Psychiatric Diagnosis in the Lab: How Far Off Are We?

Jeffrey A. Lieberman, MD


September 28, 2011

This feature requires the newest version of Flash. You can download it here.

Hello. This is Dr. Jeffrey Lieberman of Columbia University speaking to you for Medscape. Today, I would like to address the issue of psychiatric diagnosis. As most of you are aware, the Diagnostic and Statistical Manual of Mental Disorders (DSM)-IV -- the "Bible" of psychiatric nosology -- is currently undergoing revisions. DSM-V promises to be the state of the art in terms of psychiatric diagnoses, and when it was initiated we anticipated that this iteration of the DSM would incorporate biological markers and laboratory-based test results to augment the historical and phenomenologic criteria that traditionally are used to establish psychiatric diagnoses. Sadly, this has proved to be beyond the reach of the current level of evidence for incorporating into this version of the DSM, and it appears that psychiatric diagnoses, which may be rearranged, consolidated, and modified in some ways, will still be based predominantly on symptomatic and historical criteria.

However, I am here to tell you that the time is not far off in the future when psychiatric diagnoses of mental disorders and behavioral disturbances will be aided by laboratory-based tests, and this will mark a milestone in the evolution of psychiatric medicine and will occasion an enormous transformation in the accuracy and the reliability of psychiatric diagnoses.

Let me give you an example of how this is likely to change what we do when we evaluate patients. Imagine that you are evaluating someone with chest pain who is thought to be having a heart attack. But you do not have a stethoscope or an ECG machine, you cannot draw blood, test for CPK isoenzymes, or you cannot perform arteriogram or a thallium scan or an angiogram to determine the specific nature of the pathology. If you are just asking the patient to report how he or she feels, and maybe doing a physical exam, how could you distinguish myocardial infarction from gastric reflux, heartburn, some type of pulmonary disturbance, some musculoskeletal problem like costochondritis, or something of that nature? You would make a diagnosis as precisely and reliably as you could, but the precision and validity of the diagnosis would be limited by the evidence that you have in terms of the methods you can use for your diagnosis.

In psychiatry this is our current state of the art. We examine patients by watching them, by talking to them, by eliciting specific responses from them, by gathering history, and by talking to informants. We can also do various types of physical examinations or laboratory tests to rule out certain diagnoses, but we largely base psychiatric diagnoses on subjective report of symptoms and historical information about natural history and course, and this has an inherent uncertainty or unreliability of diagnoses.

In recent years, however, we have seen the emergence and refinement of a number of different technologies that I predict will, within our professional lifetimes and hopefully within the next 5 years, lead to the incorporation of laboratory-based tests for psychiatric diagnosis. When these will be proved to a satisfactory level of evidence and when they will be reimbursable by third-party payers, we can't know specifically, but I predict this will happen fairly soon. We are seeing the evidence of that even now.

What will these diagnostic tests be? The tests that appear to be emerging as the first to be marketed are ones that are based on the proteomic or metabolomic or biochemical analyses of plasma or cerebrospinal fluid. A series of different types of microarray panels have been developed that examine the profile of a series of analytes in plasma, serum, or cerebrospinal fluid. Using these, a certain profile, a biochemical signature if you will, has been found to correspond to specific psychiatric diagnoses. The first diagnoses to have been associated with these biochemical signatures are schizophrenia, manic depressive illness, and depression. Several companies have developed data, applied for approval from the US Food and Drug Administration, and are beginning to market these tests. [Some of these tests] are being used by a few psychiatrists, even though they still have not been accepted as a standard of care in the field and they are not paid for by third-party reimbursement agencies or organizations. Nevertheless, this is the cusp of the implementation of this method of diagnosis: proteomic, metabolomic-based analytes that yield a certain diagnostic signature.

A second modality that is likely to be implemented for psychiatric diagnosis is that of imaging techniques; here we're talking about both nuclear medicine imaging with PET and MR imaging with either structural, spectroscopic, or functional imaging applications. These have been used in a variety of disorders. They yield clear differences between, diagnostic groups such as schizophrenia or depression on one hand and healthy volunteer controls or nonaffected individuals on the other. The problem is that the distributions of the values of the control vs patient groups still have too much overlap and are not sufficiently differentiated as to provide high enough positive predictive value at the individual patient or subject level. But I predict that it won't be too long before these are refined, the results will become more robust, and these will contribute to a profile or augment the information that clinicians have to establish their diagnosis.

Finally, genetic testing will also come into play. As you probably know, commercial companies already are marketing DNA testing. They provide a "readout" of your genotypes for all of the known coded human genes along with associations with specific diseases in the different organ systems that these correspond to, to the best level of evidence that currently exists. These are not accepted as medically validated and are not used routinely in clinical practice but there is no reason psychiatry cannot begin to use these as other fields of medicine have done. Because all mental disorders will almost certainly prove to be polygenic or multigenic, we will need a gene profile to utilize in term of diagnostic information.

So, I encourage you to stay tuned, follow this literature, and be thinking about the fact that in the not-too-distant future, psychiatry will have laboratory-based methods to assist in our diagnoses. This should be an enormous benefit to our field in terms of enhancing the validity and precision of our diagnosis as well as elevating the scientific quality of clinical practice for the benefit of our patients.

Speaking to you today from Medscape, I'm Dr. Jeffrey Lieberman of Columbia University saying thank you and see you again soon.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.