What AI Can Actually Help With in Clinical Practice - and What It Still Cannot
Authors: SM Ahsanuzzaman
Institution: CUET
Introduction
Artificial Intelligence (AI) is rapidly transforming healthcare, offering new opportunities to improve clinical efficiency, documentation, decision-making, and patient outcomes. However, its practical application in routine clinical practice, particularly in resource-constrained settings like Bangladesh, requires careful consideration of clinical safety, data privacy, reliability, and local validation.
Methods
This presentation explores the current capabilities and limitations of AI in everyday medical practice, with particular emphasis on breast cancer and multidisciplinary oncology care. It examines three major applications of AI – perception, language, and prediction – highlighting emerging clinical evidence and practical uses of generative AI in patient communication, clinical documentation, literature review, and research. The presentation also introduces MediStack, a Bangladesh-developed digital healthcare ecosystem integrating AI into routine clinical workflows. Its capabilities include Cortex AI for conversation-to-clinical documentation, AI-powered clinical report analysis, synthesis of longitudinal patient information, and multidisciplinary Tumour Board support. These capabilities connect with oncology protocol management, chemotherapy day care workflows, and structured cancer registry development to support continuity of care from consultation and treatment planning through follow-up.
Results
Beyond improving clinical efficiency, structured digital records offer opportunities to generate locally relevant evidence on cancer presentation, treatment patterns, and outcomes in Bangladesh. Importantly, the presentation emphasises that AI should augment rather than replace clinical judgement. Responsible implementation requires clinician oversight, patient data protection, transparent limitations, and validation using local clinical data.
Conclusion
Ultimately, the goal is to demonstrate how thoughtfully integrated AI can reduce administrative burden, strengthen multidisciplinary collaboration, and support safer, more coordinated, evidence-driven cancer care in Bangladesh.