Artificial intelligence and machine learning for early detection and diagnosis of colorectal cancer in sub-Saharan Africa

Abstract

With the growing resources and investments in AI/ML-based tools in SSA, one could envision a CRC surveillance and diagnosis pipeline that employs MAAA for population-based surveillance and pattern recognition and computer vision algorithms to guide diagnostic recommendations and prognosis. These tools will need to be tailored to local needs based on available resources and testing approaches (eg, sequential testing with MAAA and then FIT) and key stakeholders will need to engage in the codesign of widespread implementation strategies (eg, community-based screening programmes, practitioner education, health policies). Future studies are required to compare the efficacy of these tools to existing CRC surveillance and diagnosis tools (eg, FIT) in SSA populations. Furthermore, these innovative solutions provide opportunities for the adaption and adoption of these approaches in high-income countries. While CRC was used as the use case, these tools could be expanded to other prevalent and emergent cancers (eg, liver, breast and cervical) or other non-communicable diseases that would benefit from lab-based MAAA and computer vision AI-based methods for automated objective assessment of disease diagnosis and prognosis.