A NEW AI-driven “digital twin” platform, CURATE.AI, enabled personalized chemotherapy dosing for advanced solid tumor patients, achieving 97.2% clinician acceptance of recommended doses and reducing average drug doses by 20% while maintaining efficacy.
Traditional chemotherapy dosing relies on population-based protocols, often leading to suboptimal outcomes due to individual variability in drug response. The CURATE.AI platform, developed by researchers at the National University of Singapore, addresses this by creating patient-specific digital twins using real-time biomarker data. This approach dynamically tailors doses to individual biological feedback, marking a paradigm shift toward precision oncology.
In a prospective feasibility trial involving 10 patients with advanced solid tumors, CURATE.AI analyzed biomarkers (CEA, CA125) and drug responses to generate personalized capecitabine dosing recommendations. Clinicians accepted 35 out of 36 AI-suggested adjustments, reducing average doses from standard-of-care levels by approximately 20% (range: ±13.8%). The platform’s quadratic regression model utilized patient-specific data—drug doses and biomarker fluctuations—to iteratively optimize treatment over a median of 3.9 cycles per patient. Notably, one participant completed eight cycles with sustained efficacy, demonstrating the system’s adaptability to evolving tumor biology.
These findings underscore CURATE.AI’s potential to enhance clinical practice by balancing efficacy with reduced toxicity. Future efforts should focus on expanding validation through randomized controlled trials and adapting the platform for immunotherapy, hypertension, and longevity medicine. Clinicians could integrate such AI tools to dynamically adjust regimens, particularly in palliative care, where minimizing side effects is critical. Standardizing real-time biomarker monitoring and refining AI algorithms for broader drug combinations will further solidify its role in precision oncology.
Reference
Blasiak A et al. Personalized dose selection platform for patients with solid tumors in the PRECISE CURATE.AI feasibility trial. npj Precision Oncology. 2025;DOI:10.1038/s41698-025-00835-7.
