Scholarlink

<p>The global prevalence of diabetic neuropathy (DN), a common and debilitating complication of diabetes mellitus, continues to rise, making early and accurate diagnosis a critical clinical priority. In recent years, artificial intelligence (AI) and machine learning (ML) techniques have emerged as promising tools for enhancing DN screening, diagnosis, and monitoring. This systematic literature review (SLR) aims to analyze and synthesize recent advances in AI-based methods for DN detection, with a particular focus on deep learning (DL) and hybrid modeling approaches. A comprehensive search was conducted across Google Scholar, PubMed, Web of Science (WoS), and Scopus to identify peer-reviewed studies published between 2014 and 2025. Studies were selected based on predefined inclusion and exclusion criteria, focusing on AI, ML, and DL-based techniques for DN diagnosis and classification. Following duplicate removal and a multi-stage screening process (titles, abstracts, and full texts), a total of 78 studies were included in the qualitative synthesis. The findings reveal several persistent challenges, including the limited availability of large, diverse, and well-annotated datasets, as well as the lack of interpretability in complex AI models. Despite these limitations, AI-driven approaches, such as ensemble learning, convolutional neural networks (CNNs), and transformer-based architectures, demonstrate strong performance in DN classification tasks. Moreover, studies incorporating multi-modal data consistently report improved diagnostic accuracy, underscoring the importance of integrating diverse patient information. However, most existing studies remain at the experimental or proof-of-concept stage, highlighting a significant gap between algorithm development and real-world clinical implementation. Future research should focus on developing explainable AI models, improving generalization across heterogeneous populations, and integrating validated AI systems into clinical workflows to fully realize their potential in advancing DN diagnosis and improving patient outcomes.</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

A Systematic Review of Artificial Intelligence Techniques for Diabetic Neuropathy Detection

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月14日

查看全文
  • SanketGulabchand Chordiya,
  • Avinash Dhole,
  • P. B. Deshmukh,
  • Mahesh Singh

摘要

The global prevalence of diabetic neuropathy (DN), a common and debilitating complication of diabetes mellitus, continues to rise, making early and accurate diagnosis a critical clinical priority. In recent years, artificial intelligence (AI) and machine learning (ML) techniques have emerged as promising tools for enhancing DN screening, diagnosis, and monitoring. This systematic literature review (SLR) aims to analyze and synthesize recent advances in AI-based methods for DN detection, with a particular focus on deep learning (DL) and hybrid modeling approaches. A comprehensive search was conducted across Google Scholar, PubMed, Web of Science (WoS), and Scopus to identify peer-reviewed studies published between 2014 and 2025. Studies were selected based on predefined inclusion and exclusion criteria, focusing on AI, ML, and DL-based techniques for DN diagnosis and classification. Following duplicate removal and a multi-stage screening process (titles, abstracts, and full texts), a total of 78 studies were included in the qualitative synthesis. The findings reveal several persistent challenges, including the limited availability of large, diverse, and well-annotated datasets, as well as the lack of interpretability in complex AI models. Despite these limitations, AI-driven approaches, such as ensemble learning, convolutional neural networks (CNNs), and transformer-based architectures, demonstrate strong performance in DN classification tasks. Moreover, studies incorporating multi-modal data consistently report improved diagnostic accuracy, underscoring the importance of integrating diverse patient information. However, most existing studies remain at the experimental or proof-of-concept stage, highlighting a significant gap between algorithm development and real-world clinical implementation. Future research should focus on developing explainable AI models, improving generalization across heterogeneous populations, and integrating validated AI systems into clinical workflows to fully realize their potential in advancing DN diagnosis and improving patient outcomes.