Q&A: Will Neuropsychiatric Imaging Lead to Overdiagnosis?

Overdiagnosis is commonly discussed in areas like breast imaging, where screening is widely conducted and media interest is high. While it’s under the radar for other diseases, it’s a big enough topic in the medical world to warrant a dedicated issue in Academic Radiology. Its August 2015 issue is devoted to overdiagnosis, a term used for disease that’s correctly diagnosed, but at the earliest stages when treatment may not be necessary and might even be harmful to the patient. This is in contrast to false positives, when the diagnostic test incorrectly indicates the possible presence of disease.

Overdiagnosis is more prevalent in modern times because the definition of disease has expanded, said Saurabh Jha, MBBS, assistant professor of radiology at the Hospital of the University of Pennsylvania, and guest editor of the Academic Radiology issue. “The rationale is the very intuitive concept that if we catch disease early on, we’ll avoid morbidity and mortality, that prevention is better than cure.”[[{"type":"media","view_mode":"media_crop","fid":"40903","attributes":{"alt":"","class":"media-image media-image-right","id":"media_crop_373722179344","media_crop_h":"0","media_crop_image_style":"-1","media_crop_instance":"4249","media_crop_rotate":"0","media_crop_scale_h":"0","media_crop_scale_w":"0","media_crop_w":"0","media_crop_x":"0","media_crop_y":"0","style":"height: 200px; width: 200px; float: right; border-width: 0px; border-style: solid; margin: 1px;","title":"Paolo Nucifora, MD, PhD","typeof":"foaf:Image"}}]]

Through screening, radiologists define the pathway to disease and are the gatekeepers of the technology often validating the diagnosis, since the gold standard of a diagnosis by pathology isn’t always an option.

With more advanced technology, radiologists can see deeper into structures in the body, discovering new (but often innocuous) abnormalities. Another reason for a growth in overdiagnosis is cultural. The division between “diseased” and “healthy” is arbitrarily set.

“There’s always going to be disputes about where the boundary takes place,” said Jha, with nothing magically happening at that threshold where disease is defined. However, in the United States, that threshold is often set at a place where doctors won’t miss any disease, erring on the side of overdiagnosing a person who doesn’t need treatment. “We live in an extremely risk averse environment.”

This Q&A series looks at a number of diseases that may be overdiagnosed in the United States, where we might be “treating health as opposed to solving sickness,” according to Jha.

Psychiatric disorders are interesting for radiology, because for a long time, people just thought it was all in the mind, but neuroimaging techniques can show changes. “Neuropsychiatric imaging has techniques that show anatomical changes...it’s good at showing changes in severe diseases, like severe depression. The danger is extrapolating that to mild changes,” Jha said.

Diagnostic Imaging talked to Paolo Nucifora, MD, PhD, assistant professor of radiology, Loyola University School of Medicine in Chicago, about overdiagnosis and psychiatry.

Can you discuss the role of imaging in detecting psychiatric disorders?

Right now, the role is still mainly experimental in terms of making a diagnosis. It’s mainly used to exclude causes of psychiatric symptoms. Usually a psychiatrist would expect the imaging to be negative, so they can just focus on common psychiatric conditions and make sure there’s not some underlying cause, like a brain tumor. This is changing as we develop newer imaging modalities and do more. We can start using imaging as a means for primary diagnosis, that’s the direction we’re going.

What imaging techniques are you using for diagnosis?

There are a couple things I use right now in the research setting. There are three main approaches to evaluating people with psychiatric disorders. The first is functional imaging, functional MRI, looking at areas of the brain that are overactive or underactive compared to a normal person. The second is diffusion tensor, diffusion imaging. These look at white matter to see if the connection between the two parts of the brain is normal. The last is structural morphometry, looking for a loss of tissue or excess tissue, usually a loss. The key is not the overall amount, but where. In the brain, location is everything. People focus on certain parts of the brain to focus on symptoms. This is still experimental. The frontal temporal lobes are most important in emotional regulation and cognitive function, therefore they’re the ones that get the most attention.

Is the main purpose of imaging early detection or correct diagnosis?

We’re looking for unusual things clinically. In most cases, there’s no imaging done at all. Imaging has a very small role to play, it’s quite unusual. Maybe if there’s a sudden change in someone’s behavior, maybe there’s a bleed in the brain causing the change in behavior. You’d expect the imaging to be negative.

Describe neuropsychology imaging’s role in the new DSM-5 classification and the NIMH’s break with the classification.

The NIMH wanted to move towards a classification that was based more on the biological, looking at imaging or other sorts of hard evidence of what’s going on in the brain. The DSM traditionally has been more descriptive in approach, not focusing on the causes of the disorder, but how it manifests. There was a little tension in the new DSM classification to see if they’d change this approach and switch to more a causal description of psychiatric disorders. In the end, they decided not to do that, and stuck to something similar as to what they already had. That’s why the NIMH broke with the American Psychiatric Association (APA).

I think the NIMH wanted to accelerate the pace of research, clinically implementing some of these ideas. It’s possible it just wasn’t ready for prime time. There’s a ton of research being done in psychiatry with neuroimaging, it just hasn’t translated yet.

If a patient had no symptoms, why would neuropsychiatric imaging be done?

It wouldn’t be done. But there are people who argue that it should be done. I’m not sure I’m one of those people. In a lot of people, we look at patients to predict what symptoms they would have (for nonpsychiatric diseases). If that’s true in psychiatry, identifying people for that risk could improve their long term outcome if you start treatment sooner. In a different way, you could give them the resources so that if they develop symptoms, they could start treatment; there are different approaches. If you find something in the brain that predicts risk for psychiatric disease, you could give some medications or follow-up to reduce long term consequences.

Do you think imaging will increase overdiagnosis of psychiatric disease?

Overdiagnosis is defined as the diagnosis of someone who has no symptoms. That doesn’t really exist to any great degree in psychiatry. The diagnosis requires symptoms. So if you’re treating someone, by definition they have symptoms, which is different from other specialties.

What is the harm in overdiagnosing a patient with a psychiatric disease?

Psychiatry is different than other medical specialties. In addition to the personal anxiety associated with getting a diagnosis, there’s also a social stigma related to the diagnosis. That’s been worse in the past maybe, but it still exists. You could be hurting the person just by making a diagnosis. In psychiatry, you must take multiple precautions when making this diagnosis.

What are the problems with treating patients who have been overdiagnosed?

There’s no way to know if early treatment would be effective. The experiments would have to be done. There’s been a little research showing that early treatment can help, but you can’t generalize that. There’s a case to be made for not making a diagnosis at all (for asymptomatic patients). You could tell someone the odds they’ll get these symptoms later in life, but don’t say they have something since they don’t have anything yet. You can have a discussion about how they want to manage their different risks.