A Systematic Review of Parkinson's Disease Cluster Analysis Research

Cluster analysis in fibromyalgia: a systematic review

BACKGROUND

The multifaceted nature of Fibromyalgia syndrome (FM) symptoms has been explored through clusters analysis.

OBJECTIVE

To synthesize the cluster research on FM (variables, methods, patient subgroups, and evaluation metrics).

METHODS

We performed a systematic review following the PRISMA recommendations. Independent searches were performed on PubMed, Embase, Web of Science, and Cochrane Central, employing the terms "fibromyalgia" and "cluster analysis". We included studies dated to January 2024, using the cluster analysis to assess any physical, psychological, clinical, or biomedical variables in FM subjects, and descriptively synthesized the studies in terms of design, cluster method, and resulting patient profiles.

RESULTS

We included 39 studies. Most with a cross-sectional design aiming to classify subsets based on the severity, adjustment, symptomatic manifestations, psychological profiles, and response to treatment, based on demographic and clinical variables. Two to four different profiles were found according to the levels of severity and adjustment to FMS. According to symptom manifestation, two to three clusters described the predominance of pain versus fatigue, and thermal pain sensitivity (less versus more sensitive). Other clusters revealed profiles of personality (pathological versus non-pathological) and psychological vulnerability (suicidal ideation). Additionally, studies identified different responses to treatment (pharmacological and multimodal).

CONCLUSION

Several profiles exist within FMS population, which point out to the need for specific treatment options given the different profiles and an efficient allocation of healthcare resources. We notice a need towards more objective measures, and the validation of the cluster results. Further research might investigate some of the assumptions of these findings, which are further discussed in this paper.

New insights on hidradenitis suppurativa phenotypes and treatment response: An exploratory automated analysis of the SUNSHINE and SUNRISE trials

BACKGROUND

Defining hidradenitis suppurativa (HS) subtypes was previously limited by small sample sizes and poor interrater reliability; no study has investigated subtype treatment responses. The objective of this analysis was to characterize HS clusters in adult patients with moderate to severe HS and evaluate secukinumab treatment responses between clusters.

METHODS

Clusters were identified via an unsupervised machine learning clustering analysis using baseline data from the randomized, placebo-controlled SUNSHINE (NCT03713619) and SUNRISE (NCT03713632) phase 3 trials. To assess treatment responses, patients received secukinumab every 2 (SECQ2W) or 4 weeks (SECQ4W) or placebo, for 16 weeks, after which, placebo patients randomly switched to SECQ2W/SECQ4W, and SECQ2W/SECQ4W patients maintained their original treatment, until week 52. Baseline outcomes included patient characteristics, disease characteristics and severity, HS-associated comorbidities and previous treatment exposures. Treatment response was assessed via the HS clinical response (HiSCR), abscess and inflammatory nodule (AN) count, flares and NRS30 (skin pain).

RESULTS

Based on baseline data, three clusters were identified from 1084 patients (Cluster 1: 54.1%, Cluster 2: 17.8%, Cluster 3: 28.1%). Cluster 1 was predominantly female (65.4%) and was characterized by milder HS. Cluster 2 had more patients from the Asia Pacific, Middle East and Africa region (58.5%) and was characterized by moderate HS. Cluster 3 had the highest rates of previous exposure to biologics (45.9%) and prior HS-related surgeries (47.5%) and was characterized by severe HS. SECQ2W and SECQ4W demonstrated efficacy versus placebo in all clusters at week 16; SECQ2W and SECQ4W efficacy was maintained to week 52. SECQ2W treatment showed a trend for greater efficacy versus SECQ4W in Cluster 3 through week 52.

CONCLUSIONS

Three HS clusters were identified. Secukinumab demonstrated benefit over placebo in all clusters. However, patients with more severe disease may take longer to respond and more frequent secukinumab dosing may be required for these patients.

TRIAL REGISTRATION

SUNSHINE (NCT03713619) and SUNRISE (NCT03713632).

Clusters of cognitive performance predict long-term cognitive impairment in elderly patients with subjective memory complaints and healthy controls

INTRODUCTION

Patients with subjective memory complaints (SMC) may include subgroups with different neuropsychological profiles and risks of cognitive impairment.

METHODS

Cluster analysis was performed on two datasets (n: 630 and 734) comprising demographic and neuropsychological data from SMC and healthy controls (HC). Survival analyses were conducted on clusters. Bayesian model averaging assessed the predictive utility of clusters and other biomarkers.

RESULTS

Two clusters with higher and lower than average cognitive performance were detected in SMC and HC. Assignment to the lower performance cluster increased the risk of cognitive impairment in both datasets (hazard ratios: 1.78 and 2.96; Plog-rank: 0.04 and <0.001) and was associated with lower hippocampal volumes and higher tau/amyloid beta 42 ratios in cerebrospinal fluid. The effect of SMC was small and confounded by mood.

DISCUSSION

This study provides evidence of the presence of cognitive clusters that hold biological significance and predictive value for cognitive decline in SMC and HC.

HIGHLIGHTS

Patients with subjective memory complaints include two cognitive clusters. Assignment to the lower performance cluster increases risk of cognitive impairment. This cluster shows a pattern of biomarkers consistent with incipient Alzheimer's disease pathology. The same cognitive cluster structure is found in healthy controls. The effect of memory complaints on risk of cognitive decline is small and confounded.

Brain network centrality and connectivity are associated with clinical subtypes and disease progression in Parkinson's disease

To investigate brain network centrality and connectivity alterations in different Parkinson's disease (PD) clinical subtypes using resting-state functional magnetic resonance imaging (RS-fMRI), and to explore the correlation between baseline connectivity changes and the clinical progression. Ninety-two PD patients were enrolled at baseline, alongside 38 age- and sex-matched healthy controls. Of these, 85 PD patients underwent longitudinal assessments with a mean of 2.75 ± 0.59 years. Two-step cluster analysis integrating comprehensive motor and non-motor manifestations was performed to define PD subtypes. Degree centrality (DC) and secondary seed-based functional connectivity (FC) were applied to identify brain network centrality and connectivity changes among groups. Regression analysis was used to explore the correlation between baseline connectivity changes and clinical progression. Cluster analysis identified two main PD subtypes: mild PD and moderate PD. Two different subtypes within the mild PD were further identified: mild motor-predominant PD and mild-diffuse PD. Accordingly, the disrupted DC and seed-based FC in the left inferior frontal orbital gyrus and left superior occipital gyrus were severe in moderate PD. The DC and seed-based FC alterations in the right gyrus rectus and right postcentral gyrus were more severe in mild-diffuse PD than in mild motor-predominant PD. Moreover, disrupted DC were associated with clinical manifestations at baseline in patients with PD and predicted motor aspects progression over time. Our study suggested that brain network centrality and connectivity changes were different among PD subtypes. RS-fMRI holds promise to provide an objective assessment of subtype-related connectivity changes and predict disease progression in PD.

Beyond shallow feelings of complex affect: Non-motor correlates of subjective emotional experience in Parkinson's disease

Affective disorders in Parkinson's disease (PD) concern several components of emotion. However, research on subjective feeling in PD is scarce and has produced overall varying results. Therefore, in this study, we aimed to evaluate the subjective emotional experience and its relationship with autonomic symptoms and other non-motor features in PD patients. We used a battery of film excerpts to elicit Amusement, Anger, Disgust, Fear, Sadness, Tenderness, and Neutral State, in 28 PD patients and 17 healthy controls. Self-report scores of emotion category, intensity, and valence were analyzed. In the PD group, we explored the association between emotional self-reported scores and clinical scales assessing autonomic dysregulation, depression, REM sleep behavior disorder, and cognitive impairment. Patient clustering was assessed by considering relevant associations. Tenderness occurrence and intensity of Tenderness and Amusement were reduced in the PD patients. Tenderness occurrence was mainly associated with the overall cognitive status and the prevalence of gastrointestinal symptoms. In contrast, the intensity and valence reported for the experience of Amusement correlated with the prevalence of urinary symptoms. We identified five patient clusters, which differed significantly in their profile of non-motor symptoms and subjective feeling. Our findings further suggest the possible existence of a PD phenotype with more significant changes in subjective emotional experience. We concluded that the subjective experience of complex emotions is impaired in PD. Non-motor feature grouping suggests the existence of disease phenotypes profiled according to specific deficits in subjective emotional experience, with potential clinical implications for the adoption of precision medicine in PD. Further research on larger sample sizes, combining subjective and physiological measures of emotion with additional clinical features, is needed to extend our findings.

Two-year clinical progression in focal and diffuse subtypes of Parkinson's disease

Heterogeneity in Parkinson's disease (PD) presents a barrier to understanding disease mechanisms and developing new treatments. This challenge may be partially overcome by stratifying patients into clinically meaningful subtypes. A recent subtyping scheme classifies de novo PD patients into three subtypes: mild-motor predominant, intermediate, or diffuse-malignant, based on motor impairment, cognitive function, rapid eye movement sleep behavior disorder (RBD) symptoms, and autonomic symptoms. We aimed to validate this approach in a large longitudinal cohort of early-to-moderate PD (n = 499) by assessing the influence of subtyping on clinical characteristics at baseline and on two-year progression. Compared to mild-motor predominant patients (42%), diffuse-malignant patients (12%) showed involvement of more clinical domains, more diffuse hypokinetic-rigid motor symptoms (decreased lateralization and hand/foot focality), and faster two-year progression. These findings extend the classification of diffuse-malignant and mild-motor predominant subtypes to early-to-moderate PD and suggest that different pathophysiological mechanisms (focal versus diffuse cerebral propagation) may underlie distinct subtype classifications.

QEEG Signatures are Associated with Nonmotor Dysfunctions in Parkinson's Disease and Atypical Parkinsonism: An Integrative Analysis

Parkinson's disease (PD) and atypical parkinsonism (AP), including progressive supranuclear palsy (PSP) and multiple system atrophy (MSA), share similar nonmotor symptoms. Quantitative electroencephalography (QEEG) can be used to examine the nonmotor symptoms. This study aimed to characterize the patterns of QEEG and functional connectivity (FC) that differentiate PD from PSP or MSA, and explore the correlation between the differential QEEG indices and nonmotor dysfunctions in PD and AP. We enrolled 52 patients with PD, 31 with MSA, 22 with PSP, and 50 age-matched health controls to compare QEEG indices among specific brain regions. One-way analysis of variance was applied to assess QEEG indices between groups; Spearman's correlations were used to examine the relationship between QEEG indices and nonmotor symptoms scale (NMSS) and mini-mental state examination (MMSE). FCs using weighted phase lag index were compared between patients with PD and those with MSA/PSP. Patients with PSP revealed higher scores on the NMSS and lower MMSE scores than those with PD and MSA, with similar disease duration. The delta and theta powers revealed a significant increase in PSP, followed by PD and MSA. Patients with PD presented a significantly lower slow-to-fast ratio than those with PSP in the frontal region, while patients with PD presented significantly higher EEG-slowing indices than patients with MSA. The frontal slow-to-fast ratio showed a negative correlation with MMSE scores in patients with PD and PSP, and a positive correlation with NMSS in the perception and mood domain in patients with PSP but not in those with PD. Compared to PD, MSA presented enhanced FC in theta and delta bands in the posterior region, while PSP revealed decreased FC in the delta band within the frontal-temporal cortex. These findings suggest that QEEG might be a useful tool for evaluating the nonmotor dysfunctions in PD and AP. Our QEEG results suggested that with similar disease duration, the cortical neurodegenerative process was likely exacerbated in patients with PSP, followed by those with PD, and lastly in patients with MSA.

Zgli: A Pipeline for Clustering by Compression with Application to Patient Stratification in Spondyloarthritis

The normalized compression distance (NCD) is a similarity measure between a pair of finite objects based on compression. Clustering methods usually use distances (e.g., Euclidean distance, Manhattan distance) to measure the similarity between objects. The NCD is yet another distance with particular characteristics that can be used to build the starting distance matrix for methods such as hierarchical clustering or K-medoids. In this work, we propose Zgli, a novel Python module that enables the user to compute the NCD between files inside a given folder. Inspired by the CompLearn Linux command line tool, this module iterates on it by providing new text file compressors, a new compression-by-column option for tabular data, such as CSV files, and an encoder for small files made up of categorical data. Our results demonstrate that compression by column can yield better results than previous methods in the literature when clustering tabular data. Additionally, the categorical encoder shows that it can augment categorical data, allowing the use of the NCD for new data types. One of the advantages is that using this new feature does not require knowledge or context of the data. Furthermore, the fact that the new proposed module is written in Python, one of the most popular programming languages for machine learning, potentiates its use by developers to tackle problems with a new approach based on compression. This pipeline was tested in clinical data and proved a promising computational strategy by providing patient stratification via clusters aiding in precision medicine.

Subtyping of early-onset Parkinson's disease using cluster analysis: A large cohort study

BACKGROUND

Increasing evidence suggests that early-onset Parkinson's disease (EOPD) is heterogeneous in its clinical presentation and progression. Defining subtypes of EOPD is needed to better understand underlying mechanisms, predict disease course, and eventually design more efficient personalized management strategies.

OBJECTIVE

To identify clinical subtypes of EOPD, assess the clinical characteristics of each EOPD subtype, and compare the progression between EOPD subtypes.

MATERIALS AND METHODS

A total of 1,217 patients were enrolled from a large EOPD cohort of the Parkinson's Disease & Movement Disorders Multicenter Database and Collaborative Network in China (PD-MDCNC) between January 2017 and September 2021. A comprehensive spectrum of motor and non-motor features were assessed at baseline. Cluster analysis was performed using data on demographics, motor symptoms and signs, and other non-motor manifestations. In 454 out of total patients were reassessed after a mean follow-up time of 1.5 years to compare progression between different subtypes.

RESULTS

Three subtypes were defined: mild motor and non-motor dysfunction/slow progression, intermediate and severe motor and non-motor dysfunction/malignant. Compared to patients with mild subtype, patients with the severe subtype were more likely to have rapid eye movement sleep behavior disorder, wearing-off, and dyskinesia, after adjusting for age and disease duration at baseline, and showed a more rapid progression in Unified Parkinson's Disease Rating Scale (UPDRS) total score (P = 0.002), UPDRS part II (P = 0.014), and III (P = 0.001) scores, Hoehn and Yahr stage (P = 0.001), and Parkinson's disease questionnaire-39 item version score (P = 0.012) at prospective follow-up.

CONCLUSION

We identified three different clinical subtypes (mild, intermediate, and severe) using cluster analysis in a large EOPD cohort for the first time, which is important for tailoring therapy to individuals with EOPD.

Stability of motor-nonmotor subtype in early-stage Parkinson’s disease

Background

The different clinical characteristics and prognostic values of the motor-nonmotor subtypes of Parkinson’s disease (PD) have been established by previous studies. However, the consistency of motor-nonmotor subtypes in patients with early-stage Parkinson’s disease required further investigation. The present study aimed to evaluate the consistency of motor-nonmotor subtypes across five years of follow-up in a longitudinal cohort.

Materials and methods

Patients were classified into different subtypes (mild-motor–predominant, intermediate, diffuse malignant; or tremor-dominant, indeterminate, postural instability and gait difficulty) according to previously verified motor-nonmotor and motor subtyping methods at baseline and at every year of follow-up. The agreement between subtypes was examined using Cohen’s kappa and total agreement. The determinants of having the diffuse malignant subtype as of the fifth-year visit were explored using logistic regression.

Results

A total of 421 patients were included. There was a fair degree of agreement between the baseline motor-nonmotor subtype and the subtype recorded at the one-year follow-up visit (κ = 0.30 ± 0.09; total agreement, 60.6%) and at following years’ visits. The motor-nonmotor subtype had a lower agreement between baseline and follow-up than did the motor subtype. The baseline motor-nonmotor subtype was the determinant of diffuse malignant subtype at the fifth-year visit.

Conclusion

Many patients experienced a change in their motor-nonmotor subtype during follow-up. Further studies of consistency in PD subtyping methods should be conducted in the future.

Clusters of conditions among US service members diagnosed with mild TBI from 2017 through 2019

Background

Many US Military Service Members (SMs) newly diagnosed with mild Traumatic Brain Injury (mTBI) may exhibit a range of symptoms and comorbidities, making for a complex patient profile that challenges clinicians and healthcare administrators. This study used clustering techniques to determine if conditions co-occurred as clusters among those newly injured with mTBI and up to one year post-injury.

Methods

We measured the co-occurrence of 41 conditions among SMs diagnosed with mTBI within the acute phase, one or three months post-mTBI diagnosis, and chronic phase, one year post-mTBI diagnosis. Conditions were identified from the literature, clinical subject matter experts, and mTBI care guidelines. The presence of conditions were based on medical encounters recorded within the military health care data system. Through a two-step approach, we identified clusters. Principal component analysis (PCA) determined the optimal number of clusters, and hierarchical cluster analyses (HCA) identified the composition of clusters. Further, we explored how the composition of these clusters changed over time.

Results

Of the 42,018 SMs with mTBI, 23,478 (55.9%) had at least one condition of interest one-month post-injury, 26,831 (63.9%) three months post-injury, and 29,860 (71.1%) one year post injury. Across these three periods, six clusters were identified. One cluster included vision, cognitive, ear, and sleep disorders that occurred one month, three months, and one year post-injury. Another subgroup included psychological conditions such as anxiety, depression, PTSD, and other emotional symptoms that co-occurred in the acute and chronic phases post-injury. Nausea and vomiting symptoms clustered with cervicogenic symptoms one month post-injury, but later shifted to other clusters. Vestibular disorders clustered with sleep disorders and headache disorders one-month post-injury and included numbness and neuropathic pain one year post-injury. Substance abuse symptoms, alcohol disorders, and suicidal attempt clustered one year post-injury in a fifth cluster. Speech disorders co-occurred with headache disorders one month and one year post-injury to form a sixth cluster.

Conclusion

PCA and HCA identified six distinct subgroups among newly diagnosed mTBI patients during the acute and chronic phases post-injury. These subgroups may help clinicians better understand the complex profile of SMs newly diagnosed with mTBI.

Efficacy of repetitive transcranial magnetic stimulation in Parkinson's disease: A systematic review and meta-analysis of randomised controlled trials

BACKGROUND

Repetitive transcranial magnetic stimulation (rTMS) is a non-invasive form of brain stimulation that positively regulates the motor and non-motor symptoms of Parkinson's disease (PD). Although, most reviews and meta-analysis have shown that rTMS intervention is effective in treating motor symptoms and depression, very few have used randomised controlled trials (RCTs) to analyse the efficacy of this intervention in PD. We aimed to review RCTs of rTMS in patients with PD to assess the efficacy of rTMS on motor and non-motor function in patients with PD.

METHODS

In this systematic review and meta-analysis, we searched PubMed, MEDLINE and Web of Science databases for RCTs on rTMS in PD published between January 1, 1988 to January 1, 2022. Eligible studies included sham-controlled RCTs that used rTMS stimulation for motor or non-motor symptoms in PD. RCTs not focusing on the efficacy of rTMS in PD were excluded. Summary data were extracting from those RCTs by two investigators independently. We then calculated standardised mean difference with random-effect models. The main outcome included motor and non-motor examination of scales that were used in PD motor or non-motor assessment. This study was registered with PROSPERO, CRD42022329633.

FINDINGS

Fourteen studies with 469 patients met the criteria for our meta-analysis. Twelve eligible studies with 381 patients were pooled to analyse the efficacy of rTMS on motor function improvement. The effect size on motor scale scores was 0.51 (P < 0.0001) and were not distinctly heterogeneous (I² = 29%). Five eligible studies with 202 patients were collected to evaluate antidepressant-like effects. The effect size on depression scale scores was 0.42 (P = 0.004), and were not distinctly heterogeneous (I² = 25%), indicating a significant anti-depressive effect (P = 0.004). The results suggest that high-frequency of rTMS on primary motor cortex (M1) is effective in improving motor symptoms; while the dorsolateral prefrontal cortex (DLPFC) may be a potentially effective area in alleviating depressive symptom.

INTERPRETATION

The findings suggest that rTMS could be used as a possible adjuvant therapy for PD mainly to improve motor symptoms, but could have potential efficacy on depressive symptoms of PD. However, further investigation is needed.

FUNDING

The National Natural Science Foundation of China (NO: 81873777, 82071414), Initiated Foundation of Zhujiang Hospital (NO: 02020318005), Scientific Research Foundation of Guangzhou (NO: 202206010005), and Science and Technology Program of Guangdong of China (NO: 2020A0505100037).

Identifying and validating subtypes of Parkinson's disease based on multimodal MRI data via hierarchical clustering analysis

Objective

We wished to explore Parkinson's disease (PD) subtypes by clustering analysis based on the multimodal magnetic resonance imaging (MRI) indices amplitude of low-frequency fluctuation (ALFF) and gray matter volume (GMV). Then, we analyzed the differences between PD subtypes.

Methods

Eighty-six PD patients and 44 healthy controls (HCs) were recruited. We extracted ALFF and GMV according to the Anatomical Automatic Labeling (AAL) partition using Data Processing and Analysis for Brain Imaging (DPABI) software. The Ward linkage method was used for hierarchical clustering analysis. DPABI was employed to compare differences in ALFF and GMV between groups.

Results

Two subtypes of PD were identified. The “diffuse malignant subtype” was characterized by reduced ALFF in the visual-related cortex and extensive reduction of GMV with severe impairment in motor function and cognitive function. The “mild subtype” was characterized by increased ALFF in the frontal lobe, temporal lobe, and sensorimotor cortex, and a slight decrease in GMV with mild impairment of motor function and cognitive function.

Conclusion

Hierarchical clustering analysis based on multimodal MRI indices could be employed to identify two PD subtypes. These two PD subtypes showed different neurodegenerative patterns upon imaging.

Cluster Analysis of Categorical Variables of Parkinson's Disease Patients

Parkinson’s disease (PD) is a chronic disease. No treatment stops its progression, and it presents symptoms in multiple areas. One way to understand the PD population is to investigate the clustering of patients by demographic and clinical similarities. Previous PD cluster studies included scores from clinical surveys, which provide a numerical but ordinal, non-linear value. In addition, these studies did not include categorical variables, as the clustering method utilized was not applicable to categorical variables. It was discovered that the numerical values of patient age and disease duration were similar among past cluster results, pointing to the need to exclude these values. This paper proposes a novel and automatic discovery method to cluster PD patients by incorporating categorical variables. No estimate of the number of clusters is required as input, whereas the previous cluster methods require a guess from the end user in order for the method to be initiated. Using a patient dataset from the Parkinson’s Progression Markers Initiative (PPMI) website to demonstrate the new clustering technique, our results showed that this method provided an accurate separation of the patients. In addition, this method provides an explainable process and an easy way to interpret clusters and describe patient subtypes.