Related Experiment Video
Updated: Jan 14, 2026
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
Progressive supranuclear palsy against everyone else: a look into the specificity of MRI markers
Vasilios C Constantinides1, Maria-Evgenia Brinia1, Ioanna Kapsali1
1First Department of Neurology, National and Kapodistrian University of Athens, School of Medicine, Eginition Hospital, Athens, Greece.
Intriduction:
Diagnostic accuracy of the clinical diagnosis of progressive supranuclear palsy (PSP) is suboptimal, due to its' clinical heterogeneity. Multiple MRI markers have been introduced in an effort to enhance clinical diagnosis.
Objectives:
This study compared the diagnostic accuracy of linear, planimetric and volumetric MRI markers for PSP diagnosis in different clinical settings.
Materials And Methods:
A total of 414 patients were included, comprising 80 PSP and 334 non-PSP patients (CBD:38; MSA:45; PD:36; AD:65; bvFTD:30; PPA:18; DLB:15; PCA:14; NPH:15; MCI:14; controls:45). PSP patients were stratified based on phenotype (PSP-RS vs. PSP variants) and level of diagnostic certainty (probable vs. possible/suggestive). Midbrain and SCP distances, surfaces and volumes, pons/midbrain (P/M) ratios and Magnetic Resonance Parkinsonism Indices were measured (MRPI). ROC curve analysis was used to identify optimal MRI markers in each clinical setting.
Results:
Midbrain surface was the optimal MRI marker in differentiating PSP from non-PSP patients (AUC = 0.930; sensitivity 88.8 %; specificity 81.1 %), followed by planimetric P/M and MRPI. MRI markers were highly accurate in differentiating PSP from MSA, PD and PCA, moderately accurate in differentiating PSP from AD, bvFTD, DLB, MCI, whereas their diagnostic accuracy was suboptimal in differentiating PSP from CBD, PPA and NPH. MRI markers were more potent in diagnosis PSP-RS over PSP variants and probable over possible/suggestive PSP patients.
Conclusions:
Diagnostic accuracy of PSP MRI markers varies significantly and depends heavily on the clinical setting, PSP phenotype and level of disease certainty. These limitations should be taken into consideration when applying MRI markers in clinical practice.
Insights
Midbrain surface MRI is the best marker for diagnosing progressive supranuclear palsy (PSP). Its accuracy varies based on the specific clinical context and patient characteristics.
Area of Science:
- Neuroimaging
- Neurology
- Radiology
Background:
- Clinical diagnosis of progressive supranuclear palsy (PSP) is challenging due to its heterogeneity.
- Magnetic Resonance Imaging (MRI) markers are being developed to improve PSP diagnosis.
Purpose of the Study:
- To compare the diagnostic accuracy of various MRI markers for PSP.
- To evaluate marker performance across different clinical settings and PSP subtypes.
Main Methods:
- Included 414 patients: 80 PSP and 334 non-PSP controls.
- Measured linear, planimetric, and volumetric MRI markers (e.g., midbrain surface, P/M ratio, MRPI).
- Used ROC curve analysis to determine optimal markers for differentiating PSP.
Main Results:
- Midbrain surface demonstrated the highest accuracy (AUC=0.930) in differentiating PSP from non-PSP.
- MRI markers showed high accuracy for PSP vs. MSA, PD, PCA; moderate for AD, bvFTD, DLB, MCI; and suboptimal for CBD, PPA, NPH.
- Markers were more effective for PSP-Rigid Spine (PSP-RS) and probable PSP.
Conclusions:
- The diagnostic accuracy of MRI markers for PSP is highly variable.
- Accuracy depends on the clinical setting, PSP phenotype (e.g., PSP-RS), and diagnostic certainty.
- Consider these limitations when applying MRI markers in clinical practice.
Related Concept Videos
01:15Parkinson's Disease: Overview
01:37Neural Regulation