Mapping the Management-to-Cure Transition in Thalassemia Research: Semantic Topic Modeling of 17,651 Publications Across Three Decades

Thalassemia syndromes affect millions worldwide, with the highest prevalence in the Mediterranean basin, the Middle East, South Asia, and southern China. For decades, standard care relied on lifelong transfusion and iron chelation, a management paradigm that extended survival without offering a cure. The first curative option, allogeneic hematopoietic stem cell transplantation (HSCT), emerged in the early 1980s; the first gene-editing therapy, Casgevy, followed in 2023, marking the latest step in this trajectory. Yet the research landscape underlying the broader management-to-cure transition has not been systematically characterized. We applied BERTopic, a transformer-based semantic topic model, to 17,651 PubMed-indexed thalassemia abstracts spanning 1990-2026, tracking topic prevalence over 37 years and comparing trajectories across seven national research communities. The model independently identified a thematic structure that separated management-oriented research (iron chelation, cardiac monitoring, prenatal screening) from curative research (hematopoietic stem cell transplantation [HSCT], gene therapy); the crossover point occurred around 2020, three years before the first gene-editing approval. By 2025, curative topics accounted for 14% of publications, versus 9% for management topics. Country-level analysis revealed marked divergence: US research focused more heavily on gene therapy (10.0% vs. 4.1% in China), while Chinese research maintained a stronger HSCT focus (6.3% vs. 4.0% in the US). These findings chart the course of a major therapeutic transition and carry direct implications for research funding allocation, global strategies for equitable access to curative therapies, and computational literature surveillance to inform clinical guideline updates.