AI based Integrated Lung Health Screening

In 2017, 544·9 million people worldwide had a chronic respiratory disease, representing an increase of 39·8% compared with 1990. Chronic respiratory diseases were the third leading cause of death in 2017, behind cardiovascular diseases and neoplasms. Additionally, 87% of TB Burden is shared by the top 30 countries ranked according to prevalence rate. A majority of these countries belong to LMIC category. Moreover, Lung cancer which is the leading cause of all cancer deaths worldwide, comprising 18.4% of all cancer deaths has shown an increasing trend in incidence in developing countries attributed to tobacco smoking and various environmental and occupational risk factors. Implementation of lung cancer screening is challenging, so organised lung cancer screening is practically non-existent.

Despite the exceedingly high burden of respiratory diseases in LMIC, there is a dearth of organised effort or infrastructure to address there. Where national programs do exist, there seem to existential structural defects ranging from lack of testing facilities, underequipped and in trained human resource and a desperate need to trained specialists to speed up the diagnosis process and bringing the patients into the treatment and care cascade. This directly effects turnaround time where in case of Lung Cancer the median time from initial presentation to specialist referral was 9.5 days; in fact some patients have to experience a wait of more than 90 days before they can be referred to a specialist. Similarly for TB, in case GeneXpert testing is available, the median TAT is 7 days. Where diagnosis depends on sputum microscopy and clinical correlation, this wait can go up to 21 days.

Hence, there is an urgent need for deep learning in tuberculosis and lung health screening especially at primary healthcare levels ro facilitate faster case finding. There is renewed interest in using chest X-rays to screen for pulmonary tuberculosis, largely due to the increasing availability of digital X-ray facilities which offer a highly sensitive, relatively inexpensive, rapid screening tool for large populations. The availability of healthcare professionals with degree of expertise needed to screen a chest X-ray is not increasing as fast as the availability of medical imaging facilities. At the same time, there is a decline in the ability to accurately read chest X-rays - both among primary care physicians as well as newly trained radiologists. Computer-assisted diagnosis (CAD) is of great value in this setting because it offers significant productivity gains and is reproducible across centers. Deep learning algorithms, a kind of artificial intelligence, have enabled computers to understand, interpret and label images by learning from examples. When trained on large datasets, these are much more scalable and accurate than traditional CAD, and continue to improve with each image that they are exposed to. These algorithms are revolutionizing computer-aided medical image interpretation.

A major problem that we at Qure are trying to solve is to Algorithm accuracy and applicability to reduce misdiagnosis and delay in treatment. There have been instances wherein, Lung Cancer incorrectly diagnosed as Tuberculosis has caused unfortunate delay in availability of treatment to the patients. We seek to minimise these misdiagnoses through our AI based detection algorithm and streamline the patient management process by integrating qTrack with the existing diagnostic process.

To scale further and increase penetration of the algorithms into more screening programs, we need to renew our approach. With 3.7 billion unique users worldwide, smartphone penetration and processing power has opened new avenues. With this project, we aim to deploy smartphone based diagnostics to screen from film images of X-rays using AI, widening the reach to include Analog X-rays.

qTrack is an end-to-end AI-powered automated disease management platform. It has been developed to be a singular, holistic data gathering and reporting system that enables effective response by providing healthcare professionals with ready access to all patient information, diagnoses trackability, test results and real-time progression monitoring. qTrack operates at the intersection of Lung Health, Public Health and AI-driven Radiology to enable a care continuum

 qXR is the integrated into qTrack which is an AI based software chest X-ray screening platform that detects signs of pulmonary, hilar, and pleural tuberculosis. The artificial intelligence algorithm underlying qXR is trained to detect not only classical primary pulmonary TB, but also atypical manifestations. It can be used to simultaneously screen for COPD, lung malignancies in high-risk populations, and certain cardiac disorders - a total of 30 different findings in a Chest X-ray. Designed for use in a real-world setting, qXR is hardware-agnostic and works with X-rays of varying quality and exposure, from all X- ray machine models - Not only TB : The same tool can also be used to screen for other chest abnormalities - Zero-footprint solution, with no extra hardware required - Artificial intelligence algorithm trained on over X-rays. - The smartphone based solution can image a chest X-ray film and give results for TB screening using the qXR deep learning platform. The platform provides integrated disease management and screening system for 30 lung conditions –

Cardiomegaly Hyperinflation/Emphysema Consolidation

Degenerative spine conditions

Pleural effusion

Prominence in Hilar region Opacity

Blunted  Costophrenic  Angle

Scoliosis

Tracheal  shift

Fibrosis

Tuberculosis screening

Atelectasis

Reticulo-nodular pattern

Nodules

Cavity

Calcification, Pneumothorax

Covid-19 risk

Elevated Hemidiaphragm

Pneumoperitoneum

Rib Fractures

Mediastinal Widening Linear Opacities

Presence of Tracheal Tube

Presence of Gastric Tube

Placement of Tracheal Tube Placement of Gastric Tube

 DICOM Input (Inclusion and Exclusion criteria)

qTrack through the qXR software  accepts  and  processes  anonymized  studies  in  a  valid  DICOM  format  (.dcm) satisfying the following criteria:

Inclusion Criteria:

-       Age 6 years and above

-       Modality CR/DR/DX

-       Study description Chest

-       Body Part Examined Chest

-       View AP/PA

-       Format DICOM (.dcm)

-       Image resolution required minimum of 1440*1440

-       Gray level at least 10 bits

Exclusion Criteria:

-       Lateral X-rays

-       X-rays that do not contain the entire lung field

-       X-rays of other body parts

-       X-rays of patients below 6 years of age.

qTrack mobile is an end-end disease management platform which provides a flexible option for low-resource settings where mainframe PCs and high grade network connectivity may not be available. This solution pacakges qXR and qTrack in a mobile app, Qure.ai’s flagship products that aid in quick diagnosis of Tuberculosis, COVID-19 and 28 other conditions and the subsequent patient data management. The app is designed to support and drive value  key stakeholders in global health programs:

QBox

QBox is a Mini-PC installed with Qure’s software qXR along with its local server. All data is processed locally within the site premises. qBox has been designed for use in remote and under-resourced areas where internet connectivity is unavailable and when data is required to be stored within client premises.

qBox Differentiator

·       Hybrid solution  - It can work both with and without the internet.

·       Easy integration - It can be integrated with any Digital X-Ray system.

·       Seamless deployment - Scans can be pushed directly to the Qure PACS node or they can be uploaded manually via the qTrack app.

 The combination of qXR and qTrack has been optimised to run on qboxes and support lung health screening in a variety of settings.

The solution focuses on improving point-of-care services and diagnostic services at the primary healthcare level. This essentially focuses on providing a means for  Frontline Health Workers and Junior Doctors to provide high quality diagnostic services and reduce the number of misdiagnosis and turnaround time to increase case finding at the most grassroot levels.

Qure, through its solutions wants to give power to these key demographic as they are usually the ones at the frontline of case detection and treatment care cascade.

Frontline staff

Frontline workers can use the app to register patients along with symptoms, risk factors, test results and diagnosis.

Physicians

Physicians / Radiologists can use the mobile app based image viewing platform to log in and view  results and comment on the AI findings on any device of their choice.

 Program managers

qTrack provides Mobile app or web-based real time dashboards for management teams to generate MIS reports for each site. There is ease of deployment at scale given the solution is localized and configurable.

Founded  in  2016,  Qure.ai is  a  breakthrough  Artificial Intelligence  (AI)  solution  provider  that  is  disrupting  the radiology ‘status quo’ by enhancing imaging accuracy and improving  health  outcomes  with  the  assistance  of machine-supported  tools. Qure.ai taps  deep  learning technology  to  provide  automated  interpretation  of radiology exams like X-rays, CTs and MRI scans for time and  resource-strapped  medical  imaging  professionals enabling faster diagnosis and speed to treatment.

o   Deployed in 50+ countries around the world

o   Impacted 10 million lives across the world

o   One of the largest imaging database in the world

o   More than 6 million scans with  corresponding radiology reports from 150+ sites, 10+ hospital and radiology groups 

o   Sites Deployed - 428+

o   Clinically validated, with peer reviewed publication

o   First AI-based chest x-ray interpretation tool to receive CE certification

o   qXR’s performance led to the WHO recommending that among individuals aged 15 years and older in populations in which TB screening is recommended, computer-aided detection software programmes may be used in place of human readers for interpreting digital chest X-rays for screening and triage for TB disease.

o   Studies conducted in partnership with StopTB indicated that qXR was able to reduce confirmatory molecular test consumption by 33%, costs by up to 66%, while operating at 95% sensitivity or higher.

o   With a 90% sensitivity rate, using qXR in conjunction with sputum microscopy can reduce the number of patients incorrectly suspected of pulmonary TB and referred for molecular tests, helping to reduce program costs significantly.

o  Multiple validation and field studies have proven qXR as a preferred AI solution for emergency settings including -

https://qure.ai/evidence/artif...

https://qure.ai/evidence/can-a...

https://qure.ai/evidence/using-artificial-intelligence-to-read-chest-radiographs-for-tuberculosis-detection-a-multi-site-evaluation-of-the-diagnostic-accuracy-of-three-deep-learning-systems/

QureAI in the course of its operations has been able to expand to 50+ countries and our AI based solutions have been deployed in more than 418 sites globally. The organisation is now seeking to expand to a truly global reach with the endeavour of helping 1 billion lives in its future. Qure has been in the continuous process of developing  AI based solutions that seek to cover the entire treatment cascade.

This requires a level of engagement with not only our implementation partners, but also with national governments, multilateral global agencies on one end of the spectrum and community based grass root organisations on the other end. This is the level of outreach QureAI is looking for in its near future to truly expand its solutions to a global level.

At the same time, we also aspire to expand our market presence by providing cutting edge, machine learning based AI tools at an affordable rate to make sure that the end user, wherever they may be, are able to overcome the barriers that may be extant for them due to geography, training or inequity. This in part will enable that only the highest quality of diagnostic services are delivered at each of our deployment, irrespective of the nature, scope or the location of the point-of-care.

Qure.ai’s flagship product - qXR is an AI solution trained on one of the largest datasets of over 8 million+ Chest X-rays from across the globe- which can screen for COVID- 19, Tuberculosis and 28 other findings from chest X-rays under a minute. As the most extensively tested AI solution developed to date, Qure.ai is uniquely positioned to lead the category in applying new solutions and answering the industry’s most difficult challenges. The technology fulfils a pertinent, unmet need in the radiology industry. Along with qXR, there is also an end-end disease management solution called qTrack that been developed to help maintain the care cascade between providers and patients.

With the aid of sophisticated AI tools that can instantly evaluate scans and x-rays to quickly prioritize actionable patient cases, clinicians and radiologists can then focus their time and advanced skill set on the most pressing diagnoses. Most importantly, the Qure team is committed to aiding public health/ radiology professionals in diagnosing illnesses faster, and with more detail and accuracy, than ever before.

The qTrack and qXR software’s are evolving modalities. We at QureAI aim at continuously improving our solutions by expanding to newer geographies and integrating the learnings into the machine learning continuum. With additional data sets, we seek to expand the conditions diagnosed by to 56 in the near future and assist in designing a more streamlined and comprehensive workflow.

With the aid of sophisticated AI tools that can instantly evaluate scans and x-rays to quickly prioritize actionable patient cases, clinicians and radiologists can then focus their time and advanced skill set on the most pressing diagnoses. Most importantly, the Qure team is committed to aiding public health/ radiology professionals in diagnosing illnesses faster, and with more detail and accuracy, than ever before.

QureAI seeks to provide cutting-edge, customised radiology solutions across a varied geography and for multiple end-users. This includes health facilities, from primary to tertiary levels of care, local governments to national governments and both public and private clients along with partnerships with global organisations such as AstraZeneca. Through these levels of engagement QureAI has ensured impact at a global level with more than 10 million people assited each year.

Impact this year –

QureAI through its AI based solutions has been able to conduct in excess of 5 million scans and has been able to diagnose 1 million critical cases by early stage diagnosis of Lung Cancer and Tuberculosis.

Impact in Next 1 year –

QureAI, through its ever expanding global footprint of 50+ countries seeks to detect 5 million scans and ensure early stage detection of 1.5-2 million critical conditions such as Lung Cancer and Tuberculosis cases.

Impact in the next 5 years –

At Qure, we seek to ultimately touch a billion lives through ourAI solutions. Persuant to the same we aim at scanning 100 million cases. Additionally we seek to detect 10 million critical alert cases including Lung Cancer and Tuberculosis

The organisation seeks to achieve these goals by undertaking a major push across the spectrum and promote holistic all-round growth through global partnership and collaboration. Some examples of the same include

-       QureAI is collaborating with AstraZenaca to promote early stage Lung Cancer detection which includes clinical trials to validate our AI based solutions.

-       We have been recipient of multiple grants including SBRI Lung Cancer Detection grant and a grant from NHS UK for evaluation of qER as a triage system for acute head abnormalities to improve patient outcomes and operational efficiency.

-       Qure solutions have been additionally deployed in NHS projects such as NHS Bolton, Manchester, East Kent and Frimley

-       The organisation is focussed on bringing AI based technological solutions for day-to-day issues faced in radiology workflow, with a special focus on Low and Middle income Countries

-       Within LMIC, we have a growing footprint in 42 sites across Africa including South Africa, Somalia, Kenya, Malawi, Mozambique, Lesotho, Zimbabwe, Zambia, Nigeria and Uganda

Qure is also stepping up its cooperation with global initiatives with partners such as Stop TB, Bill and Melinda Gates Foundation and World Health Organizations. We have partnered up with the on ground implementation partners of the initiatives and aim to work towards the collective goal of reducing morbidity due to diseases of global concern

General Screening Data points -

·      Number of new hospitals/facilities equipped with AI

Nature of facilities deployed at

urban/semi-urban/rural

·      Number of new countries/states/districts qTrack presence expanded to

o   Government/private

o   Screening camps/fixed facility/screening vans

·      Number of user trained and equipped for using digital based workflow

·      Number/proportion of frontline health workers/non-specialists empowered to make patient management decisions using qTrack

·      Number of individuals screened using qTrack

·      Number of Individuals flagged for any lung abnormality

o   Number of TB Patients diagnosed

o   Number of TB patients referred for confirmatory test

o   Number of presumptive lung cancer patients detected

o   Number of patients referred confirmatory CT testing

·      Average reduction in Turnaround Time (TAT) from baseline (abnormality specific)

o   TAT reduced for presumptive TB cases

o   TAT reduction for presumptive Lung Cancer cases

We also hope to collect qualitative feedback from doctors and their medical teams on their experience of using qXR and resulting value added (for eg. with respect to active case finding of suspects, early detection among others), as well as feedback on patient experience where possible.

Please find our Theory of Change document in the following link

In case of any issues accessing it please reach out at vishakh.saraf@qure.ai

https://qure1-my.sharepoint.co...

·      qTrack is an end-to-end AI-powered automated disease management platform. It has been developed to be a singular, holistic data gathering and reporting system that enables effective response by providing healthcare professionals with ready access to all patient information, diagnoses trackability, test results and real-time progression monitoring. qTrack operates at the intersection of Lung Health, Public Health and AI-driven Radiology to enable a care continuum

·      qXR which forms the AI component of qTrack has been trained on a 8 million plus global training data set which is the largest data set in the world for medical images.

With its solutions, Qure AI focuses in providing seamless workflow experience with an overall aim of ensuring an extremely easy to use user interface, thereby reducing the turn around time (TAT) of diagnosis of lung abnormalities in low resource settings and giving the decision making power to the frontline medical staff which includes -

Frontline staff and Physicians –

·       Disease Management

-       Helps record cases from registration to treatment outcomes

-       Creates a comprehensive case document with demographic and clinical details

-       Highly configurable system Optimized for data collection in low connectivity settings

·       Care Coordination

-       Brings stakeholders to a single platform to track patient journey

-       Supports prioritized lists, notifications, and reminders

-       Enable co-ordination between primary & referral sites

·       Radiology Workflow

-       Triage X-rays as normal or abnormal using AI

-       Supports teleradiology workflows and integrations with various PACS and RIS systems

-       Can also read analog X-rays when used against a lightbox and image taken on a smartphone

Program Manager –

-       Equipped with dashboards, hotspot maps & export features for efficient program monitoring & management

-       Interoperable and can support integrations with national systems or program stakeholder systems

Qure.ai will treat all individuals fairly and impartially, without prejudice, and not tolerate harassment in any form. We are committed to developing a working culture that is fair and 'inclusive' -enabling all employees to make their distinctive contributions to the benefit of the business. We are also determined to ensure that we extend this same openness to our suppliers, business partners and all our customers. We expect our managers to exercise leadership in this field by discouraging prejudice and by role modelling appropriate behaviour.

Qure.ai is an “equal opportunity” organization that prohibits discrimination or harassment based on race, color, religion, national origin, sex, age, sexual orientation, marital status, citizenship status, or disability.

Commercial Costing Models

 Qure.ai follows the Pay-Per-Use Model with the aim of ensuring affordable access to even the most basic level of health facility. This allows low volume clinics and radiology centres in suburban and rural areas to utilise cutting edge AI based diagnostic solutions in an affordable manner. The commercial licencing model is described below. We also believe that no one size fits all and hence offer multiple subscription options that includes –

-       In case of Immigrant Scanning that QureAI is facilitating for Government of United Arab Emirates and in Australia, the pay-per-scan model is offered

-       Conversely in our deployments in Africa, we offer an annual or biennial costing package wherin cost of total scans for the entire year or two year is charged together.

-       Keeping in mind the kind of geographies and settings we cover, QureAI has a rational costing structure that seeks to provide high end AI based teleradiology solutions even in the most remote, resource-poor regions.

Particulars

Amount (in USD)

Price of Product/Service

Price of the One time SOFTWARE Deployment costs –starting 3.75 Product/Service  Lakhs/Site- Operational Usage price to run scans- function of volumes ( upto $0.5-1.5 / scan)

Manufacturing Cost

N/A- For a software device, it's primarily support and Cost  maintenance costs + hosting costs.

Distribution and Marketing Cost

N/A

Gross Margin

N/A

Equity

 Qure.AI utilises a broad source of investments which include both commercial investments in form of equity and philanthropic sources, including grants and public health funds. Some of our equity partners have been -

- Fractal 

- SCI INVESTMENTS VI 

- REDWOOD TRUST 

- MassMutual Ventures Southeast Asia I LLC 

- Employee Round  

- Pooja Rao, ESOP exercise 

Philanthropic Sources and Grants

 Additionally, QureAI is also supported by philanthropic grants that enable deployment of AI based solutions in low-resource settings with the stated aim of assisting diagnosis and streamlining the workflow process as per the project needs. The process unfolds in the following manner –

-       QureAI is allotted a grant

-       Operationalisation in the given geography

-       Leads to direct procurement of the software solutions

-       Qure assists in integrating in the presently available diagnostic infrastructure through its deployment

Source of total philanthropic revenue as of last financial year

Confluence for Health Action and Transformation Foundation- India Funding Health Fund- qtrack  (January 2020 – December 2022)

ACT Grant 1-BMC- 6 months - 19th May to 31st October 2020 - qXR

ACT Grant 2 -Rural Hospital scale up - 6 months from Jan 2021 July 2021- qXR

Secretary of State for Health and Social care - UK- qER (head CT solution) (July 2021 – June 2023)

As done previously we want to use the funding we receive to establish the best practices around the innovation we built. Once the practice becomes well-adopted and has generated evidence and RoI (Return-on-investment) for the end-user, we would want to go for formal commercial procurement processes or find a sustained funding source like Global Fund to sustain the efforts post-implementation and successful adoption.

We have had multiple instances in the past of scaling external-funded projects to become self-sustainable. Examples of the funding obtained and self-sufficient scalable projects as of today are listed below

CIDRZ Zambia (Stop TB Partnership)

We deployed the project under scope of Wave 6 grant which has now been established as a regular SOP after due commercial procurement process

MCGM India (Stop TB Partnership)

Received $99,000 USD funding as part of grant to establish passive surveillance of TB cases in Mumbai which had now standard practice. The procurement is currently under RFP process post which it would become a self-sustained project.