imaging in psychiatric patients

This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.

Neuro111

Full Member
10+ Year Member
Joined
Aug 9, 2010
Messages
52
Reaction score
35
In grand rounds today (Neurology), we had an interesting case of psychosis in a Multiple Sclerosis patient. 25 year old guy with known schizophrenia (diagnosed at age 21) presented to our psychiatric emergency services. He complained of blurry vision and one of our senior psych residents ordered an MRI (could have been a psych attending, I am not sure). Anyways, it turned out that MRI had abnormal findings and they subsequently consulted Neurology who later diagnosed him with MS. Apparently there has been a few cases of MS patients with psychiatric symptoms in literature. Obviously it is not sure whether MS caused the psychosis or whether it is his schizophrenia and MS was an incidental finding. When I talked to one of my senior Neurology residents, he said that given what he heard on the presentation he would not have ordered an MRI. This started me thinking when do you guys actually order imaging or other testing modalities on psychiatric patients? From reading some literature it seems like 1. psychiatric symptoms in patients with known head injury/stroke 2. New onset psychosis in patients below or above the usual age range (eg: 5 year old with psychotic symptoms or 55 year old with new onset psychiatric symptoms) or 3. any abnormal neurological findings on examination.

What are your guys' thoughts on when to order imaging on our psych patients?

Members don't see this ad.
 
  • Like
Reactions: 1 user
In grand rounds today (Neurology), we had an interesting case of psychosis in a Multiple Sclerosis patient. 25 year old guy with known schizophrenia (diagnosed at age 21) presented to our psychiatric emergency services. He complained of blurry vision and one of our senior psych residents ordered an MRI (could have been a psych attending, I am not sure). Anyways, it turned out that MRI had abnormal findings and they subsequently consulted Neurology who later diagnosed him with MS. Apparently there has been a few cases of MS patients with psychiatric symptoms in literature. Obviously it is not sure whether MS caused the psychosis or whether it is his schizophrenia and MS was an incidental finding. When I talked to one of my senior Neurology residents, he said that given what he heard on the presentation he would not have ordered an MRI. This started me thinking when do you guys actually order imaging or other testing modalities on psychiatric patients? From reading some literature it seems like 1. psychiatric symptoms in patients with known head injury/stroke 2. New onset psychosis in patients below or above the usual age range (eg: 5 year old with psychotic symptoms or 55 year old with new onset psychiatric symptoms) or 3. any abnormal neurological findings on examination.

What are your guys' thoughts on when to order imaging on our psych patients?

The literature bears out what you are suggesting, namely for atypical cases it can be worth it. Certainly with focal neurologic signs, head injury, stroke but neuro would already be getting that test.

Re: your case above. I had a 40 yo modestly successful shoe salesman who presented with a known history of untreated MS but no psychiatric history admitted to psychiatry for frank paranoia and thought disorder. Pt was in denial about his MS after he got treatment for his first flare and subsequently had very minor neuro findings on exam (like very subtle differences in reflexes, maybe 4+ strength at 1 spot). Neuro said no way his presentation was related to MS, they would recommend outpatient treatment. MRI was absurd, lit up like a christmas tree with all his plaques. We said this is a pretty atypical presentation for schizophrenia and got him IV steroids. All of his sx resolved within a week and he could not figure out why he was in a psychiatric hospital. Not sure there is much like this in the literature but I was the primary resident for this case and it was pretty profound.
 
  • Like
Reactions: 5 users
There isn't any real clear consensus on this. 5-10% of all first break episodes may be due to organic causes, and some recommend imaging for all patients. It may potentially make sense, but there seems to be limited evidence to support the clinical usefulness and cost effectiveness of this. For me, a scan would be contingent on clinical findings suggestive of an underlying structural cause, atypical age, or maybe if the family had a strong preference.
 
Last edited:
Members don't see this ad :)
Obviously it is not sure whether MS caused the psychosis or whether it is his schizophrenia and MS was an incidental finding.

Neuro text books say that MS caused psychosis is very rare. At least I've seen this said in multiple places.
 
  • Like
Reactions: 1 user
From reading some literature it seems like 1. psychiatric symptoms in patients with known head injury/stroke 2. New onset psychosis in patients below or above the usual age range (eg: 5 year old with psychotic symptoms or 55 year old with new onset psychiatric symptoms) or 3. any abnormal neurological findings on examination.

What are your guys' thoughts on when to order imaging on our psych patients?

I certainly think number 3 is reasonable, although the heuristic should be something more than "psych + focal neuro = imaging", and more that the presence of neurologic symptoms should prompt a more thorough history into other symptoms, environmental exposures, a more complete neuro exam, and then perhaps a differential diagnosis. In my system this would then prompt me to consult neuro rather than order an MRI but in either event I think it helps to think about some of the things you might be expecting before making the request.

As to your number 1, I don't see why you would do imaging for behavioral symptoms in a patient with a known head injury or stroke - it is common for these patients to have behavioral symptoms and repeated imaging at the time of onset of the behavioral symptoms won't inform clinical management. An exception could be if they have some kind of acute change and you are worried about a new stroke.

Regarding psychosis in the very young, again I would not see this as a ready indication for imaging - it is true that psychotic symptoms in a 5 year old almost never indicate schizophrenia but the differential is ASD, intellectual disability, trauma, severe mood or anxiety disorders, toxic exposures, and probably in some case report you will find that this was the first manifestation of a mass lesion or inborn error of metabolism, but I certainly wouldn't see imaging as universally indicated in these cases.
 
Neuro text books say that MS caused psychosis is very rare. At least I've seen this said in multiple places.

I've read the same, but I also have a MS patient in her late 40's who only recently started having psychotic symptoms.
 
Neuro text books say that MS caused psychosis is very rare. At least I've seen this said in multiple places.

Yup that's what I've always been told which is why I presented my ginormous N=1 above. If you see lesions, treat them, they go away and concurrently psychiatric sx go away, just seems hard to not attribute this to MS.
 
Given all the various encephalitis cases I'm hearing about these days it's kind of tempting to think every psychotic person deserves 1 MRI at some point. (Not that I do this, but it's something I wonder a lot about)
 
Given all the various encephalitis cases I'm hearing about these days it's kind of tempting to think every psychotic person deserves 1 MRI at some point. (Not that I do this, but it's something I wonder a lot about)
They really don't. This would be an incredible waste of money with fairly low utility. A good history and neurobehavioral mental status examination should point to organcity in the majority of cases. We should not be relying on diagnostic tests (particularly the wrong ones) as a replacement for clinical skills. BTW, MRI is normal in 25% of cases of LE. So you have the issue of false negatives too. I'm looking for a good story to consider such diagnoses more seriously in the differential dx.

Even if you did an MRI on everyone, the average psychiatrist wouldn't necessary know what sequences to get, and the average radiologist wouldn't know what to look for.
 
  • Like
Reactions: 1 user
This is kind of off topic, but the topic of whether MRIs could be a waste of money makes me think of something I've thought of before.

MRI machines are extremely expensive to buy and situate in a hospital.

The cost of running an MRI is high, but from what I can tell, that seems to be mostly in service of recouping the initial costs of the machine rather than the massive amount of electricity they use, as electricity is relatively inexpensive.

From what I can tell the risks involved in an MRI without contrast are virtually none, and the risks with contrast are still virtually none for the vast majority of patients.

On the other hand, CT scans, the use of which have grown in huge amounts in the last few decades, expose people to a great deal of ionizing radiation and dyes that more commonly cause kidney damage.

Where I live, we have two hospitals, one that has an MRI machine and one that does not. The one that has an MRI machine is 1.5 Tesla. In the hospital with the MRI machine, it is never used in emergency situations.

However, in more urban areas with better machines and more radiologists on staff, MRI machines are starting to be used for emergency situations, such as virtual pulmonary and coronary angiograms.

I read recently that 2% of fatal cancers in the next 10-20 years will have been caused by CT scan exposure.

It makes me wonder if even "frivolous" use of MRIs could ever be a waste of money, if it is in service of paying off the machines, and/or making it possible for hospitals without them to purchase them, and starting to expand staffing and use of them such that they are used in situations currently reserved for CT scans?

No matter the indication, the money is going toward supporting machines and staffing that can replace machines that cause cancer in a not insignificant number of people.

I'm not sure if that shift will take place without doctors ordering lots of MRIs in the same way over the last few decades they started ordering millions more CT scans than they used to. What is the most healthy and logical does not seem to be the deciding factor. Cost and the way things have been done seems to be the factor.
 
From what I can tell the risks involved in an MRI without contrast are virtually none, and the risks with contrast are still virtually none for the vast majority of patients.
This is because you don't have any medical training and thus are making assumptions based on the technology used. The main risk of scanning everyone is you will find all sorts of abnormalities that probably have nothing to do with anything. These are called "incidentalomas" (i.e. incidental abnormalities completely unrelated to the patient's symptoms). When these lead to harm in the patient, the patient is called VOMIT (Victim of Medical Imaging Technology).

If someone's problems are behavioral, psychological, or due to drugs, some incidental finding then becomes the "cause" of all their problems, and the patient no longer has any incentive to address their problems.

It is very common for headache patients who get (inappropriately) scanned, to get worse when some incidental tumor or spots that are totally unrelated to the headache are unearthed. In some cases the location of the headache shifts hemisphere or other region to "match" the incidental imaging findings

Lots of people have tumors - including brain tumors since we're talkng about neuroimaging - that will never cause a problem for them. But you unearth some tiny meningioma that was never going to cause problems, and tell someone they have a brain tumor. It may not go over well. They may become excessively anxious or distressed. They may undergo repeated imaging because now we've found this tumor we can't just ignore it - or the patient can't seem to forget about it.

You might see white matter hyperintensities etc that trigger additional, potentially invasive workup that is totally unnecessary. In the process of this workup (LP, biopsy etc), the patient may experience physical injury, infection, bleeding, pain etc etc.

As for contrast, we are realizing that gadolinium seems to pool in the brains of some people who have had repeated MRIs... the significance of this is not known but is a potential issue for MS patients who undergo lots of serial MRI.

More pts than you might expect require conscious sedation for MRI which has some inherent risks.
 
  • Like
Reactions: 1 users
This is because you don't have any medical training and thus are making assumptions based on the technology used. The main risk of scanning everyone is you will find all sorts of abnormalities that probably have nothing to do with anything. These are called "incidentalomas" (i.e. incidental abnormalities completely unrelated to the patient's symptoms). When these lead to harm in the patient, the patient is called VOMIT (Victim of Medical Imaging Technology).

If someone's problems are behavioral, psychological, or due to drugs, some incidental finding then becomes the "cause" of all their problems, and the patient no longer has any incentive to address their problems.

It is very common for headache patients who get (inappropriately) scanned, to get worse when some incidental tumor or spots that are totally unrelated to the headache are unearthed. In some cases the location of the headache shifts hemisphere or other region to "match" the incidental imaging findings

Lots of people have tumors - including brain tumors since we're talkng about neuroimaging - that will never cause a problem for them. But you unearth some tiny meningioma that was never going to cause problems, and tell someone they have a brain tumor. It may not go over well. They may become excessively anxious or distressed. They may undergo repeated imaging because now we've found this tumor we can't just ignore it - or the patient can't seem to forget about it.

You might see white matter hyperintensities etc that trigger additional, potentially invasive workup that is totally unnecessary. In the process of this workup (LP, biopsy etc), the patient may experience physical injury, infection, bleeding, pain etc etc.

As for contrast, we are realizing that gadolinium seems to pool in the brains of some people who have had repeated MRIs... the significance of this is not known but is a potential issue for MS patients who undergo lots of serial MRI.

More pts than you might expect require conscious sedation for MRI which has some inherent risks.

You're right that I don't have medical training, but I was actually aware of the risks of fishing expeditions when it comes to any type of test that can result in an obligation to do further testing/treatment that can be damaging. So, you're right that saying that the risk is low is not accurate when you include that facet. But there have been a huge increase in the number of CT scans performed that have immediately damaging effects, and some of those are fishing expeditions or are done in the name of defensive medicine. So, OK, in the face of what you said, I change my opinion that there's no such thing as a not useful MRI. I do still think there's a . . . not epidemic, that's too strong . . . a cavalier attitude toward liberal CT scans.

The point you made about incidentolomas made me think about troponin testing, and how any time you have a new protocol where you have more data, you need new algorithms.

Europe has used high-sensitivity troponin tests for well over a decade now in assessing chest pain. These tests are such that every person tested has detectable troponin. It requires a different algorithm than what is used in the US. The tests can detect anything on the spectrum from stable angina to myocardial infarction, to non-cardiac issues, along with normal levels of troponin that every person has. It presents a more complicated clinical picture than what we currently have in the US, which is one of the reasons the FDA hasn't approved these tests.

I don't know how Europe deals with these test results. They must have different cut-off levels than we do, and possibly some more complicated algorithms. I know they have a much higher NPV, but not sure how they deal with equivocal results.

But if it is possible to deal with new technology that gives you more information when it comes to a blood test, I wonder if it would be with neuro-imaging, as well. I mean I know within my own family someone who had an MRI and had a brain plaque that was deemed part of normal aging even given a family history of MS and that was given no further workup, so I guess that speaks to the fact that doctors already have discretion with imaging. I had a number of incidentalomas after my CT scan for appendicitis: trace plaque in abdominal aorta, some tiny thing in a lung (didn't worry me enough to even remember what it was), some sort of bone island in my pelvis.

I guess I should re-state my desire and observation that led to my idea about no-bad MRI: Many of the things that are done with CT scans today are possible to do with MRIs, but the health of patients has not been a large enough impetus to move to MRI en masse quickly.

I don't believe in handing out MRIs like candy (except to the end that I optimistically espoused: that it might hasten a move away from CT scans) and I accept what you're saying about fishing expeditions.

As for the pooling in the brain of gadinolium, I was not aware of that. I knew about NSF and the risk for that in renally impaired patients. I am not medically trained, but I do try to educate myself. When I had my most recent CT scan and contrast for the first time, I told the doctor I should probably delay my beta blocker dose as it increases the risk of an allergic reaction to the iodinated contrast material. The doctor was medically trained but was not aware of that increase in risk. I know that I can't see medicine from as broad of a scope as a physician, but my interest can sometimes result in specific knowledge. And there are medically trained physicians who are very concerned about the number of CT scans being performed, which hasn't seemed to result in any more judiciousness. So my spitballing was order more MRIs to change the balance. That is in no way based on specific knowledge, just spitballing on what might hasten a change from CT scans to MRIs (increase the demand, lower the costs, increase the ubiquity and normalcy of MRI as default over CT in situations where it's equal or better from an imaging standpoint since whether it's equal or better from an adverse event standpoint hasn't tipped the scales).

EDIT: I would also add that is precisely medically trained professionals who have offered whole body CT scans as a preventative medical fad for the very wealthy. I remember Oprah back when she had a TV show had one and had a doctor on promoting it. It takes a judicious mind, beyond just a medically trained one, to know radiating your entire body when there are no presenting symptoms is a bad idea.

EDIT 2: Here is a medically trained professional who understands things much better than I do.

He suggests 20 mSv for everyone! You get 20 mSv, and you get 20 mSv, and you get 20 mSv!

And I'm not talking about Oprah; I'm talking about the profiteer doctor who came on the show to hawk this:

 
Last edited:
I'm in favor of everyone getting an MRI once. Because if we don't look for it, we won't find it. Freudenreich has some good papers on initial workups of first episode psychosis.
 
More practical reasons why we are a ways off from MRI replacing CT:

A) MR scans take approximately an order of magnitude longer for any given modality you care to name than an equivalent CT.

B) MR machines require a frequently replenished supply of liquid nitrogen under great pressure to keep the superconducting coils cold enough to have zero effective resistance and allow the induction of the intense magnetic fields involved. If the temperature is allowed to fluctuate by even a few degrees, resistance suddenly becomes nonzero, accompanied by a huge surge in current demand not easily accommodated by most powerful grids, a rapid spike in temperature, followed by a very violent outgassing of nitrogen that blows holes in the walls of building.

MR is a fantastic technology, but it has its drawbacks.
 
Even if you did an MRI on everyone, the average psychiatrist wouldn't necessary know what sequences to get, and the average radiologist wouldn't know what to look for.

Ugh. All psych residents go through at least two months of neuro training. How could you afford to not learn imaging?
 
Top