An Original Framework for Integrating Artificial Intelligence in Radiographic Image Interpretation for Resource-Limited Nigerian Radiology Departments

Artificial intelligence tools for radiographic image interpretation offer substantial potential for addressing the radiologist shortage in Nigeria, yet their deployment in resource-limited departments is complicated by infrastructure constraints, training data representativeness, workflow integration challenges, and governance gaps that require a theoretically grounded integration framework specific to the Nigerian context. This study develops an original framework for integrating AI image interpretation tools in resource-limited Nigerian radiology departments. A five-component research design generates the framework: systematic review of AI radiographic interpretation performance literature from sub-Saharan African populations; technical AI readiness assessment across 20 Nigerian radiology departments documenting infrastructure, connectivity, and EHR compatibility; AI implementation stakeholder analysis using structured interviews with 60 radiographers, radiologists, hospital ICT managers, and health regulators; algorithm performance validation testing using a 5,000-image Nigerian chest X-ray dataset for pneumonia and TB detection against radiologist reference readings; and iterative expert panel framework refinement through three rounds. AI performance validation achieves an AUC of 0.91 for pulmonary TB detection, demonstrating clinical potential for Nigerian applications. The original Integration Framework specifies five domains: technical readiness criteria, clinical governance standards, radiographer role evolution for AI-assisted practice, patient communication requirements, and regulatory compliance pathways. Available AI radiology literature from sub-Saharan Africa identifies training data under-representation of African pathology patterns as the primary AI performance risk. The Responsible AI in Healthcare Framework and the WHO AI in Health Regulatory Guidelines provide the normative reference. Keywords: artificial intelligence, radiographic interpretation, resource-limited, Nigeria, integration framework.