Revolutionizing Cancer Treatment in Nigeria Using Machine Learning and Deep Learning Algorithms
Ikenna David Ofordum, Chizoba Geraldine Okabuonye, Jerry Kaka Okoh, Okechukwu Paul-Chima Ugwu, Ndinyelum Onyebuchi Miracle, Ndifor Kelly Bojor · International Journal of Research Publication and Reviews · 2025
This study covered the development of machine learning (ML) and deep learning (DL)-based model designed to revolutionize cancer care in Nigeria by addressing core challenges.The model integrates patient-specific clinical data, histopathology images, and multi-omics data (genomics, proteomics, and metabolomics) to achieve three key objectives: early cancer detection, personalized treatment planning, and outcome prediction.Results from the study evaluated the performance of various artificial intelligence (AI) models applied to cancer detection, therapy response prediction, survival forecasting and real-world diagnostic applications.A Random Forest model demonstrated exceptional performance on a clinical dataset with an accuracy of 92%, precision of 91%, recall of 90%, and an F1-score of 90.5%, highlighting its balanced ability to distinguish between cancerous and non-cancerous cases.Similarly, a Convolutional Neural Network (CNN) trained on diagnostic imaging data achieved a validation accuracy of 94.2% and an area under the curve (AUC) of 0.96, showcasing its reliability in cancer diagnostics.A Multi-Omics Integration Model employing a Dense Neural Network (DNN) achieved 89.5% accuracy, 88% sensitivity, and 91% specificity, effectively predicting therapy responses using genomic and proteomic datasets.The Survival Prediction Model, utilizing a Survival Regression framework, achieved a concordance index (C-index) of 82% with a low error rate of 12%, underscoring its robustness in forecasting treatment outcomes.In a real-world pilot study at a Nigerian hospital, an AI-based diagnostic framework significantly outperformed traditional methods, reducing diagnosis time from 2 weeks to 2 days and improving diagnostic accuracy from 78% to 94%.Feedback from oncologists reflected an 88% satisfaction rate, showing the clinical viability of AI systems in resource-constrained settings.These findings underscore the transformative potential of AI in enhancing cancer care, particularly in developing countries.