Breaking through the radiation dilemma: development and clinical translation of anti-radiation drugs.

The need for effective prevention and treatment of nuclear radiation injuries is underscored by historical nuclear incidents, ongoing challenges such as nuclear wastewater management, and the expanding use of radiation in medicine and industry. Despite the approval of a limited set of drugs such as potassium iodide, Prussian blue, cytokines for hematopoietic acute radiation syndrome (H-ARS), current countermeasures are hindered by a narrow scope of application, significant side effects, and a profound mechanistic knowledge gap. While novel drug development has expanded into various mechanisms such as targeting DNA damage repair and anti-inflammation, yielding promising directions like novel nano-delivery systems, the overall clinical translation rate remains low. To address these challenges, this review synthesizes literature from the past 3 decades with the following aims: (1) to provide an updated analysis of the molecular mechanisms of radiation injury, highlighting newly discovered targets; (2) to critically evaluate drugs across clinical, trial, and preclinical stages; and (3) to introduce a transformative paradigm. The application of artificial intelligence (AI) in drug discovery, is a prospect not systematically explored in prior reviews. We posit that integrating mechanistic insights with AI-driven approaches represents a promising path forward. Finally, we propose future directions aimed at overcoming the specific challenges facing AI in this field, including the development of strategies to mitigate model "black-box" effects, the establishment of secure and ethical frameworks for sharing sensitive radiation injury data, and the creation of specialized, high-quality databases to address the critical issue of data scarcity.