Predictive Nano–Bio Interface Engineering of Curcumin–Gadolinium Hierarchical Nanocrystals for Precision Oncology using AI and ML

Abstract Magnetic nanocrystals exhibit size- and surface-dependent physicochemical properties that critically influence their biomedical functionality. Here, we report the rational design of curcumin-functionalized gadolinium (Gd) nanocrystals engineered for enhanced cancer theranostics. Leveraging artificial intelligence–assisted surface optimization and machine learning–based nano–bio interaction modelling, we investigated how nanoparticle size, surface chemistry, and curcumin conjugation influence cellular uptake, cytotoxic selectivity, and magnetic resonance imaging (MRI) contrast performance. Computational modelling predicted enhanced tumor-selective affinity driven by surface-functional group distribution and charge density optimization. Experimental validation in cancerous and non-cancerous human cell lines demonstrated preferential cytotoxicity toward malignant cells, while maintaining reduced toxicity in normal cells. In parallel, ML-assisted relaxivity prediction indicated improved T1-weighted MRI contrast efficiency compared to unconjugated Gd nanocrystals. Mechanistically, AI-enabled clustering of cellular response profiles revealed oxidative stress modulation and apoptosis pathway activation as dominant anticancer drivers. Together, these findings position curcumin–Gd nanocrystals as a data-guided theranostic platform integrating targeted therapy and diagnostic imaging.