Can X-ray and qXR AI be used for improvement in case detection for COVID?

In the wake of COVID-19 pandemic, most health facilities shifted their focus towards mitigating the spread of the virus among the people, and this meant the disease management of other acute health issues such as Tuberculosis (TB) taking a backseat. It was a challenging situation for many countries to attend to both the respiratory illnesses with the same urgency. Wielding an intervention for the case identification of both TB and COVID was the optimal way ahead to ensure care is reaching everyone. The response mechanism by Indus Hospitals in Karachi to employ their existing efforts in TB case management, with qXR for automating CXRs to screen and triage TB and COVID cases simultaneously in communities, is one of the success stories that countries can adopt for challenging times.

qXR | Integrating case finding for Tuberculosis and COVID-19

Qure.ai's flagship product, qXR is an automated chest X-ray (CXR) interpretation software that can prioritize abnormal X-rays and identifies 29 clinically relevant lung abnormalities including indications for Tuberculosis and COVID-19. CE certified, qXR software, is trained on over 3 million chest X-rays using deep learning, and accurately detects clinically relevant findings suggestive of lung diseases such as Pneumonia, Tuberculosis, COPD.

As a robust mechanism to ensure that TB diagnosis and management activities proceed in synergy with efforts to manage this COVID pandemic, qXR bridged these individual disease detection and monitoring activities into one integrated solution.

A Public Health Program was initiated in South East Asia in collaboration with different public and private sectors for mass screening of population for lung ailments. By mobilizing TB screening vans in communities, public and private hospitals across the country, a large population was screened for Tuberculosis and provided testing services for COVID-19. For optimizing the complete workflow of mass screening for integrated TB and COVID-19 using Artificial Intelligence (AI), Qure.ai collaborated with this Public Health Program during the early onset of the COVID pandemic, mid-end of March 2020. Qure deployed its AI solution, qXR in two Emergency Room (ER) settings and 29 mobile vans for drive through screening and testing.

The mobile vans fitted with X-ray units, with adequate staff, qXR powered medical vans have been deployed at multiple sites for active case finding in communities using clinical signs and Chest X-ray screening for COVID-19 and TB. The van is also equipped to perform RT-PCR testing, thus enabling the physician to make diagnosis and corresponding treatment decision. This helps for early diagnosis of cases and mitigate transmission, thus reducing burden on the healthcare system. The facilities that were once available only in hospital establishments were successfully replicated in a van based set-up offering a point-of-care system with reduced turn-around-time.

qXR detects lung abnormalities suggestive of Tuberculosis and radiological manifestations of COVID-19. It also detects findings not indicative of COVID, helping in accurately predicting the likelihood and severity of COVID-19 infection. The likelihood of the disease analyzed by the software as COVID Risk: High/Medium/Low and None, enables the healthcare staff to prioritize patients who need immediate attention in hospitals. Individuals with a likelihood score of Medium and High are triaged for RT-PCR testing, for optimizing the testing capacity in labs. Even in asymptomatic individuals, if AI showed indications of COVID in CXR, they were kept under supervision for further testing.

Taken together, the results of qXR used in conjunction with other clinical manifestations can aid in immediate decision making for collecting swab sample for COVID and/or collecting sputum sample for GeneXpert test for TB from suspects.

Out of 58,163 chest X-rays analyzed so far, 42% of them were analyzed with symptoms, risk factors and exposure, and were categorized for TB/COVID, if present.

Out of 58,163 X-rays analyzed using qXR -TB advised – 1%,
Radiological signs for COVID – 8% and
both TB and COVID – 8%

The analysis and predictions from qXR AI for X-ray screening is as below:

Preliminary analysis | COVID-19
1st April – 30th September 2020

Total 24,843 samples were considered for preliminary analysis for which 42% had symptoms, exposure and X-rays captured. Out of that, 6% of the total individuals had symptoms, exposure and a positive X-ray, 84% of them had symptoms as well as exposure and 10% of them only had presence of radiological signs for COVID analyzed using AI; but no symptoms or risk factor.
Out of 84% of individuals having symptoms and exposure, 16% were confirmed positive, out of 10% X-rays with presence of radiological COVID signs, 14% were confirmed positive and out of 6% with of individuals having symptoms, exposure and presence of radiological COVID signs, 24% were confirmed positive by RT-PCR testing.

Preliminary analysis | Tuberculosis
23rd March – 30th September 2020

Out of 42 thousand individuals screened for Tuberculosis using qXR, 7.7% were advised for TB and 92.3% were not advised for TB using AI. Out 7.7% of TB advised, 63% of individual’s sputum was collected for sputum testing of which 11.8% were biologically confirmed for Tuberculosis. Out 92.3%, 2.9% of individual’s sputum was collected who showed some symptoms, of which only 1.2% were biologically confirmed for Tuberculosis using CB-NAAT.

qXR for COVID 19- Case detection

Symptomatic cases: For individuals with symptoms presentation, and X-ray and subsequently AI analysis will not affect the decision on RT PCR testing. However, it would be an essential information treatment decision. Additionally, it can also correct the false negativity of RT PCR testing methodology.

Asymptomatic cases: For those individuals without symptoms, Chest X-ray + qXR AI is shown to have identified ~1.4% additional cases from general population, who would have been missed and increased the risk of transmission otherwise.

qXR for Tuberculosis: The yield for biological confirmatory test i.e. CB-NAAT is more in the cases which were advised for Tuberculosis qXR AI, helping in the cost reduction of GeneXpert cartridges for confirmation of Tuberculosis.

MUMBAI – the financial capital of the country
The first COVID-19 case was recorded on March 11, 2020, and by March 31st, the city recorded 64 cases & 9 deaths. While Mumbai has 1% population of the country, it has 3% of the Drug sensitive total TB cases and about 14% of total MDR-TB cases registered in country.

What makes Mumbai a vulnerable city for both COVID-19 AND TB?
Most populous Indian city
Home to more than 10lac migrant workers
Presence of some of the largest slums in the world with severe space constraints

At the onset of the pandemic, Qure was requested by the then Municipal Commissioner to deploy qXR at various hospitals across the city. While the initial focus was on testing and diagnosis, there was a gradual shift to patient care and management of admitted positive patients – in Critical Care units as well as dedicated jumbo COVID facilities in the city where qXR was used for progression monitoring. Though Mumbai saw a steep rise in cases more than any other city in India, MCGM’s efforts across various touchpoints in the city were augmented using Qure’s AI-based X-ray interpretation tool – qXR – and the extension of its capabilities and benefits. This project was ACT-IHF grant funded to employ qXR for chest X-ray-based triaging and patient management. MCGM later launched the MISSION ZERO initiative, a public-private partnership supported by the Bill & Melinda Gates Foundation, Bharatiya Jain Sanghatana (BJS) and Desh Apnayen and CREDAI-MCHI.

The approach adopted by MCGM for the pandemic management process are:
Triaging of high-risk and vulnerable and increase in case-detection in a mass screening setting to contain community transmission
Patient management in critical care units to manage mortality rates
Support the existing healthcare framework by launching MISSION ZERO initiative and using chest X-ray based screening for optimum utilization of beds at quarantine centres
Qure deployed its AI solutions as follows, in the MCGM sites which included hotspots and containment zones, Critical Care Units and COVID centres for the management of the whole pandemic-
qXR software – our AI-powered chest X-ray interpretation tool provided the COVID-19 risk score based on the condition of the patient’s lungs
qTrack – our newly launched disease management platform

MCGM sites – Hotspot and containment zones for Active Case Finding among general population/mass screening

Sites: Poddar Van 1, Van 2 [Dr. Shegde, Om Gagangiri Hospital Van]
(mobile vans equipped with qXR)

Quarantine Centres for screening contacts of COVID-19 positive individuals (symptomatic and asymptomatic)

Sites: Bhartiya Jain Sanghatana Van 1, 2 , 3, Poddar Quarantine Centre

BMC COVID centres for Patient Management of COVID-19 positive cases admitted for treatment

Sites: Poddar Isolation Centre, BKC COVID Centre, Seven Hills COVID centre,
NESCO COVID centre, NSCI COVID centre, Richardson Cruddas COVID centre

Designated COVID units for triaging COVID suspects for RT-PCR testing and determining subsequent patient management

Sites: HBT Trauma Centre, Hinduja Hospital, Hiranandani Hospital, Kasturbha Hospital, Chinchpada Hospital

Protocol for screening of general population in high risk areas such as hotspots and containment zones for active case finding of COVID-19

Mobile vans with qXR installed digital X-ray systems were stationed outside various quarantine centers in the city. Individuals identified to be at high-risk of COVID-19 infection by on-site physicians from various camps were directed to these vans for further examination. Based on the clinical and radiological indications of the individuals thus screened, they were requested to proceed for isolation, RT-PCR testing, or continue isolation in the quarantine facility.

Our objective was to reduce the load on the centers by continuously monitoring patients and discharging those who had recovered, making room for new patients to be admitted, and ensuring optimal utilization of resources.