Neuroimaging has become a powerful biomarker and helps researchers study mental illness.
A Google search of the word “neuroimaging” yields several hundred thousand hits. Metadata that pop up include PET (Positron Emission Tomography) scans, CAT (Computed Axial Tomography) scans, fluoradopa, etc. And many of these sites have very complex definitions for “neuroimaging”. For example, Wikipedia had the following definition: “Neuroimaging includes the use of techniques to either directly or indirectly image the structure, function/pharmacology of the brain.” This definition is somewhat vague in that it does not indicate exactly what is being measured.
But what is being measured by a neuroimager depends on which technology is being implemented. For example, according to the same Wikipedia page, the PET scan “measures emissions from radioactively-labeled, metabolically-active chemicals that reach the brain following injection into the bloodstream.” Specialized sensors in the PET scanner detect the radiotracer as it accumulates in certain areas of the brain. Finally, the onboard computer integrates the data to produce a multicolored, two-dimensional (or three-dimensional) image of the area of the brain in which the tracer accumulates.
The specificity of a PET scan comes from the radiotracer, which is typically a molecule that can be localized to the neurons under study. This molecule comes in many forms, be it a source of energy (e.g. glucose), a ligand (or drug compound) with affinity for a specific plasma membrane receptor, or a neurotransmitter. The labeled molecule acting as a radiotracer has the potential to interact with biomarkers for diseases such as Alzheimer’s (see cover story, p 18), allowing PET scans to be used for diagnosis and monitoring of such neurodegenerative diseases.
For example, in 18F-fluoradopa-PET scanning, the radiotracer interacts with the enzymatic conversion of levodopa to dopamine in dopaminergic neurons, a process that may be aberrant in individuals with Parkinson’s disease (Brooks et al., Exp Neurol. 2003 Nov; 184 Suppl 1:S68-79). That is why 18F-fluoradopa-PET scanning is used to evaluate Parkinson’s disease. Other neuroimaging approaches used to examine this biochemical process include (+)-[11C]-dihydrotetrabenazine-PET (for examining the storage of dopamine in synaptic vesicles via the vesicular monoamine transporter 2) and 123I-beta-CIT SPECT (for examining reuptake of dopamine into axons via the dopamine transporter). Single Photon Emission Computer Tomography (SPECT) is similar to PET except that it requires a radiotracer that emits gamma radiation.
PET for Parkinson’s
Peter Lewitt, MD, professor of neurology at Wayne State University School of Medicine, and Director of Movement Disorders at Henry Ford Hospital, Southfield, Mich., had this to say about neurobiomarkers: “We want something that is more biologically attuned to the disease processes in some instances, something that might provide more precision than simply asking questions and using rating scales that are, at best, semi-quantitative.” Lewitt points out the use of fluoradopa-PET scanning to look for markers of loss of dopaminergic neurons from the nigra striatal system in Parkinson’s patients as an example.
But using neuroimaging to detect biomarkers of disease is not a new idea. “The fluoradopa-PET scan goes back to the 1980s,” says LeWitt, “and there are a number of companies that having been
using these technologies as tools for more than two decades—validating them, using them in animal studies, getting to the level of expertise that they can to diagnose Parkinson’s disease.” In an abstract from one of his papers (Lewitt. 2006. J Neurol Transm Suppl. 71:113-22), Lewitt wrote “With surrogate biomarkers such as radiotracer neuroimaging of the dopaminergic system, the pace of clinical investigation can be increased.”
Some researchers are interested in using neuroimaging as a method for making diagnoses. For example, in a March, 2007 paper, Chertkow et al. wrote: “Functional imaging (SPECT and PET) is best suited to tracking symptomatic therapy response [in Alzheimer’s Disease], and anatomic (MRI volumetric) imaging or amyloid PET are more suited to reflect dementia modulation studies.” (Can J Neurol Sci. 2007 Mar; 34 Suppl 1: S77-83.) In the same paper, the authors expressed that “the potential for imaging with respect to pharmacological studies of dementia—to provide surrogate markers for drug studies, to improve diagnosis, to speed evaluation of outcomes, and to decrease sample sizes—is huge.”
Another example of neuroimaging as a diagnostic tool comes from a recent paper that discusses the use of neuroimaging in distinguishing between the overlapping pathologies, Parkinson’s disease with dementia (PDD), and dementia with Lewy bodies (DLB). In this paper, the authors write that “at present, PDD and DLB are distinct entities by definition” and that “future studies, including clinical trials and biomarkers such as neuroimaging and cerebral spinal fluid, will help further define the clinical and therapeutic implications of this distinction.” (Camicioli et al. Can J Neurol Sci. 2007 Mar;34 Suppl 1:S109-17.)
PET in the pipeline?
So has any drug discovery and/or drug development come out of these neurobiomarker studies? “I don’t know if they have had a major impact,” says Lewitt. “I know there have always been concerns that you cannot power a study adequately with these and expect the FDA [US Food and Drug Administration] to accept these as the primary end point . . . so they have become secondary end points.” However, Lewitt does think of neurobiomarkers as disease-specific indicators that can guide investigators in how to best carry out clinical trials, i.e., with the fewest number of patients, shortest duration of exposure, shortest duration of the study, etc.
Furthermore, in their 2004 paper in NeuroRx, Rosas et al. wrote: “Neuroimaging methods offer the potential to provide noninvasive, reproducible, and objective methods not only to better understand the disease process but also to follow in clinical studies to determine if a drug is effective in slowing down disease progression or perhaps even in delaying onset.” (NeuroRx. 2004 Apr; 1(2):263-272.)
Other than the fact that neurobiomarkers have been relegated to secondary end point status, the other problem they face is that their presence in patients with disease X is not necessarily predictive of disease X. LeWitt gives this example. “Let’s say there is a 10% change in your biomarker and you find that scientifically exciting, they [still] may not have any clinical impact at all. It might be too small a change to be clinically important.” He concludes that at the end of the day the neurobiomarker is, at best, a surrogate marker that one can only hope is so closely related to a specific disease that it can be used as a marker for that disease.
Neurobiomarkers are powerful tools for predicting, diagnosing, and treating neurological disease. However, their specificity and significance seem to always be in question. Luckily, using imaging as a neurobiomarker has made it possible to examine the neurochemistry and neuron activity of specific brain regions in situ and in real-time so that the mysteries of neurological disease may soon be unraveled.
The future is coming and neuroimaging seems to be leading the way.
|A Different View
Another type of neuroimaging is Functional Magnetic Resonance Imaging (fMRI), which works by tracking changes in blood flow in the brain that are associated with neural activity. The most important thing about fMRI is that it can be used to track changes in neural activity over time, an advantage according to Melissa P. DelBello, MD, MS, vice-chair for Clinical Research, Department of Psychiatry and associate professor of Psychiatry and Pediatrics, University of Cincinnati College of Medicine, Cincinnati, Ohio.
“What we’re doing specifically is looking at using imaging to produce biomarkers of specific treatment response,” says DelBello. She uses both fMRI or Magnetic Resonance Spectroscopy (MRS), either alone or in combination to answer the following questions: How does activation or level of activation in a specific brain region, or a specific concentration of a neurochemical, or levels of those neurochemicals in combination with activation predict which bipolar disorder patients will respond to which medication?
According to DelBello, the advantages of MRI are that it is noninvasive, non-radioactive, can be used to investigate specific brain regions, and can be used to conduct temporal studies of these regions. The fact that time-course experiments can be performed with MRI is the reason Delbello says it is what’s needed to look at biomarkers over time. “You could just scan somebody at one time point, but ideally you need to see what effects the medication has over time and what our marker responds to.”
DelBello is also looking at using neuroimaging biomarkers of disease, for which she has just received a large NIH grant. One major objective of this project is to determine if children with parents that have bipolar disorder have early markers of the illness.
“Some of our preliminary data suggest that as people develop bipolar disorder during normal adolescent development have big activation in specific regions of the prefrontal cortex and that does not happen normally,” DelBello says. “So what we are looking for is when does that occur? Is it there before the illness onset? Can we predict who is going to develop the illness?”
Because the project is still in its infancy, the success of these neuroimaging technologies is not certain yet. However, bigger things may be on the horizon, possibly diagnostic biomarkers for bipolar disorder. But there’s still a lot of work left to do.
This article was published in Drug Discovery & Development magazine: Vol. 10, No. 10, October, 2007, pp. 24-28.
Filed Under: Drug Discovery