Dr. Tathagato Rai Dastidar elucidates the driving force that ignited the creation of his medical technology startup, SigTuple. The company has ingeniously developed an AI-powered apparatus aimed at a paradigm shift in the realm of pathology, striving to enable the early and accurate diagnosis of diseases while significantly mitigating the occurrence of human errors.
How It Operates?
In traditional laboratory procedures, a blood sample is applied to a glass slide and subsequently examined under a microscope.
However, the AI100 device offers a revolutionary approach. Instead of a pathologist manually inspecting the slide under the microscope, the slide is inserted into the AI100. This innovative machine automatically captures highly magnified microscopic images of the pertinent sections of the slide, which are then transmitted to a cloud-based platform.
“The AI module resides within the cloud platform. It meticulously analyzes these images, emulating the actions of a pathologist conducting a microscopic review. The outcome of this analysis is then presented to the pathologist via a web browser interface, making the report accessible from any location and at any time. The pathologist evaluates the AI-generated report and subsequently endorses it,” explains Dr. Dastidar.
This method brings numerous advantages to the current workflow. Firstly, the pathologist no longer needs to be physically present in the lab to analyze the sample.
“Imagine a critical patient requiring urgent attention in the middle of the night. Previously, the pathologist had to rush to the hospital or the patient’s diagnosis would be delayed. Now, this is no longer necessary. The pathologist can provide a report remotely, while the sample is being processed by the AI100 machine in the laboratory. Moreover, since the AI largely handles the tasks of the pathologist, it enhances efficiency by reducing the review time. The pathologist can now focus more on critically ill patients, rather than spending excessive time on routine samples,” elaborates Dr. Dastidar.
Furthermore, AI avoids errors associated with fatigue, which can affect human pathologists. It excels in identifying rare and easily overlooked abnormalities, ultimately leading to improved patient outcomes. Thanks to AI-powered automated digital microscopy, the sample review process, which used to take 5 to 10 minutes, is now completed within just 30 seconds. Additionally, this technology has dramatically increased a pathologist’s daily workload, enabling the review of around 300 slides per day compared to the mere 30 slides possible with manual microscopic methods, as per Dr. Dastidar.
“This amalgamation of AI, robotics, and cloud technology enhances the speed and accuracy of diagnostics. In addition to blood samples, AI100 can also analyze urine samples. Urine microscopy constitutes the second most frequent application of microscopes in clinical laboratories. Soon, the platform will support other specialized tests,” adds Dr. Dastidar.
Key Components of AI100 The AI100 device comprises an optical column housing a microscopic lens (with 400X magnification), an LED illumination unit, and a CMOS (complementary metal oxide semiconductor) camera.
“It features a mechanical stage responsible for positioning and focusing the sample under the lens. This stage operates with micron-level precision. An electronic printed circuit board (PCB) controls both the mechanical stage and LED unit. Additionally, it includes a comprehensive computing unit containing an Intel processor and NVIDIA GPU,” outlines Dr. Dastidar.
“SigTuple possesses the capability to internally assemble the hardware. We’ve also established a partnership with a certified contract manufacturer specializing in medical devices. While several internal components are imported, others are sourced locally. The plastic cover and machined components are produced within the local region,” states Pranat Bhadani, President of Business Growth at SigTuple.
