Imagine cancer, 2040. A 45-year-old woman who has never smoked develops lung cancer, which today kills more people than any other kind. (By then, of course, cigarettes will have gone the way of the buggy whip, and lung cancer rates will be cut in half.) She undergoes outpatient surgery, and her doctors quickly scrutinize the tumor's genes and feed the data, along with other information from her electronic medical record, into a desktop computer that crunches out a treatment plan all but certain to work. At subsequent checkups, her blood is tested for the earliest hint of a tumor recurrence—though were such news to come, it would bring no doom and gloom. Her doctor would simply analyze a few of the cells that even the tiniest tumors shed and prescribe a suitable next round of therapy. Throughout, her busy life is barely interrupted, and her hair stays wonderfully intact.
Sound like pure fantasy? It isn't.
Yes, cancer as we know it remains a fearsome foe, killing well over half a million Americans each year despite the well-funded, 40-year-old war against it. It still erupts in healthy tissue out of the blue, and—programmed to grow perpetually—mercilessly invades surrounding tissue and spreads to distant sites. Surgery and radiation and chemotherapy can stymie its growth and, if still early, cure it altogether. But they can ravage healthy tissue, too, and all too often the tumor is deadly by the time it appears. Advanced disease brings discouraging failures as often as not. Most patients facing invasive cancer get vague guesses as answers to questions like "Will I respond to this treatment?" "Will my tumor come back?"
Epiphany. But thanks to an epiphany in medicine about the very nature of cancer, most oncologists now believe in a not-too-distant future where cancer will be curable or a manageable chronic disease. What has changed radically in recent years is the view that cancers of different types—breast, lung, colon—are relatively homogeneous based on how they look under the microscope and how far they have burrowed into the body. In fact, even malignancies that seem identical are highly complex and varied where it counts—in their genetic makeup. This reality dictates an entirely new approach to diagnosis and treatment: "personalized medicine," which looks to a patient's unique genetic differences to better understand his or her illness and tailor care.
Stephen Jay Gould, the late Harvard biologist, fathomed the implications of biological variation in 1985, ahead of his time, when he wrote "The Median Isn't the Message," an essay pointing out that cancer survival statistics tend to obscure a wide range of patient outcomes. Gould, diagnosed with a deadly cancer at age 40, faced a gloomy median survival time of just eight months. But he decided that optimism was justified when he realized that "the median" meant that half the people with his tumor survived longer. Indeed, he lived happily for 20 more years. As a scientist, Gould surely would have loved to have had access to technology that could have given him a better idea of what lay ahead.
That powerful technology, which so rapidly and successfully mapped out all 20,000 genes of the human genome, is now being aimed at decoding the genetic underpinnings of cancer. The sheer volume and variability of mutations and resulting interconnected biological processes contributing to malignancies are mind-boggling, but the Cancer Genome Atlas project, a program underway at the National Institutes of Health since 2006, is a monumental step toward order. Its goal is to map the full genome—the comprehensive DNA instruction book—of virtually all human cancers.
What does this really mean for individual patients, you might ask? A lot. A reference manual that catalogs the diversity of the genetic mix underlying individual cancers sets the stage for flagging a tumor's characteristics and behavior—witness our 2040 lung cancer patient—and offering personalized treatments. The mighty Atlas has already shown impressive results. In September, researchers reported the project's first findings, the gene profile of the most common and deadliest form of brain cancer, glioblastoma. Not only did researchers catalog the major known DNA blunders in this disease, but they also provided actionable insights into why gliomas are so often resistant to conventional chemotherapy. And Atlas uncovered three significant and previously unreported mutations, common to most of the more than 200 tumors analyzed, that point to new targets for treatment.
Comprehensive genomic information of this kind will also be critical to evaluating new therapies. And here we have to humbly admit a gaping deficit: The gold standard of today, the randomized controlled clinical trial, assumes that randomly chosen test subjects have identical tumors, based on the old thinking. That just doesn't work for cancer anymore. If in fact those tumors are markedly different from one another, little wonder that—as Gould reminds us—a drug that extends life a mere two months on average fails many people but gives precious years of life to some. Better answers would come from smaller patient studies comparing genetically matched cancers. Such trials would be faster, cheaper, and vastly improved for all patients—both for those who should, and should not, partake in the new therapy.