Scholarlink

<p>Digital twins are virtual, data-driven replicas of real patients, continuously synced with individual data (e.g., continuous glucose monitoring [CGM], insulin doses, diet, activity) to enable “what-if” simulations. Originally from industry and aerospace, they now personalize precision therapy by forecasting disease progression, comparing treatments, and tailoring strategies. In diabetology, mechanistic models and artificial intelligence (AI) merge CGM, pump, dietary, and activity logs to simulate patient-specific glucose trajectories. As decision-support tools, they predict the impact of medication combinations or dose changes on glycated hemoglobin (HbA<sub>1</sub>c) and time in range. Patients can directly simulate therapy adjustments (e.g., basal-rate aggressiveness) in AID systems, and emerging projects use twins to enhance fully closed-loop bolus delivery. Despite promising pilot data, prospective interventional and real-world studies are still needed. With growing datasets and improved AI, digital twins are poised to become ever more precise and personalized, with the potential to revolutionize diabetes care and self-management.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Künstliche Intelligenz als „decision support system“

  • 期刊论文

  • 发表日期：

    2025年6月20日

查看全文
  • Dominic Ehrmann,
  • Bernhard Kulzer,
  • Andreas Schmitt,
  • Laura Klinker,
  • Norbert Hermanns

摘要

Digital twins are virtual, data-driven replicas of real patients, continuously synced with individual data (e.g., continuous glucose monitoring [CGM], insulin doses, diet, activity) to enable “what-if” simulations. Originally from industry and aerospace, they now personalize precision therapy by forecasting disease progression, comparing treatments, and tailoring strategies. In diabetology, mechanistic models and artificial intelligence (AI) merge CGM, pump, dietary, and activity logs to simulate patient-specific glucose trajectories. As decision-support tools, they predict the impact of medication combinations or dose changes on glycated hemoglobin (HbA1c) and time in range. Patients can directly simulate therapy adjustments (e.g., basal-rate aggressiveness) in AID systems, and emerging projects use twins to enhance fully closed-loop bolus delivery. Despite promising pilot data, prospective interventional and real-world studies are still needed. With growing datasets and improved AI, digital twins are poised to become ever more precise and personalized, with the potential to revolutionize diabetes care and self-management.