Development and clinical validation of Swaasa AI platform for screening and prioritization of pulmonary TB

Gayatri Devi Yellapu, Gowrisree Rudraraju, Narayana Rao Sripada, Baswaraj Mamidgi, Charan Jalukuru, Priyanka Firmal, Venkat Yechuri, Sowmya Varanasi, Venkata Sudhakar Peddireddi, Devi Madhavi Bhimarasetty, Sidharth Kanisetti, Niranjan Joshi, Prasant Mohapatra, Kiran Pamarthi, Gayatri Devi Yellapu, Gowrisree Rudraraju, Narayana Rao Sripada, Baswaraj Mamidgi, Charan Jalukuru, Priyanka Firmal, Venkat Yechuri, Sowmya Varanasi, Venkata Sudhakar Peddireddi, Devi Madhavi Bhimarasetty, Sidharth Kanisetti, Niranjan Joshi, Prasant Mohapatra, Kiran Pamarthi

Abstract

Acoustic signal analysis has been employed in various medical devices. However, studies involving cough sound analysis to screen the potential pulmonary tuberculosis (PTB) suspects are very few. The main objective of this cross-sectional validation study was to develop and validate the Swaasa AI platform to screen and prioritize at risk patients for PTB based on the signature cough sound as well as symptomatic information provided by the subjects. The voluntary cough sound data was collected at Andhra Medical College-India. An Algorithm based on multimodal convolutional neural network architecture and feedforward artificial neural network (tabular features) was built and validated on a total of 567 subjects, comprising 278 positive and 289 negative PTB cases. The output from these two models was combined to detect the likely presence (positive cases) of PTB. In the clinical validation phase, the AI-model was found to be 86.82% accurate in detecting the likely presence of PTB with 90.36% sensitivity and 84.67% specificity. The pilot testing of model was conducted at a peripheral health care centre, RHC Simhachalam-India on 65 presumptive PTB cases. Out of which, 15 subjects truly turned out to be PTB positive with a positive predictive value of 75%. The validation results obtained from the model are quite encouraging. This platform has the potential to fulfil the unmet need of a cost-effective PTB screening method. It works remotely, presents instantaneous results, and does not require a highly trained operator. Therefore, it could be implemented in various inaccessible, resource-poor parts of the world.

Conflict of interest statement

The authors declare no competing interests.

© 2023. The Author(s).

Figures

Figure 1
Figure 1
Data distribution in the derivation phase, validation phase and pilot testing.
Figure 2
Figure 2
Illustration of the combined logic—combining feedforward artificial neural network (FFANN) model and convolutional neural network (CNN) outputs.
Figure 3
Figure 3
Block Diagram illustrating the flow of the TB prediction model.
Figure 4
Figure 4
The representative graph for ROC curve, best among tenfold validation of TB prediction model built using derivation data.
Figure 5
Figure 5
The provided graph shows the best ROC curve of a TB prediction model constructed using validation data.
Figure 6
Figure 6
The provided ROC curve illustrates the performance of a TB prediction model constructed using pilot data.

References

    1. Al Lawati R, et al. COVID-19 and pulmonary Mycobacterium tuberculosis coinfection. Oman Med. J. 2021;36:e298. doi: 10.5001/omj.2022.23.
    1. Pai M, et al. Tuberculosis. Nat. Rev. Dis. Prim. 2016;2:1–23.
    1. WHO. Global tuberculosis report 2021. Geneva: World Health Organization; 2021. Licence: CC BY-NC-SA 3.0 IGO (2021).
    1. WHO . Systematic Screening for Active Tuberculosis. World Health Organization; 2013.
    1. Migliori GB, et al. Reducing tuberculosis transmission: A consensus document from the World Health Organization Regional Office for Europe. Eur. Respir. J. 2019;53:1–18. doi: 10.1183/13993003.00391-2019.
    1. Gill CM, Dolan L, Piggott LM, McLaughlin AM. New developments in tuberculosis diagnosis and treatment. Breathe. 2022;18:1–15. doi: 10.1183/20734735.0149-2021.
    1. Chung KF, Pavord ID. Prevalence, pathogenesis, and causes of chronic cough. Lancet. 2008;371:1364–1374. doi: 10.1016/S0140-6736(08)60595-4.
    1. Simonsson BG, Jacobs FM, Nadel JA. Role of autonomic nervous system and the cough reflex in the increased responsiveness of airways in patients with obstructive airway disease. J. Clin. Investig. 1967;46:1812–1818. doi: 10.1172/JCI105671.
    1. Turner RD, Bothamley GH. Cough and the transmission of tuberculosis. J. Infect. Dis. 2015;211:1367–1372. doi: 10.1093/infdis/jiu625.
    1. Swarnkar V, et al. Automatic identification of wet and dry cough in pediatric patients with respiratory diseases. Ann. Biomed. Eng. 2013;41:1016–1028. doi: 10.1007/s10439-013-0741-6.
    1. Kaplan AG. Chronic cough in adults: Make the diagnosis and make a difference. Pulm. Ther. 2019;5:11–21. doi: 10.1007/s41030-019-0089-7.
    1. Proaño A, et al. Protocol for studying cough frequency in people with pulmonary tuberculosis. BMJ Open. 2016;6:1–9. doi: 10.1136/bmjopen-2015-010365.
    1. Botha GHR, et al. Detection of tuberculosis by automatic cough sound analysis. Physiol. Meas. 2018;39:045005. doi: 10.1088/1361-6579/aab6d0.
    1. Pahar M, et al. Automatic cough classification for tuberculosis screening in a real-world environment. Physiol. Meas. 2021;42:105014. doi: 10.1088/1361-6579/ac2fb8.
    1. Armstrong S. The computer will assess you now. BMJ. 2016;355:1–2.
    1. Lancet T. Artificial intelligence in health care: Within touching distance. Lancet. 2017;390:2739. doi: 10.1016/S0140-6736(17)31540-4.
    1. Topalovic M, et al. Artificial intelligence outperforms pulmonologists in the interpretation of pulmonary function tests. Eur. Respir. J. 2019;53:1–11.
    1. Ijaz A, et al. Towards using cough for respiratory disease diagnosis by leveraging artificial intelligence: A survey. Inform. Med. Unlocked. 2022;29:1–28. doi: 10.1016/j.imu.2021.100832.
    1. Kik SV, Denkinger CM, Casenghi M, Vadnais C, Pai M. Tuberculosis diagnostics: Which target product profiles should be prioritised? Eur. Respir. J. 2014;44:537–540. doi: 10.1183/09031936.00027714.
    1. Larson S, et al. Validation of an automated cough detection algorithm for tracking recovery of pulmonary tuberculosis patients. PLoS ONE. 2012;7:1–10. doi: 10.1371/journal.pone.0046229.
    1. Pourhoseingholi MA, Vahedi M, Rahimzadeh M. Sample size calculation in medical studies. Gastroenterol. Hepatol. Bed Bench. 2013;6:14–17.
    1. National TB elimination programme Central TB Division. National TB Prevalence Survey in India 2019–2021. Ministry of Health and Family Welfare (2021).
    1. Jones PW, Quirk FH, Baveystock CM. The St George’s respiratory questionnaire. Respir. Med. 1991;85:25–31. doi: 10.1016/S0954-6111(06)80166-6.
    1. Barry SJ, Dane AD, Morice AH, Walmsley AD. The automatic recognition and counting of cough. Cough. 2006;2:8. doi: 10.1186/1745-9974-2-8.
    1. Rudraraju G, et al. Cough sound analysis and objective correlation with spirometry and clinical diagnosis. Inform. Med. Unlocked. 2020;19:1–11. doi: 10.1016/j.imu.2020.100319.
    1. Ribeiro, M., Singh, S. & Guestrin, C. “Why should I trust you?”: Explaining the predictions of any classifier. In Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Demonstrations 1135–1144 (Association for Computational Linguistics, 2016). 10.18653/v1/N16-3020.
    1. Clopper CJ, Pearson ES. The Use of confidence or fiducial limits illustrated in the case of the binomial. Biometrika. 1934;26:404–413. doi: 10.1093/biomet/26.4.404.
    1. Belkacem AN, Ouhbi S, Lakas A, Benkhelifa E, Chen C. End-to-end AI-based point-of-care diagnosis system for classifying respiratory illnesses and early detection of COVID-19: A theoretical framework. Front. Med. 2021;8:1–13. doi: 10.3389/fmed.2021.585578.
    1. Abeyratne UR, Swarnkar V, Setyati A, Triasih R. Cough sound analysis can rapidly diagnose childhood pneumonia. Ann. Biomed. Eng. 2013;41:2448–2462. doi: 10.1007/s10439-013-0836-0.
    1. Xu X, et al. Listen2Cough: Leveraging end-to-end deep learning cough detection model to enhance lung health assessment using passively sensed audio. Proc. ACM Interact. Mob. Wear. Ubiquitous Technol. 2021;5:1–22.
    1. Klco P, Kollarik M, Tatar M. Novel computer algorithm for cough monitoring based on octonions. Respir. Physiol. Neurobiol. 2018;257:36–41. doi: 10.1016/j.resp.2018.03.010.
    1. Laguarta J, Hueto F, Subirana B. COVID-19 artificial intelligence diagnosis using only cough recordings. IEEE Open J. Eng. Med. Biol. 2020;1:275–281. doi: 10.1109/OJEMB.2020.3026928.
    1. Pahar M, Klopper M, Warren R, Niesler T. COVID-19 cough classification using machine learning and global smartphone recordings. Comput. Biol. Med. 2021;135:1–10. doi: 10.1016/j.compbiomed.2021.104572.

Source: PubMed

Sponsoren und Mitarbeiter

Krankheiten

Drogeninterventionen

3
Abonnieren