FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2133111154> ?p ?o ?g. }

• W2133111154 endingPage "S178" @default.
• W2133111154 startingPage "S177" @default.
• W2133111154 abstract "Purpose: A possible explanation for the seeming complexity of osteoarthritis is that it is not one disease, but a collection of disease sub-types with different aetiology and clinical characteristics. In this study we used latent class cluster analysis to find subtypes of knee osteoarthritis based on OA characteristics in the baseline data of the Osteoarthritis Initiative (OAI). Methods: Latent class cluster analysis (LCA) is a model based clustering approach that assumes the existence of a latent variable (the various clusters) that influences a set of observed variables, in this case OA characteristics. The optimal number of clusters was determined based on the BIC (Bayesian Information Criterion) that informs on the goodness of fit of the model, and the AWE (Approximate Weight of Evidence) which informs on the ability to classify subjects into the various clusters. The OA characteristics that served as input for the LCA was formed by the set of x-ray scores of different features, quantitative MRI cartilage measures and various clinical scores like WOMAC and VAS from 520 subjects of the OAI progression cohort at baseline. Since OA increases in severity over time, a cluster analysis can only be meaningful when performed independent of disease severity. To accomplish this we included a severity variable into the model, such that a model fit was obtained adjusted for severity. The severity variable was obtained by performing Latent Class Factor analysis, which is similar to principle component analysis, but works for both continuous and categorical data. The same set of OA characteristics was used as input. Results: Our method resulted in 4 clusters, of which the 1st contained 47% of the subjects, the 2nd 27%, the 3rd 15% and the 4th 12% of the subjects. Clusters 1, 2 and 4 showed mainly medial cartilage degeneration, while cartilage degeneration in the 3rd cluster occurred laterally. Interestingly, clusters 3 and 4 showed a strong increase in area of denuded bone with increasing OA severity as measured on MRI, while cluster 1 did not show any denuded bone, independent of OA severity. Cluster 2 showed a very mild increase in denuded bone area with increasing OA severity. Pain scores increased with increasing severity in all clusters. Pain scores were lower in the 3rd cluster, significantly so in the VAS scores reflecting pain in the last month (p<0.05). We further examined if there were difference in the presence of risk-factors between the clusters. There were no differences in the presence of heberden's nodes or BMI. However, clusters varied with respect to reported trauma. Prevalence of trauma was lowest in cluster 1 (33%) and highest in cluster 4 (56%). Finally, using the x-ray scores at 48 months, we looked at differences in progression, characterized by the prevalence of subjects that increased in JSN scores, medially or laterally. Progression was compartment specific for the various clusters. Medial progression occurred only in clusters 1, 2 and 4, most dramatically so in cluster 4 (48%). Cluster 3 again only showed lateral progression (61%). Conclusions: We found clear differences between the various clusters. Cluster 1 could be described as a mild type of OA, with no denuded bone and limited OA progression. Cluster 3 and 4 were more severe and erosive types of OA, showing strong increases in area of denuded bone with increasing OA severity, and a high prevalence of progression. Interestingly, cluster 3 was purely a lateral type of OA, while cluster 4 was purely medial in all respects. The causes that led to differences between the clusters remain speculative. The higher prevalence of reported trauma in cluster 3 and 4, the more aggressive OA types might suggest a difference in OA initiation. We kept the complexity of the models low to reduce the risk of overfitting and increase the likelihood that our findings are generalizable to other cohorts. The fact that clusters also differed on data not included in cluster finding (trauma and progression) indicates that the found clusters are meaningful. In conclusion, latent class cluster analysis confirmed the suspected existence of distinct subtypes of knee OA." @default.
• W2133111154 created "2016-06-24" @default.
• W2133111154 creator A5013980875 @default.
• W2133111154 creator A5025679642 @default.
• W2133111154 creator A5077676327 @default.
• W2133111154 date "2012-04-01" @default.
• W2133111154 modified "2023-10-16" @default.
• W2133111154 title "Latent class cluster analysis shows four distinct subtypes of knee OA: data from the osteoarthritis initiative" @default.
• W2133111154 doi "https://doi.org/10.1016/j.joca.2012.02.272" @default.
• W2133111154 hasPublicationYear "2012" @default.
• W2133111154 type Work @default.
• W2133111154 sameAs 2133111154 @default.
• W2133111154 citedByCount "1" @default.
• W2133111154 countsByYear W21331111542021 @default.
• W2133111154 crossrefType "journal-article" @default.
• W2133111154 hasAuthorship W2133111154A5013980875 @default.
• W2133111154 hasAuthorship W2133111154A5025679642 @default.
• W2133111154 hasAuthorship W2133111154A5077676327 @default.
• W2133111154 hasBestOaLocation W21331111541 @default.
• W2133111154 hasConcept C105795698 @default.
• W2133111154 hasConcept C107673813 @default.
• W2133111154 hasConcept C142724271 @default.
• W2133111154 hasConcept C154945302 @default.
• W2133111154 hasConcept C164866538 @default.
• W2133111154 hasConcept C177264268 @default.
• W2133111154 hasConcept C1862650 @default.
• W2133111154 hasConcept C199360897 @default.
• W2133111154 hasConcept C204787440 @default.
• W2133111154 hasConcept C2776164576 @default.
• W2133111154 hasConcept C2779286237 @default.
• W2133111154 hasConcept C33923547 @default.
• W2133111154 hasConcept C41008148 @default.
• W2133111154 hasConcept C51167844 @default.
• W2133111154 hasConcept C5274069 @default.
• W2133111154 hasConcept C70727504 @default.
• W2133111154 hasConcept C71924100 @default.
• W2133111154 hasConcept C73555534 @default.
• W2133111154 hasConcept C94124525 @default.
• W2133111154 hasConceptScore W2133111154C105795698 @default.
• W2133111154 hasConceptScore W2133111154C107673813 @default.
• W2133111154 hasConceptScore W2133111154C142724271 @default.
• W2133111154 hasConceptScore W2133111154C154945302 @default.
• W2133111154 hasConceptScore W2133111154C164866538 @default.
• W2133111154 hasConceptScore W2133111154C177264268 @default.
• W2133111154 hasConceptScore W2133111154C1862650 @default.
• W2133111154 hasConceptScore W2133111154C199360897 @default.
• W2133111154 hasConceptScore W2133111154C204787440 @default.
• W2133111154 hasConceptScore W2133111154C2776164576 @default.
• W2133111154 hasConceptScore W2133111154C2779286237 @default.
• W2133111154 hasConceptScore W2133111154C33923547 @default.
• W2133111154 hasConceptScore W2133111154C41008148 @default.
• W2133111154 hasConceptScore W2133111154C51167844 @default.
• W2133111154 hasConceptScore W2133111154C5274069 @default.
• W2133111154 hasConceptScore W2133111154C70727504 @default.
• W2133111154 hasConceptScore W2133111154C71924100 @default.
• W2133111154 hasConceptScore W2133111154C73555534 @default.
• W2133111154 hasConceptScore W2133111154C94124525 @default.
• W2133111154 hasLocation W21331111541 @default.
• W2133111154 hasOpenAccess W2133111154 @default.
• W2133111154 hasPrimaryLocation W21331111541 @default.
• W2133111154 hasRelatedWork W2185353617 @default.
• W2133111154 hasRelatedWork W2377248513 @default.
• W2133111154 hasRelatedWork W2560684739 @default.
• W2133111154 hasRelatedWork W2753315906 @default.
• W2133111154 hasRelatedWork W2803470807 @default.
• W2133111154 hasRelatedWork W3116499819 @default.
• W2133111154 hasRelatedWork W4200141944 @default.
• W2133111154 hasRelatedWork W4224104073 @default.
• W2133111154 hasRelatedWork W4282929412 @default.
• W2133111154 hasRelatedWork W620660060 @default.
• W2133111154 hasVolume "20" @default.
• W2133111154 isParatext "false" @default.
• W2133111154 isRetracted "false" @default.
• W2133111154 magId "2133111154" @default.
• W2133111154 workType "article" @default.