The Big Data Pathologist Looking at healthcare, Pandey came across serious findings, based on data, that he knew could cause significant disruption using AI and ML.

Bengaluru-based Rohit Kumar Pandey is a hardcore programming and artificial intelligence (AI) guy. Moreover, having risen from the ranks at the financial services multinational company, American Express, as director for its big data labs, choosing big data for his start-up was a no brainer. But instead of testing the waters in e-commerce and finance, where the role of big data and AI has already gained some ground, the potential of disruption in healthcare got him more excited to launch SigTuple – an AI and machine learning (ML) backed product start-up for medical diagnosis.

Pandey, after quitting American Express in early 2015 had even tried to sell artistic and creative products through ImagineValley Online Solutions that he had co-founded. Around the same time, he along with two of his colleagues at American Express, Tathagato Rai Dastidar and Apurv Anand, launched SigTuple. ImagineValley, however, for multiple reasons, including cash crunch, was pulled down in a year's time. The trio for the first five months worked out of Anand's home until they raised seed round of $740k in October 2015. Looking at healthcare, Pandey came across serious findings, based on data, that he knew could cause significant disruption using AI and ML. One, data in verticals like radiology and ophthalmology were already digitized, for e.g., X-rays, MRIs, and retinal scans. But the pathology sector was yet to be touched.

"You have to first digitize the images before looking at applying AI to it," says Pandey. Secondly, under pathology sector, the blood and urine tests are among the most voluminous screening tests done in India. Hence, the company focused on the two apart from semen analysis as well. "Semen analysis is not voluminous but because of the increasing fertility problem and as there is a dearth of specialists to analyse semen samples, is why we chose it as our third area of focus," says Pandey.

However, the results of such tests often have variability as its analysis happens to be a function of not only medical professionals' expertise but also of their state of mind. A minor distraction can cause a major impact on the quality of the report. And unlike other sectors, healthcare, to adopt new technology, needs new hardware and devices which means cost. "That's the reason, several labs are averse to adopt these solutions, despite major advancement in the sector," maintains Pandey.

Replacing Pathologists?

Since variability exists in multiple tests and it would take years for finding solution to every test, SigTuple came up with a platform centric approach and developed AI platform Manthana to support automated multiple tests analysis including peripheral blood smear, urine and semen sample, retinal scans and chest X-rays.

"This is how we are powering tele-pathology, a process which doesn't require a doctor or a medical expert to be present near the patient, even the sample or the patient does not need to travel," adds Pandey. The solution for blood smear analysis called Shonit will be available for user adoption soon. It is an amalgamation of cloud based AI models powered by Manthana and a low-cost hardware that includes a standard microscope and electromechanical components integrated with it along with a smartphone and a smartphone holder that transforms it into a digital scanner. The solution is currently under pilot phase.

However, this by no means is aiming to make pathologists obsolete, instead SigTuple will increase their efficiency in terms of time and cost. The start-up is one of its kinds focusing on these tests in India and wants to represent the country on the global map. There are very few product companies in healthcare that have emerged out of India. So for Pandey exiting the business in near future is completely out of question.

(This article was first published in the April issue of Entrepreneur Magazine. To subscribe, click here)