Dr Tathagato Rai Dastidar grew up in an surroundings the place science reigned supreme, he recollects. His mother and father have been professors on the Indian Affiliation for the Cultivation of Science, which is the oldest analysis institute in India.
“So, science has been my bread and butter since early childhood. I’ve all the time been fascinated and interested in science and know-how,” says Dr Dastidar (45) in dialog with The Higher India.
The 22-year veteran of the tech business, who holds 11 United States patents, is the founding father of Bengaluru medtech startup SigTuple, which automates the guide microscopic evaluation of organic samples utilizing synthetic intelligence (AI) for quicker turnaround time for sufferers, making the lives of pathologists simpler.
Dr Dastidar’s love for science and know-how took him to the corridors of the famed Indian Institute of Know-how (IIT), Kharagpur. There, he accomplished his BTech and PhD in pc science and engineering earlier than embarking on a profession constructing software program and holding technical management positions for a few of the premier know-how corporations on the earth.
Nevertheless, when he began his skilled profession again in 2000 at Nationwide Semiconductor, an American semiconductor producer, tragedy struck again dwelling.
Dr Dastidar’s father handed away the identical 12 months after his remedy for acute myeloid leukaemia was delayed as a consequence of an earlier misdiagnosis. Alongside his day job at varied tech corporations, for a couple of decade and a half, Dr Dastidar started engaged on the issue of learn how to stop such deadly situations of medical misdiagnosis. He based SigTuple in 2015.
SigTuple was, partly, the results of this seek for an answer, and with the corporate, he has launched into a mission to revolutionise pathology via robotics and superior AI.
“At SigTuple, we’re about making use of synthetic intelligence into an space which positively impacts human lives. Correct and early prognosis of illness situations is one such approach, and this trigger resonated with me personally,” he says.
Mitigating the lack of human lives
Taking a multi-disciplinary strategy, SigTuple combines AI, robotics, microfluidics and cloud computing to create clever diagnostic options that make high quality healthcare supply reasonably priced and accessible.
Right here’s how Dr Dastidar explains the issue assertion.
“In lots of circumstances, essential illnesses go undiagnosed until it’s too late. This might be due to the shortage of entry to high quality healthcare. It is also as a result of inefficiencies within the healthcare supply system. Regardless of the trigger is perhaps, the impact is finally tragic — pointless struggling and lack of human lives. The concept behind SigTuple was to construct synthetic intelligence that might assist early screening and prognosis, probably in underserved areas, which might make life higher for the top affected person.”
Extra particularly, nonetheless, he notes that there’s one pathologist for a inhabitants of greater than 65,000 folks. With the goal of constructing the lifetime of the pathologist simpler and thereby enhancing affected person outcomes, their focus lies in automating the inefficient and error susceptible guide microscopic evaluation course of for the commonest exams — blood and urine microscopy.
Navigating the ‘largest bottleneck’
One key machine they’ve developed on this regard is the AI100, which presents an answer combining software program and {hardware} parts. The {hardware} is an automatic slide scanner, able to changing any bodily specimen into digital pictures.
The AI platform analyses these pictures and supplies medical insights into the pattern, which brings down slide evaluation time for the pathologist and permits them to seamlessly work remotely or collaborate with their colleagues throughout geographies.
“Microscopic evaluation of blood samples is a gold normal for prognosis of many essential illnesses. This consists of one thing as severe as the assorted types of leukaemia. This course of stays predominantly guide in additional than 96% laboratories globally and is thus the largest bottleneck by way of turnaround time for pattern reporting,” explains Dr Dastidar.
“Being guide, it is usually susceptible to errors. Most significantly, blood microscopy is by far the most important use case of the microscope in a medical laboratory. The concept behind AI100 is to automate the microscopic evaluation of blood samples via robotics and AI, to make the method standardised, cut back turnaround time, and enhance high quality of reporting,” he provides.
Whereas the thought of AI100 originated on the time of the delivery of SigTuple, Dr Dastidar and his crew tried out totally different approaches on the {hardware} facet earlier than zeroing on the present design. He says his crew knew that they needed to construct the {hardware} themselves, as there was no cost-effective answer obtainable for automated digital microscopy out there.
“Our largest problem was the founding crew’s lack of expertise with {hardware} design, particularly {hardware} that requires micron degree precision, as in microscopy. We had huge expertise in creating synthetic intelligence and knowledge evaluation instruments, however hardly any to develop {hardware}. Consequently, we made many rookie errors alongside the way in which, which delayed the ultimate deployment of AI100 into the market. These have been good classes to study,” he notes.
AI100 was prepared for beta testing beginning early 2019. SigTuple deployed the beta model into many labs, and used the chance to collect buyer suggestions and smoothen out any tough edges.
COVID-19, nonetheless, dampened their commercialisation plans. They began business deployments of AI100 after the second wave of the pandemic subsided in late 2021.
How does it work?
In regular laboratory observe, a blood pattern is smeared on a glass slide after which reviewed beneath the microscope.
With this machine, as an alternative of a pathologist reviewing the slide beneath the microscope, the slide is inserted into AI100. The machine routinely captures excessive magnification microscopic pictures of the related parts of the slide and uploads the photographs to a cloud platform.
“The AI module resides on the cloud platform. It analyses the photographs, mimicking what a pathologist would do whereas reviewing it beneath the microscope. The results of this evaluation is then offered to the pathologist on an internet browser, which makes the report accessible from wherever and anytime. The pathologist evaluations the AI prompt report after which approves it.
This course of helps the present workflow in some ways,” says Dr Dastidar.
To start with, the pathologist needn’t be within the lab to report on the pattern.
“Think about a essential affected person requiring quick consideration within the useless of the night time. Earlier, the pathologist needed to rush to the hospital or the affected person needed to wait. Now, that’s now not vital, because the pathologist can report from dwelling, whereas the pattern is within the laboratory the place the AI100 machine is current. Additional, because the AI does many of the job of the pathologist, it makes her extra environment friendly, by lowering the evaluation time. She will now focus extra on the actually essential sufferers, relatively than losing time on close to regular samples,” he explains.
Lastly, the AI is free from fatigue-related errors that impression human pathologists. It’s higher at detecting these rarer and straightforward to overlook abnormalities. This finally improves the top affected person consequence. With AI-enabled automated digital microscopy, the pattern evaluation course of by a pathologist, which used to take 5 to 10 minutes, has come right down to 30 seconds at the moment. Additionally, due to this know-how, a pathologist can now evaluation roughly 300 slides per day as in comparison with 30 slides with guide microscopic options, in line with Dr Dastidar.
“That is how AI, robotics and cloud all come collectively to enhance the pace and accuracy of prognosis. Additionally, aside from blood, AI100 may analyse urine samples. Urine microscopy is the second commonest use of the microscope within the medical laboratory. Quickly, different and extra specialised exams will probably be enabled on the identical platform,” he provides.
Parts of AI100
AI100 consists of an optical column with a microscopic lens (400X magnification), an LED illumination unit and a CMOS (complementary steel oxide semiconductor) digicam.
“It has a mechanical stage used for transferring the pattern beneath the lens, and in addition to focus it, which operates at micron degree precision. There’s an digital printed circuit board (PCB) to regulate the mechanical stage and the LED unit. It additionally accommodates a full-fledged compute unit consisting of an Intel processor and NVIDIA GPU,” states Dr Dastidar.
“SigTuple has the aptitude to assemble the {hardware} in home. We even have a tie-up with an authorized contract producer for medical gadgets. Whereas most of the inside parts are imported, some are regionally sourced. The plastic cowl and the machined parts are manufactured regionally,” Pranat Bhadani, President of Enterprise Development, SigTuple.
Bettering care in Tier 2 and three cities
SigTuple can be centered on equipping Tier 2 and three cities by putting in their machine in pathology labs and introducing the necessity for superior automated digital microscopy, which can assist deal with the dearth of pathologists.
“Most laboratory chains are stunted of their potential to open up high quality diagnostic centres in Tier 2 and three cities (and even past), by the lack to discover a good pathologist for such labs. So, they usually function in a hub-and-spoke mannequin, the place samples from smaller cities and cities are transported to their major centres in bigger cities. This has its apparent drawbacks — elevated turnaround time, pattern degeneration whereas transport, and so forth,” explains Bhadani.
Microscopy requires the bodily presence of the pathologist within the lab. Most different capabilities within the lab will be carried out digitally, with technicians on the lab operating the machines.
“AI100 permits the labs to carry out microscopy digitally over the web, obviating the necessity to have the pathologist bodily current within the lab. Given the elevated effectivity of the pathologist as a result of AI, the identical pathologist can now have a look at many extra samples in a day than what she may do earlier. That is how SigTuple helps laboratories of their foray into smaller cities and cities. We rely almost all the highest laboratory and hospital chains in India as our prospects — HCG, Manipal, Krsnaa Diagnostics, SRL, Aarthi Scans, and so on,” says Bhadani.
What’s the street forward for SigTuple?
“We’ll proceed our journey of innovation within the medical area and give you newer options which resolve actual medical issues. We will even proceed our journey to increase commercially, each in India and internationally,” says Dr Dastidar.
(Edited by Divya Sethu)