Machine learning based algorithms to impute PaO2 from SpO2 values and development of an online calculator
Abstract We created an online calculator using machine learning (ML) algorithms to impute the partial pressure of oxygen (PaO2)/fraction of delivered oxygen (FiO2) ratio using the non-invasive peripheral saturation of oxygen (SpO2) and compared the accuracy of the ML models we developed to published...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2022-05-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-022-12419-7 |
| _version_ | 1828822557021700096 |
|---|---|
| author | Shuangxia Ren Jill A. Zupetic Mohammadreza Tabary Rebecca DeSensi Mehdi Nouraie Xinghua Lu Richard D. Boyce Janet S. Lee |
| author_facet | Shuangxia Ren Jill A. Zupetic Mohammadreza Tabary Rebecca DeSensi Mehdi Nouraie Xinghua Lu Richard D. Boyce Janet S. Lee |
| author_sort | Shuangxia Ren |
| collection | DOAJ |
| description | Abstract We created an online calculator using machine learning (ML) algorithms to impute the partial pressure of oxygen (PaO2)/fraction of delivered oxygen (FiO2) ratio using the non-invasive peripheral saturation of oxygen (SpO2) and compared the accuracy of the ML models we developed to published equations. We generated three ML algorithms (neural network, regression, and kernel-based methods) using seven clinical variable features (N = 9900 ICU events) and subsequently three features (N = 20,198 ICU events) as input into the models. Data from mechanically ventilated ICU patients were obtained from the publicly available Medical Information Mart for Intensive Care (MIMIC III) database and used for analysis. Compared to seven features, three features (SpO2, FiO2 and PEEP) were sufficient to impute PaO2 from the SpO2. Any of the ML models enabled imputation of PaO2 from the SpO2 with lower error and showed greater accuracy in predicting PaO2/FiO2 ≤ 150 compared to the previously published log-linear and non-linear equations. To address potential hidden hypoxemia that occurs more frequently in Black patients, we conducted sensitivity analysis and show ML models outperformed published equations in both Black and White patients. Imputation using data from an independent validation cohort of ICU patients (N = 133) showed greater accuracy with ML models. |
| first_indexed | 2024-12-12T13:17:14Z |
| format | Article |
| id | doaj.art-9d4935ba198f45efa9af4ef277ca6956 |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-12-12T13:17:14Z |
| publishDate | 2022-05-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-9d4935ba198f45efa9af4ef277ca69562022-12-22T00:23:24ZengNature PortfolioScientific Reports2045-23222022-05-0112111010.1038/s41598-022-12419-7Machine learning based algorithms to impute PaO2 from SpO2 values and development of an online calculatorShuangxia Ren0Jill A. Zupetic1Mohammadreza Tabary2Rebecca DeSensi3Mehdi Nouraie4Xinghua Lu5Richard D. Boyce6Janet S. Lee7Intelligent Systems Program, University of PittsburghDivision of Pulmonary, Allergy, and Critical Care Medicine, University of PittsburghDivision of Pulmonary, Allergy, and Critical Care Medicine, University of PittsburghDivision of Pulmonary, Allergy, and Critical Care Medicine, University of PittsburghDivision of Pulmonary, Allergy, and Critical Care Medicine, University of PittsburghIntelligent Systems Program, University of PittsburghIntelligent Systems Program, University of PittsburghDivision of Pulmonary, Allergy, and Critical Care Medicine, University of PittsburghAbstract We created an online calculator using machine learning (ML) algorithms to impute the partial pressure of oxygen (PaO2)/fraction of delivered oxygen (FiO2) ratio using the non-invasive peripheral saturation of oxygen (SpO2) and compared the accuracy of the ML models we developed to published equations. We generated three ML algorithms (neural network, regression, and kernel-based methods) using seven clinical variable features (N = 9900 ICU events) and subsequently three features (N = 20,198 ICU events) as input into the models. Data from mechanically ventilated ICU patients were obtained from the publicly available Medical Information Mart for Intensive Care (MIMIC III) database and used for analysis. Compared to seven features, three features (SpO2, FiO2 and PEEP) were sufficient to impute PaO2 from the SpO2. Any of the ML models enabled imputation of PaO2 from the SpO2 with lower error and showed greater accuracy in predicting PaO2/FiO2 ≤ 150 compared to the previously published log-linear and non-linear equations. To address potential hidden hypoxemia that occurs more frequently in Black patients, we conducted sensitivity analysis and show ML models outperformed published equations in both Black and White patients. Imputation using data from an independent validation cohort of ICU patients (N = 133) showed greater accuracy with ML models.https://doi.org/10.1038/s41598-022-12419-7 |
| spellingShingle | Shuangxia Ren Jill A. Zupetic Mohammadreza Tabary Rebecca DeSensi Mehdi Nouraie Xinghua Lu Richard D. Boyce Janet S. Lee Machine learning based algorithms to impute PaO2 from SpO2 values and development of an online calculator Scientific Reports |
| title | Machine learning based algorithms to impute PaO2 from SpO2 values and development of an online calculator |
| title_full | Machine learning based algorithms to impute PaO2 from SpO2 values and development of an online calculator |
| title_fullStr | Machine learning based algorithms to impute PaO2 from SpO2 values and development of an online calculator |
| title_full_unstemmed | Machine learning based algorithms to impute PaO2 from SpO2 values and development of an online calculator |
| title_short | Machine learning based algorithms to impute PaO2 from SpO2 values and development of an online calculator |
| title_sort | machine learning based algorithms to impute pao2 from spo2 values and development of an online calculator |
| url | https://doi.org/10.1038/s41598-022-12419-7 |
| work_keys_str_mv | AT shuangxiaren machinelearningbasedalgorithmstoimputepao2fromspo2valuesanddevelopmentofanonlinecalculator AT jillazupetic machinelearningbasedalgorithmstoimputepao2fromspo2valuesanddevelopmentofanonlinecalculator AT mohammadrezatabary machinelearningbasedalgorithmstoimputepao2fromspo2valuesanddevelopmentofanonlinecalculator AT rebeccadesensi machinelearningbasedalgorithmstoimputepao2fromspo2valuesanddevelopmentofanonlinecalculator AT mehdinouraie machinelearningbasedalgorithmstoimputepao2fromspo2valuesanddevelopmentofanonlinecalculator AT xinghualu machinelearningbasedalgorithmstoimputepao2fromspo2valuesanddevelopmentofanonlinecalculator AT richarddboyce machinelearningbasedalgorithmstoimputepao2fromspo2valuesanddevelopmentofanonlinecalculator AT janetslee machinelearningbasedalgorithmstoimputepao2fromspo2valuesanddevelopmentofanonlinecalculator |