Development and In Vivo Evaluation of pH-Responsive Chitosan/Alginate Hydrogel-Based Films Co-Loaded With Carissa Carandas L. Anthocyanins and Gallic acid for Wound Healing and Monitoring

Smart wound dressings capable of simultaneously accelerating healing and providing real-time feedback on wound status are highly desirable for advanced wound care. In this study, we designed pH-responsive chitosan/alginate (CS/ALG) hydrogel-based films co-loaded with anthocyanins extracted from Carissa carandas L. (serve as pH indicator) and gallic acid (GA) (serves as model drug), and integrated them with a newly developed deep-learning model (WoundCareAI) for noninvasive wound healing and pH prediction.

The films were prepared by simple solvent casting method, characterized in terms of structure, thickness (~100 μm), area (~55 cm²), morphology (homogeneous under SEM analysis), chemical interactions, and thermal behavior (enhanced thermal stability under TGA analysis), pH-responsiveness, drug release rate, antibacterial and antioxidant activities, in vivo murine wound healing process. A new AI model named WoundCareAI to predict pH from images of the pH-sensitive films applied to wounds was also developed.

The films exhibited pH-dependent color changes and a moderate Korsmeyer-Peppas-kinetics drug release profile that are suitable for wound healing and monitoring. Biologically, the co-loaded films demonstrated notable in vitro antibacterial activity on both S. aureus (ZOI = 5.7 mm) and E. coli (ZOI = 6.0 mm), and enhanced antioxidant capacity compared with single-component films. Moreover, in in vivo murine excisional wound healing model, the film dressing significantly accelerated wound closure and supported restoration of skin pH toward physiological values, compared with the commercial products. Lastly, the WoundCareAI, built on a YOLOv11-Convolution neural network trained on images of the developed pH-sensitive films, automatically detected the indicator region and predicted wound pH from film color with promising accuracy.

Conclusively, the present integrated approach provides a low-cost, biopolymer-based colorimetric dressing uniquely coupled to a deep-learning model, offering personalized, noninvasive wound healing and pH monitoring.