Expanding the Role of Medical Physics: Building Robust System Architectures to Facilitate AI in Clinical Trials
Description
As artificial intelligence becomes increasingly integral to clinical oncology and clinical trials, medical physicists are uniquely positioned to lead the development of the robust system architectures required to integrate AI safely and effectively. This session will explore how the profession is evolving from local experts in dosimetry and QA to institutional and network-level leaders in AI strategy, infrastructure, and ethics, with a particular focus on enabling next-generation clinical trials through harmonized data and automated QA pipelines. Through concrete examples and emerging frameworks, speakers will illustrate how harmonized, secure data architectures can enable AI tools to query across EMRs, treatment management systems, and clinical trial databases. At the system level, the session will highlight approaches to building architectures that support large-scale AI applications—such as RadOnc-GPT for multi-institutional data harmonization—while ensuring regulatory compliance and data security. At the patient level, it will examine how existing tools like RapidPlan and AI-based contouring can support real-time QA of clinical trial data submissions, accelerating research while maintaining rigor. The session will also spotlight future directions: the use of AI to identify eligible patients and reduce enrollment bias, chatbot interfaces for patient education, and strategic partnerships with vendors to shape the next generation of clinical AI platforms. By leading the development of these architectures, medical physicists can ensure that AI deployment in clinical trials is ethical, efficient, and impactful.