Differentiation of neurodegenerative parkinsonian syndromes by volumetric magnetic resonance imaging analysis and support vector machine classification

Huppertz, Hans‐Jürgen; Möller, Leona; Südmeyer, Martin; Hilker, Rüdiger; Hattingen, Elke; Egger, Karl; Amtage, Florian; Respondek, Gesine; Stamelou, María; Schnitzler, Alfons; Pinkhardt, Elmar H.; Oertel, Wolfgang H.; Knake, Susanne; Kassubek, Jan; Höglinger, Günter U.

doi:10.1002/mds.26715

Cited by 132 publications

(190 citation statements)

References 63 publications

Supporting

Mentioning

179

Contrasting

Unclassified

Order By: Relevance

“…However, other kernels might be more suitable, leading to higher accuracy rates and an improved sensitivity and specificity. In a recent study, Huppertz et al (2016) performed SVM classification using a radial basis function (RBF) kernel. However, the extraction of SVM weighting factors is mathematically only defined for the linear kernel and not for an RBF kernel.…”

Section: Discussionmentioning

confidence: 99%

“…However, the extraction of SVM weighting factors is mathematically only defined for the linear kernel and not for an RBF kernel. Therefore, the SVM approach was repeated with a linear kernel to identify the most relevant regions for classification (Huppertz et al, 2016). Motivated by a recent paper demonstrating the advantage of using a polynomial kernel showing an improved accuracy when dissociating mild cognitive impairment from Alzheimer's disease (Belmokhtar and Benamrane, 2012), we also used a polynomial kernel for PSP disease classification.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Disease-Specific Regions Outperform Whole-Brain Approaches in Identifying Progressive Supranuclear Palsy: A Multicentric MRI Study

et al. 2017

Self Cite

View full text Add to dashboard Cite

To identify progressive supranuclear palsy (PSP), we combined voxel-based morphometry (VBM) and support vector machine (SVM) classification using disease-specific features in multicentric magnetic resonance imaging (MRI) data. Structural brain differences were investigated at four centers between 20 patients with PSP and 20 age-matched healthy controls with T1-weighted MRI at 3T. To pave the way for future application in personalized medicine, we applied SVM classification to identify PSP on an individual level besides group analyses based on VBM. We found a major decline in gray matter density in the brainstem, insula, and striatum, and also in frontomedian regions, which is in line with current literature. Moreover, SVM classification yielded high accuracy rates above 80% for disease identification in imaging data. Focusing analyses on disease-specific regions-of-interest (ROI) led to higher accuracy rates compared to a whole-brain approach. Using a polynomial kernel (instead of a linear kernel) led to an increased sensitivity and a higher specificity of disease detection. Our study supports the application of MRI for individual diagnosis of PSP, if combined with SVM approaches. We demonstrate that SVM classification provides high accuracy rates in multicentric data—a prerequisite for potential application in diagnostic routine.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Disease-Specific Regions Outperform Whole-Brain Approaches in Identifying Progressive Supranuclear Palsy: A Multicentric MRI Study

et al. 2017

Self Cite

View full text Add to dashboard Cite

show abstract

“…Quantitative assessment with MR volumetry using region-of-interests (ROI) in patients with MSA showed atrophy of the putamen, caudate, brainstem and cerebellum (Burk et al, 2004; Ghaemi et al, 2002; Huppertz et al, 2016; Sako et al, 2014; Schulz et al, 1999). The MR Parkinsonism Index (MRPI), taking into consideration the volume of the pons, midbrain, and cerebellar peduncles, appears to have high sensitivity and specificity to distinguish between PSP and MSA-P or PD (Hussl et al, 2010; Quattrone et al, 2008).…”

Section: Brain and Cardiac Neuroimagingmentioning

confidence: 99%

“…Age at onset, prevalence of cardiovascular autonomic dysfunction, sleep disorders, and retinal abnormalities are similar in both phenotypes (Mendoza-Santiesteban et al, 2015; Palma et al, 2015; Roncevic et al, 2014). Specific neuroimaging markers differ between the cerebellar and parkinsonian phenotypes (Deguchi et al, 2015; Huppertz et al, 2016; Lee et al, 2015), as well as the degree of sudomotor dysfunction which may be more severe in patients with MSA-P (Coon et al, 2016) and urogenital dysfunction which may occur earlier in patients with MSA-C (Zheng et al, 2017). …”

Section: Introductionmentioning

confidence: 99%

Diagnosis of multiple system atrophy

Palma

Norcliffe‐Kaufmann

Kaufmann

2018

Autonomic Neuroscience

126

114

View full text Add to dashboard Cite

Multiple system atrophy (MSA) may be difficult to distinguish clinically from other disorders, particularly in the early stages of the disease. An autonomic-only presentation can be indistinguishable from pure autonomic failure. Patients presenting with parkinsonism may be misdiagnosed as having Parkinson disease. Patients presenting with the cerebellar phenotype of MSA can mimic other adult-onset ataxias due to alcohol, chemotherapeutic agents, lead, lithium, and toluene, or vitamin E deficiency, as well as paraneoplastic, autoimmune, or genetic ataxias. A careful medical history and meticulous neurological examination remain the cornerstone for the accurate diagnosis of MSA. Ancillary investigations are helpful to support the diagnosis, rule out potential mimics, and define therapeutic strategies. This review summarizes diagnostic investigations useful in the differential diagnosis of patients with suspected MSA. Currently used techniques include structural and functional brain imaging, cardiac sympathetic imaging, cardiovascular autonomic testing, olfactory testing, sleep study, urological evaluation, and dysphagia and cognitive assessments. Despite advances in the diagnostic tools for MSA in recent years and the availability of consensus criteria for clinical diagnosis, the diagnostic accuracy of MSA remains sub-optimal. As other diagnostic tools emerge, including skin biopsy, retinal biomarkers, blood and cerebrospinal fluid biomarkers, and advanced genetic testing, a more accurate and earlier recognition of MSA should be possible, even in the prodromal stages. This has important implications as misdiagnosis can result in inappropriate treatment, patient and family distress, and erroneous eligibility for clinical trials of disease-modifying drugs.

show abstract

“…In patients with parkinsonism, disease-specific volume loss compared to controls was demonstrated in pontocerebellar structures including the middle cerebellar peduncles for MSA-C and mainly in the putamen for MSA-P, while PSP patients presented pronounced atrophy in the midbrain, the midsagittal midbrain tegmentum plane, and superior cerebellar peduncles and PD patients showed a subtle gray matter (GM) reduction modestly peaking in the caudate nucleus [12,13]. …”

Section: Introductionmentioning

confidence: 99%

Patterns of Eye Movement Impairment Correlate with Regional Brain Atrophy in Neurodegenerative Parkinsonism

Vintonyak

Gorges²,

Müller³

et al. 2017

Neurodegener Dis

Self Cite

View full text Add to dashboard Cite

Background: One common feature of neurodegenerative parkinsonism including Parkinson's disease (PD), multisystem atrophy (MSA), and progressive supranuclear palsy (PSP) is altered eye movement control. Characteristic regional structural atrophy patterns in MRI can be observed in PD, MSA, and PSP. Objective: To investigate the association between eye movement disturbances and regional brain atrophy in patients with PD, MSA, and PSP. Methods: High-resolution 3-dimensional T1-weighted MRI images and video-oculographic recordings (EyeLink®) were obtained from 39 PD, 32 PSP, and 18 MSA patients and 24 matched healthy control subjects. Automatic regional volumetric assessment was performed using atlas-based volumetry (ABV). Results: The prevalence of saccadic intrusions as a measure of inhibitory control was significantly increased in PD patients compared to controls (p < 0.001) and negatively correlated with whole brain volume, cerebral brain volume, and occipital lobe volume (p = 0.0057, p = 0.0049, and p = 0.0059, respectively; all p values are false discovery rate corrected). In MSA, smooth pursuit was disturbed by characteristic “catch-up” saccades (p < 0.001) and it was significantly correlated with cerebellar volume (p = 0.004) and pontine volume (p < 0.001). The hallmark of PSP was pathologically slowed vertical peak eye velocities (p < 0.001); the lower the peak eye velocity, the more marked midbrain atrophy (p = 0.007). Conclusions: Foci of regional atrophy correlated with disease-specific eye movement alterations in all investigated parkinsonian syndromes. Oculomotor impairment in PD, predominantly the result of executive dysfunction, was linked to cerebral atrophy. Impairment in the corresponding oculomotor pathways was associated with atrophy of pontocerebellar oculomotor structures in MSA and midbrain atrophy in PSP.

show abstract

Differentiation of neurodegenerative parkinsonian syndromes by volumetric magnetic resonance imaging analysis and support vector machine classification

Abstract: Brain volumetry combined with support vector machine classification allowed for reliable automated differentiation of parkinsonian syndromes on single-patient level even for MRI acquired on different scanners. © 2016 International Parkinson and Movement Disorder Society.

Cited by 132 publications

References 63 publications

Disease-Specific Regions Outperform Whole-Brain Approaches in Identifying Progressive Supranuclear Palsy: A Multicentric MRI Study

Disease-Specific Regions Outperform Whole-Brain Approaches in Identifying Progressive Supranuclear Palsy: A Multicentric MRI Study

Diagnosis of multiple system atrophy

Patterns of Eye Movement Impairment Correlate with Regional Brain Atrophy in Neurodegenerative Parkinsonism

Contact Info

Product

Resources

About