Tag Archives: imaging

Imaging in the diagnosis of PSP

Posted on by

A very active area of research right now is how various imaging techniques (“scans” in English) can and cannot assist in distinguishing the atypical Parkinsonian disorders from other conditions and from one another.  Yes, this is important for clinical care and counseling.  But even more important right now is that until we have specific treatments for these diseases, we need accurate diagnosis in living people.  This is important for laboratory researchers who want to know the true diagnosis of the patient who supplied a fluid sample, and to the designers of clinical trials who want to make sure the patients in their trials have the disease for which the treatment was designed. 

Dr. Jennifer L. Whitwell, a radiology researcher at the Mayo Clinic in Rochester, MN has just published a very helpful review of that topic in Current Opinion in Neurology.  It gets pretty technical, but here are the takeaways with, of course, my own editorial contributions:

  • Measurement of atrophy by MRI: 
    • The magnetic resonance Parkinsonism index (MRPI), especially its updated version, the MRPI 2.0, gives excellent differentiation of PSP from non-PSP, with an area under the receiver operating curve of 0.98. (That statistic is 1.00 for perfect accuracy and 0.50 for no better than a flip of a coin. For a bit more explanation of the AUROC and a graph, see this post from last year.) 
    • Routine MRI can also be used to compare the various PSP variants with regard to atrophy of specific brain structures.  Atrophy of the brainstem is worse in PSP-Richardson syndrome, PSP-CBS and PSP-frontal than in the others.  This information could be useful in treatment trial design because some PSP variants progress faster than others. So, a trial of a disease-slowing treatment could potentially determine which patients have which variants and adjust the statistical analysis of the treatment outcomes accordingly.
    • MRI shows that atrophy in PSP-Richardson’s syndrome starts in the midbrain, (the upper part of the brainstem where the substantia nigra and the vertical gaze centers reside), followed by a succession of other areas of the brainstem, cerebellum and basal ganglia, before finally reaching the cerebral cortex.  But two PSP subtypes involving relatively more cortical function (the frontal behavioral type and the corticobasal syndrome type) spread into the frontal cortex much earlier in the process, although like PSP-RS, they start in the midbrain.  These observations could help guide designers of other imaging-based diagnostic tests for PSP.
  • Measurement of metabolism by fluorodeoxyglucose positron emission tomography (FDG PET):  The pattern of reduced brain tissue metabolic activity in PSP can be distinguished from normal with an AUROC of 0.99.  Distinguishing PSP from corticobasal syndrome, multiple system atrophy and Parkinson’s disease is more difficult but still useful at 0.90.
  • Measurement of tau deposition by PET:  A PET technique that reveals the location of abnormal tau in Alzheimer’s is already in standard clinical use for AD, but it doesn’t work well for other tauopathies.  Two techniques designed for PSP are expected to enter pivotal clinical trials in the next few months.  In small studies, the two ligands called PI-2620 and APN-1607 (formerly called PM-PBB3) have shown good results in distinguishing PSP from the other tau disorders and Parkinsonian disorders.  (A PET ligand is the chemical injected intravenously that sticks to the brain chemical of interest, allowing its location to be mapped.) But these two ligands can occasionally give conflicting results when given to the same patient, so the AUROCs from the upcoming trials are eagerly awaited.
  • Measurement of iron deposition by MRI: This technique, called quantitative susceptibility mapping (QSM), is not part of a routine MRI, but it can be performed with the same machine.  Its PSP-diagnostic results are not quite as accurate of the ones above, with an AUROC of 0.83 for distinguishing PSP-Richardson’s syndrome from Parkinson’s disease.  I assume the AUROCs for other kinds of PSP, especially PSP-Parkinsonism, are worse, but work continues on this technique.
  • Functional MRI: The spread through the brain of abnormal tau, and with it the damage of PSP, proceeds along “functional networks.”  That means that after first affecting the substantia nigra in the midbrain, the damage proceeds to areas most closely yoked to it in terms of simultaneous electrical activity.  Those physiological relationships have been delineated by “functional MRI” technique, where a person is given various mental or motor tasks to perform while in the MRI machine.  The MRI software is set to measure not physical or chemical structure as for routine MRI, but blood flow, which correlates exquisitely with brain tissue electrical activity.  These observations could potentially allow researchers to assess the effects of experimental drugs on the immediate or longer-term pattern of brain cell activation in people with PSP.

Exciting developments all, and I apologize for nerding out on you yet again.  But as one of this blog’s commenters said a few years ago, “Thanks, Dr. Golbe, for respecting our intelligence.”  Still, I’ll try softer stuff next time.

Solving our image problem

Posted on by

Günter Höglinger of Hannover University in Germany is probably the world’s most productive PSP researcher right now.  A few years ago, he organized a PSP Study Group as part of the International Parkinson and Movement Disorder Society.  Most of the 51 members are European – I’m one of the 11 US members.  The PSPSG’s main accomplishment to date is developing and publishing a new set of diagnostic criteria for PSP.  The group meets for a couple of hours in person every year in conjunction with the IPMDS’s conference, but of course, the past two meetings have been on Zoom.  The agenda is to informally discuss our recent research activities and ideas.

This year’s meeting was held on October 4, 2021.  Here’s a boiled-down, edited and explicated version of the proceedings.  The topics were classified into imaging, longitudinal studies, fluid markers and treatment.  As per last week, it will be the installment plan: Each of those four topics will be a separate post here on PSP Blog.

Imaging

James Rowe of Cambridge University, a legitimate rival to Günter as the world’s current leading PSP researcher, described the value that 7-Tesla MRI brings to PSP research.

Most standard MRI scans for medical care use a magnetic field strength of 1.5 Tesla and a growing number use 3 T for additional resolution.  But about 100 research MRI machines world-wide are capable of 7 T imaging.  This provides, for the first time, a clear image of the locus ceruleus (LC), a cylindrical cluster of blue pigmented cells in the brainstem that uses noradrenaline as its neurotransmitter.  It averages 14.5 mm in length but only 2.0 mm in diameter, making it difficult to see with conventional 1.5 T or 3.0 T MRI.  It supplies input to many other brain areas and degenerates in PSP and other neurodegenerative disorders.  Dr. Rowe hopes that the rate of worsening of atrophy of the LC on 7 T MRI may be usable as an outcome measure in PSP neuroprotective treatment trials.

A technique called magnetic resonance spectroscopy (MRS) uses existing MRI machines to provide not an anatomical image, but a measure of levels of some kinds of chemicals in specified areas of brain tissue.  It’s currently used mostly in brain tumor diagnosis. (Side note: MR spectroscopy long antedates MR imaging, which essentially takes MRS measurements of multiple pencil-shaped volumes of tissue sharing a slice of brain and then uses a computer to reconstruct those numbers into a two-dimensional image.) 7T MRI provides greater resolution here as well.  Dr. Rowe reported that he is studying the effects on circuits in the cortex of tiagabine (brand name Gabitril), an approved epilepsy drug that increases levels of the inhibitory neurotransmitter gamma-amino-butyric acid (GABA).  A similar drug is atomoxetine (brand name Strattera), which is approved for attention-deficit hyperactivity disorder.  Dr. Rowe is leading a clinical trial of that drug for disinhibited behavior, apathy and impulsivity in PSP.  A secondary outcome measure in that PSP trial, i.e., one that will not be critical to the study’s conclusions because it’s still an experimental test, is using 7T MRS to assess GABA levels in selected brain areas.   

Adam Boxer of University of California San Francisco, yet another very prolific PSP researcher, described the progress of his NIH-supported project, “4-repeat tau neuroimaging initiative,” or 4RTNI (pronounced “Fortney”).  The study is following patients with PSP or CBS every 6 months using MRI to track atrophy and tau PET to track tau aggregate accumulation.  The study also includes clinical evaluations, plasma levels of Ptau217 (tau with a phosphate group attached at amino acid number 217) and PET scans for beta-amyloid to detect Alzheimer’s disease (AD), which in an atypical form is the pathology underlying many cases of CBS.  The goal is to develop better diagnostic tests and progression markers for use in future PSP and CBD treatment trials.  While the Richardson syndrome clinical picture is almost always explained by underlying PSP pathology, an especially pressing issue is to distinguish CBS caused by CBD pathology (CBS-CBD) from CBS caused by AD pathology (CBS-AD).  Plasma levels of Ptau217 are very high in people with AD pathology, either as classic clinical AD or as CBS-AD, but normal in CBS-CBD and CBS with other pathologies.  A commonly used statistical measure of accuracy, the “area under the receiver operating characteristic (AUC),” for plasma Ptau217 in distinguishing CBS-AD from CBS-CBD is 0.96, very close to the theoretical ideal of 1.0.  However, that’s for advanced cases.  The test’s utility in early cases, where it’s likely to be needed most, is much less so far.  Dr. Boxer tentatively concludes that in distinguishing CBS-CBD from CBS-AD, plasma Ptau217 is almost as accurate as amyloid PET, the current standard, a much more difficult and costly procedure.

Dr. Boxer discussed another project in progress within the 4RTNI umbrella to help distinguish CBS-AD from CBS-CBD or CBS-PSP (i.e., to allow patients with CBS to participate in anti-tau treatment trials).  His research group combined a measure of cortical atrophy with one of midbrain atrophy using a Bayesian logistic regression.  That’s a technique that allows one to create a statistical “model” of a phenomenon, or to dissect its component parts, by successively trying different solutions and tweaking each based on the previous result.  This is different from traditional statistical models, which use event frequencies rather than successive refinements of an a priori hypothesis.  Look it up.  They were able to achieve an AUC for CBS-CBD vs. CBS-PSP to 0.95 for patients presenting with motor signs and 0.91 for those presenting with non-motor signs. 

Next post: Longitudinal PSP studies