Soon, precision medicine could help you avoid expensive, risky procedures that you may not need.
Each day in the clinic, urologist Peter Carroll can count on seeing three or four patients who don’t belong in his office.
Carroll, the chair of the department of urology at the University of California, San Francisco, is an expert on prostate cancer, and many of the men who come to see him truly need his help. Others do not: They’d do fine on their own, and the treatments he has to offer, such as surgery and radiation, will do them more harm than good.
Even though prostate cancer is the second-most deadly cancer for men, about 40 percent of patients who get the diagnosis have slow-growing cases that will never threaten their lives. It’s those patients Carroll would like to stop meeting. They come to his office after a getting a blood test for prostate-specific antigen (PSA) that comes back high. It’s a notoriously unreliable index of cancer — some men with high PSA do in fact have dangerous cancer, but many simply have enlarged prostates, or tumors that will never grow big enough to do damage.
The problem is in telling which is which. Today’s medical system is built on a one-size-fits-all platform. Right now, there is no simple, accurate way to separate out the men who should get only minimal treatment from those who really need a cancer doctor’s help.
The good news is that a savior may be on the horizon: precision medicine, or the idea of targeting medical care based on individual physiology and genetics. From the beginning, it was intended to do two things: show doctors exactly which treatments will be best for which individual patients, and also reveal which people need little or no treatment.
The individualized-treatment side of precision medicine is already credited with improving the outlook for some people with otherwise deadly conditions such as cancer and cystic fibrosis. It’s also under attack for promising too much — hyping the potential and overstating the possible cost savings.
But the other promise of this approach might soon deliver on some of its early promises. Precision medicine doesn’t always mean more. In the near future, it might also be able to keep the doctor away.
80 of every 100 men diagnosed with prostate cancer will undergo radiation or have their prostate removed. Only 1.3 will be saved from a deadly cancer.
Some cancer patients are already avoiding onerous chemotherapy or surgeries that won’t help them. Soon, everyone might benefit: Fewer mammograms, fewer false-positive tests that lead to expensive follow-up biopsies or procedures, fewer cycles of trial and error to find effective medications for things like inflammatory bowel disease.
“The challenge now is to really have people do more asking of this question,” says Gregory Gibson, a genomics researcher at Georgia Tech. “Can we use genomics to reduce medication use?”
Less is more
Nobody should be treated for prostate cancer if he doesn’t really need it. But today, that’s what medicine does. According to the American Academy of Family Physicians, about 80 of every 100 men diagnosed with prostate cancer will undergo radiation or have their prostate removed, which can cost anywhere from about $20,000 to about $50,000. Fifty of the men treated will develop sexual dysfunction and 15 will become incontinent due to treatment, but only 1.3 will be saved from a cancer that would have otherwise killed them.
Prostate cancer isn’t the only area where current best practices result in too many tests and too much treatment. A 2017 survey of physicians found that an estimated 20 percent of medical care is unnecessary. That’s more than one in five pills, a quarter of medical tests and more than 10 percent of actual procedures. The Institute of Medicine estimates that the cost of all this unneeded care is not just measured in hassle and harm for patients, but in real dollars: 210 billion of them per year.
In oncology, the “treat everyone the same” mentality has caused real harm. Prostate cancer is an egregious example, but breast cancer is similar. Mammograms save lives, but at a cost. A woman who starts yearly mammograms at 40 has a 61 percent chance of getting a false-positive result before she’s 50, according to the U.S. Preventative Services Task Force Recommendation on mammograms. (A false positive is a result that suggests cancer, but on further testing turns out to be wrong.) The same report found that between 10.7 percent and 19 percent of women diagnosed with breast cancer have cases that would never have threatened their lives, and yet they’re likely to undergo treatments ranging from radiation and chemotherapy to mastectomy.
Beyond cancer, even widely used treatments come with a heaping side of waste. Medicine relies on standard cut-offs, like blood pressure less than 140/90 or total cholesterol under 200. Fall on the wrong side of the numbers, and you’re likely looking at a new prescription, more testing, or some other medical intervention.
The cholesterol-lowering drug Crestor is proven to reduce strokes, heart attacks, bypass surgery, and death. But for every person who takes Crestor and avoids one of those fates, 19 see no benefit, according to a 2009 study. In 2015, Nicholas Schork, the director of human biology at the J. Craig Venter Institute, gathered up the numbers on the top 10 blockbuster drugs in the U.S. For every person helped by drugs like Abilify (for schizophrenia) and Nexium (for heartburn), between three and 24 people might as well have skipped the pharmacy. These statistics can’t differentiate between drugs that help only a small subgroup of patients, leaving the rest with no benefit at all, and those that only confer a tiny amount of benefit for each person who takes them. Either way, their use outpaces their utility.
Already, some tools do help separate out those who definitely need treatment. Factors like prostate volume, patient age, and Gleason score — a scoring system for prostate tumors — can provide a pretty solid prediction of how dangerous a given prostate cancer is likely to be, Carroll says. The predictions, in turn, can guide treatment decisions. But that information isn’t always routinely used.
And such predictions could become far more accurate. Today’s “medium” and “high” risk groups of prostate cancer patients include cancers that grow at vastly different rates. Better genomic screening could better stratify those groups, Carroll says. So could advanced imaging, like multiparametric magnetic resonance imaging, a new type of MRI that provides far more detailed pictures of the prostate than ultrasound, the usual method. Doctors can inject contrast dyes into the tissue and track it on multiparametric MRI. Because cancerous tissues absorb the dyes faster than non-cancerous tissues, this method can help confirm or rule out whether a tumor is truly cancerous, according to a 2015 review of the technique.
Improved diagnostic methods could provide more certainty about how individuals should be treated. For many patients in the lowest-risk groups, precision medicine might mean never diagnosing a slow-growing cancer at all, Carroll says. Eventually, doctors might be able to ditch PSA screenings altogether in favor of urine screens or tests that look for DNA fragments in the blood that only flag genetic markers unique to dangerous cancers.
Breast cancer researchers also see a less-medicated future on the horizon. Family history, demographics, and genetics all intertwine to determine how likely it is that a given woman will get breast cancer, but right now, no one really knows how to make sense of all of these factors.
A test of less testing
A study launched in 2016 by UCSF oncologist Laura Esserman is investigating whether we can start changing the way doctors screen for breast cancer. The idea, Esserman has said, is to apply the idea of precision medicine, which has mostly focused on treatment, to screening and diagnosis. Some women in the study will get the usual annual mammogram. Others will get screening based on their personal risk, which often means less-frequent mammograms.
The study will track how many Stage 2 (growing, but not yet metastatic) cancers get diagnosed in each group. If risk-based assessment catches as many of these potentially dangerous cancers as standard assessment, the findings could change the way screening recommendations are made for everyone. The hope is that many women may need to get mammograms only once every two years. Some may even be able to put mammograms off altogether until their 50s.
Precision medicine could reduce uncertainty, with life-changing consequences. Are your risk factors sensitive to diet and exercise? Great — you can avoid popping pills.
Other fields are also experimenting with this approach. Precision medicine is a hot topic in cardiovascular medicine, says North Carolina cardiologist Hadley Wilson, secretary of the American College of Cardiology Board of Trustees. Already there are genomic and proteomic tests that analyze genes and proteins to predict whether a given person will respond to particular blood thinners, anti-platelet drugs, and anti-arrhythmia drugs that prevent stroke and heart disease, Wilson says. For example, a gene called CYP2C9 codes for an enzyme that metabolizes the blood thinner warfarin. People with certain variants of this gene can have an exaggerated response to the drug, leading to bleeding episodes.
Clinicians are still untangling how to use this information to help their patients. For instance, a 2013 study using CYP2C9 and another gene variant to fine-tune warfarin dosing found that incorporating genetic information didn’t reduce bleeding episodes or other nasty side effects. But a study in 2017 that analyzed three genes did find that gene-based dosing could reduce problems like bleeding or deep vein thrombosis. The conflicting results point to one of the many complexities of precision medicine: The 2013 study included more than 25 percent African-Americans, while the 2017 study was 91 percent white. The frequency of genetic variants differs between populations, researchers have found, so genetic risk calculations derived from one race or ethnicity may not translate well to others.
That said, precision medicine in cardiology could eventually reduce uncertainty, with potentially life-changing consequences. Are your risk factors sensitive to diet and exercise? Great — you can avoid popping pills. Have the kind of high cholesterol that will never go down, no matter how many bran muffins you eat? Skip the diet drama and just get on the right medication right away. Precision medicine may someday even give statins the boot, at least for some patients. These cholesterol-lowering drugs don’t prevent heart disease in everyone, and drugs in a different category might work better for some.
Already, genetic markers have been identified that may predict cardiovascular disease better than cholesterol levels, Georgia Tech’s Gibson says; these DNA variants might sort people into more precise risk categories than is possible by looking at cholesterol levels alone.
A precision medicine approach could be valuable not just to predict who will get sick, but how serious their symptoms will be. Gibson focuses on Crohn’s disease, an inflammatory bowel disease that can cause a range of complications, some life-threatening. To date, most Crohn’s genomics research has tried to predict who will get the disease, Gibson says. Very little has asked how genes influence disease progression or whether drug treatments will work.
“It turns out that, at least in Crohn’s disease, the genetics of getting the disease does not necessarily predict disease progression,” he says.
A deeper understanding of risk and response could save the medical system billions and let doctors and patients weigh the risks and benefits of treatment.
In a 2017 study, Gibson and his colleagues combined genomic information with standard blood and tissue tests for Crohn’s in children to predict who might benefit from drugs that block an inflammatory compound called tumor necrosis factor (TNF)-alpha. These anti-TNF drugs can cost $50,000 a year, Gibson says. For a subset of patients with inflammatory and autoimmune diseases, they’re transformative, but don’t help as many as 80 percent of Crohn’s patients.
According to the team’s calculations, about a third of the 913 study participants were at little risk of progressing to penetrating Crohn’s disease, a particularly dangerous form of the disorder that eats holes in the bowel walls. The results still need to be replicated in bigger studies, but it suggests some patients could avoid anti-TNF drugs in favor of less expensive therapies.
Crohn’s is just one example of how a deeper understanding of risk and response could save the medical system billions. It could also give doctors and patients alike the chance to weigh the risks and benefits of treatment, truly assessing whether a new medicine or test is worth the potential risks or side effects. “I see it as giving patients more control over their decision-making,” Gibson says.
Less treatment, more control? Sounds like precisely what we need.