[Updated 5/7/18, 5pm ET. See below.] Genetic testing that helps guide more precise cancer treatment is making its way into standard practice in major U.S. medical centers. Could the same eventually be true for depression?
One test maker thinks it has finally gathered enough evidence to convince psychiatrists to order its product and make more informed prescriptions based on a patient’s DNA. At the American Psychiatric Association conference today in New York, Assurex Health, a division of diagnostics firm Myriad Genetics (NASDAQ: [[ticker:MYGN]]), is releasing data from a study of 1,167 people with serious depression, officially known as major depressive disorder, whom previous treatments have failed to help.
“What we currently do is trial and error based on judgment,” says Stanford University Medical Center psychiatrist Charles DeBattista, one of several academic investigators who participated in the study. “There are no tools to even slightly increase the patient’s chances.”
The upshot of the study: The people whose prescriptions were based in part on the results of Assurex’s GeneSight genetic test fared better than those guided solely by a psychiatrist’s judgment. The study did not quite hit its main goal, but it did well in other goals that are arguably of greater importance to the practice of depression treatment. And in a first for the nascent field, the test was designed with enough rigor—randomized and double-blinded—for psychiatrists unaffiliated with the study to give it cautious praise.
“This is really welcome data for us,” says Jehannine Austin, a psychiatric genetic counselor and researcher at the University of British Columbia in Vancouver. “It’s a pressing clinical problem finding medicine that works without side effects.”
GeneSight is what’s known as a pharmacogenomic test. It provides a look at the status of 12 genes that, depending on their variations, affect a patient’s ability to metabolize drugs prescribed for depression.
Just under half of the study participants had their antidepressant prescription guided by results of GeneSight. The other half of the study participants received prescriptions based on a psychiatrist’s best judgment—in other words, the current standard to determine which among the constellation of approved depression drugs is right for a patient.
There is little argument that the standard is not adequate. Major depression affects nearly 7 percent of adult Americans and nearly 13 percent of teenagers, according to a federal 2016 survey. It is the leading cause of disability around the world, according to the World Health Organization.
Anyone familiar with the now-household names—Prozac, Paxil, Wellbutrin, Cymbalta, and dozens more—knows the carousel of hit-and-miss attempts, with some drugs causing side effects and others simply not working. A person with major depression has less than 40 percent chance of getting better on the first try with the pharmacopoeia of antidepressants, antipsychotics, and other drugs that psychiatrists use.
Tests like GeneSight have been available for a few years. GeneSight costs about $2,500 and has been covered in full by Medicare and Medicaid since 2014. But Assurex’s previous five studies were small and did not provide reassuring evidence to many psychiatrists and private insurers. They wanted to see a study with more than 1,000 participants, blinded, and showing longer-term effects, says Myriad executive vice president Bryan Dechairo, and he is confident that “this study has answered all their questions.”
Critics of the earlier studies like psychiatrist Robert Howland of the University of Pittsburgh Medical Center, who took Assurex and others to task in a 2014 paper, were measured in their praise for the new study. “It’s good that a large randomized controlled study has been conducted,” Howland says now upon reviewing the new data.
But he wants to see the new study, which was funded by Assurex and led by John Greden, executive director of the University of Michigan’s Comprehensive Depression Center, replicated by another party without Assurex’s involvement. And while most of the improvements in patient outcomes are statistically sound, Howland isn’t convinced that they will be meaningful enough to make a real-world difference.
Conducted in ten academic centers including Stanford, the University of Michigan, Harvard University, and Washington University in St. Louis, the study reported improvements in three major measurements of depression after eight weeks.
GeneSight alerts a doctor which among 56 drugs a patient should avoid, use with caution, or use as directed (red, yellow, and green). It’s not a pinpoint prescription; there could be several drugs in the green and yellow categories. But ruling out the red-category drugs helps narrow the guessing game. Dechairo says it’s akin to looking for a needle in a haystack by removing part of the hay.
People whose treatment was prescribed with GeneSight guidance were in remission—no more depression—15 percent of the time, a 50 percent improvement over the 10 percent remission in the “treatment as usual” arm. In other words, for every 100 people, five more would have the best result possible using GeneSight. (Assurex promises patients they won’t pay more than $330 for the test if it’s not covered completely.)
Those showing a response but not as dramatic as remission: 26 percent versus 20 percent, a 30 percent improvement.
The improvements in remission and response were both statistically significant results.
A third measure was symptom improvement, based on the Hamilton Rating Scale for Depression (HAM-D17), and it was in fact the main goal of the study. There was improvement, but it was not statistically significant. (27.2 percent to 24.4 percent.)
[Updated with analyst note and stock price.] “[Myriad] has consistently highlighted response and remission as the driving factor for GeneSight,” wrote Leerink analyst Puneet Souda in a research note after Monday’s presentation. “But since the study failed to hit the primary endpoint, questions still remain.” Investors nonetheless boosted Myriad shares 3.4 percent Monday to close at $28.98.
Missing the main goal of a study would seem to be a major flaw. Dechairo and Greden counter that the American Psychiatry Association’s treatment guidelines underscore that the goal of treatment is remission, not just improvement in certain symptoms.
Howland calls the miss “disappointing” because changing a patient’s HAM-D score “is a low bar” and the study design, with plenty of recruits, should have done the trick. (It was disappointing to Assurex shareholders, too. Because of the miss, they did not collect a multimillion dollar payout from Myriad as part of its deal to buy Assurex in 2016.)
Dechairo says the study reported “in reverse”—hitting its marks on the tougher measures but not quite getting there on the easier one—due to the nature of the participants. Coming into the study, they were already struggling to find the right medication, having failed one or more drugs.
Dechairo says a subset of 213 patients joined already taking medication that didn’t line up with their genetic profiles. They either knew this right away, because they were randomized into the GeneSight arm and got test results. Or, if they were assigned to the “treatment as usual” arm, they found out after eight weeks, when the study was unblinded. These people—whom Dechairo argues would be the true target population in the real world—had even better improvements across the board, in symptoms, response, and remission.
Austin says she was a “major skeptic” of GeneSight because of the inadequate previous tests but is impressed by the new study. These data “give us a lot more to interpret the usefulness of this kind of testing,” but she says she wants to see the full paper, which is not out yet. (Lead investigator Greden is only presenting selected data in a poster this week.)
Like Howland, Austin wants more evidence that the improvements were not just statistically significant but meaningful when making treatment decisions. “You can make any unimportant difference between two groups statistically significant if you recruit enough people to participate,” says Austin.
The poster also gives tantalizing but incomplete longer term results, showing data at 24 weeks—but only for the participants who had GeneSight-guided care. (At 24 weeks, 30 percent were in remission, twice the rate at eight weeks, and about 45 percent had a strong response.) The curves showing the “treatment as usual” patients won’t be available until the full paper is published.
Skepticism about GeneSight from people like Howland and Austin in the field hasn’t stopped Assurex, and now Myriad, from selling it. Diagnostics aren’t subject to the same regulatory approval hurdles as drugs. The test brought in $60.5 million in the second half of 2017, about 16 percent of Myriad’s total revenues, and is growing. The latest quarterly GeneSight sales were nearly 50 percent higher than the year-ago quarter.
Based in Salt Lake City, Myriad bought Assurex in 2016 and promised to keep it in its hometown of Mason, OH. Myriad paid $225 million, with $185 million more promised under certain conditions, such as meeting the primary goal of the new study. Dechairo says Myriad will discuss the extra payouts during its quarterly earnings call tomorrow.
Myriad is best known for its test for the breast cancer genes BRCA1 and BRCA2. It famously lost a landmark Supreme Court case to the American Civil Liberties Union, which successfully argued that human genes could not be patented.
Photo by Ryan Melaugh via Creative Commons 2.0.
