Can precision medicine do for depression what it’s done for cancer? It won’t be easy

At a growing number of research centers across the country, scientists are scanning brains of patients with depression, drawing their blood, asking about their symptoms, and then scouring that data for patterns. The goal: pinpoint subtypes of depression, then figure out which treatments have the best chance of success for each particular variant of the disease.

The idea of precision medicine for depression is quickly gaining ground — just last month, Stanford announced it is establishing a Center for Precision Mental Health and Wellness. And depression is one of many diseases targeted by All of Us, the National Institute of Health campaign launched this month to collect DNA and other data from 1 million Americans. Doctors have been treating cancer patients this way for years, but the underlying biology of mental illness is not as well understood.

“There’s not currently a way to match people with treatment,” said Dr. Madhukar Trivedi, a depression researcher at the University of Texas Southwestern Medical Center. “That’s why this is a very exciting field to research.”

A precision approach would be welcome news for many patients with depression. There’s a well-documented cycle of trial and error for these patients, who wait weeks for drugs to kick in, only to find out they don’t work. Then they might have to repeat the process, often more than once.

But it’s not an easy task to break down the many factors that contribute to depression into clean categories with clear treatments. While some in the field are excited about the promise of precision medicine to better tailor treatments for depression, others are worried the idea is being overhyped.

“It remains to be shown that depression coalesces into neat subcategories, as opposed to being a fuzzy set,” said Dr. Steven Hyman, director of the Stanley Center for Psychiatric Research at the Broad Institute of MIT and Harvard.

Bringing precision medicine to depression would be difficult for the same reason it would be useful — depression is a heterogeneous disease that varies wildly from one patient to the next.

Take someone with a family history of depression, who first experiences symptoms during adolescence and has several instances of depression over her lifetime, and compare her to a 70-year-old man who is in the early stages of Alzheimer’s disease and is experiencing symptoms of depression for the first time.

“The [Diagnostic and Statistical Manual] would give you the same diagnosis, and clearly there’s reason to think those are very, very different,” said Hyman, a former director of the National Institute of Mental Health.

Leanne Williams, the director of Stanford’s new precision mental health center, acknowledged the task will be tricky, but she believes it’s worth trying to bring more order to depression treatment by making a coordinated push to gather as much data as possible. The center — which will be the focus of a dedicated funding drive — will build on $5 million in grants awarded to the researchers involved. Williams hopes to launch a study to follow patients with depression for years, which would cost much more.

The center’s roughly 35 faculty collaborators — including psychiatrists, engineers, geneticists, and data scientists — are using functional and structural MRI scans to analyze how depression disrupts neural circuits in research subjects’ brains. They’re also sequencing patients’ genomes to find common mutations that might play a role, as well as gathering clinical data on their symptoms.

“We want to use that information [to] guide the treatment by subtype,” she said.

Once researchers define depression subtypes, however, there’s still the question of whether — and how — doctors could use data from genetic studies and brain scans to guide treatment. Williams is planning a pilot project involving roughly 75 patients with depression to answer that question. All of the patients will undergo genetic testing, structural and functional MRIs, diagnostic interviews, and other clinical assessments.

For about half of those patients, the psychiatric team will get that information before coming up with a treatment plan. For the other half, their doctors will get the data after 12 weeks of treatment. Ultimately, Williams wants to find tests that physicians can use to guide their depression treatment decisions. But the onus is on researchers to show whether such tests “can shift the needle enough” to make them worthwhile for doctors to use — and insurers to pay for, she said.

Elsewhere, researchers are on the hunt for genetic mutations and biomarkers in the blood that could be used to diagnose depression — and guide treatment. In his lab at UT Southwestern, Trivedi has his eye on several potential depression biomarkers, including C-reactive protein, which is a marker of inflammation.

In a study published last year, Trivedi and his colleagues randomly assigned 100 patients with depression to receive one of two depression treatments. Patients with lower CRP levels had a higher remission rate on one treatment, while patients with higher CRP levels had a higher remission rate on the other.

But while the finding is intriguing, it’s just a small study. And that, experts say, is part of the problem.

Hyman, for his part, is skeptical — not of the idea of precision medicine for depression, but of the quality of the data.

“My skepticism is not that it’s a bad project, but we better not get ahead of ourselves and overhype,” he said. “I just think people are just a bit ahead of what the data will permit.”

He pointed to several concerns. With fMRI scans, for example, it’s difficult to tease out what might be caused by depression, what might be a cause of depression, what might be due to prior treatment attempts, and what’s just noise. It’s also challenging to collect brain scans from enough patients to have a well-designed study that includes enough patients to make the findings translatable to the general population.

And stratifying patients based on genetic mutations is messy. Depression is highly polygenic, meaning there are likely thousands of genetic variations in different combinations that can contribute to the condition. And depression can’t be explained solely by genetics: Environmental factors, particularly during development, also seem to play a role.

“It’s some grab bag of these thousands of [genetic] variants, plus bad luck as the brain develops, plus lived environments,” Hyman said.

Williams said she hopes to set up a large-scale study to create the kind of data set needed to better understand the factors at play in depression. She sees the Framingham Heart Study — a long-term study that began in 1948 and has followed patients for decades, as a way to identify risk factors for heart disease — as a model for depression research.

“We’ll start with what we can and keep building on it,” she said.

Trivedi, the UT researcher, said he’s well aware that there’s relatively little research on biomarkers that could be used to guide depression treatments — but he and colleagues say each study inches the field toward answers.

“Right now, we are throwing everything at the wall and seeing what sticks,” he said. “Precision medicine is deciding what to throw at the wall so the chances of sticking are better.”