PT - JOURNAL ARTICLE AU - Saboury, Babak AU - Bradshaw, Tyler AU - Boellaard, Ronald AU - Buvat, Irene AU - Dutta, Joyita AU - Hatt, Mathieu AU - Jha, Abhinav AU - Li, Quanzheng AU - Liu, Chi AU - McMeekin, Helena AU - Morris, Michael AU - Scott, Peter AU - Siegel, Eliot AU - Sunderland, John AU - Wahl, Richard AU - Zuehlsdorff, Sven AU - Rahmim, Arman TI - <strong>Artificial Intelligence Ecosystem in Nuclear Medicine: Opportunities, Challenges, and Responsibilities</strong> DP - 2022 Aug 01 TA - Journal of Nuclear Medicine PG - 2733--2733 VI - 63 IP - supplement 2 4099 - http://jnm.snmjournals.org/content/63/supplement_2/2733.short 4100 - http://jnm.snmjournals.org/content/63/supplement_2/2733.full SO - J Nucl Med2022 Aug 01; 63 AB - 2733 Introduction: Integration and implementation of artificial intelligence (AI) applications in the practice of nuclear medicine require careful analysis of potential opportunities and critical challenges. This comprehensive evaluation is inevitable in order to enhance patient care through innovation on one hand and to address concerns of all relevant stakeholders on the other.The AI ecosystem contains the total life-cycle of the application including data acquisition, model training and prototyping, production/testing, validation/evaluation, implementation and development, and post-deployment surveillance. Attention to all these steps through the lens of trustworthiness is essential. This educational exhibit explores the elements of trust in the healthcare ecosystem in the AI era while reviewing potential opportunities and critical challenges.Methods: This presentation summarizes the discussions of the Society of Nuclear Medicine and Molecular Imaging (SNMMI) AI Task Force, which consists of various stakeholders and experts including physicists, computational imaging scientists, physicians, statisticians, and representatives from industry &amp; regulatory agencies.Results: A.Opportunities:Diagnostic Imaging Image Generation Image Analysis Emerging Nuclear Imaging Approaches Radiopharmaceutical Therapies (RPTs) AI-driven theranostic drug discovery and labeling Precision Dosimetry Predictive Dosimetry and Digital Twins Clinical workflow: Increase throughput while maintaining excellenceB.Challenges: Development of AI Applications/Medical Devices Data Optimal Network Architecture Measurement and Communication of Uncertainty Clinically Impactful Use Cases Team Science Validation (Verification of performance Performance Profiling Through Task-Based Evaluations Guidelines for Validation Multi-Center Clinical Trial Network Ethical, Regulatory, and Legal Ambiguities Implementation of Clinical AI solutions &amp; Post-Implementation Monitoring AI-Platform Barriers of Dissemination and Implementation of AI Technology in Medicine Post-implementation: change management &amp; performance monitoring Trust and Trustworthiness C.Key Elements of Trustworthy AI Ecosystems:1) Human Agency, 2) Oversight, 3) Technical Robustness, 4) Safety and Accountability, 5) Security and Data Governance, 6) Predetermined Change Control Plan, 7) Diversity, Bias-awareness, Non-discrimination, and Fairness, 8) Stakeholder Participation, 9) Transparency and Explainability, 10) Sustainability of societal wellbeing, 11) Privacy, 12) Fairness and Supportive Context of ImplementationConclusions: The SNMMI AI Task Force has identified valuable opportunities to enhance the practice of nuclear medicine through AI-based innovation. In addition, critical pitfalls that commonly afflict AI algorithm development, evaluation, and implementation have been recognized. In the end, Task Force elaborated on the responsibilities of SNMMI and the nuclear medicine community to ensure the trustworthiness of these tools.