FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386906243> ?p ?o ?g. }

• W4386906243 endingPage "S7" @default.
• W4386906243 startingPage "S7" @default.
• W4386906243 abstract "We have published the use of convolutional neural networks (CNNs) to estimate the probability of structural heart diseases including hypertrophic cardiomyopathy (HCM), low ejection fraction (EF), and amyloidosis from the electrocardiogram (ECG). These studies used a 10 second 12-lead ECG as inputs to train the network. To compare the performance of a single heartbeat to a 10 second ECG when training neural networks for structural heart disease screening. Unique research authorized patients were identified who had an ECG within 14 days of a transthoracic echocardiogram (TTE) to create a pool of ECG-TTE pairs. ECG-TTE pairs belonging to patients previously identified in our published studies with amyloidosis (2,095 pairs) and HCM (1,812 pairs) formed the basis of our cohort. Remaining pairs were labeled based on EF (low EF defined as <=40%) for age and sex matching. Amyloid and HCM ECG-TTE pairs were then each age and sex matched in a desired ratio of 1:1:12 (disease: low EF: normal EF). The final cohort of 51,080 patients was 63% male with an average age of 59 ± 14 years with a low EF prevalence of 8%. This cohort was then split once randomly into training, validation, and testing tests in a ratio of 7:1:2. Splitting the cohort once ensured that the same ECGs were used for training, validation, and testing of all the ECG median models. The median beat 12-lead models estimated probabilities of amyloidosis, low EF, and HCM with an area under the curve (AUC) of 0.94 (95% Confidence Interval (CI) 0.92-0.95), 0.89 (95% CI 0.88-0.90), and 0.94 (95% CI 0.92-0.95), respectively. Patient sex was estimated with AUC of 0.93 (95% CI 0.92-0.93). The 12-lead median models show comparable performance to the 10 second 12-lead models. This finding suggests that the salient features for ECG model predictions are encompassed just in a single heartbeat, which may guide future model explainability studies." @default.
• W4386906243 created "2023-09-21" @default.
• W4386906243 creator A5001963115 @default.
• W4386906243 creator A5020289692 @default.
• W4386906243 creator A5023936819 @default.
• W4386906243 creator A5024430748 @default.
• W4386906243 creator A5046460674 @default.
• W4386906243 date "2023-10-01" @default.
• W4386906243 modified "2023-09-27" @default.
• W4386906243 title "12-LEAD MEDIAN BEAT IS ALL YOU NEED FOR STRUCTURAL HEART DISEASE SCREENING" @default.
• W4386906243 doi "https://doi.org/10.1016/j.cvdhj.2023.08.015" @default.
• W4386906243 hasPublicationYear "2023" @default.
• W4386906243 type Work @default.
• W4386906243 citedByCount "0" @default.
• W4386906243 crossrefType "journal-article" @default.
• W4386906243 hasAuthorship W4386906243A5001963115 @default.
• W4386906243 hasAuthorship W4386906243A5020289692 @default.
• W4386906243 hasAuthorship W4386906243A5023936819 @default.
• W4386906243 hasAuthorship W4386906243A5024430748 @default.
• W4386906243 hasAuthorship W4386906243A5046460674 @default.
• W4386906243 hasBestOaLocation W43869062431 @default.
• W4386906243 hasConcept C126322002 @default.
• W4386906243 hasConcept C164705383 @default.
• W4386906243 hasConcept C2777607188 @default.
• W4386906243 hasConcept C2778198053 @default.
• W4386906243 hasConcept C2779951007 @default.
• W4386906243 hasConcept C2780040984 @default.
• W4386906243 hasConcept C2780074459 @default.
• W4386906243 hasConcept C2780185194 @default.
• W4386906243 hasConcept C71924100 @default.
• W4386906243 hasConcept C72563966 @default.
• W4386906243 hasConcept C78085059 @default.
• W4386906243 hasConceptScore W4386906243C126322002 @default.
• W4386906243 hasConceptScore W4386906243C164705383 @default.
• W4386906243 hasConceptScore W4386906243C2777607188 @default.
• W4386906243 hasConceptScore W4386906243C2778198053 @default.
• W4386906243 hasConceptScore W4386906243C2779951007 @default.
• W4386906243 hasConceptScore W4386906243C2780040984 @default.
• W4386906243 hasConceptScore W4386906243C2780074459 @default.
• W4386906243 hasConceptScore W4386906243C2780185194 @default.
• W4386906243 hasConceptScore W4386906243C71924100 @default.
• W4386906243 hasConceptScore W4386906243C72563966 @default.
• W4386906243 hasConceptScore W4386906243C78085059 @default.
• W4386906243 hasIssue "5" @default.
• W4386906243 hasLocation W43869062431 @default.
• W4386906243 hasOpenAccess W4386906243 @default.
• W4386906243 hasPrimaryLocation W43869062431 @default.
• W4386906243 hasRelatedWork W2051712573 @default.
• W4386906243 hasRelatedWork W2077083067 @default.
• W4386906243 hasRelatedWork W2091139515 @default.
• W4386906243 hasRelatedWork W2167036398 @default.
• W4386906243 hasRelatedWork W2355594703 @default.
• W4386906243 hasRelatedWork W2361407492 @default.
• W4386906243 hasRelatedWork W2538662566 @default.
• W4386906243 hasRelatedWork W2965474825 @default.
• W4386906243 hasRelatedWork W3033076790 @default.
• W4386906243 hasRelatedWork W4382048704 @default.
• W4386906243 hasVolume "4" @default.
• W4386906243 isParatext "false" @default.
• W4386906243 isRetracted "false" @default.
• W4386906243 workType "article" @default.