Cluster Analysis Of Patients Research Articles

Serum potassium levels and in-hospital mortality in distinct phenotypes of myocardial infarction patients by machine learning clustering

Abstract Background Abnormal serum potassium levels are associated with an increased risk of mortality in patients with myocardial infarction (MI); however, it is unclear whether the impact of potassium levels on mortality risk varies depending on the characteristics of MI patients. Purpose We aimed to classify MI patients according to their clinical characteristics and explore the relationship between potassium levels and in-hospital mortality in distinct clusters of MI. Methods Using Medical Information Mart for Intensive Care (MIMIC) IV database, we identified 3,159 patients who admitted for the first time to the intensive care unit with a primary diagnosis of acute MI. Machine learning clustering was performed on this population based on available clinical and laboratory data. Results Three distinct clusters of MI patients were identified using clustering analysis: 1,455 (46%) in cluster 1, 587 (19%) in cluster 2, and 1,117 (35%) in cluster 3. Patients in cluster 2 were older, had the highest comorbidity burden, Sequential Organ Failure Assessment (SOFA) score, peak lactate, creatinine, peak troponin, N-terminal B-type natriuretic peptide (NT-proBNP) levels and the lowest hemoglobin levels among the 3 clusters. Patients in cluster 3 were younger, had the lowest comorbidity burden, SOFA score, peak lactate, creatinine, NT-proBNP levels. Patients belonging to cluster 1 had clinical characteristics intermediate to those of cluster 2 and 3, except for the lowest troponin level. In-hospital mortality was highest in cluster 2, followed by cluster 1 and then cluster 3. Patients in cluster 2 had the highest in-hospital mortality risk with hypokalemia (OR 3.71, 95% 1.81­­-7.63) and patients in cluster 3 had highest in-hospital mortality risk with hyperkalemia (OR 2.77, 95% CI 0.84-9.06). Patients in cluster 1 did not exhibit a linearly increasing mortality with hyperkalemia. Conclusions The use of clustering analysis in patients with acute MI could characterize baseline clinical data into clinically distinct clusters with different in-hospital mortality across the dyskalemia continuum. The results suggest that the optimal range of potassium levels in patients with MI may vary depending on patient characteristics.Figure - clusteringTable - K and In-hospital mortality

Clinical application of cluster analysis in patients with newly diagnosed type 2 diabetes.

Early prevention and treatment of type 2 diabetes mellitus (T2DM) is still a huge challenge for patients and clinicians. Recently, a novel cluster-based diabetes classification was proposed which may offer the possibility to solve this problem. In this study, we report our performance of cluster analysis of individuals newly diagnosed with T2DM, our exploration of each subtype's clinical characteristics and medication treatment, and the comparison carried out concerning the risk for diabetes complications and comorbidities among subtypes by adjusting for influencing factors. We hope to promote the further application of cluster analysis in individuals with early-stage T2DM. In this study, a k-means cluster algorithm was applied based on five indicators, namely, age, body mass index (BMI), glycosylated hemoglobin (HbA1c), homeostasis model assessment-2 insulin resistance (HOMA2-IR), and homeostasis model assessment-2 β-cell function (HOMA2-β), in order to perform the cluster analysis among 567 newly diagnosed participants with T2DM. The clinical characteristics and medication of each subtype were analyzed. The risk for diabetes complications and comorbidities in each subtype was compared by logistic regression analysis. The 567 patients were clustered into four subtypes, as follows: severe insulin-deficient diabetes (SIDD, 24.46%), age-related diabetes (MARD, 30.86%), mild obesity-related diabetes (MOD, 25.57%), and severe insulin-resistant diabetes (SIRD, 20.11%). According to the results of the oral glucose tolerance test (OGTT) and biochemical indices, fasting blood glucose (FBG), 2-hour postprandial blood glucose (2hBG), HbA1c, total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C) and triglyceride-glucose index (TyG) were higher in SIDD and SIRD than in MARD and MOD. MOD had the highest fasting C-peptide (FCP), 2-hour postprandial C-peptide (2hCP), fasting insulin (FINS), 2-hour postprandial insulin (2hINS), serum creatinine (SCr), and uric acid (UA), while SIRD had the highest triglycerides (TGs) and TyG-BMI. Albumin transaminase (ALT) and albumin transaminase (AST) were higher in MOD and SIRD. As concerms medications, compared to the other subtypes, SIDD had a lower rate of metformin use (39.1%) and a higher rate of α-glucosidase inhibitor (AGI, 61.7%) and insulin (74.4%) use. SIRD showed the highest frequency of use of sodium-glucose cotransporter-2 inhibitors (SGLT-2i, 36.0%) and glucagon-like peptide-1 receptor agonists (GLP-1RA, 19.3%). Concerning diabetic complications and comorbidities, the prevalence of diabetic kidney disease (DKD), cardiovascular disease (CVD), non-alcoholic fatty liver disease (NAFLD), dyslipidemia, and hypertension differed significantly among subtypes. Employing logistic regression analysis, after adjusting for unmodifiable (sex and age) and modifiable related influences (e.g., BMI, HbA1c, and smoking), it was found that SIRD had the highest risk of developing DKD (odds ratio, OR = 2.001, 95% confidence interval (CI): 1.125-3.559) and dyslipidemia (OR = 3.550, 95% CI: 1.534-8.215). MOD was more likely to suffer from NAFLD (OR = 3.301, 95%CI: 1.586-6.870). Patients with newly diagnosed T2DM can be successfully clustered into four subtypes with different clinical characteristics, medication treatment, and risks for diabetes-related complications and comorbidities, the cluster-based diabetes classification possibly being beneficial both for prevention of secondary diabetes and for establishment of a theoretical basis for precision medicine.

Unsupervised Cluster Analysis in Patients with Cardiorenal Syndromes: Identifying Vascular Aspects.

Background/Objectives: Cardiorenal syndrome (CRS) is a disorder of the heart and kidneys, with one type of organ dysfunction affecting the other. The pathophysiology is complex, and its actual description has been questioned. We used clustering analysis to identify clinically relevant phenogroups among patients with CRS. Methods: Data for patients admitted from 1 January 2012 to 31 December 2012 were collected from the French national medico-administrative database. Patients with a diagnosis of heart failure and chronic kidney disease and at least 5 years of follow-up were included. Results: In total, 13,665 patients were included and four clusters were identified. Cluster 1 could be described as the vascular-diabetes cluster. It comprised 1930 patients (14.1%), among which 60% had diabetes, 94% had coronary artery disease (CAD), and 80% had peripheral artery disease (PAD). Cluster 2 could be described as the vascular cluster. It comprised 2487 patients (18.2%), among which 33% had diabetes, 85% had CAD, and 78% had PAD. Cluster 3 could be described as the metabolic cluster. It comprised 2163 patients (15.8%), among which 87% had diabetes, 67% dyslipidemia, and 62% obesity. Cluster 4 comprised 7085 patients (51.8%) and could be described as the low-vascular cluster. The vascular cluster was the only one associated with a higher risk of cardiovascular death (HR: 1.48 [1.32-1.66]). The metabolic cluster was associated with a higher risk of kidney replacement therapy (HR: 1.33 [1.17-1.51]). Conclusions: Our study supports a new classification of CRS based on the vascular aspect of pathophysiology differentiating microvascular or macrovascular lesions. These results could have an impact on patients' medical treatment.

Cluster analysis of patients with alcohol use disorder featuring alexithymia, depression, and diverse drinking behavior.

This study aimed to identify subgroups of alcohol use disorder (AUD) based on a multidimensional combination of alexithymia, depression, and diverse drinking behavior. We recruited 176 patients with AUD, which were initially divided into non-alexithymic (n = 130) and alexithymic (n = 46) groups using a cutoff score of 61 on the Toronto Alexithymia Scale (TAS-20). Subsequently, the profiles of the two groups were compared. Thereafter, a two-stage cluster analysis using hierarchical and K-means methods was performed with the Z-scores from the TAS-20, the Quick Inventory of Depressive Symptomatology Self-Report Japanese Version, the 12-item questionnaire for quantitative assessment of depressive mixed state, and the 20-item questionnaire for drinking behavior pattern. In the first analysis, Alexithymic patients with AUD showed greater depressive symptoms and more pathological drinking behavior patterns than those without alexithymia. Cluster analysis featuring alexithymia, depression, and drinking behavior identified three subtypes: Cluster 1 (core AUD type) manifesting pathological drinking behavior highlighting automaticity; Cluster 2 (late-onset type) showing relatively late-onset alcohol use and fewer depressive symptoms or pathological drinking behavior; and Cluster 3 (alexithymic type) characterized by alexithymia, depression, and pathological drinking behavior featuring greater coping with negative affect. The multidimensional model with alexithymia, depression, and diverse drinking behavior provided possible practical classification of AUD. The alexithymic subtype may require more caution, and additional support for negative affect may be necessary due to accompanying mood problems and various maladaptive drinking behaviors.

Clinical features and complication rates in type 2 diabetes mellitus clusters on five variables: glycated hemoglobin, age at diagnosis, body mass index, HOMA-IR, HOMA-B

Diabetes mellitus (T2DM) is a serious medical and social problem. Now they are studying the possibility of a new stratification of diabetes. The possibility of cluster analysis for different durations of diabetes, in different cohorts to identify phenotypic clusters of T2DM and validation by reproducing clusters is relevant. Identify clusters of type 2 diabetes mellitus in patients with different disease duration based on five variables: HbA1c, age at diagnosis, BMI, HOMA-IR, HOMA-B and study the clinical features and complication rates in each cluster in the Novosibirsk region. Cluster analysis of K-means was performed in 2131 patients with T2DM, aged 44 to 70 years, with a duration of diabetes of 6.42±5.66 years, living in the Novosibirsk region based on 5 variables - HbA1c, age at -diagnosis, BMI, HOMA-IR, HOMA-B. All patients a complete clinical and laboratory examination. The insulin resistance index in the HOMA (HOMA-IR, u) and the β-cell function assessment index (HOMA-B) were calculated using the calculator -version 2.2.3 at www.dtu.ox.ac.uk. Cluster analysis revealed: Cluster 1 included 455 patients with preserved β-cell function (HOMA-B 82.97±23.28%), moderate insulin resistance (HOMA-IR 5.57±4.72) and higher diastolic BP; Cluster 2 in 1658 patients with reduced function of β-cells (HOMA-B 21.71±12.51%), the lowest indices of insulin resistance (HOMA-IR 3.50±2.48) and was characterized by a longer duration of diabetes, high fasting glycemia , HbA1c, higher eGFR and MAU, men compared with women had a 31% higher risk of developing diabetic neuropathy and 28% more diabetic nephropathy; Cluster 3 in 18 patients with high function of β-cells (HOMA-B 228.53±63.32%), severe insulin resistance (HOMA-IR 6.92±4.77), features were high incidence of men, shorter duration of diabetes, lower fasting glycemia and HbA1c, lower diastolic BP and eGFR, high incidence of early development of diabetic retinopathy after 4.00±3.6 years. Cluster analysis in patients with different durations of diabetes mellitus confirmed the possibility of using cluster analysis to identify T2DM phenotypes in the Russian population. The clusters differed in the clinical characteristics of patients, the frequency and risk of diabetic complications. These results have potential value for early stratification of therapy.

Unsupervised learning to characterize patients with known coronary artery disease undergoing myocardial perfusion imaging

PurposePatients with known coronary artery disease (CAD) comprise a heterogenous population with varied clinical and imaging characteristics. Unsupervised machine learning can identify new risk phenotypes in an unbiased fashion. We use cluster analysis to risk-stratify patients with known CAD undergoing single-photon emission computed tomography (SPECT) myocardial perfusion imaging (MPI).MethodsFrom 37,298 patients in the REFINE SPECT registry, we identified 9221 patients with known coronary artery disease. Unsupervised machine learning was performed using clinical (23), acquisition (17), and image analysis (24) parameters from 4774 patients (internal cohort) and validated with 4447 patients (external cohort). Risk stratification for all-cause mortality was compared to stress total perfusion deficit (< 5%, 5–10%, ≥10%).ResultsThree clusters were identified, with patients in Cluster 3 having a higher body mass index, more diabetes mellitus and hypertension, and less likely to be male, have dyslipidemia, or undergo exercise stress imaging (p < 0.001 for all). In the external cohort, during median follow-up of 2.6 [0.14, 3.3] years, all-cause mortality occurred in 312 patients (7%). Cluster analysis provided better risk stratification for all-cause mortality (Cluster 3: hazard ratio (HR) 5.9, 95% confidence interval (CI) 4.0, 8.6, p < 0.001; Cluster 2: HR 3.3, 95% CI 2.5, 4.5, p < 0.001; Cluster 1, reference) compared to stress total perfusion deficit (≥10%: HR 1.9, 95% CI 1.5, 2.5 p < 0.001; < 5%: reference).ConclusionsOur unsupervised cluster analysis in patients with known CAD undergoing SPECT MPI identified three distinct phenotypic clusters and predicted all-cause mortality better than ischemia alone.

Unsupervised Machine Learning with Cluster Analysis in Patients Discharged after an Acute Coronary Syndrome: Insights from a 23,270-Patient Study

Characterization and management of patients admitted for acute coronary syndromes (ACS) remain challenging, and it is unclear whether currently available clinical and procedural features can suffice to inform adequate decision making. We aimed to explore the presence of specific subsets among patients with ACS. The details on patients discharged after ACS were obtained by querying an extensive multicenter registry and detailing patient features, as well as management details. The clinical outcomes included fatal and nonfatal cardiovascular events at 1-year follow-up. After missing data imputation, 2 unsupervised machine learning approaches (k-means and Clustering Large Applications [CLARA]) were used to generate separate clusters with different features. Bivariate- and multivariable-adjusted analyses were performed to compare the different clusters for clinical outcomes. A total of 23,270 patients were included, with 12,930 cases (56%) of ST-elevation myocardial infarction (STEMI). K-means clustering identified 2 main clusters: a first 1 including 21,998 patients (95%) and a second 1 including 1,282 subjects (5%), with equal distribution for STEMI. CLARA generated 2 main clusters: a first 1 including 11,268 patients (48%) and a second 1 with 12,002 subjects (52%). Notably, the STEMI distribution was significantly different in the CLARA-generated clusters. The clinical outcomes were significantly different across clusters, irrespective of the originating algorithm, including death reinfarction and major bleeding, as well as their composite. In conclusion, unsupervised machine learning can be leveraged to explore the patterns in ACS, potentially highlighting specific patient subsets to improve risk stratification and management.

Subtypes of type 2 diabetes and their association with outcomes in Korean adults - A cluster analysis of community-based prospective cohort

BackgroundLittle is known about the subtypes of type 2 diabetes (T2D) and their association with clinical outcomes in Asians. MethodsWe performed data-driven cluster analysis in patients with newly diagnosed drug-naive T2D (n = 756) from the Korean Genome and Epidemiology Study. Clusters were based on five variables (age at diagnosis, BMI, HbA1c, and HOMA2 β-cell function, and insulin resistance). ResultsWe identified four clusters of patients with T2D according to k-means clustering: cluster 1 (22.4 %, severe insulin-resistant diabetes [SIRD]), cluster 2 (32.7 %, mild age-related diabetes [MARD]), cluster 3 (32.7 %, mild obesity-related diabetes [MOD]), and cluster 4 (12.3 %, severe insulin-deficient diabetes [SIDD]). During 14 years of follow-up, individuals in the SIDD cluster had the highest risk of initiation of glucose-lowering therapy compared to individuals in the other three clusters. Individuals in the MARD and SIDD clusters showed the highest risk of chronic kidney disease and cardiovascular disease, and individuals in the MOD clusters showed the lowest risk after adjusting for other risk factors (P < 0.05). ConclusionsPatients with T2D can be categorized into four subgroups with different glycemic deterioration and risks of diabetes complications. Individualized management might be helpful for better clinical outcomes in Asian patients with different T2D subgroups.

Cluster analysis of patients with chronic rhinosinusitis and asthma after endoscopic sinus surgery

BackgroundPatients with chronic rhinosinusitis with nasal polyps and asthma (CRSwAS) are highly heterogenous in severity and prognosis. The clinical phenotypes and inflammatory endotypes of CRSwAS and their association with outcomes of endoscopic sinus surgery (ESS) have not been fully studied yet. ObjectiveWe aimed to find out the clinical phenotypes of CRSwAS and explore their relationship with ESS outcomes using cluster analysis. MethodsWe recruited 103 consecutive adult patients with CRSwAS who had undergone ESS and been followed up for more than 1 year. For cluster analysis, we collected the data from 63 variables pertaining to demographic characteristics, preoperative disease status, surgical techniques, postoperative medical treatment, and outcomes. Eosinophilic CRS was defined as greater than or equal to 10 eosinophils/high-power field, and sinus computed tomography was evaluated by Lund-Mackay sinus computed tomography score (LM score). ResultsWe screened 92 eligible patients and 13 preoperative variables for balanced iterative reducing and clustering using hierarchies cluster analysis. Patients with CRSwAS were divided into 4 clusters with distinct ESS outcomes: (1) cluster 1, characterized by aspirin-exacerbated respiratory disease, eosinophilic CRS, high preoperative LM score, moderate-to-severe asthma, and uncontrolled CRS after ESS; (2) cluster 2, characterized as having female dominance (66.67%), non–aspirin-exacerbated respiratory disease, eosinophilic CRS, high preoperative LM score, moderate-to-severe asthma, and uncontrolled CRS after ESS; (3) cluster 3, characterized as having female dominance (95.83%), noneosinophilic CRS, low preoperative LM score, moderate asthma, and controlled CRS after ESS; and (4) cluster 4, characterized as men-only, smoker, noneosinophilic CRS, low preoperative LM score, mild asthma, and controlled CRS after ESS. ConclusionCRSwAS has distinct clusters, each corresponding to unique clinical and inflammatory characteristics and ESS outcomes.

Identification by cluster analysis of patients with asthma and nasal symptoms using the MASK-air® mHealth app

BackgroundThe self-reporting of asthma frequently leads to patient misidentification in epidemiological studies. Strategies combining the triangulation of data sources may help to improve the identification of people with asthma. We aimed to combine information from the self-reporting of asthma, medication use and symptoms to identify asthma patterns in the users of an mHealth app. MethodsWe studied MASK-air® users who reported their daily asthma symptoms (assessed by a 0-100 visual analogue scale – “VAS Asthma”) at least three times (either in three different months or in any period). K-means cluster analysis methods were applied to identify asthma patterns based on: (i) whether the user self-reported asthma; (ii) whether the user reported asthma medication use and (iii) VAS asthma. Clusters were compared by the number of medications used, VAS asthma levels and Control of Asthma and Allergic Rhinitis Test (CARAT) levels. FindingsWe assessed a total of 8,075 MASK-air® users. The main clustering approach resulted in the identification of seven groups. These groups were interpreted as probable: (i) severe/uncontrolled asthma despite treatment (11.9-16.1% of MASK-air® users); (ii) treated and partly-controlled asthma (6.3-9.7%); (iii) treated and controlled asthma (4.6-5.5%); (iv) untreated uncontrolled asthma (18.2-20.5%); (v) untreated partly-controlled asthma (10.1-10.7%); (vi) untreated controlled asthma (6.7-8.5%) and (vii) no evidence of asthma (33.0-40.2%). This classification was validated in a study of 192 patients enrolled by physicians. InterpretationWe identified seven profiles based on the probability of having asthma and on its level of control. mHealth tools are hypothesis-generating and complement classical epidemiological approaches in identifying patients with asthma.

IMPULSE OSCILLOMETRY IDENTIFIES DISTINCT ASTHMA PHENOTYPES ON CLUSTER ANALYSIS

IMPULSE OSCILLOMETRY IDENTIFIES DISTINCT ASTHMA PHENOTYPES ON CLUSTER ANALYSIS

Mortality Risk Stratification Using Cluster Analysis in Patients With Myositis-Associated Interstitial Lung Disease Receiving Initial Triple-Combination Therapy.

ObjectiveTo stratify patients with polymyositis/dermatomyositis-associated interstitial lung disease (ILD) who were initially treated with an intensive regimen consisting of high-dose corticosteroids, a calcineurin inhibitor, and intravenous cyclophosphamide (triple-combo therapy) into subgroups based on mortality outcomes by a cluster analysis using a large-scale multicenter retrospective cohort of Japanese patients with myositis-associated ILD (JAMI).MethodsTwo-step cluster analysis of preclustering and subsequent hierarchical clustering was conducted in 185 patients who received triple-combo therapy in an unbiased manner. Initial predictors for mortality previously reported in patients with myositis-associated ILD were used as variables and included age, sex, disease duration, classification of myositis, requirement of supplemental oxygen, anti-aminoacyl tRNA synthetase (ARS) antibody, anti-melanoma differentiation-associated gene 5 (MDA5) antibody, and serum levels of C-reactive protein (CRP) and Krebs von den Lungen-6 (KL-6). The cluster model was further applied to 283 patients who received conventional regimens consisting of corticosteroids with or without a single immunosuppressive agent (dual-combo therapy or monotherapy). Cumulative survival rates were compared using Kaplan-Meier analysis, and the log-rank test was used to test for significant differences between two groups.ResultsWe developed a cluster model consisting of 6 clusters, which were categorized by age at onset, clinically amyopathic dermatomyositis, CRP, KL-6, requirement of supplemental oxygen, anti-ARS antibody, and anti-MDA5 antibody. This model was judged to be of good quality based on the silhouette measure of cohesion and separation of 0.6. These clusters were regrouped into three subsets based on low (<10%), moderate (10-50%), and high (>50%) mortality rates. The performance of the clustering was generally replicated in patients who received initial dual-combo therapy or monotherapy. Survival benefits of triple-combo therapy over dual-combo therapy or monotherapy were not observed in any of the clusters.ConclusionWe successfully developed a cluster model that stratified patients with myositis-associated ILD who were treated with initial triple-combo therapy into subgroups with different prognoses, although this model failed to identify a patient subgroup that showed survival benefits from triple-combo therapy over dual-combo therapy or monotherapy.

MO083: Phenotypes of Patients With Combined Hypertension and COPD Depending on Renal Function

Abstract BACKGROUND AND AIMS Chronic kidney disease and hypertension are closely linked to cause and effect. Patients with COPD, in turn, have a 1.61 times higher risk of developing CKD than patients without COPD, after adjusting for clinical risk factors. Comorbidity of hypertension and COPD contributes to a greater decrease in renal function. The aim of this study was to identify the phenotypes of patients with a combined course of hypertension and COPD using cluster analysis. METHOD In a one-time study, 32 patients with hypertension and chronic obstructive pulmonary were examined. Patients of three groups underwent general clinical examination, collected complaints, tested on the questionnaires of mMRC and CAT, determined nutritional status by anthropometric study and bioimpedancemetry, conducted daily monitoring of blood pressure and ventilatory function, assessed the functional state of the kidneys by laboratory examination of blood. Statistical methods were used, including cluster analysis, using the "Statistics" program. RESULTS During the cluster analysis in patients with combined hypertension and chronic obstructive pulmonary disease, 5 large clusters of clinical and laboratory data of patients were identified. Cluster 1: male and female patients with COPD and hypertension, middle age 43 (1,2) years, smokers. The average duration of COPD is 10 years, hypertension is 20 years. Results for CAT-9, mMRC: 1, the number of exacerbations of COPD per year: 2. The average content of adipose tissue: 22.7% of muscle tissue: 36%, visceral fat: 9%. BMI: 29.5 kg/m2. FEV1: 49%, GFR: 68 ml/min. Cluster 2: male and female patients with COPD and hypertension, middle age 61 (1,3) years, smokers. The average duration of COPD is 27.5 years, and hypertension is 20 years. Results for CAT: 13.5, mMRC: 2.5, the number of exacerbations of COPD per year: 1. The average content of adipose tissue: 31.4% of muscle tissue: 30.4%, visceral fat: 17.5%. BMI: 31.3 kg/m2. FEV1: 44%, GFR: 63 mL/min. Cluster 3: male and female patients with COPD and hypertension, middle age 61 (1,3) years, smokers. The average duration of COPD is 13 years, and hypertension is 5 years. Results for CAT: 12.5, mMRC: 1.5, the number of exacerbations of COPD per year: 1. The average content of adipose tissue: 20.6% of muscle tissue: 34.3%, visceral fat: 7.0%. BMI: 23.9 kg/m2. FEV1: 32%, GFR: 58 ml/min. Cluster 4: male and female patients with COPD and hypertension, middle age 61 (1,3) years, smokers. The average duration of COPD is 10 years, and hypertension is 20 years. Results for CAT: 9.0, mMRC: 1, the number of exacerbations of COPD per year: 1. The average content of adipose tissue: 41.3% of muscle tissue: 28.6%, visceral fat: 11.0%. BMI: 35.0 kg/m2. FEV1: 80%, GFR: 81 mL/min. Cluster 5: male and female patients with COPD and hypertension, mean age 58.5 (1.2) years, smokers. The average duration of COPD is 27.5 years, and hypertension is 20 years. Results for CAT: 8.5, mMRC: 1.5, the number of exacerbations of COPD per year: 1.5. The average content of adipose tissue is 36.7% of muscle tissue: 30.5%, visceral fat: 13%. BMI: 30.1 kg/m2. FEV1: 61%, GFR: 77 mL/min. CONCLUSION Cluster analysis allows to identify five clusters of patients with combined hypertension and COPD, which can be combined into two phenotypes: with greater bronchial obstruction and lower GFR, and with less bronchial obstruction and greater GFR.

Comprehensive Analysis of Subtypes and Identification of Key lncRNAs Based on Glutamine Metabolism-Related Long Noncoding RNAs.

Background Long noncoding RNAs (lncRNAs) are becoming a critical class of metabolic regulate molecule in cancer. Glutamine is a regulator that contributes to each of the core metabolic tasks in proliferating tumor cells. Thus, we aimed to evaluate the association of lncRNAs with glutamine metabolism in lung adenocarcinoma (LUAD). Methods Using single-sample gene set enrichment analysis (ssGSEA), LUAD specimens were assigned scores based on glutamine metabolism-related genes, and the shared common glutamine metabolism-related lncRNAs in three different LUAD data cohorts were identified. ConsensusClusterPlus was used to perform unsupervised clustering analysis in patients with LUAD. Key glutamine metabolism-related lncRNAs were identified by first-order partial correlation analysis. Results A total of 11 shared glutamine metabolism-associated lncRNAs were identified in three LUAD data cohorts, and LUAD patients were classified into three glutamine metabolism subtypes based on the expressions of the related genes. C1 exhibited shorter overall survival (OS), poor genomic instability, and inadequate infiltration of immune cell types in the tumor microenvironment (TME) and was representative of the immunodeficiency phenotype. C2 represented the immunosuppressive phenotype while C3 represented the immune activation phenotype, exhibiting the highest sensitivity to immunotherapy. Nine of the 11 lncRNAs were localized to the nucleus. Finally, three key lncRNAs, significantly enriched in multiple metabolic pathways, were screened and found to be remarkably related to the OS of LUAD. Conclusion We identified three glutamine metabolism subtypes of LUAD, which reflected different OS, genomic, and TME features, and identified three key glutamine metabolism-associated lncRNAs may contribute to further study of lncRNAs in cancer metabolism.

Real world data collection and cluster analysis in patients with sciatica due to lumbar disc herniation

ObjectiveThe aim of this study was to identify and describe clusters of patients with similar characteristics presenting with sciatica caused by a lumbar disc herniation in secondary care. MethodsForty-six percent (n = 163) of the eligible patients (n = 352) completed all questionnaires and were included in a hierarchical cluster analysis. The analysis was based on baseline characteristics for pain, pain awareness and catastrophizing, disability and quality of life (QOL). Clusters were compared for the use of pain medication, employment status and allocated treatment. ResultsThree significant clusters were identified.Cluster 2 (n = 49), coined the painfulness cluster, reported the lowest baseline characteristics for pain (>5) and disability together with a higher health-related QOL. Patients in cluster 3, labeled the painfulness and suffering cluster, had relatively high pain scores for back and leg pain (>6), high pain awareness and catastrophizing, i.e. suffering, but relatively limited disability and maintained QOL. Cluster 1 (n = 71), the painfulness-suffering and disability cluster, was characterized by the most severe back and leg pain (>7), high pain awareness and catastrophizing with the lowest QOL and highest disability. Patients in cluster 1 underwent significantly more surgery and used the most extensive pain medication (WHO III). ConclusionThis research gives insight in the complex population with sciatica and is of added value to the recent, sparsely existing literature on relevant patient subgroups in the low back and leg pain population. The data suggest that VAS scores < 6 do not lead to suffering and VAS scores < 7 not to disability.

Construction of survival-related co-expression modules and identification of potential prognostic biomarkers of osteosarcoma using WGCNA.

BackgroundOsteosarcoma (OS) is a primary malignant bone tumor. Patients with different immune characteristics respond differently to chemotherapy and have a lower chance of survival. The potential pathogenesis and therapeutic targets of OS must be investigated further.MethodsOS expression profile data and clinical information were downloaded from the Therapeutically Applicable Research to Generate Effective Treatments (TARGET) and the Gene Expression Omnibus (GEO) databases. The immune-related gene set was obtained from the ImmPort database, and the immune-related gene expression profiles were used for non-negative matrix factorization (NMF) cluster analysis of patients in the 2 databases to find the best clustering number. In the TARGET database, OS patients were classified into low-risk and high-risk groups based on the differences in their survival rates. Weighted correlation network analysis (WGCNA) was applied to the low-risk and high-risk groups to determine the module with the lowest conservatism in order to differentiate the prognosis of the 2 groups.ResultsA total of 500 key genes associated with poor prognosis were identified. Gene Ontology (GO) enrichment analysis revealed that the biological processes of these genes were primarily focused on the regulation of small guanosine triphosphatase (GTPase) mediated signal transduction, collagen-containing extracellular matrix, and Rho GTPase binding. A random survival forest identified EPHB3, TEAD1, and KRR1P1 as key genes. Their expression level was linked to overall survival. We discovered that the core genes were associated to immune cell infiltration. Simultaneously, paired survival analysis of two genes revealed differences in survival. We also reverse-predicted the main genes and developed their competitive endogenous RNA (ceRNA) network. Finally, utilizing the CellMiner database, we observed that the genes TEAD1 and EPHB3 were connected to drug sensitivity.ConclusionsIn this study, we identified the modules and key genes related to the poor prognosis of OS patients by using WGCNA, and verified their impact on the prognosis of OS patients and their role in the immune microenvironment of OS. In addition, targeted gene related antitumor drugs were screened out. The discoveries may lead to novel molecular targets and treatment methods for OS patients.KeywordsOsteosarcoma (OS); weighted gene co-expression network analysis (WGCNA); gene

The cluster analysis of patients with morbid obesity in bariatric and metabolic surgery

INTRODUCTION. Existing classifications of various forms of obesity are poorly adapted to the practical goals of bariatric surgery from the standpoint of personalizing the choice of method and volume of intervention.The OBJECTIVE was develop new approaches to stratification of patients with morbid obesity based on cluster analysis; to evaluate of the effectiveness and safety of metabolic surgical interventions, depending on the belonging of patients to a particular bariatric cluster.METHODS AND MATERIALS. The study included 745 bariatric patients. To form clinical and demographic groups, a two-stage cluster analysis procedure was used using the SPSS 16.0 package with 11 categorical and 2 quantitative variables. The efficacy and safety of metabolic surgical interventions depending on the patients' belonging to a particular bariatric cluster was evaluated.RESULTS. As a result of cluster analysis in the general population of patients with morbid obesity, 3 unique clinical and demographic groups (clusters) of patients were identified, characterized by an initial set of similar indicators that differed from that in patients of other groups (clusters): the1st cluster: «premorbid obesity»; the 2nd cluster: «morbid obesity without metabolic syndrome and complications»; the 3rd cluster: «morbid obesity with metabolic syndrome and/ or complications». The best long-term results on the BAROS scale were generally observed in patients of the 1st cluster (87.8 % of excellent and good results compared to 66.7 % in patients of the 2nd clusters and 65.5 % in patients of the 3rd cluster, p&lt;0.001). The total risk of serious complications in patients of the 3rd cluster was 2.4 times higher than in patients of the 1st cluster (P&lt;0.05) and 1.9 times higher than in patients of the 2nd cluster (P&gt;0.05).CONCLUSION. The procedure of clustering patients with obesity allows individualizing the indications for conducting and choosing bariatric interventions.

Cluster analysis of patients with granulomatosis with polyangiitis (GPA) based on clinical presentation symptoms: a UK population-based cohort study

BackgroundGranulomatosis with polyangiitis (GPA) is small vessel vasculitis with heterogeneous clinical presentation. In the present population-based cohort study, we classified patients with GPA based on clinical features at presentation using an unsupervised clustering approach and compared their mortality, infections and frequency of comorbidities.MethodsIn this open cohort study, de-identified primary care data of patients with GPA included in the IQVIA Medical Research Data database between 1 January 1995 and 25 September 2019 was analysed retrospectively. Latent class analysis was performed to create symptom clusters of patients based on 16 categories of symptoms representing various organ involvement. All-cause mortality of resultant clusters was compared after adjusting for age, sex, Townsend deprivation quintile and smoking status at index date using extended Cox proportional hazards models. Prescription of antibiotics, considered as an indirect indicator of recurrent bacterial infection, was compared using a recurrent event model, after adjusting for quarterly use of steroid as a time-dependent covariate. Cumulative frequencies of common comorbidities were compared among the clusters at index visit, 1-year and 3-year follow-up.ResultsAltogether, 649 patients with GPA [median age 60.0 (IQR: 49.6–70.1)] were included. Three clusters were identified: patients with limited disease mainly with involvement of ENT and cough were classified into cluster 1 (n = 426); cluster 2 had generalised non-renal disease (n = 176); while patients in cluster 3 had renal-predominant disease (n = 47). Many patients in cluster 1 developed generalised disease at the end of 1 year. Mortality in clusters 2 and 3 was higher compared with cluster 1. Mortality in cluster 1 itself was 68% higher than the general population without GPA. The duration of antibiotics prescription and frequency of coexisting medical illnesses was also higher in clusters 2 and 3.ConclusionsIn a primary care setting, patients with GPA can be classified into three distinct clusters with different prognosis, susceptibility to recurrent infections and presence of comorbidities. The tendency of cluster 1 to evolve into a more generalised disease raises questions about current immunosuppressive treatment approaches in these patients.

Cluster analysis of deterioration sites after first-line epidermal growth factor receptor-tyrosine kinase inhibitor in epidermal growth factor receptor mutated non-small cell lung cancer.

IntroductionFor epidermal growth factor receptor (EGFR) mutated non-small cell lung cancer (NSCLC), no studies have treated the site of recurrence after first-line tyrosine kinase inhibitor (TKI) treatment as a “metastasis pattern”. This study aims to assess whether these patients have a specific “metastasis pattern” at the site of recurrence after the treatment.Material and methodsData were collected from all consecutive EGFR mutated NSCLC patients between 2009 and 2021. Metastatic patterns were analyzed using cluster analysis in patients with EGFR mutated NSCLC.ResultsDuring the study period, 83 EGFR mutated NSCLC patients were treated with EGFR-TKI. Patients who had no metastases at the time of diagnosis were divided into two groups according to the presence or absence of recurrence of metastases after TKI administration. Patients with metastases at diagnosis were divided into 4 groups by cluster analysis. A statistically significant difference in metastasis frequency was confirmed among these 6 groups (χ2 test, p = 0.0001). Furthermore, when the frequency of metastasis recurrence after TKI administration in these 6 groups was examined, a statistically significant difference was confirmed (χ2 test, p = 0.0001).ConclusionsEven in EGFR mutation-positive patients, the knowledge of the recurrent patterns might be useful for clinical practice in the foreseeable future, as it enables more efficient detection of metastatic disease through imaging, and more effective treatment at predicted metastatic sites.

Distant organ metastasis patterns in lung cancer patients with pancreatic metastasis - a cluster analysis.

Pancreatic metastasis of lung cancer is rare, but the narrowing or obstruction of the biliary tract from pancreatic metastases limits the choice of antitumour agents. Lung cancer patients with pancreatic metastasis often have metastasis of other organs. For these patients, however, no studies have been conducted to evaluate distant metastasis sites as metastasis patterns. This study aims to assess whether these patients have specific metastasis patterns at the time of diagnosis of lung cancer. Data were collected from consecutive lung cancer patients with pancreatic metastasis between April 2012 and March 2022. Metastatic patterns were analysed using cluster analysis in patients with lung cancer. During the study period, 33 (2.0%) of 1659 patients were diagnosed as having pancreatic metastasis. Of the 33 patients, 28 (84.8%) were male. Eighteen, 14, and one patient had small cell lung cancer (SCLC), lung adenocarcinoma, and large cell neuroendocrine carcinoma, respectively. They were divided into 3 groups by cluster analysis. A statistically significant difference in metastasis frequency was confirmed among these 3 groups (χ2 test, p = 0.001). In lung cancer patients with pancreatic metastasis, knowledge of the metastatic patterns might be useful for clinical practice in the foreseeable future because it enables more efficient detection of metastatic disease through imaging, and more effective treatment at predicted metastatic sites.