Scholarlink

<p>This review examines the convergence of wearable biosensors and artificial intelligence (AI) in personalized diabetes care. It addresses the limitations of traditional glucose monitoring and underscores the need for continuous, multi-analyte physiological surveillance. The manuscript evaluates multi-biofluid sensing platforms, specifically those utilizing interstitial fluid (ISF), sweat, saliva, tears, and urine. ISF, an extracellular medium and a plasma ultrafiltrate, exhibits low protein content, a property that reduces sensor biofouling. ISF glucose demonstrates a strong correlation with blood glucose (R² &gt; 0.95) and can achieve high analytical sensitivity and specificity in clinically validated systems; however, diffusion-based time lags of 5–10 min present a kinetic challenge. Consequently, AI correction is necessary to ensure real-time accuracy, which is often achieved through minimally invasive microneedle arrays. Sweat analysis allows for non-invasive, multi-parameter measurements. Nevertheless, challenges such as pH instability and analyte loss due to evaporation complicate this sensing approach. Therefore, microfluidic techniques are essential for maintaining sample stability. A primary finding indicates that clinically validated Continuous Glucose Monitoring (CGM) systems yield substantial improvements in glycemic control, increasing Time in Range (TIR) by 10–15% and reducing the incidence of hypoglycemic events by 30–40%. AI-based predictive algorithms can forecast glucose excursions 30–60 min in advance, exhibiting an accuracy exceeding 94%. Key barriers to implementation include sensor calibration challenges, algorithmic bias, and significant healthcare equity issues. Future research should prioritize the development of multi-analyte implantable devices, leverage federated learning frameworks, and incorporate additional biomarkers to deliver continuous, multi-analyte, skin-conformal monitoring.</p> Graphical Abstract <p></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

Nanostructured electrode materials and flexible-substrate engineering for wearable multi-analyte biosensors in diabetes monitoring and personalized care: a comprehensive review

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月1日

查看全文
  • Roghieh Sodeify,
  • Mir Amirhossein Seyednazari,
  • Amir Mohammad Dorosti,
  • Alireza Nourazarian

摘要

This review examines the convergence of wearable biosensors and artificial intelligence (AI) in personalized diabetes care. It addresses the limitations of traditional glucose monitoring and underscores the need for continuous, multi-analyte physiological surveillance. The manuscript evaluates multi-biofluid sensing platforms, specifically those utilizing interstitial fluid (ISF), sweat, saliva, tears, and urine. ISF, an extracellular medium and a plasma ultrafiltrate, exhibits low protein content, a property that reduces sensor biofouling. ISF glucose demonstrates a strong correlation with blood glucose (R² > 0.95) and can achieve high analytical sensitivity and specificity in clinically validated systems; however, diffusion-based time lags of 5–10 min present a kinetic challenge. Consequently, AI correction is necessary to ensure real-time accuracy, which is often achieved through minimally invasive microneedle arrays. Sweat analysis allows for non-invasive, multi-parameter measurements. Nevertheless, challenges such as pH instability and analyte loss due to evaporation complicate this sensing approach. Therefore, microfluidic techniques are essential for maintaining sample stability. A primary finding indicates that clinically validated Continuous Glucose Monitoring (CGM) systems yield substantial improvements in glycemic control, increasing Time in Range (TIR) by 10–15% and reducing the incidence of hypoglycemic events by 30–40%. AI-based predictive algorithms can forecast glucose excursions 30–60 min in advance, exhibiting an accuracy exceeding 94%. Key barriers to implementation include sensor calibration challenges, algorithmic bias, and significant healthcare equity issues. Future research should prioritize the development of multi-analyte implantable devices, leverage federated learning frameworks, and incorporate additional biomarkers to deliver continuous, multi-analyte, skin-conformal monitoring.

Graphical Abstract