PROJECT

Development and Validation of Artificial Intelligence Tool for Screening/detection of Pulmonary TB and other lung diseases using Chest X-rays

June 2022 - June 2027

Objectives:

• To develop a computer-assisted screening system to differentiate clinically normal chest X-rays from clinically abnormal types.
• To develop a computer-aided detection system that enables auto-differentiation of TB from other chest diseases/ other lung diseases using X-rays
• To further develop the computer-aided detection system for auto-identification of various presentations of pulmonary tuberculosis.

Study Overview:

The work related to training and development of the AI tool will be primarily done at the Institute for Plasma Research, and the data sets would be provided by participating Institutes. ICMR, NIRT and IPR already have an MOU for development of the AI tool for TB. This project is being initiated by ICMR, headquarters as the central coordinating institute, and other ICMR Institutes/participating Institutes would provide the data. The X-ray dataset already available with Institutes would be used for Phase 1 of development of the tool. Prospective images collected through all sites will be used for Phase 2 of Validation/testing of the tool.

Study Approach:

• Phase 1: Development of tool: (learning and training)
• Phase 2: (Validation/testing of tool)

Expected Public Health Impact:

The aim is to develop a robust AI tool with high sensitivity and specificity (as per WHO reference) for detection/screening of TB using chest X-rays and other lung diseases. The final version of the tool will greatly facilitate bridging the gap between diagnosis and treatment, more so in economically or geographically difficult populations, and hence significantly cater to the National Goal of Eliminating Tuberculosis by 2025 as well as the END TB Strategy of the Sustainable Development Goals