Deep learning system with Grad-CAM shows potential for tuberculosis screening on chest X-raysCan a simple phone app help spot tuberculosis on chest X-rays?

medRxiv Published April 3, 2026 Study authors: Agumba, J.; Erick, S.; Pembere, A.; Nyongesa, J. DOI ↗ Editorial oversight: Dr. Lars van Dijk, PhD · Surgical, Procedural & Diagnostic

AI-generated summary of the cited source, checked by automated accuracy review. How we work

Key Takeaway

Interpret this AI system for TB screening as a technical proof-of-concept awaiting clinical validation.

This technical evaluation assessed a deployable deep learning system with Gradient-weighted Class Activation Mapping (Grad-CAM) for tuberculosis screening from chest radiographs. The system was developed and tested using publicly available chest X-ray datasets containing images labeled as Normal and Tuberculosis. The intended setting is resource-constrained, high-burden environments, with offline deployment on mobile and desktop platforms.

The primary outcomes were classification performance and deployment feasibility. The model demonstrated strong classification performance on an independent test dataset, with high accuracy and AUC values indicating effective discrimination between Normal and Tuberculosis cases. However, no specific numerical results for accuracy, precision, recall, F1 score, or AUC were reported. Grad-CAM visualizations showed the model focused primarily on anatomically relevant lung regions, particularly the upper and mid-lung fields in Tuberculosis cases.

Deployment testing confirmed consistent prediction outputs and Grad-CAM visualizations across both Windows desktop and Flutter-based mobile applications. No safety or tolerability data were reported, as this was a technical system evaluation. Key limitations include the lack of reported sample size, comparator, specific performance metrics, and clinical validation. The practice relevance is speculative, suggesting potential as an AI-assisted screening tool in resource-limited settings, but this remains unproven in clinical practice.

Imagine a clinic with no radiologist on staff, trying to figure out if a patient's cough is tuberculosis. A new study tested whether an artificial intelligence system could help. The system was built to look at chest X-rays and flag ones that might show signs of tuberculosis. It also creates a simple heatmap, showing which parts of the lung it's focusing on, so a healthcare worker can understand its reasoning.

The researchers trained and tested the system using publicly available collections of X-rays labeled as either 'Normal' or 'Tuberculosis.' They report the AI showed strong performance in telling these two groups apart. Crucially, the heatmaps it generated tended to highlight the upper and middle parts of the lungs—areas often affected by TB—suggesting it's looking at the right places.

A key goal was making this tool usable where it's needed most. The team successfully packaged the system to run offline on both Windows computers and mobile phones, with the AI giving the same results on both platforms. This means it could potentially work in remote areas with poor internet.

It's important to remember this is a test of the technology itself, not yet a test in a live clinic. We don't know the exact accuracy numbers from this study, and the system was only tested on existing image datasets. The next, crucial step is to see how well it assists real healthcare workers with real patients in those resource-constrained settings it was designed for.

What this means for you:

An AI tool for spotting TB on X-rays works on phones, but real-world testing is still needed.

Study Details

EvidenceLevel 5

PublishedApr 2026

View Original Abstract ↓

Abstract Objectives: To develop and evaluate a deployable deep learning system with Gradient-weighted Class Activation Mapping (Grad-CAM) for tuberculosis screening from chest radiographs and to assess its classification performance and explainability across desktop and mobile deployment platforms. Materials and methods: This study used publicly available chest X-ray datasets containing Normal and Tuberculosis images. A DenseNet121-based transfer learning model was trained using stratified training, validation, and test splits with data augmentation and class weighting. Model performance was evaluated using accuracy, precision, recall, F1 score, receiver operating characteristic (ROC) curve, and area under the ROC curve (AUC). Grad-CAM was used to visualize regions influencing model predictions. The trained model was converted to TensorFlow Lite and deployed in both a Windows desktop application and a Flutter-based mobile application for offline inference and visualization. Results: The model demonstrated strong classification performance on the independent test dataset, with high accuracy and AUC values indicating effective discrimination between Normal and Tuberculosis cases. Grad-CAM visualizations showed that the model focused primarily on anatomically relevant lung regions, particularly the upper and mid-lung fields in Tuberculosis cases. Deployment testing confirmed consistent prediction outputs and Grad-CAM visualizations across both Windows and mobile platforms. Conclusion: The proposed deployable deep learning system with Grad-CAM provides accurate and interpretable tuberculosis screening from chest radiographs and demonstrates feasibility for offline mobile and desktop deployment. This approach has potential as an artificial intelligence-assisted screening and decision support tool in radiology, particularly in resource-limited and remote healthcare settings.

Deep learning system with Grad-CAM shows potential for tuberculosis screening on chest X-raysCan a simple phone app help spot tuberculosis on chest X-rays?

Study Details

Meta-analysis reveals CMR and biomarkers key in ICI myocarditis

Heart MRI and Blood Tests Help Detect Immunotherapy Heart Inflammation

Clinical research that matters. Delivered to your inbox.

Deep learning system with Grad-CAM shows potential for tuberculosis screening on chest X-raysCan a simple phone app help spot tuberculosis on chest X-rays?

More on Tuberculosis

Study Details

Meta-analysis reveals CMR and biomarkers key in ICI myocarditis

Heart MRI and Blood Tests Help Detect Immunotherapy Heart Inflammation

Clinical research that matters. Delivered to your inbox.

Related in Radiology & Imaging

From Other Specialties