Matches in SemOpenAlex for { <https://semopenalex.org/work/W3014626030> ?p ?o ?g. }
Showing items 1 to 53 of
53
with 100 items per page.
- W3014626030 endingPage "S385" @default.
- W3014626030 startingPage "S385" @default.
- W3014626030 abstract "Purpose The extreme difficulty of carrying out randomized studies in lung transplantation leads us to consider statistical methods derived from artificial intelligence. Previous studies in kidney and liver transplantations, as well as in heart surgery, showed better predictive ability than traditional statistical analysis to predict postoperative outcomes. Methods We analyzed a prospective database of all 410 double lung transplantations performed in our center, from January 2012 to June 2018. We used a Random Forest approach over 284 variables, in order to predict one-year mortality. Performance of the predictive model is evaluated at successive temporal stages of the transplantation process. Variables are incrementally acquired during the process, starting with patient-only variables at stage 1, ending with patient, donor and surgery-related measurements at stage 12. At each stage of the process, a machine-learning model has been trained based on available variables. A 80-20-cross-validation procedure has been performed at each stage, and repeated 40 times, resulting in a set of 40 area under ROC curve scores, whose distribution has been summarized using boxplots (median, 25 and 75 percentiles). Gini score allowed to each variable its weight in the model. Results The AUROC performance starts at 0.65 for step 1 and reaches 0.75 (figure) at the final step. Gini score found Lung Allocation Score as the best predictor of one-year mortality (3.17), followed by hyperlactatemia at second lung implantation (2.62), PaO2/FiO2 ratio at end-surgery (2.46) and age (2.35). Second lung ischemic time was a the 20th position in Gini Score (1.6). Conclusion Conclusion: Machine-learning approach is feasible to predict one-year mortality after lung transplantation. This study is encouraging to go further in the analysis of our database." @default.
- W3014626030 created "2020-04-10" @default.
- W3014626030 creator A5029215808 @default.
- W3014626030 creator A5039476981 @default.
- W3014626030 creator A5059969352 @default.
- W3014626030 creator A5091571501 @default.
- W3014626030 date "2020-04-01" @default.
- W3014626030 modified "2023-10-17" @default.
- W3014626030 title "Machine Learning in Lung Transplantation" @default.
- W3014626030 doi "https://doi.org/10.1016/j.healun.2020.01.497" @default.
- W3014626030 hasPublicationYear "2020" @default.
- W3014626030 type Work @default.
- W3014626030 sameAs 3014626030 @default.
- W3014626030 citedByCount "3" @default.
- W3014626030 countsByYear W30146260302022 @default.
- W3014626030 crossrefType "journal-article" @default.
- W3014626030 hasAuthorship W3014626030A5029215808 @default.
- W3014626030 hasAuthorship W3014626030A5039476981 @default.
- W3014626030 hasAuthorship W3014626030A5059969352 @default.
- W3014626030 hasAuthorship W3014626030A5091571501 @default.
- W3014626030 hasConcept C126322002 @default.
- W3014626030 hasConcept C177713679 @default.
- W3014626030 hasConcept C2777714996 @default.
- W3014626030 hasConcept C2781448352 @default.
- W3014626030 hasConcept C2911091166 @default.
- W3014626030 hasConcept C71924100 @default.
- W3014626030 hasConceptScore W3014626030C126322002 @default.
- W3014626030 hasConceptScore W3014626030C177713679 @default.
- W3014626030 hasConceptScore W3014626030C2777714996 @default.
- W3014626030 hasConceptScore W3014626030C2781448352 @default.
- W3014626030 hasConceptScore W3014626030C2911091166 @default.
- W3014626030 hasConceptScore W3014626030C71924100 @default.
- W3014626030 hasIssue "4" @default.
- W3014626030 hasLocation W30146260301 @default.
- W3014626030 hasOpenAccess W3014626030 @default.
- W3014626030 hasPrimaryLocation W30146260301 @default.
- W3014626030 hasRelatedWork W1986842757 @default.
- W3014626030 hasRelatedWork W2012591382 @default.
- W3014626030 hasRelatedWork W2033140325 @default.
- W3014626030 hasRelatedWork W2318299045 @default.
- W3014626030 hasRelatedWork W2417519090 @default.
- W3014626030 hasRelatedWork W2471678635 @default.
- W3014626030 hasRelatedWork W2795716061 @default.
- W3014626030 hasRelatedWork W3031286556 @default.
- W3014626030 hasRelatedWork W3141986935 @default.
- W3014626030 hasRelatedWork W4210884284 @default.
- W3014626030 hasVolume "39" @default.
- W3014626030 isParatext "false" @default.
- W3014626030 isRetracted "false" @default.
- W3014626030 magId "3014626030" @default.
- W3014626030 workType "article" @default.