From conventional learning to emergence of large language and vision-language models (LLM/VLM) and Agentic AI models..., this decade shows AI transforming in biomedical imaging.

Adaptive AI-based frameworks has potentials for end-to-end workflow automation for clinicians to perform multiple tasks involoved in medical theranostics — diagnosis, therapy, management, and outcome evaluation. With advances in MRI, CT/PET, and Ultrasound imaging technologies, clinicians’ experience improved anatomy and pathology visualization abilities with high spatial and temporal resolution. However, real-time interpretation of data from diverse sources could be complex and cognitively demanding while new AI-enable paradigms could facilitate visual scene parsing, heterogeneous source integration, human-computer interaction, and multi-task procedure execution to support decision making.