls, “he spoke to us as if we were children but without the care that a sensible adult takes in choosing what words to use with a child.”

  The sheer inflexibility of that approach and the arrogance of its finality was a nearly fatal blow for Sontag. Hopelessness became breathlessness, especially for a woman who wanted to live twice as energetically, to breathe the world in twice as fast as anyone else—for whom stillness w
as mortality. It took months before Sontag found another doctor whose attitude was vastly more measured and who was willing to negotiate with her psyche. Dr. A. was right, of course, in the formal, statistical sense. A moody, saturnine leukemia eventually volcanoed out of Sontag’s marrow, and, yes, there were few medical options. But Sontag’s new physician also told her precisely the same information, without ever choking off the possibility of a miraculous remission. He moved her in succession from standard drugs to experimental drugs to palliative drugs. It was all masterfully done, a graded movement toward reconciliation with death, but a movement nonetheless—statistics without stasis.

  Of all the clinicians I met during my fellowship, the master of this approach was Thomas Lynch, a lung cancer doctor, whom I often accompanied to clinic. Clinics with Lynch, a youthful-looking man with a startling shock of gray hair, were an exercise in medical nuance. One morning, for instance, a sixty-six-year-old woman, Kate Fitz, came to the clinic having just recovered from surgery for a large lung mass, which had turned out to be cancerous. Sitting alone in the room, awaiting news of her next steps, she looked nearly catatonic with fear.

  I was about to enter the room when Lynch caught me by the shoulder and pulled me into the side room. He had looked through her scan and her reports. Everything about the excised tumor suggested a high risk of recurrence. But more important, he had seen Fitz folded over in fear in the waiting room. Right now, he said, she needed something else. “Resuscitation,” he called it cryptically as he strode into her room.

  I watched him resuscitate. He emphasized process over outcome and transmitted astonishing amounts of information with a touch so slight that you might not even feel it. He told Fitz about the tumor, the good news about the surgery, asked about her family, then spoke about his own. He spoke about his child who was complaining about her long days at school. Did Fitz have a grandchild? he inquired. Did a daughter or a son live close by? And then, as I watched, he began to insert numbers here and there with a light-handedness that was a marvel to observe.

  “You might read somewhere that for your particular form of cancer, there is a high chance of local recurrence or metastasis,” he said. “Perhaps even fifty or sixty percent.”

  She nodded, tensing up.

  “Well, there are ways that we will tend to it when that happens.”

  I noted that he had said “when,” not “if.” The numbers told a statistical truth, but the sentence implied nuance. “We will tend to it,” he said, not “we will obliterate it.” Care, not cure. The conversation ran for nearly an hour. In his hands, information was something live and molten, ready to freeze into a hard shape at any moment, something crystalline yet negotiable; he nudged and shaped it like glass in the hands of a glassblower.

  An anxious woman with stage III breast cancer needs her imagination to be repossessed before she will accept chemotherapy that will likely extend her life. A seventy-six-year-old man attempting another round of aggressive experimental chemotherapy for a fatal, drug-resistant leukemia needs his imagination to be reconciled to the reality that his disease cannot be treated. Ars longa, vita brevis. The art of medicine is long, Hippocrates tells us, “and life is short; opportunity fleeting; the experiment perilous; judgment flawed.”

  For cancer therapeutics, the mid and late 1980s were extraordinarily cruel years, mixing promise with disappointment, and resilience with despair. As physician-writer Abraham Verghese wrote, “To say this was a time of unreal and unparalleled confidence, bordering on conceit, in the Western medical world is to understate things. . . . When the outcome of treatment was not good, it was because the host was aged, the protoplasm frail, or the patient had presented too late—never because medical science was impotent.

  “There seemed to be little that medicine could not do. . . . Surgeons, like Tom Starzl . . . were embarking on twelve- to fourteen-hour ‘cluster operations’ where liver, pancreas, duodenum and jejunum were removed en bloc from a donor and transplanted into a patient whose belly, previously riddled with cancer, had now been eviscerated, scooped clean in preparation for this organ bouquet.

  “Starzl was an icon for that period in medicine, the pre-AIDS days, the frontier days of every-other-night call.”

  Yet even the patients eviscerated and reimplanted with these “organ bouquets” did not make it: they survived the operation, but not the disease.

  The chemotherapeutic equivalent of that surgical assault—of eviscerating the body and replacing it with an implant—was a procedure known as autologous bone marrow transplant, or ABMT, which roared into national and international prominence in the mid-1980s. At its core, ABMT was based on an audacious conjecture. Ever since high-dose, multidrug regimens had succeeded in curing acute leukemia and Hodgkin’s disease in the 1960s, chemotherapists had wondered whether solid tumors, such as breast or lung cancer, had remained recalcitrant to chemotherapeutic obliteration simply because the bludgeon of drugs used was not powerful enough. What if, some had fantasized, one could tip the human body even closer to the brink of death with even higher doses of cytotoxic drugs? Might it be dragged back from that near-lethal brink, leaving cancer behind? What if one could double, or even quadruple, the dosage of drugs?

  The dose limit of a drug is set by its toxicity to normal cells. For most chemotherapy drugs, that dose limit rested principally on a single organ—the bone marrow, whose whirring cellular mill, as Farber had found, was so exquisitely sensitive to most drugs that patients administered drugs to kill cancer were left with no normal blood-forming cells. For a while, then, it was the bone marrow’s sensitivity to cytotoxic drugs that had defined the outer horizon of chemotherapeutic dosage. The bone marrow represented the frontier of toxicity, an unbreachable barrier that limited the capacity to deliver obliterative chemotherapy—the “red ceiling” as some oncologists called it.

  But by the late 1960s, even that ceiling had seemed to lift. In Seattle, one of Farber’s early protégés, E. Donnall Thomas, had shown that bone marrow, much like a kidney or liver, could be harvested from one patient and transplanted back—either into the same patient (called autologous transplantation) or into another patient (termed allogeneic transplantation).

  Allogeneic transplantation (i.e., transplanting foreign marrow into a patient) was temperamental—tricky, mercurial, often deadly. But in some cancers, particularly leukemias, it was potentially curative. One could, for instance, obliterate a marrow riddled with leukemia using high-dose chemo and replace it with fresh, clean marrow from another patient. Once the new marrow had engrafted, the recipient ran the risk of that foreign marrow turning and attacking his or her own body as well as any residual leukemia left in the marrow, a deadly complication termed graft-versus-host disease or GVHD. But in some patients, that trifecta of assaults—obliterative chemotherapy, marrow replacement, and the attack on the tumor by foreign cells—could be fashioned into an exquisitely potent therapeutic weapon against cancer. The procedure carried severe risks. In Thomas’s initial trial at Seattle, only twelve out of a hundred patients had survived. But by the early 1980s, doctors were using the procedure for refractory leukemias, multiple myeloma, and myelodysplastic syndrome—diseases inherently resistant to chemotherapy. Success was limited, but at least some patients were eventually cured.

  Autologous bone marrow transplantation was, if conceivable, the lighter fraternal twin of allogeneic transplantation. Here, the patient’s own marrow was harvested, frozen, and transplanted back into his or her body. No donor was needed. The principal purpose was not to replace diseased marrow (using a foreign marrow) but to maximize chemotherapeutic dosage. A patient’s own marrow, containing blood-forming cells, was harvested and frozen. Then blisteringly high levels of drugs were administered to kill cancer. The frozen marrow was thawed and implanted. Since the frozen marrow cells were spared the brunt of chemotherapy, transplantation allowed doctors, theoretically at least, to push doses of chemo to their ultimate end.

  For advocates of megadose chemotherapy, ABMT
breached a final and crucial roadblock. It was now possible to give five- or even tenfold the typical doses of drugs, in poisonous cocktails and combinations once considered incompatible with survival. Among the first and most fervent proponents of this strategy was Tom Frei—cautious, levelheaded Frei, who had moved from Houston to Boston as the director of Farber’s institute. By the early 1980s, Frei had convinced himself that a megadose combination regimen, bolstered by marrow transplantation, was the only conceivable solution in cancer therapy.

  To test this theory, Frei hoped to launch one of the most ambitious trials in the history of chemotherapy. With his ear for catchy acronyms, Frei christened the protocol the Solid Tumor Autologous Marrow Program—or STAMP. Crystallized in that name was the storm and rage of cancer medicine; if brute force was needed, then brute force would be summoned. With searing doses of cytotoxic drugs, STAMP would trample its way over cancer. “We have a cure for breast cancer,” Frei told one of his colleagues in the summer of 1982. Uncharacteristically, he had already let his optimism fly to the far edge of brinkmanship. The first patient had not even been enrolled on trial.