Identifying and predicting Parkinson’s disease subtypes through trajectory clustering via bipartite networks

Krishnagopal, Sanjukta; Coelln, Rainer von; Shulman, Lisa M.; Girvan, Michelle

doi:10.1371/journal.pone.0233296

Cited by 23 publications

(32 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Trajectory Profile Clustering algorithm [24] is designed to group together patients based on the similarities of their disease trajectories. In essence, it uses graphical tools to generate trajectory profiles for each individual that track their evolution of symptoms across time, then clusters them into communities of similarly behaving individuals that define a recovery subtype.…”

Section: Methodsmentioning

confidence: 99%

“…Recently, Krishnagopal et al [24] introduced a network-based approach called Trajectory Profile Clustering (TPC) that groups patients based on similar patterns of symptom evolution. The intuitiveness and ability of TPC to integrate variables on multiple different scales make it especially useful for studying disease severity, progression, and recovery.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Stroke recovery phenotyping through network trajectory approaches and graph neural networks

Krishnagopal

Lohse

Braun

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Stroke is a leading cause of neurological injury characterized by impairments in multiple neurological domains including cognition, language, sensory and motor functions. Clinical recovery in these domains is tracked using a wide range of measures that may be continuous, ordinal, interval or categorical in nature, which presents challenges for standard multivariate regression approaches. This has hindered stroke researchers' ability to achieve an integrated picture of the complex time-evolving interactions amongst symptoms. Here we use tools from network science and machine learning that are particularly well-suited to extracting underlying patterns in such data, and may assist in prediction of recovery patterns. To demonstrate the utility of this approach, we analyzed data from the NINDS tPA trial using the Trajectory Profile Clustering (TPC) method to identify distinct stroke recovery patterns for 11 different neurological domains at 5 discrete time points. Our analysis identified 3 distinct stroke trajectory profiles that align with clinically relevant stroke syndromes, characterized both by distinct clusters of symptoms, as well as differing degrees of symptom severity. We then validated our approach using graph neural networks to determine how well our model performed predictively for stratifying patients into these trajectory profiles at early vs. later time points post-stroke. We demonstrate that trajectory profile clustering is an effective method for identifying clinically relevant recovery subtypes in multidimensional longitudinal datasets, and for early prediction of symptom progression subtypes in individual patients. This paper is the first work introducing network trajectory approaches for stroke recovery phenotyping, and is aimed at enhancing the translation of such novel computational approaches for practical clinical application.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Stroke recovery phenotyping through network trajectory approaches and graph neural networks

Krishnagopal

Lohse

Braun

2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Categorical variables of gender, family history, and dominant symptom side were reviewed as part of post-analysis of the clusters in a limited number of studies. Only one published study, [ 30 ], included gender in the clustering.…”

Section: Patient Variables Included In Clusteringmentioning

confidence: 99%

“…Prior to the Z-transformations in [ 14 ], the assumption was that disease feature severities increase with longer disease durations, and hence, each clinical variable was adjusted for disease duration by obtaining its residual value from a linear regression with the clinical feature as the dependent variable and disease duration as the independent variable. In study [ 30 ], data was transformed data such that for each non-binary variable, a direction was determined in that higher values were associated with greater disease severity, defining its direction as = +1, otherwise direction = -1.…”

Section: Pre-processing and Reduction Of Variablesmentioning

confidence: 99%

A Systematic Review of Parkinson’s Disease Cluster Analysis Research

Hendricks

Khasawneh

2021

Aging and disease

View full text Add to dashboard Cite

One way to understand the Parkinson’s disease (PD) population is to investigate the similarities and differences among patients through cluster analysis, which may lead to defined, patient subgroups for diagnosis, progression tracking and treatment planning. This paper provides a systematic review of PD patient clustering research, evaluating the variables included in clustering, the cluster methods applied, the resulting patient subgroups, and evaluation metrics. A search was conducted from 1999 to 2021 on the PubMed database, using various search terms including: Parkinson’s disease, cluster, and analysis. The majority of studies included a variety of clinical scale scores for clustering, of which many provide a numerical, but ordinal, categorical value. Even though the scale scores are ordinal, these were treated as numerical values with numerical and continuous values being the focus of the clustering, with limited attention to categorical variables, such as gender and family history, which may also provide useful insights into disease diagnosis, progression, and treatment. The results pointed to two to five patient clusters, with similarities among the age of onset and disease duration. The studies lacked the use of existing clustering evaluation metrics which points to a need for a thorough, analysis framework, and consensus on the appropriate variables to include in cluster analysis. Accurate cluster analysis may assist with determining if PD patients’ symptoms can be treated based on a subgroup of features, if personalized care is required, or if a mix of individualized and group-based care is the best approach.

show abstract

“…In this context, researchers have turned to data-driven perspectives, where patient subtyping is transformed into a typical clustering problem 11 – 16 . These works have focused on several different types of data collected from patients besides just clinical assessments, which include neuroimaging data 17 – 19 , genomic data 20 , and neurophysiological assessment data 21 .…”

Section: Introductionmentioning

confidence: 99%

Comprehensive subtyping of Parkinson’s disease patients with similarity fusion: a case study with BioFIND data

Brendel

Hou

et al. 2021

npj Parkinsons Dis.

View full text Add to dashboard Cite

Parkinson’s disease (PD) is a complex neurodegenerative disorder with diverse clinical manifestations. To better understand this disease, research has been done to categorize, or subtype, patients, using an array of criteria derived from clinical assessments and biospecimen analyses. In this study, using data from the BioFIND cohort, we aimed at identifying subtypes of moderate-to-advanced PD via comprehensively considering motor and non-motor manifestations. A total of 103 patients were included for analysis. Through the use of a patient-wise similarity matrix fusion technique and hierarchical agglomerative clustering analysis, three unique subtypes emerged from the clustering results. Subtype I, comprised of 60 patients (~58.3%), was characterized by mild symptoms, both motor and non-motor. Subtype II, comprised of 20 (~19.4%) patients, was characterized by an intermediate severity, with a high tremor score and mild non-motor symptoms. Subtype III, comprised of 23 (~22.3%) patients, was characterized by more severe motor and non-motor symptoms. These subtypes show statistically significant differences when looking at motor (on and off medication) clinical features and non-motor clinical features, while there was no clear difference in demographics, biomarker levels, and genetic risk scores.

show abstract

Identifying and predicting Parkinson’s disease subtypes through trajectory clustering via bipartite networks

Cited by 23 publications

References 32 publications

Stroke recovery phenotyping through network trajectory approaches and graph neural networks

Stroke recovery phenotyping through network trajectory approaches and graph neural networks

A Systematic Review of Parkinson’s Disease Cluster Analysis Research

Comprehensive subtyping of Parkinson’s disease patients with similarity fusion: a case study with BioFIND data

Contact Info

Product

Resources

About