A deep learning model enables accurate prediction and quantification of pulmonary edema from chest X-rays
Abstract Background A quantitative assessment of pulmonary edema is important because the clinical severity can range from mild impairment to life threatening. A quantitative surrogate measure, although invasive, for pulmonary edema is the extravascular lung water index (EVLWI) extracted from the tr...
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2023-05-01
|
| Series: | Critical Care |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s13054-023-04426-5 |
| _version_ | 1797818019756376064 |
|---|---|
| author | Dominik Schulz Sebastian Rasch Markus Heilmaier Rami Abbassi Alexander Poszler Jörg Ulrich Manuel Steinhardt Georgios A. Kaissis Roland M. Schmid Rickmer Braren Tobias Lahmer |
| author_facet | Dominik Schulz Sebastian Rasch Markus Heilmaier Rami Abbassi Alexander Poszler Jörg Ulrich Manuel Steinhardt Georgios A. Kaissis Roland M. Schmid Rickmer Braren Tobias Lahmer |
| author_sort | Dominik Schulz |
| collection | DOAJ |
| description | Abstract Background A quantitative assessment of pulmonary edema is important because the clinical severity can range from mild impairment to life threatening. A quantitative surrogate measure, although invasive, for pulmonary edema is the extravascular lung water index (EVLWI) extracted from the transpulmonary thermodilution (TPTD). Severity of edema from chest X-rays, to date is based on the subjective classification of radiologists. In this work, we use machine learning to quantitatively predict the severity of pulmonary edema from chest radiography. Methods We retrospectively included 471 X-rays from 431 patients who underwent chest radiography and TPTD measurement within 24 h at our intensive care unit. The EVLWI extracted from the TPTD was used as a quantitative measure for pulmonary edema. We used a deep learning approach and binned the data into two, three, four and five classes increasing the resolution of the EVLWI prediction from the X-rays. Results The accuracy, area under the receiver operating characteristic curve (AUROC) and Mathews correlation coefficient (MCC) in the binary classification models (EVLWI < 15, ≥ 15) were 0.93 (accuracy), 0.98 (AUROC) and 0.86(MCC). In the three multiclass models, the accuracy ranged between 0.90 and 0.95, the AUROC between 0.97 and 0.99 and the MCC between 0.86 and 0.92. Conclusion Deep learning can quantify pulmonary edema as measured by EVLWI with high accuracy. |
| first_indexed | 2024-03-13T09:02:04Z |
| format | Article |
| id | doaj.art-d25d726d59194ba085e11a819e301134 |
| institution | Directory Open Access Journal |
| issn | 1364-8535 |
| language | English |
| last_indexed | 2024-03-13T09:02:04Z |
| publishDate | 2023-05-01 |
| publisher | BMC |
| record_format | Article |
| series | Critical Care |
| spelling | doaj.art-d25d726d59194ba085e11a819e3011342023-05-28T11:18:06ZengBMCCritical Care1364-85352023-05-012711410.1186/s13054-023-04426-5A deep learning model enables accurate prediction and quantification of pulmonary edema from chest X-raysDominik Schulz0Sebastian Rasch1Markus Heilmaier2Rami Abbassi3Alexander Poszler4Jörg Ulrich5Manuel Steinhardt6Georgios A. Kaissis7Roland M. Schmid8Rickmer Braren9Tobias Lahmer10Klinik und Poliklinik für Innere Medizin II, Klinikum rechts der IsarKlinik und Poliklinik für Innere Medizin II, Klinikum rechts der IsarKlinik und Poliklinik für Innere Medizin II, Klinikum rechts der IsarKlinik und Poliklinik für Innere Medizin II, Klinikum rechts der IsarInnere Medizin - Gastroenterologie, Krankenhaus AgathariedKlinik und Poliklinik für Innere Medizin II, Klinikum rechts der IsarInstitute for Diagnostic and Interventional Radiology, Klinikum rechts der Isar, School of Medicine, Technical University of MunichInstitute for Diagnostic and Interventional Radiology, Klinikum rechts der Isar, School of Medicine, Technical University of MunichKlinik und Poliklinik für Innere Medizin II, Klinikum rechts der IsarInstitute for Diagnostic and Interventional Radiology, Klinikum rechts der Isar, School of Medicine, Technical University of MunichKlinik und Poliklinik für Innere Medizin II, Klinikum rechts der IsarAbstract Background A quantitative assessment of pulmonary edema is important because the clinical severity can range from mild impairment to life threatening. A quantitative surrogate measure, although invasive, for pulmonary edema is the extravascular lung water index (EVLWI) extracted from the transpulmonary thermodilution (TPTD). Severity of edema from chest X-rays, to date is based on the subjective classification of radiologists. In this work, we use machine learning to quantitatively predict the severity of pulmonary edema from chest radiography. Methods We retrospectively included 471 X-rays from 431 patients who underwent chest radiography and TPTD measurement within 24 h at our intensive care unit. The EVLWI extracted from the TPTD was used as a quantitative measure for pulmonary edema. We used a deep learning approach and binned the data into two, three, four and five classes increasing the resolution of the EVLWI prediction from the X-rays. Results The accuracy, area under the receiver operating characteristic curve (AUROC) and Mathews correlation coefficient (MCC) in the binary classification models (EVLWI < 15, ≥ 15) were 0.93 (accuracy), 0.98 (AUROC) and 0.86(MCC). In the three multiclass models, the accuracy ranged between 0.90 and 0.95, the AUROC between 0.97 and 0.99 and the MCC between 0.86 and 0.92. Conclusion Deep learning can quantify pulmonary edema as measured by EVLWI with high accuracy.https://doi.org/10.1186/s13054-023-04426-5Pulmonary edemaTranspulmonary thermodilutionTPTDExtravascular lung waterEVLWIChest X-ray |
| spellingShingle | Dominik Schulz Sebastian Rasch Markus Heilmaier Rami Abbassi Alexander Poszler Jörg Ulrich Manuel Steinhardt Georgios A. Kaissis Roland M. Schmid Rickmer Braren Tobias Lahmer A deep learning model enables accurate prediction and quantification of pulmonary edema from chest X-rays Critical Care Pulmonary edema Transpulmonary thermodilution TPTD Extravascular lung water EVLWI Chest X-ray |
| title | A deep learning model enables accurate prediction and quantification of pulmonary edema from chest X-rays |
| title_full | A deep learning model enables accurate prediction and quantification of pulmonary edema from chest X-rays |
| title_fullStr | A deep learning model enables accurate prediction and quantification of pulmonary edema from chest X-rays |
| title_full_unstemmed | A deep learning model enables accurate prediction and quantification of pulmonary edema from chest X-rays |
| title_short | A deep learning model enables accurate prediction and quantification of pulmonary edema from chest X-rays |
| title_sort | deep learning model enables accurate prediction and quantification of pulmonary edema from chest x rays |
| topic | Pulmonary edema Transpulmonary thermodilution TPTD Extravascular lung water EVLWI Chest X-ray |
| url | https://doi.org/10.1186/s13054-023-04426-5 |
| work_keys_str_mv | AT dominikschulz adeeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT sebastianrasch adeeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT markusheilmaier adeeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT ramiabbassi adeeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT alexanderposzler adeeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT jorgulrich adeeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT manuelsteinhardt adeeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT georgiosakaissis adeeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT rolandmschmid adeeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT rickmerbraren adeeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT tobiaslahmer adeeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT dominikschulz deeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT sebastianrasch deeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT markusheilmaier deeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT ramiabbassi deeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT alexanderposzler deeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT jorgulrich deeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT manuelsteinhardt deeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT georgiosakaissis deeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT rolandmschmid deeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT rickmerbraren deeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays AT tobiaslahmer deeplearningmodelenablesaccuratepredictionandquantificationofpulmonaryedemafromchestxrays |