Computer-Aided Detection (CAD)

Uses

Benefits

Comparisons

Background

Evidence quality

Highlighted claims

• The WHO endorses CAD as a triage tool for TB only in persons aged 15 years and older, and at least 16 commercial adult CAD products are currently on the market. — Catalysing Artificial Intelligence for Paediatric Tuberculosis Research (CAPTURE): protocol for a global multicentre study establishing a paediatric chest X-ray repository to evaluate computer-aided detection algorithms
• No CAD products exist for children under 2 years, and vendor claims that products can be used in children lack independent evaluation. — Catalysing Artificial Intelligence for Paediatric Tuberculosis Research (CAPTURE): protocol for a global multicentre study establishing a paediatric chest X-ray repository to evaluate computer-aided detection algorithms
• Adult-trained CAD algorithms perform suboptimally in children because paediatric chest anatomy and TB radiological patterns differ substantially from adult patterns. — Catalysing Artificial Intelligence for Paediatric Tuberculosis Research (CAPTURE): protocol for a global multicentre study establishing a paediatric chest X-ray repository to evaluate computer-aided detection algorithms
• In the CAPTURE evaluation, each CAD product is installed on an isolated server and developer access is removed to ensure fully manufacturer-independent evaluation. — Catalysing Artificial Intelligence for Paediatric Tuberculosis Research (CAPTURE): protocol for a global multicentre study establishing a paediatric chest X-ray repository to evaluate computer-aided detection algorithms
• CAD software powered by deep learning has achieved performance comparable to human readers for adults with microbiologically confirmed TB. — Catalysing Artificial Intelligence for Paediatric Tuberculosis Research (CAPTURE): protocol for a global multicentre study establishing a paediatric chest X-ray repository to evaluate computer-aided detection algorithms