FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2890104787> ?p ?o ?g. }

• W2890104787 endingPage "e430" @default.
• W2890104787 startingPage "e430" @default.
• W2890104787 abstract "Abstract.Abstract1: To develop and validate an artificial intelligence (AI) system that directly predicts the probability that an embryo can lead to a pregnancy with a fetal heart (FH) following embryo transfer (ET), through the automated analysis of raw time-lapse (TL) video sequences. The study used AI to analyse TL videos. All embryos were included regardless of their development stage, grade, ploidy, fertilization status, culture method and duration. IVF and ICSI cycles were both included. The TL cycles were collected from 8 laboratories in 4 countries between 2014 and early 2018. All patients were included with a mean age of 35.6 years (age range: 22-50 years). The primary outcome measure was the presence of a FH. The two possible outcomes were defined as: (a) YES (FH detected at 6-8 weeks of gestation), or (b) NO (ET with a negative FH or discarded embryos due to cleavage arrest, abnormal fertilization, aneuploidy or very poor morphology); single ET and double ET with either a negative FH or two FHs were included The AI is a deep neural network trained with a large multicentre, multinational dataset of TL videos to perform binary classification task of predicting FH outcome. From training, it empirically derived spatial-temporal features from the TL videos that are effective in predicting FH outcome independent of maternal age and specific culture method. Once trained, the neural network took a raw TL video as the input and produced a continuous probabilistic score (range 0-100%) as the output. This score was calibrated to the probability that the given embryo would develop into a FH. A total of 10,208 embryos from 1,603 patients were extracted. 2000 embryos were transferred, 73% of which were single ET, with 670 embryos resulting in a FH. 8,389 embryos had known FH outcome data and were used for training or validation. The other 1,819 embryos are still frozen or were a part of a double ET resulting in only one FH and were not used. The full 8,389 embryos were split into a 80% training and a 20% validation dataset, using 5-fold stratified cross-validation, with embryos used in training never used in validation. The AI discriminative power was evaluated using the Receiver Operator Characteristic (ROC) analysis. Mean Area Under the Curve (AUC) from 5-fold cross-validation was 0.93, 95% CI [0.92, 0.94] for predicting FH outcome on the validation set; indicating that 93% of the time, the AI will score embryos that lead to a FH higher than embryos that will not. The AI is fully automated, requiring no human input and hence is not subjected to any inter-grader or intra-grader variability. The AI presented is an automated, non-invasive, objective, and highly discriminative embryo ranking tool for TL generated embryos to predict the probability that an embryo will result in a FH. The resulting probability score provides a ranking for a cohort of embryos giving an optimized order for transfer. Multicentre RCT validation of this AI is ongoing." @default.
• W2890104787 created "2018-09-27" @default.
• W2890104787 creator A5004472184 @default.
• W2890104787 creator A5014865393 @default.
• W2890104787 creator A5050479075 @default.
• W2890104787 creator A5085008925 @default.
• W2890104787 date "2018-09-01" @default.
• W2890104787 modified "2023-09-30" @default.
• W2890104787 title "Artificial intelligence as a novel approach for embryo selection" @default.
• W2890104787 doi "https://doi.org/10.1016/j.fertnstert.2018.08.034" @default.
• W2890104787 hasPublicationYear "2018" @default.
• W2890104787 type Work @default.
• W2890104787 sameAs 2890104787 @default.
• W2890104787 citedByCount "8" @default.
• W2890104787 countsByYear W28901047872019 @default.
• W2890104787 countsByYear W28901047872020 @default.
• W2890104787 countsByYear W28901047872021 @default.
• W2890104787 countsByYear W28901047872023 @default.
• W2890104787 crossrefType "journal-article" @default.
• W2890104787 hasAuthorship W2890104787A5004472184 @default.
• W2890104787 hasAuthorship W2890104787A5014865393 @default.
• W2890104787 hasAuthorship W2890104787A5050479075 @default.
• W2890104787 hasAuthorship W2890104787A5085008925 @default.
• W2890104787 hasBestOaLocation W28901047871 @default.
• W2890104787 hasConcept C154945302 @default.
• W2890104787 hasConcept C160099875 @default.
• W2890104787 hasConcept C196843134 @default.
• W2890104787 hasConcept C2779234561 @default.
• W2890104787 hasConcept C41008148 @default.
• W2890104787 hasConcept C54355233 @default.
• W2890104787 hasConcept C86803240 @default.
• W2890104787 hasConceptScore W2890104787C154945302 @default.
• W2890104787 hasConceptScore W2890104787C160099875 @default.
• W2890104787 hasConceptScore W2890104787C196843134 @default.
• W2890104787 hasConceptScore W2890104787C2779234561 @default.
• W2890104787 hasConceptScore W2890104787C41008148 @default.
• W2890104787 hasConceptScore W2890104787C54355233 @default.
• W2890104787 hasConceptScore W2890104787C86803240 @default.
• W2890104787 hasIssue "4" @default.
• W2890104787 hasLocation W28901047871 @default.
• W2890104787 hasOpenAccess W2890104787 @default.
• W2890104787 hasPrimaryLocation W28901047871 @default.
• W2890104787 hasRelatedWork W1892544028 @default.
• W2890104787 hasRelatedWork W1970709224 @default.
• W2890104787 hasRelatedWork W2005627939 @default.
• W2890104787 hasRelatedWork W2047779449 @default.
• W2890104787 hasRelatedWork W2138547054 @default.
• W2890104787 hasRelatedWork W2169340257 @default.
• W2890104787 hasRelatedWork W2415301716 @default.
• W2890104787 hasRelatedWork W2509789438 @default.
• W2890104787 hasRelatedWork W2596000028 @default.
• W2890104787 hasRelatedWork W2596232480 @default.
• W2890104787 hasVolume "110" @default.
• W2890104787 isParatext "false" @default.
• W2890104787 isRetracted "false" @default.
• W2890104787 magId "2890104787" @default.
• W2890104787 workType "article" @default.