Scholarlink

<p>Breast cancer continues to be one of the primary health threats, with more than 2.26&#xa0;million new cases diagnosed in 2020, making the importance of accurate early diagnosis extremely vital. The application of AI methods, including classic ML and DL techniques, has been shown to significantly increase diagnostic precision and reduce misdiagnosis in addition to optimizing the clinical workflow of diagnosing breast cancer. The purpose of this systematic review is to provide an overview of recent findings related to applying AI in breast cancer detection and analysis based on peer reviewed articles published from 2022 to 2024. A total number of 62 publications were included in this systematic analysis. Findings of this research show that CNN models are capable of achieving 99.43% accuracy, whereas hybrid solutions and XAI have proven to increase model reliability and explainability. The most pressing issues include data imbalances, poor model generalizability, as well as ethical issues with regard to privacy and transparency. Future directions discussed in this paper include data fusion, federated learning, thermal imaging, and standardizing clinical testing protocols.</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 Review on Artificial Intelligence Techniques for Breast Cancer Detection: Advances, Challenges, and Future Directions

  • 期刊论文

  • 发表日期：

    2026年6月19日

查看全文
  • Syed Mohsin Bukhari,
  • Ishan Kumar,
  • Dipen Saini

摘要

Breast cancer continues to be one of the primary health threats, with more than 2.26 million new cases diagnosed in 2020, making the importance of accurate early diagnosis extremely vital. The application of AI methods, including classic ML and DL techniques, has been shown to significantly increase diagnostic precision and reduce misdiagnosis in addition to optimizing the clinical workflow of diagnosing breast cancer. The purpose of this systematic review is to provide an overview of recent findings related to applying AI in breast cancer detection and analysis based on peer reviewed articles published from 2022 to 2024. A total number of 62 publications were included in this systematic analysis. Findings of this research show that CNN models are capable of achieving 99.43% accuracy, whereas hybrid solutions and XAI have proven to increase model reliability and explainability. The most pressing issues include data imbalances, poor model generalizability, as well as ethical issues with regard to privacy and transparency. Future directions discussed in this paper include data fusion, federated learning, thermal imaging, and standardizing clinical testing protocols.