When technology is disrupting all the major industries, one of the most deeply affected areas has been healthcare. With a slew of startups and companies bringing innovative AI and analytics-driven solutions, the healthcare industry has seen drastic developments. It has empowered the care-givers through exponential technology, knowledge, resource and infrastructure.
This week we got in touch with one such entrepreneur who has been building diagnostic software based on emerging technologies to detect complex diseases from medical images — making it faster, cheaper and robust. Mausumi Acharyya is the founder and CEO of Chandigarh-based Advenio, that is providing state-of-the-art computer vision and AI solutions. Currently, 10-member strong, they are aggressively working to demonstrate their technology in as many operational environments as possible. Here are the interview excerpts.
Analytics India Magazine: How has AI disrupted the healthcare space in India?
Mausumi Acharyya: In India, we are now moving from curative healthcare to a predictive one. The offering of today’s technology to capture big data and analyse it is truly disrupting. It has enriched healthcare deliveries for more accurate diagnosis and patient-specific treatment. Technology is also increasing the general awareness of people on personal health and providing self-monitoring through plenty of health and wellness apps. Having said that, the majority of India is rural population and with highly skewed doctor-patient ratio, the disarray in public health can only be solved through technology.
Remote patient monitoring, early disease screening are actually transforming public health delivery. Exponential technologies like artificial intelligence, social media, smartphone, IoT, cloud computing, big data and predictive analytics, imaging, genomics, internet penetration and mobility, blockchain etc., are rapidly transforming the healthcare pathway in India. Although this technological inclusion in healthcare in India is in a nascent stage but it would soon have enormous growth and disrupt the healthcare ecosystem.
AIM: Please tell us about Advenio. How is it leveraging AI to be a major influence on the Indian healthcare space? How does this AI technology work in detecting diseases?
MA: The AI technique by Advenio uses structured data such as digitized medical images for the purpose of disease detection. It identifies the characteristics of diseases from the images while appending the clinical symptoms of the subject and other clinical and genetic data. The AI system is trained through a large collection of clinical data (training data) from different subjects. The training phase extracts patterns characterizing of the diseases and descriptive features from the images so that the model can learn a similar group of information, associate the features and patterns and assign an outcome.
During training we design and code algorithms for building sophisticated mathematical functions by extracting meaningful patterns from the training data. Once the training is completed, the AI model so generated tries to detect the patterns and features learnt in unknown images. Efficacy of the AI model is determined by the accuracy with which the model detects the disease condition in the unknown data. The AI model learns only those patterns present in the training set. It is to be noted that this is a very naive explanation of the methodology while there are several other advanced and cutting edge development of the technology which is being actively researched.
AIM: What are the various diseases that you provide a diagnosis for? How has it been
received so far by the end-users?
MA: Advenio’s flagship AI platform analyzes retinal (fundus) images in real-time. It detects several prevalent ocular diseases and currently, we are actively researching to detect a host of lifestyle-related systemic diseases as well. At present, our prototype is in a phase of demonstration in an operational environment.
AIM: How did the idea of venturing into disruptive, affordable, equitable healthcare space
occur to you?
MA: I decided to embark on an entrepreneurial journey while I was professionally engaged with the med-tech giants (Siemens and GE). The fact that healthcare is grossly disjointed and needs technology to solve the unmet problems; has pushed me to start my venture where my vision has been to build health-tech solutions for an equitable world. My ideas primarily were around building healthcare solutions at the point-of-care, which would be highly cost-effective and simple to use. At the same time, I wanted to build innovative solutions for differential diagnosis to detect crucial and complex diseases from medical images and mitigate uncertainties in interpretation — making it faster, cheaper and more robust than traditional methods.
AIM: What have been the challenges of transitioning from the corporate world to startup space?
MA: Establishing Advenio has been an uphill journey for me professionally as well as personally. They say that the brightest day follows the darkest night; this summarizes my journey towards creating Advenio. The decision to start my own venture was not an easy one. I was at the height of my career. To leave all that and a cushy salary, all for an uncertain future was the greatest risk I have taken in life.
When I embarked into this journey the Indian economy was not in the best conditions. To start a new venture alone, without a capital, team or a product in hand was a big struggle. Another issue I faced is financing myself and sustaining alone. Keeping me motivated required tremendous determination and will power. I also had to overcome personal and emotional loss and pursue a new direction. However, I received immense support from my family and friends during my whole entrepreneurial journey. It took me two years of working on my own with limited social resources and support, to come up with a product, and another two years to start a proper team.
AIM: How does the technology stack at Advenio look like? What are the various technologies and tools that you use?
MA: We use core image processing, machine learning and deep learning for building our technology. There is considerable use of statistical learning and data science too. We use several state-of-the-art machine learning libraries, cloud, server, databases and cutting edge web technologies to build our platform.
AIM: How has been the growth story so far? What have been some of the important milestones?
MA: Since the healthcare ecosystem is yet to be established, there is a lack of data and statistics as reference or precedence; hence the adoption of new technological solutions is a challenge and monetization is difficult. Gaining the trust of the target audience and prove reliability on any new technology is time-consuming and is another trial. After years of such struggle, the situation has started changing and there is not the only trend of adoption and acceptance to our solution but also there are interests in strategic collaboration and partnerships with other renowned and established med-tech enablers and stakeholders amongst the healthcare fraternity for both complementing and supplementing respective offerings. This, in my opinion, is definitely a step towards success.