頭蓋顔面外科における人工知能導入の倫理的および規制上の考慮事項

Artificial intelligence (AI) continues to enhance surgical care, with growing applications in craniofacial surgery, including automated cephalometric analysis, diagnostics for craniosynostosis, virtual surgical planning, perioperative decision support, and prediction of postoperative outcomes. As these technologies transition from experimental systems to clinical tools that influence diagnosis, planning, and counseling, the specialty faces increasingly complex ethical and regulatory challenges. In this manuscript, the authors review and synthesize the ethical principles guiding the responsible adoption of AI in craniofacial surgery, including informed consent in predominantly pediatric populations, the risks of algorithmic bias, the heightened privacy concerns associated with facial biometrics, the requirements for transparency and explainability, and the evolving concepts of clinician accountability in AI-augmented decision-making. The authors also examine the global regulatory landscape, focusing on the US FDA's framework for AI/ML-based software as a medical device (SaMD), the European Union's medical device and AI regulations, and international oversight models. The implications of continuous-learning systems for real-world performance monitoring and postmarket governance are discussed in detail. Finally, the authors propose a pragmatic implementation framework emphasizing governance structures, pilot evaluation, structured local validation, performance auditing, and equity-centered deployment. Establishing rigorous ethical guardrails, robust regulatory pathways, and specialty-specific governance structures will be essential for realizing AI's potential while safeguarding patient trust, safety, and equity in craniofacial care.