The Experts below are selected from a list of 3990 Experts worldwide ranked by ideXlab platform
Knut Magne Augestad - One of the best experts on this subject based on the ideXlab platform.
-
Predicting colorectal surgical complications using heterogeneous clinical data and kernel methods
Journal of biomedical informatics, 2016Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Arthur Revhaug, Rolv-ole Lindsetmo, Inma Mora-jiménez, Kim Erlend Mortensen, Knut Magne AugestadAbstract:Display Omitted Exploitation of clinical heterogeneous data for Anastomosis Leakage (AL) prediction.Use of clinical narrative in free text form, blood tests and physiological data.Weighted kernel summation for fusion of heterogeneous data was studied.AL risk assessment with posterior probabilities as an aid to raise alertness.The AL risk score improved when using multisource information. ObjectiveIn this work, we have developed a learning system capable of exploiting information conveyed by longitudinal Electronic Health Records (EHRs) for the prediction of a common postoperative complication, Anastomosis Leakage (AL), in a data-driven way and by fusing temporal population data from different and heterogeneous sources in the EHRs. Material and methodsWe used linear and non-linear kernel methods individually for each data source, and leveraging the powerful multiple kernels for their effective combination. To validate the system, we used data from the EHR of the gastrointestinal department at a university hospital. ResultsWe first investigated the early prediction performance from each data source separately, by computing Area Under the Curve values for processed free text (0.83), blood tests (0.74), and vital signs (0.65), respectively. When exploiting the heterogeneous data sources combined using the composite kernel framework, the prediction capabilities increased considerably (0.92). Finally, posterior probabilities were evaluated for risk assessment of patients as an aid for clinicians to raise alertness at an early stage, in order to act promptly for avoiding AL complications. DiscussionMachine-learning statistical model from EHR data can be useful to predict surgical complications. The combination of EHR extracted free text, blood samples values, and patient vital signs, improves the model performance. These results can be used as a framework for preoperative clinical decision support.
-
Support Vector Feature Selection for Early Detection of Anastomosis Leakage From Bag-of-Words in Electronic Health Records
IEEE Journal of Biomedical and Health Informatics, 2016Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Kim Mortensen, Arthur Revhaug, Rolv-ole Lindsetmo, Knut Magne Augestad, Robert JenssenAbstract:The free text in electronic health records (EHRs) conveys a huge amount of clinical information about health state and patient history. Despite a rapidly growing literature on the use of machine learning techniques for extracting this information, little effort has been invested toward feature selection and the features' corresponding medical interpretation. In this study, we focus on the task of early detection of Anastomosis Leakage (AL), a severe complication after elective surgery for colorectal cancer (CRC) surgery, using free text extracted from EHRs. We use a bag-of-words model to investigate the potential for feature selection strategies. The purpose is earlier detection of AL and prediction of AL with data generated in the EHR before the actual complication occur. Due to the high dimensionality of the data, we derive feature selection strategies using the robust support vector machine linear maximum margin classifier, by investigating: 1) a simple statistical criterion (leave-one-out-based test); 2) an intensive-computation statistical criterion (Bootstrap resampling); and 3) an advanced statistical criterion (kernel entropy). Results reveal a discriminatory power for early detection of complications after CRC (sensitivity 100%; specificity 72%). These results can be used to develop prediction models, based on EHR data, that can support surgeons and patients in the preoperative decision making phase.
-
Bootstrap resampling feature selection and Support Vector Machine for early detection of Anastomosis Leakage
IEEE-EMBS International Conference on Biomedical and Health Informatics (BHI), 2014Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Kim Mortensen, Arthur Revhaug, Rolv-ole Lindsetmo, Inma Mora-jiménez, Knut Magne AugestadAbstract:We propose a Bootstrap resampling approach for Feature Selection (FS) using the weights obtained by a linear Support Vector Machine (SVM) when it is applied to high-dimensional input spaces. We build our approach on a practical application with an extremely high-dimensional input space. The application is the detection of Anastomosis Leakage (AL) after colorectal cancer surgery using free text Bag-of-Words in Electronic Health Records (EHRs). Colorectal cancer is the third most common cancer type, and surgery is the only curative treatment, making the detection of AL of prime importance. The reduced input space obtained by the proposed FS strategy in combination with the linear SVM provided a much improved performance for early detection AL after colorectal cancer (earlier/final sensitivity 97%/100% and specificity 47%/89%). Further extensions of the method can be the basis for a principled FS strategy in high-dimensional input spaces.
-
BHI - Bootstrap resampling feature selection and Support Vector Machine for early detection of Anastomosis Leakage
IEEE-EMBS International Conference on Biomedical and Health Informatics (BHI), 2014Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Arthur Revhaug, Rolv-ole Lindsetmo, Inma Mora-jiménez, Kim Erlend Mortensen, Knut Magne AugestadAbstract:We propose a Bootstrap resampling approach for Feature Selection (FS) using the weights obtained by a linear Support Vector Machine (SVM) when it is applied to high-dimensional input spaces. We build our approach on a practical application with an extremely high-dimensional input space. The application is the detection of Anastomosis Leakage (AL) after colorectal cancer surgery using free text Bag-of-Words in Electronic Health Records (EHRs). Colorectal cancer is the third most common cancer type, and surgery is the only curative treatment, making the detection of AL of prime importance. The reduced input space obtained by the proposed FS strategy in combination with the linear SVM provided a much improved performance for early detection AL after colorectal cancer (earlier/final sensitivity 97%/100% and specificity 47%/89%). Further extensions of the method can be the basis for a principled FS strategy in high-dimensional input spaces.
Cristina Soguero-ruiz - One of the best experts on this subject based on the ideXlab platform.
-
Predicting colorectal surgical complications using heterogeneous clinical data and kernel methods
Journal of biomedical informatics, 2016Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Arthur Revhaug, Rolv-ole Lindsetmo, Inma Mora-jiménez, Kim Erlend Mortensen, Knut Magne AugestadAbstract:Display Omitted Exploitation of clinical heterogeneous data for Anastomosis Leakage (AL) prediction.Use of clinical narrative in free text form, blood tests and physiological data.Weighted kernel summation for fusion of heterogeneous data was studied.AL risk assessment with posterior probabilities as an aid to raise alertness.The AL risk score improved when using multisource information. ObjectiveIn this work, we have developed a learning system capable of exploiting information conveyed by longitudinal Electronic Health Records (EHRs) for the prediction of a common postoperative complication, Anastomosis Leakage (AL), in a data-driven way and by fusing temporal population data from different and heterogeneous sources in the EHRs. Material and methodsWe used linear and non-linear kernel methods individually for each data source, and leveraging the powerful multiple kernels for their effective combination. To validate the system, we used data from the EHR of the gastrointestinal department at a university hospital. ResultsWe first investigated the early prediction performance from each data source separately, by computing Area Under the Curve values for processed free text (0.83), blood tests (0.74), and vital signs (0.65), respectively. When exploiting the heterogeneous data sources combined using the composite kernel framework, the prediction capabilities increased considerably (0.92). Finally, posterior probabilities were evaluated for risk assessment of patients as an aid for clinicians to raise alertness at an early stage, in order to act promptly for avoiding AL complications. DiscussionMachine-learning statistical model from EHR data can be useful to predict surgical complications. The combination of EHR extracted free text, blood samples values, and patient vital signs, improves the model performance. These results can be used as a framework for preoperative clinical decision support.
-
Support Vector Feature Selection for Early Detection of Anastomosis Leakage From Bag-of-Words in Electronic Health Records
IEEE Journal of Biomedical and Health Informatics, 2016Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Kim Mortensen, Arthur Revhaug, Rolv-ole Lindsetmo, Knut Magne Augestad, Robert JenssenAbstract:The free text in electronic health records (EHRs) conveys a huge amount of clinical information about health state and patient history. Despite a rapidly growing literature on the use of machine learning techniques for extracting this information, little effort has been invested toward feature selection and the features' corresponding medical interpretation. In this study, we focus on the task of early detection of Anastomosis Leakage (AL), a severe complication after elective surgery for colorectal cancer (CRC) surgery, using free text extracted from EHRs. We use a bag-of-words model to investigate the potential for feature selection strategies. The purpose is earlier detection of AL and prediction of AL with data generated in the EHR before the actual complication occur. Due to the high dimensionality of the data, we derive feature selection strategies using the robust support vector machine linear maximum margin classifier, by investigating: 1) a simple statistical criterion (leave-one-out-based test); 2) an intensive-computation statistical criterion (Bootstrap resampling); and 3) an advanced statistical criterion (kernel entropy). Results reveal a discriminatory power for early detection of complications after CRC (sensitivity 100%; specificity 72%). These results can be used to develop prediction models, based on EHR data, that can support surgeons and patients in the preoperative decision making phase.
-
Bootstrap resampling feature selection and Support Vector Machine for early detection of Anastomosis Leakage
IEEE-EMBS International Conference on Biomedical and Health Informatics (BHI), 2014Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Kim Mortensen, Arthur Revhaug, Rolv-ole Lindsetmo, Inma Mora-jiménez, Knut Magne AugestadAbstract:We propose a Bootstrap resampling approach for Feature Selection (FS) using the weights obtained by a linear Support Vector Machine (SVM) when it is applied to high-dimensional input spaces. We build our approach on a practical application with an extremely high-dimensional input space. The application is the detection of Anastomosis Leakage (AL) after colorectal cancer surgery using free text Bag-of-Words in Electronic Health Records (EHRs). Colorectal cancer is the third most common cancer type, and surgery is the only curative treatment, making the detection of AL of prime importance. The reduced input space obtained by the proposed FS strategy in combination with the linear SVM provided a much improved performance for early detection AL after colorectal cancer (earlier/final sensitivity 97%/100% and specificity 47%/89%). Further extensions of the method can be the basis for a principled FS strategy in high-dimensional input spaces.
-
BHI - Bootstrap resampling feature selection and Support Vector Machine for early detection of Anastomosis Leakage
IEEE-EMBS International Conference on Biomedical and Health Informatics (BHI), 2014Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Arthur Revhaug, Rolv-ole Lindsetmo, Inma Mora-jiménez, Kim Erlend Mortensen, Knut Magne AugestadAbstract:We propose a Bootstrap resampling approach for Feature Selection (FS) using the weights obtained by a linear Support Vector Machine (SVM) when it is applied to high-dimensional input spaces. We build our approach on a practical application with an extremely high-dimensional input space. The application is the detection of Anastomosis Leakage (AL) after colorectal cancer surgery using free text Bag-of-Words in Electronic Health Records (EHRs). Colorectal cancer is the third most common cancer type, and surgery is the only curative treatment, making the detection of AL of prime importance. The reduced input space obtained by the proposed FS strategy in combination with the linear SVM provided a much improved performance for early detection AL after colorectal cancer (earlier/final sensitivity 97%/100% and specificity 47%/89%). Further extensions of the method can be the basis for a principled FS strategy in high-dimensional input spaces.
Rolv-ole Lindsetmo - One of the best experts on this subject based on the ideXlab platform.
-
Predicting colorectal surgical complications using heterogeneous clinical data and kernel methods
Journal of biomedical informatics, 2016Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Arthur Revhaug, Rolv-ole Lindsetmo, Inma Mora-jiménez, Kim Erlend Mortensen, Knut Magne AugestadAbstract:Display Omitted Exploitation of clinical heterogeneous data for Anastomosis Leakage (AL) prediction.Use of clinical narrative in free text form, blood tests and physiological data.Weighted kernel summation for fusion of heterogeneous data was studied.AL risk assessment with posterior probabilities as an aid to raise alertness.The AL risk score improved when using multisource information. ObjectiveIn this work, we have developed a learning system capable of exploiting information conveyed by longitudinal Electronic Health Records (EHRs) for the prediction of a common postoperative complication, Anastomosis Leakage (AL), in a data-driven way and by fusing temporal population data from different and heterogeneous sources in the EHRs. Material and methodsWe used linear and non-linear kernel methods individually for each data source, and leveraging the powerful multiple kernels for their effective combination. To validate the system, we used data from the EHR of the gastrointestinal department at a university hospital. ResultsWe first investigated the early prediction performance from each data source separately, by computing Area Under the Curve values for processed free text (0.83), blood tests (0.74), and vital signs (0.65), respectively. When exploiting the heterogeneous data sources combined using the composite kernel framework, the prediction capabilities increased considerably (0.92). Finally, posterior probabilities were evaluated for risk assessment of patients as an aid for clinicians to raise alertness at an early stage, in order to act promptly for avoiding AL complications. DiscussionMachine-learning statistical model from EHR data can be useful to predict surgical complications. The combination of EHR extracted free text, blood samples values, and patient vital signs, improves the model performance. These results can be used as a framework for preoperative clinical decision support.
-
Support Vector Feature Selection for Early Detection of Anastomosis Leakage From Bag-of-Words in Electronic Health Records
IEEE Journal of Biomedical and Health Informatics, 2016Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Kim Mortensen, Arthur Revhaug, Rolv-ole Lindsetmo, Knut Magne Augestad, Robert JenssenAbstract:The free text in electronic health records (EHRs) conveys a huge amount of clinical information about health state and patient history. Despite a rapidly growing literature on the use of machine learning techniques for extracting this information, little effort has been invested toward feature selection and the features' corresponding medical interpretation. In this study, we focus on the task of early detection of Anastomosis Leakage (AL), a severe complication after elective surgery for colorectal cancer (CRC) surgery, using free text extracted from EHRs. We use a bag-of-words model to investigate the potential for feature selection strategies. The purpose is earlier detection of AL and prediction of AL with data generated in the EHR before the actual complication occur. Due to the high dimensionality of the data, we derive feature selection strategies using the robust support vector machine linear maximum margin classifier, by investigating: 1) a simple statistical criterion (leave-one-out-based test); 2) an intensive-computation statistical criterion (Bootstrap resampling); and 3) an advanced statistical criterion (kernel entropy). Results reveal a discriminatory power for early detection of complications after CRC (sensitivity 100%; specificity 72%). These results can be used to develop prediction models, based on EHR data, that can support surgeons and patients in the preoperative decision making phase.
-
Bootstrap resampling feature selection and Support Vector Machine for early detection of Anastomosis Leakage
IEEE-EMBS International Conference on Biomedical and Health Informatics (BHI), 2014Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Kim Mortensen, Arthur Revhaug, Rolv-ole Lindsetmo, Inma Mora-jiménez, Knut Magne AugestadAbstract:We propose a Bootstrap resampling approach for Feature Selection (FS) using the weights obtained by a linear Support Vector Machine (SVM) when it is applied to high-dimensional input spaces. We build our approach on a practical application with an extremely high-dimensional input space. The application is the detection of Anastomosis Leakage (AL) after colorectal cancer surgery using free text Bag-of-Words in Electronic Health Records (EHRs). Colorectal cancer is the third most common cancer type, and surgery is the only curative treatment, making the detection of AL of prime importance. The reduced input space obtained by the proposed FS strategy in combination with the linear SVM provided a much improved performance for early detection AL after colorectal cancer (earlier/final sensitivity 97%/100% and specificity 47%/89%). Further extensions of the method can be the basis for a principled FS strategy in high-dimensional input spaces.
-
BHI - Bootstrap resampling feature selection and Support Vector Machine for early detection of Anastomosis Leakage
IEEE-EMBS International Conference on Biomedical and Health Informatics (BHI), 2014Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Arthur Revhaug, Rolv-ole Lindsetmo, Inma Mora-jiménez, Kim Erlend Mortensen, Knut Magne AugestadAbstract:We propose a Bootstrap resampling approach for Feature Selection (FS) using the weights obtained by a linear Support Vector Machine (SVM) when it is applied to high-dimensional input spaces. We build our approach on a practical application with an extremely high-dimensional input space. The application is the detection of Anastomosis Leakage (AL) after colorectal cancer surgery using free text Bag-of-Words in Electronic Health Records (EHRs). Colorectal cancer is the third most common cancer type, and surgery is the only curative treatment, making the detection of AL of prime importance. The reduced input space obtained by the proposed FS strategy in combination with the linear SVM provided a much improved performance for early detection AL after colorectal cancer (earlier/final sensitivity 97%/100% and specificity 47%/89%). Further extensions of the method can be the basis for a principled FS strategy in high-dimensional input spaces.
Arthur Revhaug - One of the best experts on this subject based on the ideXlab platform.
-
Predicting colorectal surgical complications using heterogeneous clinical data and kernel methods
Journal of biomedical informatics, 2016Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Arthur Revhaug, Rolv-ole Lindsetmo, Inma Mora-jiménez, Kim Erlend Mortensen, Knut Magne AugestadAbstract:Display Omitted Exploitation of clinical heterogeneous data for Anastomosis Leakage (AL) prediction.Use of clinical narrative in free text form, blood tests and physiological data.Weighted kernel summation for fusion of heterogeneous data was studied.AL risk assessment with posterior probabilities as an aid to raise alertness.The AL risk score improved when using multisource information. ObjectiveIn this work, we have developed a learning system capable of exploiting information conveyed by longitudinal Electronic Health Records (EHRs) for the prediction of a common postoperative complication, Anastomosis Leakage (AL), in a data-driven way and by fusing temporal population data from different and heterogeneous sources in the EHRs. Material and methodsWe used linear and non-linear kernel methods individually for each data source, and leveraging the powerful multiple kernels for their effective combination. To validate the system, we used data from the EHR of the gastrointestinal department at a university hospital. ResultsWe first investigated the early prediction performance from each data source separately, by computing Area Under the Curve values for processed free text (0.83), blood tests (0.74), and vital signs (0.65), respectively. When exploiting the heterogeneous data sources combined using the composite kernel framework, the prediction capabilities increased considerably (0.92). Finally, posterior probabilities were evaluated for risk assessment of patients as an aid for clinicians to raise alertness at an early stage, in order to act promptly for avoiding AL complications. DiscussionMachine-learning statistical model from EHR data can be useful to predict surgical complications. The combination of EHR extracted free text, blood samples values, and patient vital signs, improves the model performance. These results can be used as a framework for preoperative clinical decision support.
-
Support Vector Feature Selection for Early Detection of Anastomosis Leakage From Bag-of-Words in Electronic Health Records
IEEE Journal of Biomedical and Health Informatics, 2016Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Kim Mortensen, Arthur Revhaug, Rolv-ole Lindsetmo, Knut Magne Augestad, Robert JenssenAbstract:The free text in electronic health records (EHRs) conveys a huge amount of clinical information about health state and patient history. Despite a rapidly growing literature on the use of machine learning techniques for extracting this information, little effort has been invested toward feature selection and the features' corresponding medical interpretation. In this study, we focus on the task of early detection of Anastomosis Leakage (AL), a severe complication after elective surgery for colorectal cancer (CRC) surgery, using free text extracted from EHRs. We use a bag-of-words model to investigate the potential for feature selection strategies. The purpose is earlier detection of AL and prediction of AL with data generated in the EHR before the actual complication occur. Due to the high dimensionality of the data, we derive feature selection strategies using the robust support vector machine linear maximum margin classifier, by investigating: 1) a simple statistical criterion (leave-one-out-based test); 2) an intensive-computation statistical criterion (Bootstrap resampling); and 3) an advanced statistical criterion (kernel entropy). Results reveal a discriminatory power for early detection of complications after CRC (sensitivity 100%; specificity 72%). These results can be used to develop prediction models, based on EHR data, that can support surgeons and patients in the preoperative decision making phase.
-
Bootstrap resampling feature selection and Support Vector Machine for early detection of Anastomosis Leakage
IEEE-EMBS International Conference on Biomedical and Health Informatics (BHI), 2014Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Kim Mortensen, Arthur Revhaug, Rolv-ole Lindsetmo, Inma Mora-jiménez, Knut Magne AugestadAbstract:We propose a Bootstrap resampling approach for Feature Selection (FS) using the weights obtained by a linear Support Vector Machine (SVM) when it is applied to high-dimensional input spaces. We build our approach on a practical application with an extremely high-dimensional input space. The application is the detection of Anastomosis Leakage (AL) after colorectal cancer surgery using free text Bag-of-Words in Electronic Health Records (EHRs). Colorectal cancer is the third most common cancer type, and surgery is the only curative treatment, making the detection of AL of prime importance. The reduced input space obtained by the proposed FS strategy in combination with the linear SVM provided a much improved performance for early detection AL after colorectal cancer (earlier/final sensitivity 97%/100% and specificity 47%/89%). Further extensions of the method can be the basis for a principled FS strategy in high-dimensional input spaces.
-
BHI - Bootstrap resampling feature selection and Support Vector Machine for early detection of Anastomosis Leakage
IEEE-EMBS International Conference on Biomedical and Health Informatics (BHI), 2014Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Arthur Revhaug, Rolv-ole Lindsetmo, Inma Mora-jiménez, Kim Erlend Mortensen, Knut Magne AugestadAbstract:We propose a Bootstrap resampling approach for Feature Selection (FS) using the weights obtained by a linear Support Vector Machine (SVM) when it is applied to high-dimensional input spaces. We build our approach on a practical application with an extremely high-dimensional input space. The application is the detection of Anastomosis Leakage (AL) after colorectal cancer surgery using free text Bag-of-Words in Electronic Health Records (EHRs). Colorectal cancer is the third most common cancer type, and surgery is the only curative treatment, making the detection of AL of prime importance. The reduced input space obtained by the proposed FS strategy in combination with the linear SVM provided a much improved performance for early detection AL after colorectal cancer (earlier/final sensitivity 97%/100% and specificity 47%/89%). Further extensions of the method can be the basis for a principled FS strategy in high-dimensional input spaces.
Fred Godtliebsen - One of the best experts on this subject based on the ideXlab platform.
-
Predicting colorectal surgical complications using heterogeneous clinical data and kernel methods
Journal of biomedical informatics, 2016Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Arthur Revhaug, Rolv-ole Lindsetmo, Inma Mora-jiménez, Kim Erlend Mortensen, Knut Magne AugestadAbstract:Display Omitted Exploitation of clinical heterogeneous data for Anastomosis Leakage (AL) prediction.Use of clinical narrative in free text form, blood tests and physiological data.Weighted kernel summation for fusion of heterogeneous data was studied.AL risk assessment with posterior probabilities as an aid to raise alertness.The AL risk score improved when using multisource information. ObjectiveIn this work, we have developed a learning system capable of exploiting information conveyed by longitudinal Electronic Health Records (EHRs) for the prediction of a common postoperative complication, Anastomosis Leakage (AL), in a data-driven way and by fusing temporal population data from different and heterogeneous sources in the EHRs. Material and methodsWe used linear and non-linear kernel methods individually for each data source, and leveraging the powerful multiple kernels for their effective combination. To validate the system, we used data from the EHR of the gastrointestinal department at a university hospital. ResultsWe first investigated the early prediction performance from each data source separately, by computing Area Under the Curve values for processed free text (0.83), blood tests (0.74), and vital signs (0.65), respectively. When exploiting the heterogeneous data sources combined using the composite kernel framework, the prediction capabilities increased considerably (0.92). Finally, posterior probabilities were evaluated for risk assessment of patients as an aid for clinicians to raise alertness at an early stage, in order to act promptly for avoiding AL complications. DiscussionMachine-learning statistical model from EHR data can be useful to predict surgical complications. The combination of EHR extracted free text, blood samples values, and patient vital signs, improves the model performance. These results can be used as a framework for preoperative clinical decision support.
-
Support Vector Feature Selection for Early Detection of Anastomosis Leakage From Bag-of-Words in Electronic Health Records
IEEE Journal of Biomedical and Health Informatics, 2016Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Kim Mortensen, Arthur Revhaug, Rolv-ole Lindsetmo, Knut Magne Augestad, Robert JenssenAbstract:The free text in electronic health records (EHRs) conveys a huge amount of clinical information about health state and patient history. Despite a rapidly growing literature on the use of machine learning techniques for extracting this information, little effort has been invested toward feature selection and the features' corresponding medical interpretation. In this study, we focus on the task of early detection of Anastomosis Leakage (AL), a severe complication after elective surgery for colorectal cancer (CRC) surgery, using free text extracted from EHRs. We use a bag-of-words model to investigate the potential for feature selection strategies. The purpose is earlier detection of AL and prediction of AL with data generated in the EHR before the actual complication occur. Due to the high dimensionality of the data, we derive feature selection strategies using the robust support vector machine linear maximum margin classifier, by investigating: 1) a simple statistical criterion (leave-one-out-based test); 2) an intensive-computation statistical criterion (Bootstrap resampling); and 3) an advanced statistical criterion (kernel entropy). Results reveal a discriminatory power for early detection of complications after CRC (sensitivity 100%; specificity 72%). These results can be used to develop prediction models, based on EHR data, that can support surgeons and patients in the preoperative decision making phase.
-
Bootstrap resampling feature selection and Support Vector Machine for early detection of Anastomosis Leakage
IEEE-EMBS International Conference on Biomedical and Health Informatics (BHI), 2014Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Kim Mortensen, Arthur Revhaug, Rolv-ole Lindsetmo, Inma Mora-jiménez, Knut Magne AugestadAbstract:We propose a Bootstrap resampling approach for Feature Selection (FS) using the weights obtained by a linear Support Vector Machine (SVM) when it is applied to high-dimensional input spaces. We build our approach on a practical application with an extremely high-dimensional input space. The application is the detection of Anastomosis Leakage (AL) after colorectal cancer surgery using free text Bag-of-Words in Electronic Health Records (EHRs). Colorectal cancer is the third most common cancer type, and surgery is the only curative treatment, making the detection of AL of prime importance. The reduced input space obtained by the proposed FS strategy in combination with the linear SVM provided a much improved performance for early detection AL after colorectal cancer (earlier/final sensitivity 97%/100% and specificity 47%/89%). Further extensions of the method can be the basis for a principled FS strategy in high-dimensional input spaces.
-
BHI - Bootstrap resampling feature selection and Support Vector Machine for early detection of Anastomosis Leakage
IEEE-EMBS International Conference on Biomedical and Health Informatics (BHI), 2014Co-Authors: Cristina Soguero-ruiz, Kristian Hindberg, José Luis Rojo-Álvarez, Stein Olav Skrøvseth, Fred Godtliebsen, Arthur Revhaug, Rolv-ole Lindsetmo, Inma Mora-jiménez, Kim Erlend Mortensen, Knut Magne AugestadAbstract:We propose a Bootstrap resampling approach for Feature Selection (FS) using the weights obtained by a linear Support Vector Machine (SVM) when it is applied to high-dimensional input spaces. We build our approach on a practical application with an extremely high-dimensional input space. The application is the detection of Anastomosis Leakage (AL) after colorectal cancer surgery using free text Bag-of-Words in Electronic Health Records (EHRs). Colorectal cancer is the third most common cancer type, and surgery is the only curative treatment, making the detection of AL of prime importance. The reduced input space obtained by the proposed FS strategy in combination with the linear SVM provided a much improved performance for early detection AL after colorectal cancer (earlier/final sensitivity 97%/100% and specificity 47%/89%). Further extensions of the method can be the basis for a principled FS strategy in high-dimensional input spaces.