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

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.

Accelerating cough-based algorithms for pulmonary tuberculosis screening: Results from the CODA TB DREAM Challenge

“…The cough sounds collected were restricted to 0.5 second recordings around the peak; the use of whole cough sounds may further improve performance. 31 As all participants were symptomatic, there are limitations in extending these models for community-wide screening, and additional data collection from screening cohorts is needed. The participants also all had cough; while solicited cough sounds may have value for those without cough, this needs to be further evaluated.…”

Screening COVID-19 by Swaasa AI platform using cough sounds: a cross-sectional study

“…), chroma features, contrast features, tonnentz features, zero-crossing rate (ZCR), energy, skewness, and kurtosis 29 . Correlation-based feature selection was used to reduce the feature size from 209 to 170 26 . When the model is unsure if COVID-19 will be detected as yes/no, it provides an inconclusive output as shown in the block diagram Fig.…”

Solicited Cough Sound Analysis for Tuberculosis Triage Testing: The CODA TB DREAM Challenge Dataset

“…9 In a study combining cough sound analysis and patients’ clinical characteristics, Yellapu et al report that ML can be used to detect TB with 90% sensitivity and 85% specificity. 10 Those pilot studies report on ML models which were designed on small datasets and were not validated in external populations. Given the potential impact on performance of local disease epidemiology and population ethnicity among other confounders, large and diverse cough datasets are needed to replicate those studies.…”

“…9 In a study combining cough sound analysis and patients' clinical characteristics, Yellapu et al report that ML can be used to detect TB with 90% sensitivity and 85% specificity. 10 Those pilot studies report on ML models which were designed on small datasets and were not validated in external populations.…”