New artificial intelligence resource center for TB diagnosis

CAD uses artificial intelligence to automatically read chest X-rays for TB-related abnormalities. These products are especially beneficial in low-resource countries and regions where there is a high burden of TB and limited access to radiologists. CAD products can be applied to TB screening and triaging interventions to quickly and efficiently screen large populations to identify individuals in need of further testing. Prevalence surveys have demonstrated that chest X-rays can identify 40−79% of people affected by TB who do not have any classical TB symptoms and would benefit from follow-on testing.

Numerous CAD products are now available in various stages of development. The Stop TB Partnership and FIND undertook a landscaping analysis, gathering intervention-relevant information on different CAD products for TB from companies known to have developed or be developing CAD products. Features of these products, made available on www.ai4hlth.org, include:

• certification
• product development and training
• deployment mechanism: online, offline
• input: chest X-ray format, image types, machine compatibility, validation requirements
• output format
• data sharing and privacy: server location, anonymization, data sharing
• operational characteristics: hardware, software, server, system integration, processing time
• price
• existing market distribution

“Use of CAD with digital chest X-ray technology is already improving access to TB diagnosis, and the speed and scalability of it are impressive. These tools allow us to rethink how TB screening programs can be managed and implemented, to save more lives,” said Dr. Catharina Boehme, CEO of FIND.

“The Stop TB Partnership has been long supporting the roll-out of innovative solutions to diagnose and treat TB and support countries and programs move away from the usual old TB tools. The CAD used with chest X-ray will increase the reach, access, and efficiency in providing critical services to all those in need, especially now when due to the COVID-19 situation we will all have to move to more virtual care," said Dr. Lucica Ditiu, Executive Director, Stop TB Partnership.

www.ai4hlth.org provides the TB community with an overview of these relatively new digital tools and provides a platform on which implementers can compare products and select those most suitable for their needs. This resource will aid in country-level decision-making for implementers of TB programs.

As the field of artificial intelligence is constantly evolving, the website will be updated regularly to reflect product updates and the entry of new products.

Source: Stop TB Partnership