SigTuple, an AI healthcare started by techies, raises $5.8M to help better diagnosis
SigTuple is an AI startup in the healthcare space that analyses medical images, scans and videos, which generates information and data to help diagnosis.
There was one thought 39-year-old Tathagato Rai Dastidar, 34-year-old Rohit Kumar Pandey, and 35-year-old Apurv Anand had that April two years ago — the visual data in terms of images, scans, and videos was simply mind-boggling.
Tatha had over 16 years of experience in leadership roles in National Semiconductor, Yahoo! Search, American Express Big Data Labs, Gracenote, and even a previous entrepreneurial experience in VLSI domain.
Rohit, on the other hand, had over 10 years of experience in various leadership roles in American Express, and was a director at Big Data Labs, while Apurv came with over 12 years of experience in leadership roles in Veveo and American Express Big Data Labs.
Building the product
The trio thus met at American Express Big Data Labs in 2012 and was given the charter to form the team and build the big data stack from scratch. After working with different forms and kinds of data, the idea of applying artificial intelligence (AI) to medical data intrigued them.
Though challenging, they felt that the effort would have a positive social impact. Thus SigTuple was born. They started exploring the opportunity of applying AI to analyse these images and videos and spent around three months talking to various doctors and medical experts.
Explaining further Rohit says,
“We started working on the data sets, which were freely available to see if something can be developed using it. We showcased the early results with the medical experts in our network and we got some very positive feedback. That was the Eureka moment and we decided to formally register the company in July 2015.”
In October, SigTuple raised $740,000 from the likes of Sachin Bansal, Binny Bansal, Ashok Bareja, Dr Nirupa Bareja, Debanjan Mukherjee, and Accel Partners. Today SigTuple also announced that it raised a Series A funding of $5.8 million led by Accel Partners and supported by other prominent institutional investors like IDG Ventures, Endiya Partners, pi Ventures, VH Capital, and Axilor Ventures.
Sachin and Binny have re-invested in the Series A round. Other prominent investors like Amit Singhal (SVP Engineering, Uber and Ex-SVP, Google Search), Neeraj Arora (VH Capital), Kris Gopalakrishnan, and SD Shibulal (Axilor Ventures) have also joined them.
Workings of the product
SigTuple’s solutions allow labs and hospitals to scale by implementing a hub-and-spoke model where the medical experts can operate from hubs and devices can be installed in spokes. Rohit adds that it will improve the efficiency of the medical experts by automating painful and fatigue driven visual medical analysis process.
The team began work on building their core product — Manthana — a continuous learning platform that uses AI for healthcare. The platform is used to churn data to generate intelligence.
Manthana allows SigTuple to ingest visual medical data from different devices and build a longitudinal memory. It has the capability to train, validate, and execute AI- and ML-powered models to classify various objects of interest, detect diseases, and compute the metrics for reporting.
The metrics provided by the platform are supported by visual evidence which eliminates the need for the medical expert to sit next to the patient, medical device, or biological sample.
The platform also exposes an interface for the medical experts to get feedback and annotations, which enables continuous learning with new and changing visual medical data.
Currently, SigTuple is using Manthana to come up provide solutions for automated analysis of peripheral blood smear, urine and semen sample, retinal scans, and chest x-rays. “The solution for blood, Shonit™, has already undergone three clinical trials and will be soon available for user adoption followed by commercials. Other solutions are in the advanced stages of development,” says Rohit.
Working along the challenges
SigTuple is working with large hospitals and labs and a group of medical experts in various fields to develop these solutions. They have filed half a dozen patents and published in medical and computer science journals. Manthana uses a pay-per-use for every report generated.
Rohit adds that the hardware is a low-cost digital scanner digitises the physical slide into a digital slide and sends it to Manthana. The digital slide is then analysed and a detailed report is generated by the AI and ML models hosted in Manthana.
Rohit adds that these reports not only contains numbers but also visual medical evidence for each and every reported metrics. This report is made available to a medical expert on his/her handheld. The medical expert reviews the report and approves it for the patient.
But starting an AI startup in the medical space isn’t easy. The trio faced several challenges, from sourcing the data and annotations, to getting the medical knowledge required for the task, to making the general medical community comfortable with the idea of being assisted by AI.
Market and future
“We interacted with a few prominent medical industry veterans and requested them to be advisors to the company. These advisors helped us in reaching out to forward-looking labs and hospitals to forge data partnerships and also introduced us to subject matter experts in the industry who helped us with annotations and medical consultancy,” adds Rohit.
Apurv adds that the team realised early in their journey that the platform they needed to achieve their vision did not exist at all, in whole or in part, in spite of the recent proliferation of numerous AI platforms by big and small players. He explains,
“Thus Manthana was created, from grounds-up, to be a continuously improving, automatically upgrading platform which enables digitisation, management and analysis of visual medical data, trains various kinds of AI models on it, and provides an insightful, interactive report to medical specialists.”
SigTuple will use the latest round of funding to expand the team, bullet-proof the platform and the product for user adoption followed by commercials, and regulatory clearances for global markets. Speaking of why Accel decided to invest in SigTuple, Barath Shankar Subramanian, Principal, Accel Partners says,
“Sigtuple is operating in a global ecosystem — our comparison is looking at the area globally and finding the right team to back. This is a cutting edge area and the work the SigTuple team is doing needs to be benchmarked globally.”
Expected to touch $79 billion in 2012, the healthcare sector is now expected to reach $160 billion by 2017, and $280 billion by 2020. Today, it is considered one of the largest sectors in India, in terms of both revenue and employment.
Ranjith Menon, Executive Director at IDG believes that diagnostics market is a multi-billion dollar market. He adds that if you look at the number of tests done in India, it goes beyond a billion every year.
“Most of these being routine tests can be automated. AI and deep learning can play a big role in addressing this automation. The idea here is not to replace a pathologist but to help her/him complete his tasks faster and more effectively,” says Ranjith.