Computer-Aided Reading of Chest Radiographs for Paediatric TB: Current Status and Future Directions

26 August 2025

Computer-aided detection (CAD) systems for automated reading of chest x-rays (CXRs) have been developed and approved for TB triage in adults but not in children. However, CXR is frequently the only adjunctive tool for clinical assessment in the evaluation of paediatric TB in primary care settings, and children would benefit from CAD models that can detect their unique clinical and radiographic features. To advance CAD for childhood TB, large, diverse paediatric CXR datasets linked to standardised TB classifications are required. These datasets would be used to train and validate paediatric-specific models for TB screening, diagnosis, and severity stratification. Previous studies on CAD algorithms for reading paediatric CXRs have highlighted promising approaches, including the use of transfer learning with existing deep learning models. Including data from children in CAD models is essential to improve equity and reduce the global burden of TB disease.

A viewpoint, published in The Lancet Digital Health, outlines the key characteristics of paediatric CXR repositories for TB, and highlights relevant previous studies and analytical considerations to guide the development of novel CAD models for paediatric TB.