- Using AI and Deep Learning, the startup helps in the screening of tuberculosis and other chest abnormalities, to enable faster diagnosis and reduce time to treatment.
- Its Chest X-ray solution is being currently utilised to detect abnormalities in lungs that are indicative of Covid-19.
- Qure.ai has also developed an AI-powered chatbot called qSCOUT, for remote COVID care.
The coronavirus pandemic has put the spotlight on the crumbling health infrastructure in the country, and the skewed doctor-patient ratio that is prevalent today. One of the most critical areas of medicine – radiology – has been grappling with an acute shortage of medical professionals for over decades now. Research suggests that there is one radiologist for every one lakh people in India.
To tackle this burgeoning issue, Prashant Warier and Pooja Rao started Qure.ai in 2016.
Using advanced new-age technologies like Artificial Intelligence (AI) and Deep Learning (DL), the startup helps in the screening of tuberculosis, triaging of trauma and stroke cases, and automated reporting of chest X-rays, to enable faster diagnosis and speed to treatment. Its Chest X-ray solution is being currently utilised to detect abnormalities in lungs that are indicative of Covid-19.
Backed by Sequoia Capital and MassMutual Ventures, the Mumbai-based startup (with operations in New York and San Francisco) has impacted more than 600,000 lives to date in over 20 countries, and aims to use AI to make healthcare more accessible and affordable through its real-world solutions.
In a freewheeling chat with The Blue Circle, the co-founder of Qure.ai, Prashant discusses how the idea of the startup came about, what makes them different from the rest, and how they are a step ahead in their fight against Covid-19.
Resonance with Radiology
Before starting Qure.ai, Prashant was running another startup, Imagna Analytics in the adtech space, which was later acquired by Fractal Analytics. As the Chief Data Scientist at Fractal, he wanted to incubate a startup within the company and discussed the idea with the CEO, Srikanth.
“We explored several different ideas and zeroed in on AI in healthcare, specifically around AI in radiology. That’s how the idea of Qure.ai came about,” he added.
Soon after, Prashant reached out to Pooja on LinkedIn. A doctor with an advanced degree in neuroscience, she was then based out of Amsterdam, and was working in the field of data science.
”Her background was very interesting. I have always been a data scientist with no medical training or healthcare background before I started Qure.ai. So I thought Pooja would be a great partner in executing this idea of AI in healthcare. Pooja agreed to move back to India to join us and set this up, so that’s how it all started in early 2016,” shared Prashant.
Initially, the duo spent a lot of time speaking to several radiologists in India, US, and other parts of the world to zero in on the kind of solutions they wanted to build. Moreover, they tried to get access to volumes of data, since large datasets are a key ingredient in training accurate machine learning models.
“Healthcare products are very different from other B2B products because of the regulatory environment around them. We are a software product but are still considered a medical device, so we have to go through the regulatory approval process in various geographies,” shared Prashant.
Resolving Healthcare Challenges
Qure.ai is trying to solve both accessibility and affordability through their path-breaking solutions across the world. In the Philippines, for instance, the startup works with an NGO that conducts tuberculosis screening, and screens about 100 patients a day.
Although they have several mobile vans replete with X-ray units; because of their small size, these vans are unable to accomodate a radiologist alongside a health worker.
“So the healthcare worker collects the X-rays generated over a couple of weeks and sends it to the radiologist to read. By the time the X-ray is read by a radiologist, it might be 2-3 weeks since the acquisition of the X-ray. This is an accessibility problem – people in rural parts of low and middle income countries have limited access to radiology services. Using AI, you can bring the interpretation time of that X-ray read from 3 weeks to a minute,” said Prashant, adding that Qure.ai’s qXR tool is being used to interpret radiology scans within a minute, and reduce time to treatment.
qXR is a CE-certified automated Chest X-ray interpretation tool that screens for tuberculosis and other abnormalities. India recorded more than 2.6 million cases of tuberculosis, and the reason for high morbidity is due to the gap between detection and proper treatment.
The startup is also working on peer-reviewed publications documenting the economic impact, and some joint studies done with customers show that they can improve productivity by 50% or more.
“This will lead to a reduced cost of radiology and healthcare delivery to the patient in the long run, and that’s how we deal with the affordability aspect. Also, by providing an AI “second reader” for radiology studies, we can substantially improve the accuracy of the reads, leading to far fewer missed diagnoses. One paper pegged the cost of misdiagnosis in the US at USD 750 billion” he said.
AI-powered Pandemic Response
Innovation has always been at the centre of Qure.ai’s vision, and this time around, the health startup has re-purposed their Chest X-ray solution qXR for COVID-19. While a confirmed diagnosis of Covid-19 on a Chest X-ray is not possible because COVID-19 resembles several other kinds of pneumonias, yet X-rays still have a use-case in triaging and progression monitoring.
“In late February, we started looking at publications coming out of China about how COVID-19 presents itself on a Chest X-ray. COVID-19 shows up as opacities and shows up as a bilateral distribution i.e. on both sides of the lungs; it presents as peripheral distribution more prevalent on the outside of the lungs. Identifying the presence, location and size of opacities was a capability that our qXR solution already had. We packaged all of this information into a COVID-19 score that can classify X-rays into High, Medium or Low-risk categories for COVID-19 and also separate out completely normal Chest X-rays,” shared Prashant.
There are several cases, wherein RT-PCR tests are in short supply, so the COVID-19 risk score on a Chest X-ray, in conjunction with symptoms and other clinical indications can be used to decide who should be prioritised for testing. Also, by precisely marking out and quantifying the percentage of lung infected, qXR can help measure progression and treatment efficacy, expressed Prashant.
Qure.ai has also developed an AI-powered chatbot called qSCOUT, for remote COVID care. qScout uses a natural language chat capability to enable whatsapp chats (they can also work with other messaging platforms) with self-isolated/ quarantined people and track the progression of their symptoms.
Currently, the government of Oman has integrated qScout in their National COVID response app (Tarassud Plus).
“With lockdowns in effect across the world and most people isolated at home, it became imperative that governments and health systems track people’s health remotely. That’s why we used our 4-years plus of experience in medical natural language processing to develop our new AI-powered chatbot qScout. This longitudinal user data is used to triage whether they need to be hospitalised or need to visit a testing center to be tested for COVID-19,” said Prashant.
Recently, their head CT scan solution – qER – received an approval from the US Food and Drug Administration (it took a year and a half), making it the first Indian artificial intelligence company to get the go-ahead. The tool can be used to triage radiology scans with intracranial bleeds, mass effect, midline shift, and cranial fractures.
This is encouraging for the startup, since it finally allows them to market their solutions in the US.
Besides, Qure.ai is already scaling up, since they now have team members in Europe, USA, Middle East and India.
“Our products can be deployed remotely anywhere in the world. Even during the last 3- months of lockdown, we have deployed our solutions at over 60+ sites in 7 countries. We also recently received funding, which is being utilised to expand the commercial team, build new solutions and invest into regulatory approvals,” said Prashant, signing off.
(Edited by Anu Choudary)
- Here’s how Dozee is Converting Beds into step-down ICUs in 2 Minutes
- This Bengaluru Based Startup’s Portable Ventilator is at the Frontline in Fighting Covid-19
- Home Care Services: A new Lease of Life in the post-Covid era
- How IoT is Boosting Productivity and Improving Lives
- Technology and Healthcare: Collaborating for a Healthier India