Towards Personalized AI-Based Diabetes Therapy: A Review

Area of Science:

• Medical Artificial Intelligence (AI)
• Diabetes Management Technologies

Background:

• Diabetes affects millions globally, necessitating advanced treatment solutions.
• Smart devices and AI are revolutionizing medical diagnostics and therapeutic strategies.
• AI methods aid physicians in diagnosis, treatment selection, and outcome prediction.

Purpose of the Study:

• To analyze the application of AI in enhancing and personalizing diabetes treatment.
• To review learning models, data types, deployment stages, and application domains of AI in diabetes care.
• To identify key trends and limitations in AI-driven diabetes management.

Main Methods:

• Systematic review of 77 original research papers on AI in diabetes treatment.
• Analysis of AI learning models, data typology, deployment stage, and application domains.
• Identification of trends in patient-based therapy personalization and therapeutic algorithm optimization.

Main Results:

• AI enables personalized diabetes treatment tailored to individual patients.
• AI optimizes therapeutic algorithms for more effective diabetes management.
• Key trends include patient-based personalization and algorithm optimization.

Conclusions:

• AI significantly enhances diabetes treatment personalization and therapeutic algorithm optimization.
• Current limitations include a lack of multimodal data analysis and interpretability.
• Future AI improvements and increased data availability promise wider clinical deployment.