Dataset Information

Multiclass classification of FDG PET scans for the distinction between Parkinson's disease and atypical parkinsonian syndromes.

ABSTRACT: Most available pattern recognition methods in neuroimaging address binary classification problems. Here, we used relevance vector machine (RVM) in combination with booststrap resampling ('bagging') for non-hierarchical multiclass classification. The method was tested on 120 cerebral (18)fluorodeoxyglucose (FDG) positron emission tomography (PET) scans performed in patients who exhibited parkinsonian clinical features for 3.5 years on average but that were outside the prevailing perception for Parkinson's disease (PD). A radiological diagnosis of PD was suggested for 30 patients at the time of PET imaging. However, at follow-up several years after PET imaging, 42 of them finally received a clinical diagnosis of PD. The remaining 78 APS patients were diagnosed with multiple system atrophy (MSA, N = 31), progressive supranuclear palsy (PSP, N = 26) and corticobasal syndrome (CBS, N = 21), respectively. With respect to this standard of truth, classification sensitivity, specificity, positive and negative predictive values for PD were 93% 83% 75% and 96%, respectively using binary RVM (PD vs. APS) and 90%, 87%, 79% and 94%, respectively, using multiclass RVM (PD vs. MSA vs. PSP vs. CBS). Multiclass RVM achieved 45%, 55% and 62% classification accuracy for, MSA, PSP and CBS, respectively. Finally, a majority confidence ratio was computed for each scan on the basis of class pairs that were the most frequently assigned by RVM. Altogether, the results suggest that automatic multiclass RVM classification of FDG PET scans achieves adequate performance for the early differentiation between PD and APS on the basis of cerebral FDG uptake patterns when the clinical diagnosis is felt uncertain. This approach cannot be recommended yet as an aid for distinction between the three APS classes under consideration.

SUBMITTER: Garraux G

PROVIDER: S-EPMC3778264 | biostudies-literature | 2013

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Multiclass classification of FDG PET scans for the distinction between Parkinson's disease and atypical parkinsonian syndromes.

Garraux Gaëtan G Phillips Christophe C Schrouff Jessica J Kreisler Alexandre A Lemaire Christian C Degueldre Christian C Delcour Christian C Hustinx Roland R Luxen André A Destée Alain A Salmon Eric E

NeuroImage. Clinical 20130614

Most available pattern recognition methods in neuroimaging address binary classification problems. Here, we used relevance vector machine (RVM) in combination with booststrap resampling ('bagging') for non-hierarchical multiclass classification. The method was tested on 120 cerebral (18)fluorodeoxyglucose (FDG) positron emission tomography (PET) scans performed in patients who exhibited parkinsonian clinical features for 3.5 years on average but that were outside the prevailing perception for Pa ...[more]

PMID: 24179839

Similar Datasets

Project description:Parkinson's disease (PD) and atypical Parkinsonian syndromes are progressive heterogeneous neurodegenerative diseases that share clinical characteristic of parkinsonism as a common feature, but are considered distinct clinicopathological disorders. Based on the predominant protein aggregates observed within the brain, these disorders are categorized as, (1) α-synucleinopathies, which include PD and other Lewy body spectrum disorders as well as multiple system atrophy, and (2) tauopathies, which comprise progressive supranuclear palsy and corticobasal degeneration. Although, great strides have been made in neurodegenerative disease research since the first medical description of PD in 1817 by James Parkinson, these disorders remain a major diagnostic and treatment challenge. A valid diagnosis at early disease stages is of paramount importance, as it can help accommodate differential prognostic and disease management approaches, enable the elucidation of reliable clinicopathological relationships ideally at prodromal stages, as well as facilitate the evaluation of novel therapeutics in clinical trials. However, the pursuit for early diagnosis in PD and atypical Parkinsonian syndromes is hindered by substantial clinical and pathological heterogeneity, which can influence disease presentation and progression. Therefore, reliable neuroimaging biomarkers are required in order to enhance diagnostic certainty and ensure more informed diagnostic decisions. In this article, an updated presentation of well-established and emerging neuroimaging biomarkers are reviewed from the following modalities: (1) structural magnetic resonance imaging (MRI), (2) diffusion-weighted and diffusion tensor MRI, (3) resting-state and task-based functional MRI, (4) proton magnetic resonance spectroscopy, (5) transcranial B-mode sonography for measuring substantia nigra and lentiform nucleus echogenicity, (6) single photon emission computed tomography for assessing the dopaminergic system and cerebral perfusion, and (7) positron emission tomography for quantifying nigrostriatal functions, glucose metabolism, amyloid, tau and α-synuclein molecular imaging, as well as neuroinflammation. Multiple biomarkers obtained from different neuroimaging modalities can provide distinct yet corroborative information on the underlying neurodegenerative processes. This integrative "multimodal approach" may prove superior to single modality-based methods. Indeed, owing to the international, multi-centered, collaborative research initiatives as well as refinements in neuroimaging technology that are currently underway, the upcoming decades will mark a pivotal and exciting era of further advancements in this field of neuroscience.

Project description:Background and objectivesImpulse control disorders (ICDs) are characterized by potentially harmful actions resulting from disturbances in the self-control of emotions and behavior. ICDs include disorders such as gambling, hypersexuality, binge eating, and compulsive buying. ICDs are known non-motor symptoms in Parkinson's disease (PD) and are associated primarily with the use of dopaminergic treatment (DRT) and especially dopamine agonists (DA). However, in atypical parkinsonism (APS), such as progressive supranuclear palsy (PSP) or multiple system atrophy (MSA), there are only single case reports of ICDs without attempts to determine the risk factors for their occurrence. Moreover, numerous reports in the literature indicate increased impulsivity in PSP. Our study aimed to determine the frequency of individual ICDs in APS compared to PD and identify potential factors for developing ICDs in APS.Materials and methodsOur prospective study included 185 patients with PD and 35 with APS (27 patients with PSP and 9 with MSA) hospitalized between 2020 and 2023 at the Neurological Department of University Central Hospital in Katowice. Each patient was examined using the Questionnaire for Impulsive-Compulsive Disorders in Parkinson's Disease (QUIP) to assess ICDs. Additionally, other scales were used to assess the advancement of the disease, the severity of depression, and cognitive impairment. Information on age, gender, age of onset, disease duration, and treatment used were collected from medical records and patient interviews.ResultsICDs were detected in 23.39% of patients with PD (including binge eating in 11.54%, compulsive buying in 10.44%, hypersexuality in 8.79%, and pathological gambling in 4.40%), in one patient with MSA (hypersexuality and pathological gambling), and in 18.52% of patients with PSP (binge eating in 3.70%, compulsive buying in 7.41%, and hypersexuality in 11.11%). We found no differences in the frequency of ICDs between individual diseases (p = 0.4696). We confirmed that the use of higher doses of DA and L-dopa in patients with PD, as well as a longer disease duration and the presence of motor complications, were associated with a higher incidence of ICDs. However, we did not find any treatment effect on the incidence of ICDs in APS.ConclusionsICDs are common and occur with a similar frequency in PD and APS. Well-described risk factors for ICDs in PD, such as the use of DRT or longer disease duration, are not fully reflected in the risk factors for ICDs in APS. This applies especially to PSP, which, unlike PD and MSA, is a tauopathy in which, in addition to the use of DRT, other mechanisms related to the disease, such as disorders in neuronal loops and neurotransmitter deficits, may influence the development of ICDs. Further prospective multicenter studies recruiting larger groups of patients are needed to fully determine the risk factors and mechanisms of ICD development in APS.

Dataset Information

Multiclass classification of FDG PET scans for the distinction between Parkinson's disease and atypical parkinsonian syndromes.

Publications

Multiclass classification of FDG PET scans for the distinction between Parkinson's disease and atypical parkinsonian syndromes.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets