What if you could just cough onto your smartphone, and it could identify if you have any lung diseases? Salcit Technologies, a Hyderabad-based AI startup, has just done that.

The startup, along with the Centre for Cellular and Molecular Platforms (C-CAMP) and UK’s Foreign, Commonwealth and Development Office (FCDO), has developed an AI platform where you can download the app “Swaasa”. The person will have to just open the app/platform, which is powered by AI and can cough into a mobile phone to assess if the user is suffering from any lung disease. A doctor can then go through the report made by the platform and determine the next course of action.

“Swaasa” is AI technology that assesses respiratory health through audiometric analysis of cough sounds with critical data from the user’s temperature, oxygen saturation, medical history, past activity, and other symptoms. This AI is an ensemble of advanced ML models to provide instant insights about respiratory health, indicative patterns, and severity of the condition.

The platform has been jointly developed by a three-member team that includes Venkat Yechuri, CEO, Narayana Rao Sripada, founder and CTO and Manmohan Jain, CIO &CSO of Salcit Technologies.

Venkat Yechuri, Narayana Rao Sripada and Manmohan Jain from Salcit Technologies.

Venkat has two decades of experience internationally strategising businesses, building products, and creating an inspiring environment for a team. Before Salcit Technologies, Venkat held various leadership roles across various industries and was an uplifting coach and mentor.

Narayana has extensive experience in the technological space, working with reputed multinational companies and playing diverse roles. He was a co-inventor for many patents and began his entrepreneurial journey with iStudymate, a mobile learning platform, which a Malaysian partner later acquired.

Manmohan has two decades of experience in technical and managerial roles, including software design, product development, technical management, and operations. He brings out his technological expertise and innovation to improve our products continuously.

Analytics India Magazine caught up with the Salcit technology team, asking them about the Swaasa and the AI platform, which can be a boon in this pandemic time.

AIM – When was the product (AI Platform) developed, and how does it work?

Salcit – Swaasa is an AI Platform not an app. It operates on an “intel inside paradigm”. We have APIs that are used to integrate Swaasa into existing products and medical apps/clinical workflows where seamless integration can be programmed within a few hours. Our product, Swaasa, is Google Photos for cough sounds. We can identify underlying respiratory lung conditions by analysing a 10 second (solicited) cough sound recording. The product is the brainchild of Salcit Technologies’ founder, Narayana Rao. The journey started in September of 2017.

AIM – What are the tech stacks used in the product (AI Platform)? Please explain in detail?

Salcit – Swaasa components include the core AI engine built-in Python, a Xamarin based cross-platform mobile app, a responsive and mobile web-friendly web app developed in Angular 10, and a React-based Web Dashboard. The cloud-based Swaasa platform is built using AWS services and hosted on AWS Cloud.

More specific details of the tech stack are as below:

Swaasa AI Engine/Platform

  • Core Swaasa algorithm is written in Python language.
  • Under the hood, it uses ML frameworks like Scikit, Pytorch, Fast.AI. 
  • FastAPI is used to create the REST APIs.

Mobile App

Xamarin – multilingual [EN, TE, HI, FR, BN] (open-source platform for building modern and performant applications for iOS and Android). We have used the Syncfusion UI toolkit to develop the UI of the application.

Web App

  • It is developed on Angular 10.
  • The RecordRTC library is used for audio recording functionality. 

Web Dashboard

It is developed on React v16.

Infrastructure

  • The platform is hosted on AWS Cloud.
  • AWS DynamoDB and AWS S3 are used for storage.
  • Application is deployed using AWS ECS services.

AIM – Have you already used the product (AI Platform) on Covid-19 patients or other patients with lung diseases? What has been the outcome?

Salcit – We have been using Swaasa’s general lung health assessment capabilities since June 2020, when we obtained a NOC from the CDSCO. The core ability of Swaasa is to operate remotely and asynchronously [patient and healthcare need not be in the same place and at the same time] and is revolutionary in its impact. For example, at the pandemic’s peak, tens of thousands of people used this capability through Apollo 24×7 to assess their risk for COVID.

AIM – Will this product (AI Platform) be a game-changer in identifying lung infections?

Salcit – Apart from the general lung assessment capabilities (the functional equivalent of a lung function test/spirometry), Salcit has been adding disease identification capability to Swaasa. At this point, COVID-19 and TB identification capabilities are available. By mid-2022, identification of Asthma, Bronchitis, COPD, ILD, and Pneumonia will be available.

This ability to screen for diseases inexpensively and without the need for expensive equipment brings the screening ability to the country’s remotest regions, putting the ability in the hands of frontline community (ASHA, ANW) workers. The benefit to patients and society at large is early identification and intervention of tremendously underdiagnosed respiratory ailments, which impose an enormous burden on society.

AIM – Is the product (AI Platform) already in use in Hyderabad? What is the feedback from all the stakeholders?

Salcit – Dr Lakshmi Shankar, Medical Practitioner, said that she had used the AI platform on her patients from her clinic and found it useful in the early detection of restrictive and obstructive disorders. It was also used in a few Covid patients to countercheck & found it to be satisfactory.

Dr Gayatri Yellapu, Associate Professor, Government Hospital for Chest and Communicable Disease/Andhra Medical College.

Dr Anand Krishnan, MD, PhD, Professor, Centre for Community Medicine – All India Institute of Medical Sciences, said, “Screening for chronic respiratory diseases has not been done extensively due to lack of expertise and lab facilities which cannot be done at primary care level centres. Any technology/tool that can facilitate Screening and identify whether the problem is attributed to airway or lung parenchyma/pleura will be a useful value proposition for the appropriate next intervention. Swaasa tool developed by Salcit Technologies addresses this need; the timing is right to validate and introduce such technologies in the country.

Dr Sai Praveen Haranath, Pulmonologist – Apollo Hospitals, said, “As a case-study by Salcit on the application of ML and AI on continuous monitoring of respiratory sounds helps get to know the lung condition. For instance, sound records on lung collapse will be useful information (ahead of X-ray), especially when the intensities are very low or sounds are absent. Temporal analysis of changes in sound pattern helps in managing the disease condition. Similarly, exceptional cases are likely to be very meaningful such

as silence and ILD pattern for early detection. 

Video from British Deputy High Commissioner talking about Swaasa

Video recording from Andhra Medical College Doctors

AIM – How do you want to take this product (AI Platform) forward? What is the future?

Salcit – Swaasa envisages initial deployments primarily in the LMICs of Asia, Africa, LatAm, Middle East – places with large populations with poor or no access to quality healthcare. Once we have traction in these markets in 18-24 months, we propose to go to the developed world.

Swaasa is focused on becoming the lung health assessment tool of choice for community health workers, rural/primary health centres, primary care physicians and pulmonologists. Swaasa’s functionality is going to be intermediated by “distributors” such as telehealth platforms (mfine), hospital systems (Apollo), other health and wellness apps (Helfie) or providers of primary health services (CureBay, Digital Healthcare Solutions). Apart from these, Governments are a very large target customer. In other words, Swaasa would be the “intel inside” model – power the lung assessments in the background.

Currently, Swaasa is used in the monitoring mode by two pharmaceutical companies to track the efficacy of their medications – Windlas Biotech & Mateon Therapeutics.

Apart from the clinical use case, Swaasa has three other markets that it seeks to explore daily health screening for employees, students etc. (RTW – return to work; Trace+ technologies); Insurance – both life and health (screen for upfront risk as well as monitor on an ongoing basis; Prudential Myanmar) and wellness (self-care and advocacy by consumers; HappilyHealth).

AIM – How is artificial intelligence evolving in healthcare? Give us an overall perspective.

Salcit – AI/ML-based technologies have the potential to transform healthcare by deriving new insights from a vast amount of data generated on a day to day healthcare needs. Example – high-value applications include earlier disease detection and triaging, monitoring, accurate diagnosis, identifying new patterns in human physiology, and developing personalised diagnostics and therapeutics. One of the benefits of AI/ML is its ability to learn from real-world use and experience and its ability to improve its performance. The ability for AI/ML software to learn from real-world feedback (training) and improve its performance (adaptation) makes these technologies uniquely situated among software as a medical device (SaMD) and a rapidly expanding area of research

and development. 

AIM – Does technology like AI/ML help reach the remotest location and the people living there?

Salcit – Swaasa is a perfect example of how AI/ML will help bring superior healthcare to remote and underserved populations. The legendary Silicon Valley VC, Vinod Khosla, in his prescient seminal post called “20 Percent Doctor Included“, cogently laid down back in 2016 why increasingly, AI will take up a bulk of the healthcare burden. Swaasa embodies all of these characteristics he listed, like – Trained personnel – Specialised equipment and Cost.

AIM – Will people trust health apps while doing a self-diagnosis?

Salcit – Self-care/self-diagnosis/self-advocacy is a growing market segment worldwide. People are voting with their pocketbooks by taking out subscriptions in droves. Also, many of them are finding support (reimbursement) from health insurance companies – which stand to benefit by keeping their members healthy.