Predicting total lung capacity from spirometry: a machine learning approach
Background and objectiveSpirometry patterns can suggest that a patient has a restrictive ventilatory impairment; however, lung volume measurements such as total lung capacity (TLC) are required to confirm the diagnosis. The aim of the study was to train a supervised machine learning model that can a...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2023-05-01
|
| Series: | Frontiers in Medicine |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fmed.2023.1174631/full |
| _version_ | 1797823599426404352 |
|---|---|
| author | Luka Beverin Marko Topalovic Armin Halilovic Paul Desbordes Wim Janssens Maarten De Vos Maarten De Vos |
| author_facet | Luka Beverin Marko Topalovic Armin Halilovic Paul Desbordes Wim Janssens Maarten De Vos Maarten De Vos |
| author_sort | Luka Beverin |
| collection | DOAJ |
| description | Background and objectiveSpirometry patterns can suggest that a patient has a restrictive ventilatory impairment; however, lung volume measurements such as total lung capacity (TLC) are required to confirm the diagnosis. The aim of the study was to train a supervised machine learning model that can accurately estimate TLC values from spirometry and subsequently identify which patients would most benefit from undergoing a complete pulmonary function test.MethodsWe trained three tree-based machine learning models on 51,761 spirometry data points with corresponding TLC measurements. We then compared model performance using an independent test set consisting of 1,402 patients. The best-performing model was used to retrospectively identify restrictive ventilatory impairment in the same test set. The algorithm was compared against different spirometry patterns commonly used to predict restriction.ResultsThe prevalence of restrictive ventilatory impairment in the test set is 16.7% (234/1402). CatBoost was the best-performing machine learning model. It predicted TLC with a mean squared error (MSE) of 560.1 mL. The sensitivity, specificity, and F1-score of the optimal algorithm for predicting restrictive ventilatory impairment was 83, 92, and 75%, respectively.ConclusionA machine learning model trained on spirometry data can estimate TLC to a high degree of accuracy. This approach could be used to develop future smart home-based spirometry solutions, which could aid decision making and self-monitoring in patients with restrictive lung diseases. |
| first_indexed | 2024-03-13T10:25:17Z |
| format | Article |
| id | doaj.art-e2b382f81bbb462cbe11aadb48004dd6 |
| institution | Directory Open Access Journal |
| issn | 2296-858X |
| language | English |
| last_indexed | 2024-03-13T10:25:17Z |
| publishDate | 2023-05-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Medicine |
| spelling | doaj.art-e2b382f81bbb462cbe11aadb48004dd62023-05-19T14:03:07ZengFrontiers Media S.A.Frontiers in Medicine2296-858X2023-05-011010.3389/fmed.2023.11746311174631Predicting total lung capacity from spirometry: a machine learning approachLuka Beverin0Marko Topalovic1Armin Halilovic2Paul Desbordes3Wim Janssens4Maarten De Vos5Maarten De Vos6Statistics Research Centre, KU Leuven, Leuven, BelgiumArtiQ NV, Leuven, BelgiumArtiQ NV, Leuven, BelgiumArtiQ NV, Leuven, BelgiumLaboratory of Respiratory Diseases and Thoracic Surgery, Department of Chronic Diseases Metabolism and Ageing, Ku Leuven, Leuven, BelgiumStadius, Department of Electrical Engineering, KU Leuven, Leuven, BelgiumArtiQ NV, Leuven, BelgiumBackground and objectiveSpirometry patterns can suggest that a patient has a restrictive ventilatory impairment; however, lung volume measurements such as total lung capacity (TLC) are required to confirm the diagnosis. The aim of the study was to train a supervised machine learning model that can accurately estimate TLC values from spirometry and subsequently identify which patients would most benefit from undergoing a complete pulmonary function test.MethodsWe trained three tree-based machine learning models on 51,761 spirometry data points with corresponding TLC measurements. We then compared model performance using an independent test set consisting of 1,402 patients. The best-performing model was used to retrospectively identify restrictive ventilatory impairment in the same test set. The algorithm was compared against different spirometry patterns commonly used to predict restriction.ResultsThe prevalence of restrictive ventilatory impairment in the test set is 16.7% (234/1402). CatBoost was the best-performing machine learning model. It predicted TLC with a mean squared error (MSE) of 560.1 mL. The sensitivity, specificity, and F1-score of the optimal algorithm for predicting restrictive ventilatory impairment was 83, 92, and 75%, respectively.ConclusionA machine learning model trained on spirometry data can estimate TLC to a high degree of accuracy. This approach could be used to develop future smart home-based spirometry solutions, which could aid decision making and self-monitoring in patients with restrictive lung diseases.https://www.frontiersin.org/articles/10.3389/fmed.2023.1174631/fullrestrictionspirometrymachine learninginterstitial lung diseasetotal lung capacity |
| spellingShingle | Luka Beverin Marko Topalovic Armin Halilovic Paul Desbordes Wim Janssens Maarten De Vos Maarten De Vos Predicting total lung capacity from spirometry: a machine learning approach Frontiers in Medicine restriction spirometry machine learning interstitial lung disease total lung capacity |
| title | Predicting total lung capacity from spirometry: a machine learning approach |
| title_full | Predicting total lung capacity from spirometry: a machine learning approach |
| title_fullStr | Predicting total lung capacity from spirometry: a machine learning approach |
| title_full_unstemmed | Predicting total lung capacity from spirometry: a machine learning approach |
| title_short | Predicting total lung capacity from spirometry: a machine learning approach |
| title_sort | predicting total lung capacity from spirometry a machine learning approach |
| topic | restriction spirometry machine learning interstitial lung disease total lung capacity |
| url | https://www.frontiersin.org/articles/10.3389/fmed.2023.1174631/full |
| work_keys_str_mv | AT lukabeverin predictingtotallungcapacityfromspirometryamachinelearningapproach AT markotopalovic predictingtotallungcapacityfromspirometryamachinelearningapproach AT arminhalilovic predictingtotallungcapacityfromspirometryamachinelearningapproach AT pauldesbordes predictingtotallungcapacityfromspirometryamachinelearningapproach AT wimjanssens predictingtotallungcapacityfromspirometryamachinelearningapproach AT maartendevos predictingtotallungcapacityfromspirometryamachinelearningapproach AT maartendevos predictingtotallungcapacityfromspirometryamachinelearningapproach |