Using AI to Aid Medical Diagnosis

Image 1.jpg

SigTuple, is an AI start-up that aids medical diagnosis by building intelligent screening solutions. Ikyatha Yerasala finds out more about the Bangalore-based company.

 

When three American Express executives with absolutely no background in the healthcare industry decided to create an AI health start-up, it was definitely a bit of a gamble. But despite the possibility of having to face challenges in a capricious field, they got together to form SigTuple, a company that utilises artificial intelligence to improve speed and accuracy of medical diagnosis. In a chat with HE, Rohit Kumar Pandey, CEO of the Bangalore-based start-up talks about their journey, their inventive products and more.

HOW WAS SIGTUPLE FORMED?

Apurv Anand, Tathagato Rai Dastidar and I met in 2012 when American Express decided to setup a Big Data Labs team in Bangalore. We were given the charter to setup the team and the big data stack so that intelligent data products can be brought to the market. We worked together for two and a half years and then started looking for interesting opportunities. The healthcare industry caught our attention because it has tremendous amount of data, both in the form of numbers and images, and the potential was hugely untapped. Additionally, there was an opportunity to touch the lives of billions. Therefore, even though we had no background in healthcare, we decided to take a plunge.

WHEN YOU STARTED OFF, HOW DID YOU INTEND TO REVOLUTIONISE THE MEDICAL DIAGNOSTIC SPACE?

We registered SigTuple in July 2015, but the actual work on the solution started only in September 2015 because we had to find a partner who could support us with data and consultancy. The healthcare industry has tremendous amount of data and we believe that every decision made by a medical expert can be augmented with data-driven intelligence. We have started our journey by developing smart disease-screening solutions which are powered by advancements in artificial intelligence, image processing and cloud computing. These solutions will change the way the common disease-screening tests are being done by improving the quality of reporting, reducing the cost and turnaround time. In the coming years, we intend to move up the value chain i.e. from screening, diagnosis to prognosis and monitoring.

TELL US ABOUT YOUR PLATFORM MANTHANA AND THE SOLUTIONS YOU OFFER. HOW DO YOU PLAN TO CHANGE THE FACE OF HEALTHCARE THROUGH YOUR PRODUCTS?

We have developed Manthana, a cloud-based continuous learning AI platform to bring multiple solutions to the market. This platform enables our data scientists to work on multiple solutions in parallel which will enable us to cover a larger spectrum of tests. Through this platform, every decision can be backed up by evidence (visual/numerical) which will help in reducing the bias in reporting. It will also help in improving the quality and turnaround time of reporting, which will enable patients to get timely and effective diagnosis. Last but not the least, it will increase the efficiency and outreach of medical experts.

AS AN AI STARTUP, WHAT CHALLENGES DID YOU FACE ALONG THE WAY AND HOW DID YOU TACKLE THEM?

We faced multiple challenges on the way. Even though there is an enormous amount of data within healthcare, it is not stored in a structured manner, which makes it very difficult for an AI startup to consume it directly. There is a lot of inter-observer variability in the manual analysis of visual medical data and therefore, the annotated data can’t be directly consumed for training AI models. Furthermore, there are a lot of good and bad things being talked about AI which makes change management critical and challenging.

HOW DID YOU GO ABOUT RAISING FUNDS FOR YOUR VENTURE? YOU RAISED $5.8 MILLION EARLIER THIS YEAR; TELL US MORE ABOUT THAT.

We wanted to scale the company, so we started looking for funds in August 2016. It took us six months to close the round and we had to go through some tough discussions because of multiple reasons– new technology, sensitive and critical industry and also, the founding team didn’t have any medical background and there were a few bad examples that industry had already seen. Also, the Indian start-up ecosystem is known for service-driven companies and the traction around the product companies is limited. However, we took it up as a challenge and went through multiple rounds of discussions where we provided demo of our solution, clinical trial reports, inputs from medical experts etc. Seeing is believing, and it worked for us too. After four months into the funding discussion, we received backing of more than 10 investors and we closed the round in February 2017.

WHAT ARE YOUR FUTURE PLANS AND WHAT MORE DO YOU INTEND TO DO IN THE MEDICAL TECHNOLOGY FIELD?

We are scratching the surface. At this point in time, our focus is to start distributing our flagship blood screening solution (ShonitTM) into the market. Gradually in the next few quarters, we plan to distribute solutions for urine (Shrava), semen (Aadi) and retinal (Drishti) scans. In the next one year, we plan to move up the value chain from screening to diagnosis.
As for Shonit, it is a complete peripheral blood smear analyser solution which automates routine tasks like differential counts. It also provides a screening solution for various parasitic infections like malaria and disorders like anaemia. It starts by capturing images of blood smear slides with a phone fitted on a microscope. The images are analysed on cloud and finally, it generates reports containing differential blood counts, visualisations on various blood metrics and suggestions about any abnormalities.

YOU'VE BEEN USING YOUR PROGRAMS IN HOSPITALS AND LABS. PLEASE THROW MORE LIGHT ON THIS.

We have completed three pilot runs with three big diagnostic chains where our solution was used for their daily workload and for a period of 45 to 60 days. Our solution not only provided results which correlated well with the existing state-of-the-art devices, but in certain cases it provided better results than the devices available in the market.

ACCORDING TO YOU, HOW CAN AI CHANGE THE FACE OF INDIAN HEALTHCARE?

AI will change the way healthcare delivery is done not only in India but even abroad. It will make healthcare delivery more accessible, affordable and with improved quality.

 
Med-TechVivek desaiAIComment