1
|
Alam MZ, Simonetti A, Brillantino R, Tayler N, Grainge C, Siribaddana P, Nouraei SAR, Batchelor J, Rahman MS, Mancuzo EV, Holloway JW, Holloway JA, Rezwan FI. Predicting Pulmonary Function From the Analysis of Voice: A Machine Learning Approach. Front Digit Health 2022; 4:750226. [PMID: 35211691 PMCID: PMC8861188 DOI: 10.3389/fdgth.2022.750226] [Citation(s) in RCA: 6] [Impact Index Per Article: 3.0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/30/2021] [Accepted: 01/14/2022] [Indexed: 11/21/2022] Open
Abstract
Introduction To self-monitor asthma symptoms, existing methods (e.g. peak flow metre, smart spirometer) require special equipment and are not always used by the patients. Voice recording has the potential to generate surrogate measures of lung function and this study aims to apply machine learning approaches to predict lung function and severity of abnormal lung function from recorded voice for asthma patients. Methods A threshold-based mechanism was designed to separate speech and breathing from 323 recordings. Features extracted from these were combined with biological factors to predict lung function. Three predictive models were developed using Random Forest (RF), Support Vector Machine (SVM), and linear regression algorithms: (a) regression models to predict lung function, (b) multi-class classification models to predict severity of lung function abnormality, and (c) binary classification models to predict lung function abnormality. Training and test samples were separated (70%:30%, using balanced portioning), features were normalised, 10-fold cross-validation was used and model performances were evaluated on the test samples. Results The RF-based regression model performed better with the lowest root mean square error of 10·86. To predict severity of lung function impairment, the SVM-based model performed best in multi-class classification (accuracy = 73.20%), whereas the RF-based model performed best in binary classification models for predicting abnormal lung function (accuracy = 85%). Conclusion Our machine learning approaches can predict lung function, from recorded voice files, better than published approaches. This technique could be used to develop future telehealth solutions including smartphone-based applications which have potential to aid decision making and self-monitoring in asthma.
Collapse
Affiliation(s)
- Md. Zahangir Alam
- Human Development and Health, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
- Department of Computer Science and Engineering, Bangladesh University of Engineering and Technology, Dhaka, Bangladesh
| | - Albino Simonetti
- Human Development and Health, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
- Department of Information and Electrical Engineering and Applied Mathematics/DIEM, University of Salerno, Fisciano, Italy
| | - Raffaele Brillantino
- Human Development and Health, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
- Department of Information and Electrical Engineering and Applied Mathematics/DIEM, University of Salerno, Fisciano, Italy
| | - Nick Tayler
- Peter Doherty Institute, The University of Melbourne, Melbourne, VIC, Australia
| | - Chris Grainge
- Hunter Medical Research Institute, The University of Newcastle, Newcastle, NSW, Australia
- Department of Respiratory Medicine, John Hunter Hospital, Newcastle, NSW, Australia
| | - Pandula Siribaddana
- Postgraduate Institute of Medicine, University of Colombo, Colombo, Sri Lanka
| | - S. A. Reza Nouraei
- Clinical Informatics Research Unit, University of Southampton, Southampton, United Kingdom
- Robert White Centre for Airway Voice and Swallowing, Poole Hospital, Poole, United Kingdom
| | - James Batchelor
- Clinical Informatics Research Unit, University of Southampton, Southampton, United Kingdom
| | - M. Sohel Rahman
- Department of Computer Science and Engineering, Bangladesh University of Engineering and Technology, Dhaka, Bangladesh
| | - Eliane V. Mancuzo
- Medical School, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil
| | - John W. Holloway
- Human Development and Health, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
- National Institute for Health Research Southampton Biomedical Research Centre, University Hospital Southampton, Southampton, United Kingdom
| | - Judith A. Holloway
- Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
- MSc Allergy, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
| | - Faisal I. Rezwan
- Human Development and Health, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
- Department of Computer Science, Aberystwyth University, Aberystwyth, United Kingdom
- *Correspondence: Faisal I. Rezwan
| |
Collapse
|
2
|
Ontiveros N, Eapen-John D, Osorio N, Song J, Li L, Sheshadri A, Tiang X, Ghosh N, Vaporciyan A, Correa A, Walsh G, Grosu HB, Ost DE. Predicting Lung Function Following Lobectomy: A New Method to Adjust for Inherent Selection Bias. Respiration 2018; 96:434-445. [PMID: 30257257 DOI: 10.1159/000490258] [Citation(s) in RCA: 3] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/13/2018] [Accepted: 05/21/2018] [Indexed: 12/25/2022] Open
Abstract
BACKGROUND Predictions that overestimate post-lobectomy lung function are more likely than underestimates to lead to lobectomy. Studies of post-lobectomy lung function have included only surgical patients, so overestimates are overrepresented. This selection bias has led to incorrect estimates of prediction bias, which has led to inaccurate threshold values for determining lobectomy eligibility. OBJECTIVE The objective of this study was to demonstrate and adjust for this selection bias in order to arrive at correct estimates of prediction bias, the 95% limits of agreement, and adjusted threshold values for determining when exercise testing is warranted. METHODS We conducted a retrospective study of patients evaluated for lobectomy. We used multiple imputations to determine postoperative results for patients who did not have surgery because their predicted postoperative values were low. We combined these results with surgical patients to adjust for selection bias. We used the Bland-Altman method and the bivariate normal distribution to determine threshold values for surgical eligibility. RESULTS Lobectomy evaluation was performed in 114 patients; 79 had lobectomy while 35 were ineligible based on predicted values. Prediction bias using the Bland-Altman method changed significantly after controlling for selection bias. To achieve a postoperative FEV1 > 30% and DLCO ≥30%, a predicted FEV1 > 46% and DLCO ≥53% were required. Compared to current guidelines, using these thresholds would change management in 17% of cases. CONCLUSION The impact of selection bias on estimates of prediction accuracy was significant but can be corrected. Threshold values for determining surgical eligibility should be reassessed.
Collapse
Affiliation(s)
- Narda Ontiveros
- Tecnologico de Monterrey, Escuela de Medicina y Ciencias de la Salud, Monterrey, Mexico
| | - David Eapen-John
- Department of Pulmonary Medicine, MD Anderson Cancer Center, Houston, Texas, USA
| | - Natasha Osorio
- Tecnologico de Monterrey, Escuela de Medicina y Ciencias de la Salud, Monterrey, Mexico
| | - Juhee Song
- Department of Biostatistics, MD Anderson Cancer Center, Houston, Texas, USA
| | - Liang Li
- Department of Biostatistics, MD Anderson Cancer Center, Houston, Texas, USA
| | - Ajay Sheshadri
- Department of Pulmonary Medicine, MD Anderson Cancer Center, Houston, Texas, USA
| | - Xin Tiang
- Department of Pulmonary Medicine, MD Anderson Cancer Center, Houston, Texas, USA
| | - Natasha Ghosh
- Department of Pulmonary Medicine, MD Anderson Cancer Center, Houston, Texas, USA
| | - Ara Vaporciyan
- Department of Thoracic Surgery, MD Anderson Cancer Center, Houston, Texas, USA
| | - Arlene Correa
- Department of Thoracic Surgery, MD Anderson Cancer Center, Houston, Texas, USA
| | - Garrett Walsh
- Department of Thoracic Surgery, MD Anderson Cancer Center, Houston, Texas, USA
| | - Horiana B Grosu
- Department of Biostatistics, MD Anderson Cancer Center, Houston, Texas, USA
| | - David E Ost
- Department of Pulmonary Medicine, MD Anderson Cancer Center, Houston, Texas, USA
| |
Collapse
|
3
|
Fernández-Rodríguez L, Torres I, Romera D, Galera R, Casitas R, Martínez-Cerón E, Díaz-Agero P, Utrilla C, García-Río F. Prediction of postoperative lung function after major lung resection for lung cancer using volumetric computed tomography. J Thorac Cardiovasc Surg 2018; 156:2297-2308.e5. [PMID: 30195604 DOI: 10.1016/j.jtcvs.2018.07.040] [Citation(s) in RCA: 12] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 08/05/2017] [Revised: 06/15/2018] [Accepted: 07/02/2018] [Indexed: 12/25/2022]
Abstract
OBJECTIVES The study objectives were to assess the accuracy of volumetric computed tomography to predict postoperative lung function in patients with lung cancer in relation to anatomic segments counting and perfusion scintigraphy, to generate specific predictive equations for each functional parameter, and to evaluate accuracy and precision of these in a validation cohort. METHODS We assessed pulmonary functions preoperatively and 3 to 4 months postoperatively after lung resection for lung cancer (n = 114). Absolute and relative lung volumes (total and upper/middle/lower) were determined using volumetric software analysis for staging thoracic computed tomography scans. Predicted postoperative function was calculated by segments counting, scintigraphy, and volumetric computed tomography. RESULTS Volumetric computed tomography achieves a higher correlation and precision with measured postoperative lung function than segments counting or scintigraphy (correlation and intraclass correlation coefficients, 0.779-0.969 and 0.776-0.969; 0.573-0.887 and 0.552-0.882; and 0.578-0.834 and 0.532-0.815, respectively), as well as greater accuracy, determined by narrower agreement coefficients for forced vital capacity, forced expiratory volume in 1 second, lung diffusing capacity, and peak oxygen uptake. After validation in an independent cohort (n = 43), adjusted linear regression including volumetric estimation of decreased postoperative ventilation for postoperative lung function parameters explains 98% to 99% of variance. CONCLUSIONS Volumetric computed tomography is a reliable and accurate method to predict postoperative lung function in patients undergoing lung resection that provides better accuracy than conventional procedures. Because lung computed tomography is systematically performed in the staging of patients with suspected lung cancer, this volumetric analysis might simultaneously provide the information necessary to evaluate operability.
Collapse
Affiliation(s)
| | - Isabel Torres
- Servicio de Radiodiagnóstico, Hospital Universitario La Paz, Madrid, Spain
| | - Delia Romera
- Servicio de Neumología, Hospital Universitario La Paz, IdiPAZ, Madrid, Spain
| | - Raúl Galera
- Servicio de Neumología, Hospital Universitario La Paz, IdiPAZ, Madrid, Spain; CIBER de Enfermedades Respiratorias, Madrid, Spain
| | - Raquel Casitas
- Servicio de Neumología, Hospital Universitario La Paz, IdiPAZ, Madrid, Spain; CIBER de Enfermedades Respiratorias, Madrid, Spain
| | - Elisabet Martínez-Cerón
- Servicio de Neumología, Hospital Universitario La Paz, IdiPAZ, Madrid, Spain; CIBER de Enfermedades Respiratorias, Madrid, Spain
| | - Prudencio Díaz-Agero
- Servicio de Cirugía Torácica, Hospital Universitario La Paz, IdiPAZ, Madrid, Spain
| | - Cristina Utrilla
- Servicio de Radiodiagnóstico, Hospital Universitario La Paz, Madrid, Spain
| | - Francisco García-Río
- Servicio de Neumología, Hospital Universitario La Paz, IdiPAZ, Madrid, Spain; CIBER de Enfermedades Respiratorias, Madrid, Spain; Facultad de Medicina, Universidad Autónoma de Madrid, Madrid, Spain.
| |
Collapse
|