Artificial Intelligence in Nuclear Medicine (2015-2025): Clinical Applications, Regulatory Governance, and Pharmacotherapeutic Implications

Artificial intelligence (AI) is becoming increasingly relevant in nuclear medicine, particularly in applications that depend on quantitative image analysis, image reconstruction, acquisition optimization, and clinical decision support. This critical narrative review examines the evolution, clinical applications, ethical and regulatory challenges, and pharmacotherapeutic implications of AI in nuclear medicine from 2015 to 2025. A structured literature search was conducted in PubMed/MEDLINE, Scopus, Web of Science, Google Scholar, and selected professional and regulatory websites, including those of the Society of Nuclear Medicine and Molecular Imaging, the European Association of Nuclear Medicine, the International Atomic Energy Agency, the US Food and Drug Administration, and the European Medicines Agency. The reviewed literature reports advances in deep learning-based reconstruction, denoising, and preservation of quantitative parameters in positron emission tomography and single-photon emission computed tomography. Clinical applications have expanded across oncology, neurology, cardiology, theranostics, and radiopharmacy-related workflows. However, the evidence remains limited by methodological heterogeneity, predominantly retrospective and single-center designs, restricted external validation, uneven regulatory development, and limited representation of Latin American settings. Overall, AI may improve diagnostic precision, personalized dosimetry, workflow efficiency, and radiopharmaceutical safety in nuclear medicine. Responsible implementation will require robust multicenter validation, transparent evaluation standards, adaptive governance, and multidisciplinary oversight. Radiopharmacists should be actively involved in this process, particularly in relation to radiopharmaceutical quality, traceability, dosimetry, and pharmacotherapeutic decision-making.