FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4384156811> ?p ?o ?g. }

• W4384156811 endingPage "184" @default.
• W4384156811 startingPage "184" @default.
• W4384156811 abstract "Regarding our Article on in-silico scores for coronary artery disease (ISCAD),1Forrest IS Petrazzini BO Duffy Á et al.Machine learning-based marker for coronary artery disease: derivation and validation in two longitudinal cohorts.Lancet. 2023; 401: 215-225Summary Full Text Full Text PDF PubMed Scopus (0) Google Scholar Linghua Yu suggests limitations with the study populations, dataset, and model. We evaluated 95 935 electronic health records across two large-scale biobanks: a diverse cohort representative of its New York City, USA, community (BioMe Biobank) and a cohort of volunteers from community centres across the UK (UK Biobank). We applied quality control steps on the dataset and imputed missing data as described in the methods of our Article. Nevertheless, we agree analyses in other populations would further increase generalisability and that misclassification is possible for electronic health record studies without quality control. We refer Yu to the appendix of our Article (pp 15, 16, 21, 22, 47, 48, 53) for extensive Shapley additive explanations, feature importance, and streamlined model analyses for model interpretability. Amanda Elliott and James P Pirruccello suggest that missing K49 and K75 procedure codes and the I25.1 diagnosis code could explain the model's high predictive performance in the UK Biobank. We clarified in the appendix of our Article (p 4) that K49 and K75 codes were in fact used to define CAD cases. Notably, we used similar diagnosis codes to other studies2Khera AV Chaffin M Aragam KG et al.Genome-wide polygenic scores for common diseases identify individuals with risk equivalent to monogenic mutations.Nat Genet. 2018; 50: 1219-1224Crossref PubMed Scopus (1412) Google Scholar, 3Pirruccello JP Bick A Wang M et al.Analysis of cardiac magnetic resonance imaging in 36 000 individuals yields genetic insights into dilated cardiomyopathy.Nat Commun. 2020; 11: 1-10Crossref PubMed Scopus (0) Google Scholar, 4Petrazzini BO Chaudhary K Marquez-Luna C et al.One-year risk prediction of coronary artery disease using clinical features from electronic health records.JACC. 2022; 79: 1155-1166Crossref Scopus (7) Google Scholar to define cases; none used I25.1. Elliott and Pirruccello also suggest groin pain is an indicator of CAD-related procedures; however, only data before first CAD diagnosis, procedure, or medication were used. Hence, groin pain occurred before and unrelatedly to coronary angiography, and instead could be due to peripheral artery disease,5Schorr EN Peden-McAlpine C Treat-Jacobson D Lindquist R Characterization of the peripheral artery disease symptom experience.Geriatr Nurs (Minneap). 2015; 36: 293-300Crossref Google Scholar a known risk factor of CAD. Moreover, groin pain was not used in any UK Biobank model. Finally, age was defined at each participant's last clinical encounter; participants were excluded if younger than 40 years or if there were no clinical encounters. We refer Elliott and Pirruccello to our previous study4Petrazzini BO Chaudhary K Marquez-Luna C et al.One-year risk prediction of coronary artery disease using clinical features from electronic health records.JACC. 2022; 79: 1155-1166Crossref Scopus (7) Google Scholar that analysed CAD controls with and without comorbidities and discusses how comorbid controls are misclassified, skew interpretability, and decrease model performance. Lastly, Justin Jee conflates the method of our previous study predicting 1-year risk for CAD4Petrazzini BO Chaudhary K Marquez-Luna C et al.One-year risk prediction of coronary artery disease using clinical features from electronic health records.JACC. 2022; 79: 1155-1166Crossref Scopus (7) Google Scholar with that of the current study, which does not predict 1-year risk for CAD nor restrict to 1 year before diagnosis. As stated in the Methods, the model analysed electronic health record data up until CAD diagnosis, procedure, or medication in cases. For quality control, all participants had at least 1 year of electronic health record data to ensure adequate data as described in the appendix of our Article (p 4). Nonetheless, we agree a time-dependent classifier should be examined. Overall, ISCAD's results show that machine learning with clinical data can quantify CAD on a spectrum. We expect refinements to the model will further bolster ISCAD's clinical translation. RD reports receiving grants from AstraZeneca; grants and non-financial support from Goldfinch Bio; being a scientific co-founder, consultant, and equity holder for Pensieve Health; and being a consultant for Variant Bio, outside of the submitted work. ISF and BOP declare no competing interests. Machine learning-based marker for coronary artery disease: derivation and validation in two longitudinal cohortsElectronic health record-based machine learning was used to generate an in-silico marker for coronary artery disease that can non-invasively quantify atherosclerosis and risk of death on a continuous spectrum, and identify underdiagnosed individuals. Full-Text PDF Machine learning-based markers for CADIain S Forrest and colleagues1 present a novel method for predicting the risk of coronary artery disease (CAD) using machine learning algorithms. However, there are several potential limitations that should be considered. Full-Text PDF Machine learning-based markers for CADWe read with interest the Article by Iain S Forrest and colleagues.1 We congratulate the authors for developing a portable model for coronary artery disease (CAD) prediction that achieved an area under the receiver operating characteristic curve of 0·89 for identifying CAD upon external validation in the UK Biobank. This model's area under the receiver operating characteristic curve was higher than those of other CAD prediction models applied in the UK Biobank,2 so we examined the findings with curiosity. Full-Text PDF Machine learning-based markers for CADIain S Forrest and colleagues1 describe a machine learning model to predict coronary artery disease (CAD). Following their previous work,2 a snapshot of health records from the CAD positive group is taken one year before their first CAD diagnosis and compared with records from controls who were not diagnosed with CAD. This snapshot provides a convenient framework for training a static machine learning classifier but makes the clinical utility of the model difficult to interpret and apply prospectively. Full-Text PDF" @default.
• W4384156811 created "2023-07-14" @default.
• W4384156811 creator A5031766503 @default.
• W4384156811 creator A5061614889 @default.
• W4384156811 creator A5075343690 @default.
• W4384156811 date "2023-07-01" @default.
• W4384156811 modified "2023-10-15" @default.
• W4384156811 title "Machine learning-based markers for CAD – Authors' reply" @default.
• W4384156811 cites W2077548502 @default.
• W4384156811 cites W2886752110 @default.
• W4384156811 cites W3023537913 @default.
• W4384156811 cites W4220894353 @default.
• W4384156811 cites W4312085640 @default.
• W4384156811 doi "https://doi.org/10.1016/s0140-6736(23)01059-0" @default.
• W4384156811 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/37453750" @default.
• W4384156811 hasPublicationYear "2023" @default.
• W4384156811 type Work @default.
• W4384156811 citedByCount "0" @default.
• W4384156811 crossrefType "journal-article" @default.
• W4384156811 hasAuthorship W4384156811A5031766503 @default.
• W4384156811 hasAuthorship W4384156811A5061614889 @default.
• W4384156811 hasAuthorship W4384156811A5075343690 @default.
• W4384156811 hasBestOaLocation W43841568111 @default.
• W4384156811 hasConcept C116567970 @default.
• W4384156811 hasConcept C119857082 @default.
• W4384156811 hasConcept C126322002 @default.
• W4384156811 hasConcept C154945302 @default.
• W4384156811 hasConcept C161191863 @default.
• W4384156811 hasConcept C2778213512 @default.
• W4384156811 hasConcept C2778805511 @default.
• W4384156811 hasConcept C2781067378 @default.
• W4384156811 hasConcept C41008148 @default.
• W4384156811 hasConcept C60644358 @default.
• W4384156811 hasConcept C71924100 @default.
• W4384156811 hasConcept C72563966 @default.
• W4384156811 hasConcept C86803240 @default.
• W4384156811 hasConcept C9357733 @default.
• W4384156811 hasConceptScore W4384156811C116567970 @default.
• W4384156811 hasConceptScore W4384156811C119857082 @default.
• W4384156811 hasConceptScore W4384156811C126322002 @default.
• W4384156811 hasConceptScore W4384156811C154945302 @default.
• W4384156811 hasConceptScore W4384156811C161191863 @default.
• W4384156811 hasConceptScore W4384156811C2778213512 @default.
• W4384156811 hasConceptScore W4384156811C2778805511 @default.
• W4384156811 hasConceptScore W4384156811C2781067378 @default.
• W4384156811 hasConceptScore W4384156811C41008148 @default.
• W4384156811 hasConceptScore W4384156811C60644358 @default.
• W4384156811 hasConceptScore W4384156811C71924100 @default.
• W4384156811 hasConceptScore W4384156811C72563966 @default.
• W4384156811 hasConceptScore W4384156811C86803240 @default.
• W4384156811 hasConceptScore W4384156811C9357733 @default.
• W4384156811 hasFunder F4320307770 @default.
• W4384156811 hasFunder F4320337338 @default.
• W4384156811 hasFunder F4320337354 @default.
• W4384156811 hasIssue "10397" @default.
• W4384156811 hasLocation W43841568111 @default.
• W4384156811 hasLocation W43841568112 @default.
• W4384156811 hasOpenAccess W4384156811 @default.
• W4384156811 hasPrimaryLocation W43841568111 @default.
• W4384156811 hasRelatedWork W2899084033 @default.
• W4384156811 hasRelatedWork W3006943036 @default.
• W4384156811 hasRelatedWork W3027068985 @default.
• W4384156811 hasRelatedWork W4200511449 @default.
• W4384156811 hasRelatedWork W4206534706 @default.
• W4384156811 hasRelatedWork W4229079080 @default.
• W4384156811 hasRelatedWork W4299487748 @default.
• W4384156811 hasRelatedWork W4385957992 @default.
• W4384156811 hasRelatedWork W4385965371 @default.
• W4384156811 hasRelatedWork W4386025632 @default.
• W4384156811 hasVolume "402" @default.
• W4384156811 isParatext "false" @default.
• W4384156811 isRetracted "false" @default.
• W4384156811 workType "article" @default.