Artificial Intelligence in Resource-Limited Healthcare Settings: Applications, Implementation Challenges, and Future Directions in Cardiology and Oncology

Abstract

 

Artificial intelligence (AI) holds transformative potential for health systems in low- and middle-income countries (LMICs), where clinician shortages, infrastructure deficits, and rising chronic disease burdens converge. This review synthesises 2019–2026 evidence on AI deployment in cardiology and oncology within resource-limited settings, examining predictive algorithms, digital twins, point-of-care diagnostics, and human-in-the-loop governance. We analyse applications ranging from ECG-based arrhythmia detection and heart-failure risk stratification to cancer screening, precision treatment selection, radiotherapy planning, and immune-checkpoint inhibitor response prediction. Despite promising performance— including HIV testing prediction models in Sierra Leone, Nigerian cardiologist-facing AI tools, and dialysis-optimising algorithms—deployment remains constrained by data scarcity, algorithmic bias, workforce gaps, regulatory ambiguity, and unsustainable energy supply. Drawing on 100 peer-reviewed sources, we propose an equity-centred implementation framework emphasising local validation, federated learning, explainable AI, sustainable energy harvesting, and additive manufacturing for medical devices. Without deliberate attention to governance, interoperability, and community engagement, AI risks exacerbating rather than ameliorating global health disparities.

 

Keywords: Artificial intelligence, resource-limited settings, low- and middle-income countries, cardiology, oncology, digital twin, precision medicine, human-in-the-loop, explainable AI, sustainable health technology