1. General

Category

Category

Health

2. Project Details

Company or Institution

Qure.ai Technologies Private Limited

Project

Strengthening of hospitals in resource limited settings by calling up AI algorithm for Chest Xrays to tackle dual burden of COVID-19 and TB

General description of the AI solution

qXR, an Artificial Intelligence (AI) based software to fully automate Chest X-rays, identifies 28 abnormalities of the chest including 20 CE certified findings in less than 1 minute. qXR can also detect TB and COVID presumptive cases, equipping the health staff for mass screening. qXR can further triage the Chest x-rays into risk categories based on the severity of the condition, equip COVID care centres to manage the patients and make quick decisions. The solution is developed using 3.7 million scans, which is considerably a high volume to test and validate a solution. The sensitivity and specificity of qXR is 95% and 80% respectively, which already meets the WHO recommended accuracy.

Qure.ai’s qXR for TB screening is a WHO endorsed product for aiding TB detection using Chest X-rays and is used in 30+ countries. qXR-TB has been deployed across many LMICs as a decision support tool for identifying TB presumptive cases. Qure.ai has also been at the forefront of the fight against the COVID-19 virus, by introducing intervention of technology into a country’s existing infrastructure to address these issues. Qure has developed a suite of solutions that streamlines operations for end-to-end management and advancement of the health system. The digital interventions in imaging and workflow optimization developed by Qure.ai place emphasis on strengthening the community level operations by simplifying the workload of the stakeholders involved at various levels.  

QXR can currently identify 28 findings with disease focus areas such as Tuberculosis, COVID, lung malignancy, COPD. The solutions by Qure are currently live in 428+ sites in 47+ countries with key partners such as AstraZeneca, Partners in Health, IRD Vietnam, NHS UK, PATH, MCGM, to name a few.

Website

https://qure.ai

Organisation

Qure.ai Technologies Private Limited

3. Aspects

Excellence and Scientific Quality: Please detail the improvements made by the nominee or the nominees’ team or yourself if your applying for the award, and why they have been a success.

qXR is a deep learning based technology developed by Qure.ai with a large dataset of 3.7 million scans. In a prospective clinical validation trial, qXR outperformed other CXR AI algorithms (CAD4TB, and Insight CXR) and exceeded World Health Organisation’s (WHO) accuracy requirement for radiologist performance in an independent evaluation by Stop TB Partnership.
Qure.ai is the only Indian company who has been recognised in building deep learning based solutions for medical imaging of global standard. Both qXR, and qER has the capability of interpreting multiple abnormalities in the scan, while some of the competing products are limited by the number of findings they can identify.
qXR is used in multiple applications – radiologist reporting assistance, worklist prioritisation, screening and triaging of presumptive disease cases, progression monitoring, placement of lines and tubes in an emergency department setting. The solution is vendor agnostic and is compatible with all types of x-ray machines and specification – analog, CR amd digital radiography. The software is at TRL-9, already in the market and proven effective.
QXR has been granted 2 US patents, and has 6 published patents in India, Europe, China and Japan. Qure has 37+ publications in its name, in leading journals such as The Lancet, Nature amongst others. qXR is also presented in leading conferences such as the RSNA, ECR.

Following are the leading media Coverage received for qXR
The New York Times coverage of Qure.ai’s work towards eliminating Tuberculosis in India
MIT Technology Review -UK National Health Service’s (NHS) use of Qure.ai’s tools for COVID-19
BBC, Wired and Forbes articles that feature Qure.ai’s technology

Scaling of impact to SDGs: Please detail how many citizens/communities and/or researchers/businesses this has had or can have a positive impact on, including particular groups where applicable and to what extent.

The COVID-19 pandemic has created huge stress on existing health infrastructure of LMICs, especially in rural areas where access to primary/secondary care and timely testing is severely limited. Currently, over 40% districts in India don’t have a single RT-PCR lab for COVID-19 testing (ICMR, 2021). COVID-19 is also derailing global efforts to eliminate T; in 2020, the World Health Organization (WHO) reported a 50% drop in the number of TB patients diagnosed, which may have resulted in up to 400,000 additional TB deaths. In this scenario, Chest X-rays act as a crucial primary bi-directional screening tool to detect TB and COVID-19, given their wide availability and affordability.
Further, while hospitals in rural areas are equipped with chest X-Ray facilities, they lack of the availability dedicated radiologists or chest physicians and overall, poor access to radiology services, leading to large turnaround times for interpretation of chest X-Rays and thus, limited use of chest X-Rays for triage. And with testing facilities located far away or not available, this resulted in not only in huge delays in diagnosis but also increased spread of infectious lung diseases such as COVID-19 and TB in the meantime.
To bridge this gap, Qure.ai partnered with rural and semi-urban secondary hospitals dedicated to delivering quality and affordable healthcare at the last mile, to support clinicians at these hospitals (handling the COVID-19 crisis in addition to the existing TB burden, despite their limited resources) to augment their capacity and support them in providing care to the most vulnerable populations. qXR is deployed and actively being used for COVID-19 & TB screening and triaging by clinicians across 20+ remote secondary hospitals across 15 states in India; over 35,000 individuals have been screened as of mid-2021 for COVID-19 and TB using AI as a result of this intervention.

Scaling of AI solution: Please detail what proof of concept or implementations can you show now in terms of its efficacy and how the solution can be scaled to provide a global impact ad how realistic that scaling is.

In 2020, qXR was used for screening and triaging individuals for COVID-19 to bolster Mumbai’s Municipal Corporation’s COVID management efforts in the city with AI. qXR was deployed across multiple health facilities for COVID-19 detection and used for active case finding using X-ray equipped mobile vans. Over 25,000 Chest X-Rays were screened for COVID through this intervention; A few impact highlights are mentioned below:

20.3% of general population were identified as COVID-19 suspects through mass screening in hotspots and screening of contacts of COVID-19 positive individuals. These included asymptomatic individuals which make up more than 50% of positive cases; These individuals were immediately directed for next steps in patient management process
90.8% correlation was observed between COVID-19 likelihood predicted by qXR and RT-PCR confirmation of disease
Early diagnosis of co-morbid individuals was reported as being especially beneficial in influencing their treatment outcome

Building on the success of this intervention, as well as recognising need for a tool such as qXR in rural areas as well as the impact of COVID-19 on TB, Qure.ai expanded its efforts towards strengthening healthcare systems in remote and resource-limited settings by scaling up usage of qXR for COVID-19 and TB screening. Over 35,000+ individuals have been reached through our work with rural and semi-urban secondary hospitals across 15 states in India. Out of those who were flagged as COVID-19 presumptives by qXR during the first few months of this intervention (for which data was available):

58.3% were admitted or isolated within the hospital/advised for home isolation/referred to another COVID centre as a next step
61.2% were referred for a COVID-19 microbiological test such as RT-PCR or Rapid Antigen test, or similar supportive test
11.2% were incidentally detected as COVID-19 presumptives as a result of active surveillance using qXR; these cases would likely be missed otherwise

Ethical aspect: Please detail the way the solution addresses any of the main ethical aspects, including trustworthiness, bias, gender issues, etc.

Qure.ai believes in the mission of using Artificial intelligence to make healthcare more accessible and affordable. Qure’s products are validated on large dataset of 4.2 million scans on diverse population, thus reducing chances of any bias in the interpretation. The products are live in 47+ countries, including 37 Global South countries. qXR is used in a lot of resource limited settings for Tuberculosis triaging, and has been found to be effective in reducing the turn-around time and identifying presumptive patients early. The x-rays processed by qXR is done consecutively without any patient information, thereby not allowing for any discrimination among people to receive equitable access.

qXR is GDPR compliant and HIPAA compliant tool, ensuring safety of patient data. The scans processed for screening by qXR is de-identified with only the x-ray available for processing. Qure conducts audits very year for quality control, and is also ISO 27001 compliant. Recently, qXR was endorsed by WHO as one of the three AI tools that can be used for tuberculosis triaging as an alternative to radiologist. Informed consent is taken from the patients before x-ray is processed by the implementation/program partners. Ethics approval is sought for all the studies conducted by Qure.

QXR is used in various applications for screening all lung ailments. With a single solution, multiple ailments such as Tuberculosis, COVID risk, Lung malignancy risk, COPD, Pneumoconiosis can be identified using qXR in under a minute. This is very useful for countries that has a high burden of infectious diseases and other lung ailments. qXR promotes for rapid diagnosis of age group of 6 years and above, allowing for more people to be screened, irrespective of the gender or ethnicity, and access to early treatment.