Pulmonary tuberculosis (TB) remains one of the most critical global health challenges, affecting millions of individuals annually and imposing substantial morbidity and mortality. Chest X-ray (CXR) examination is an essential tool for TB screening, triage, and diagnosis. Accurate disease diagnosis and identification of TB lesions on CXRs, such as consolidation and cavity, are crucial for optimizing treatment, monitoring patient outcomes, and allocating resources efficiently, particularly in resource-constrained settings . Radiological features, such as cavity presence and lung area involvement, are established indicators of TB severity(Nachiappan et al., 2017). However, manual interpretation of chest X-rays is time-intensive and dependent on radiologist expertise, which may not always be available. This challenge addresses these gaps by introducing two innovative tasks. Task 1 focuses on cavity presence detection and segmentation, using advanced deep learning techniques to identify and quantify cavities in CXRs. Task 2 expands the scope to a multimodal approach for TB diagnosis, combining radiological data with clinical information to develop robust, explainable, and generalizable AI models. From a biomedical perspective, the challenge aims to enhance diagnostic accuracy and improve disease monitoring by providing consistent and reproducible assessments. Technically, it leverages state-of-the-art machine learning models to develop automated, explainable, and scalable tools suitable for diverse healthcare environments. The anticipated impact includes facilitating personalized treatment planning, reducing dependency on radiological expertise, and addressing disparities in TB care globally.