Cold atmospheric plasma for wound healing: Toward AI-guided therapeutic optimization

Acute and chronic wounds are still a major clinical problem and the therapeutic response to cold atmospheric plasma (CAP) is highly dependent on the device architecture, the delivered dose, the wound characteristics and the treatment protocol. This review critically discusses the relationships between the physicochemical properties of CAP, the generation of reactive oxygen and nitrogen species (RONS). It further examines how these species contribute to antimicrobial, anti-inflammatory, angiogenic, and regenerative effects. It also discusses direct CAP therapy, plasma-activated liquids and the current evidence from human clinical studies. Particular emphasis is placed on the standardization of dosimetry and treatment safety. The review also distinguishes between direct evidence from wound-healing studies and transferable findings from related biomedical applications. The review also points out emerging uses of AI in CAP research, such as plasma diagnostics, prediction of plasma response, analysis of wound images, and optimization of preclinical treatments. Although these studies demonstrate the potential of data-driven monitoring and predictive modeling, prospectively validated autonomous AI-controlled CAP wound therapy has not yet been achieved. To address this gap, this study proposes a clinician-supervised closed-loop framework. The framework integrates patient and wound characteristics, plasma diagnostics, longitudinal treatment response, uncertainty estimation, and device-specific safety constraints. Future clinical translation will require standardized dose reporting, integrated CAP–wound datasets, rigorous external validation, and long-term safety assessment to enable reliable AI-assisted personalized plasma therapy.