Depression: PET Scan or Scan Pet?

Alvin B. Lin, MD, FAAFP
Dr. Lin is an associate professor of family and community medicine at University of Nevada School of Medicine and an adjunct professor of family medicine and geriatrics at Touro University Nevada College of Medicine. He also serves as an advisory medical director for Infinity Hospice Care and as medical director of Lions HealthFirst Foundation. Dr. Lin maintains a small private practice in Las Vegas, NV. The posts represent the views of Dr. Lin, and in no way are to be construed as representative of the above listed organizations. Dr. Lin blogs about current medical literature and news at

The medical specialty of psychiatry has been the news quite a bit lately. Last month, the National Institute of Mental Health withdrew their support of the new Diagnostic & Statistical Manual 5 of Mental Disorders, which has been called the Bible of psychiatry. Earlier this spring, a psychiatric hospital here in Las Vegas was accused of "patient dumping", that is the discharge of patients with mental illnesses to places where they have no family or friends for support.

On a more upbeat note, I wanted to point out a small short randomized trial published early online two days ago in JAMA Psychiatry (aka Archives of General Psychiatry) in which the authors concluded that positron emission tomography (PET) scan was potentially useful in determining those whose depression might improve as a result of escitalopram vs cognitive behavioral therapy. Currently, it's near impossible to predict who will respond to which antidepressant drug. At most one in three will respond to the first choice. Of the remaining two thirds, another one third will respond to the second choice. Of those remaining, another third will respond to the third choice (after failing to respond to the first two). But some will improve with cognitive behavioral therapy. The difficulty lies in determining who will respond best to what.

So here you have a study of just 67 men and women followed for 24 weeks total. However, of that group, only 38 had interpretable PET scans upon which the authors based their conclusions. But you have to admit, the differences in imaging between responders & non-responders was like night and day. More importantly, we still haven't developed a way to determine who will respond to escitalopram vs citalopram vs sertraline vs paroxetine vs fluoxetine vs cognitive behavioral therapy. Don't get me started because dosing might make a difference along with the number of CBT sessions; perhaps longer or shorter sessions at more or less frequent intervals.

For now, this study, while it made the press, is just an interesting theory. It's very doubtful that any insurer will pay for this test, not at least for the foreseeable future. It's less expensive to just go through a series of trials and errors. Which is a sad commentary on how we treat our mentally ill (see "patient dumping" above).