For my week-long blog for Powell's, I'll be attempting something novel. Every day, I will pick a single recent case from my files that highlights a major new discovery in cancer science or cancer medicine. As always, the names and dates will be scrambled to protect the identities of patients. And I will provide a link to the major medical or scientific study that lies behind the case. My goal here is to demonstrate how every case in medicine is a story in its own right — and grand ideas are also concealed within each story.
÷ ÷ ÷
In the summer of 2009, a friend of mine awoke one night as if from a nightmare — flushed, hot, and short of breath, with a powerful fluttering in his chest. He thought he was having a heart attack. He was 43, an acclaimed writer about cities and urbanism, and in otherwise excellent health. That evening, he had made himself a particularly high-voltage version of the chili that he loved, spiced with the hottest Indian peppers in the world (in large part, our friendship is based on the mutual admiration of chili peppers).
He went to see his doctor the next week, and was tested with an EKG, which was normal. "It's probably heartburn," she said, reassuringly. But then, almost as an afterthought, she thought she would scan his chest. And there it was in the scan: a spiculated, two-inch shadow in his chest that, biopsied, turned out to be the earliest stages of lung cancer. A few weeks later, a surgeon removed the mass, carefully excising it out of the lung with wide margins around.
My friend has likely been cured — but by a strange confluence of luck and spices. Cases like his are fleetingly rare. Lung tumors are almost always only detected in advanced stages, and even with chemotherapy, surgery, and radiation, cure rates for advanced lung cancer have remained unchanged for decades. My friend's cancer was obviously detected by chance, but, in the spirit of a true gastronome, he attributes its detection to his beloved peppers. "Add yet another virtue to them," he said recently, "they literally saved my life."
Lung cancer, oncologists know, is an inherently insidious disease. Locked away in the airy reticulations of the lung, most tumors become symptomatic only in late stages — often after they've transambulated to the bones, or the brain. And yet, caught in its earliest stages, lung cancer is curable. (The numbers tell the full story: patients with stage I lung cancer have a 75% survival at 5 years; patients with stage IV cancer have a 5% survival rate.) The stage-specificity of cure raises an important question: What if we could screen the entire asymptomatic population of men and women and thus catch lung cancer in its earliest stages?
By the mid-1980s, several trials had shown that mammograms, which can detect early breast cancer, decreased the mortality from breast cancer. Extending that logic to lung cancer made obvious sense. Trial upon trial was thus launched to use X-rays to detect early lung cancer. But every trial was negative. For unknown reasons, the strategy that had worked in breast cancer — detecting early tumors using radiography — simply failed to work in lung cancer.
Frustrated, researchers added an important variation to this study: rather than screening the population in general, they thought they would screen men and women at highest risk: long-term smokers. But even narrowing the focus to the highest risk group didn't succeed. The technology was too crude to work.
Last month, at the end of nearly three decades of frustration, the National Cancer Institute announced a strikingly positive trial. In the so-called National Lung Screening Trial (NLST), more than 50,000 high risk smokers (men and women with more than 30 pack-years of smoking) were screened with CT scans. The first patients were randomized in 2002. The initial plan was to continue the study over nearly a decade.
But the trial observers had to stop the trial before it was over. Clinical trials often get stopped for two reasons: either the drug or method being tested is found to be unexpectedly toxic, or the drug or method is found to be unexpectedly effective. In the NLST (fortunately) it was the latter. By late October 2010, 354 lung cancer deaths had occurred among those who had been screened with CT scan versus 442 deaths among those who were not screened by CT scanning — a 20% reduction in lung cancer mortality. Researchers could no longer justify continuing the trial. The CT-unscreened group would need to be screened as well.
What do we do with this information? Should we all start queuing for CT scans in the hopes of detecting pre-symptomatic tumors (like my friend's)? Obviously not. This screening program was launched on a very high-risk group of smokers and cannot be legitimately used for nonsmokers. But generalizing such a screening plan nationwide even among smokers would carry formidable costs — overburdening an already overburdened health care system.
Should smokers bear the costs of screening? Obviously not, again: we don't expect men and women with high blood pressure to pay for their coronary surgery. Insurance, by its very nature, works on the distribution of risk. But what if that risk is increased by a particular behavior — smoking, in this case — and the technology to reduce the risk turns out to be laborious, time-consuming, and expensive? In a sense, the NLST has pushed the once-moribund field of lung cancer screening into astonishingly thorny territory. Its moral is illuminating. In medicine, it might be worthwhile to be careful what one wishes for. If the wish comes true — in this case, by reducing mortality with CT scanning — the problems raised might be even more acute than the problems solved.