Résumé : Scientific research trends and interests evolve over time. The ability to identify and forecast these trends is vital for educational institutions, practitioners, investors, and funding organizations. In this study, we predict future trends in scientific publications using heterogeneous public sources, including historical publications from PubMed, research and review articles, and patents. We demonstrate that scientific trends can be predicted five years in advance, with preceding publications and future patents serving as leading indicators for emerging scientific topics. We found that the ratio of reviews to original research articles is an informative feature for identifying increasing or declining topics, with declining topics having an excess of reviews. We find that language models provide improved insights and predictions into topic temporal dynamics. Our findings suggest that similar dynamics apply to molecular, technological, and conceptual topics across biomedical research.