The rapid digitalization of healthcare has enabled unprecedented opportunities for cross-border telemedicine and collaborative artificial intelligence (AI) research. However, the development of robust, generalized diagnostic models requires access to diverse, large-scale datasets that are often fragmented across national borders. This necessity clashes directly with increasingly stringent data sovereignty and privacy regulations, most notably the General Data Protection Regulation (GDPR) in the European Union and the Act on the Protection of Personal Information (APPI) in Japan. While mutual adequacy decisions exist between these jurisdictions, the transfer of raw medical records remains legally complex, technically risky, and ethically sensitive. This paper proposes a novel Federated Learning (FL) framework designed specifically to bridge the regulatory gap between German and Japanese healthcare institutions. Unlike traditional centralized learning, which requires aggregating patient data into a single repository, our proposed framework enables the collaborative training of deep learning models without raw data ever leaving the local institution's firewalls. We present a dual-node architecture connecting the University of Tokyo and the Technical University of Munich, utilizing a privacy-preserving aggregation mechanism that ensures compliance with both GDPR and APPI standards. By keeping sensitive patient information decentralized while sharing only model updates, this approach preserves patient privacy while unlocking the potential for global-scale medical AI development.
Keywords: Federated Learning, Cross-Border Telemedicine, GDPR, APPI, Privacy- Preserving Machine Learning, Health Informatics.