Hybrid Explainable Deep Learning Framework for Tuberculosis Detection Using Chest X-Ray Images and Clinical Data Fusion

Abstract

Tuberculosis (TB) remains one of the leading infectious diseases worldwide, particularly in low- and middle-income countries where rapid diagnosis is still challenging due to limited healthcare infrastructure and shortage of expert radiologists. This paper proposes a hybrid explainable deep learning framework for automated tuberculosis detection using chest X-ray images integrated with clinical information and secure healthcare data management principles. The proposed framework combines a customized Convolutional Neural Network (CNN), EfficientNetB0 feature extraction, and an Extreme Learning Machine (ELM) classifier to improve diagnostic accuracy and computational efficiency. In addition, lung segmentation using UNet++ is incorporated to enhance region-of-interest extraction and reduce irrelevant image features. Explainable Artificial Intelligence (XAI) techniques including Grad-CAM, SHAP, and LIME are integrated to improve interpretability and clinical trustworthiness. Experimental evaluation was conducted using publicly available tuberculosis chest X-ray datasets including Shenzhen, Montgomery, and TB Chest X-ray datasets. The proposed model achieved an accuracy of 99.21%, precision of 98.74%, recall of 99.03%, and F1-score of 98.88%, outperforming existing deep learning approaches such as DenseNet121, ResNet50, VGG16, and standalone EfficientNetB0 models. Comparative analysis demonstrates that the integration of segmentation, hybrid classification, and clinical data fusion significantly improves TB detection performance while reducing false positives. The proposed system can support early diagnosis, improve clinical decision-making, and assist healthcare professionals in resource-constrained environments.