WHO Policy on Computer-Aided Detection Software for Tuberculosis Screening Based on FIND Research

FIND led the evidence generation for the World Health Organization (WHO) policy recommendation on the use of computer-aided detection (CAD) software for chest X-rays in tuberculosis (TB) screening which was released earlier this week. The new policy doubles the number of recommended CAD products from 4 to 8. This milestone highlights FIND’s critical role in driving TB innovation and translating new tools into policy recommendations.

FIND’s validation platform was used to thoroughly test and evaluate potential AI tools designed to interpret X-ray images for the detection of TB. Software developers submit their programs to the FIND platform, which then tests the application against FIND’s proprietary library of images. This allows the platform to determine whether the proposed AI tool is valid and accurate in diagnosing TB, particularly in low- and middle-income country settings. To date, only approximately half of the applications submitted have been validated as being accurate and reliable for the detection of TB.

The evidence generated through this FIND’s validation platform informed WHO’s updated policy guidance. It prepares the way for the WHO Prequalification Department to undertake a more comprehensive review of the CAD products for TB screening, supporting a global scale-up in access to these tools.

WHO estimated that globally in 2023, 8.2 million people were reported as being diagnosed with a new episode of tuberculosis (TB). This is the highest number of TB notifications registered for a single year since WHO began compiling data from all countries.