Abstract
Background Diagnosis of asthma in primary care is challenged by a multistep pathway with variable adherence leading to significant misdiagnosis, late diagnosis and poor patient outcomes. This is driven by a lack of rapid, accurate, easy-to-use diagnostic tests that can reliably diagnose asthma with both high sensitivity and specificity. The objective of this work was to develop and evaluate a multivariate machine learning classifier for asthma diagnosis. The classifier was built using interpretable data processing and machine learning techniques applied to 75-second tidal breathing CO2 recordings captured on TidalSense’s N-Tidal® hand-held capnometer. The target population comprises patients with clinically suspected asthma who have no contraindications to performing capnometry.
Methods Capnograms were collected from 138 asthmatic and 132 non-asthmatic participants (including healthy volunteers, and those with chronic obstructive pulmonary disease (COPD), heart failure, and other cardiores-piratory conditions) recruited from both primary and secondary care. Each high-resolution CO2 recording was transformed into 82 features (using the N-Tidal® Diagnose 1 v1.0 software) that characterise the constituent breathing cycles. A logistic regression model was trained on these features and performance metrics generated from an unseen test set of 64 participants. Model performance was evaluated using discrimination, measured by the area under the receiver operating characteristic curve (AUROC), as well as clinically relevant predictive accuracy metrics, including positive predictive value (PPV) and negative predictive value (NPV). This was repeated 20 times with different training and testing participants for additional statistical robustness; the average and variability of these metrics were recorded.
Results The classification model achieved an AUROC of 0.91 ± 0.03%, sensitivity of 83 ± 4%, specificity of 85 ± 6%, positive predictive value (PPV) of 87 ± 4%, and negative predictive value (NPV) of 81 ± 4% in detecting asthma from a single breath recording. The model demonstrated diagnostic stability, with 95.8% of each participant’s recordings over the course of data collection being classified correctly on average. No model bias was observed with regards to sex, but performance did improve with age, possibly reflecting increasing severity of disease with age.
Conclusion This study introduces a highly accurate and interpretable multivariate diagnostic model capable of classifying asthma from a single breath recorded using the N-Tidal® Handset. It achieves high sensitivity and specificity compared with current methods, such as spirometry, and could enable point-of-care diagnosis in patients suspected of having asthma.
Competing Interest Statement
HB, LT, RHL, JM, GL and AXP are currently employed, or were employed at the time of the research, by TidalSense Limited. GH and HFA are funded by the National Institute for Health Research (NIHR) Community Healthcare MedTech and In Vitro Diagnostics Co-operative at Oxford Health NHS Foundation Trust. The views expressed in this publication are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care.
Funding Statement
The studies which provided the data for this report were funded by NIHR (i4i grant), Innovate UK, and Pfizer OpenAir. The authors had sole responsibility for the study design, data collection, data analysis, data interpretation and report writing.
Author Declarations
I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.
Yes
The details of the IRB/oversight body that provided approval or exemption for the research described are given below:
Ethical approval was obtained from the South Central - Berkshire Research Ethics Committee for GBRS and ABRS, the Yorkshire and the Humber Research Ethics Committee for CBRS and the West Midlands Solihull Research Ethics Committee for CBRS2. All participants provided written informed consent to participate.
I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.
Yes
I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).
Yes
I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.
Yes
Data Availability
The datasets generated during and/or analysed, and the algorithms developed, during the current study are not publicly available for data protection, confidentiality, and commercial sensitivity reasons.
