NVIDIA AI India Boost: Diagnosing Tuberculosis by X-ray

India has the highest prevalence of tuberculosis in the world, with 10 million new cases of tuberculosis worldwide each year, accounting for more than a quarter of india’s total. The Un’s goal is to eradicate tuberculosis by 2030, while India’s target is 2025. Chest X-ray imaging is the most sensitive screening tool for tuberculosis, helping clinicians determine which patients should be referred for further laboratory tests, but two-thirds of the world’s population does not even have access to basic radiation therapy, in part because of the high cost of treatment and inadequate infrastructure.

DeepTek, an Indian healthcare start-up that is developing products to diagnose tuberculosis, is also a member of the NVIDIA Inception Entrepreneurship Acceleration Program, which provides basic tools, expertise, and marketing support from NVIDIA.

DeepTek hopes to significantly enhance global capabilities in the field of medical imaging diagnostics through an AI-driven radiation platform by developing a new DxTB tool to screen for X-ray images of tuberculosis and mark cases for priority examination by medical experts.

NVIDIA AI India Boost: Diagnosing Tuberculosis by X-ray

DeepTek AI model detects calcification of the right upper leaf lung

AI can play an important role because radiologists typically spend more time on CT or MRI scans, and there is a shortage of imaging experts who can read X-ray scans.

They also use NVIDIA GPUs and train and reason through their deep learning algorithms through Google Cloud and Amazon Web Services, which handle nearly 25,000 imaging studies per month on cloud platforms.

To further improve its deep learning network, it also uses data enhancement and data normalization capabilities and uses the user’s radiology reports as feedback to refine and retrain AI.

NVIDIA AI India Boost: Diagnosing Tuberculosis by X-ray

The DeepTek DxTB tool has worked with the Chennai municipality of India to analyze more than 70,000 chest X-rays and is supported by the Clinton Health Initiative.

The tool can also be deployed in mobile vans equipped with digital X-ray machines, transformed into mobile clinics, facilitates tuberculosis screening for high-risk populations, and scans of chest X-ray images are then safely transmitted to the cloud for reasoning, speeding up case classification and immediate additional examinations.

This work used to take a month, but now with AI, it can be done in a few hours and can greatly reduce the chance of missing a job.

In addition, about 50 imaging centers and hospitals in India use the DeepTek AI model, and a network of hospitals will soon deploy the DeepTek ICU Chest tool to diagnose a range of conditions related to intensive care patients.