Scholarlink

<p>Accurate fibrosis quantification is essential for understanding muscle and cardiac disease, yet current manual and semi‑automated methods remain slow, subjective, and poorly reproducible. We introduce FibroTrack, a standalone deep learning platform with a graphical user interface (GUI) that streamlines fibrosis analysis across Sirius Red (SR), Masson’s Trichrome (MT), and immunohistochemistry (IHC) stainings. FibroTrack uniquely integrates LAB (lightness, green–red, blue–yellow) color space normalization with a You Only Look Once version 11 (YOLOv11) segmentation model trained on 2,034 histological images. This approach achieved 99.5% mask precision for muscle segmentation and demonstrated excellent concordance with blinded pathologists (Spearman correlation, <i>r</i> = 0.87–0.96). Automated outputs include segmented images and structured spreadsheets, ensuring high reproducibility and scalability. By combining advanced color analysis with state‑of‑the‑art segmentation in an accessible tool, FibroTrack provides a novel, accurate, and clinically relevant solution for high‑throughput fibrosis quantification in both preclinical research and pathology practice.</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

FibroTrack: a standalone deep learning platform for automated fibrosis quantification in muscle and cardiac histology

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月26日

查看全文
  • Anas Odeh,
  • Rahaf Salem,
  • Maher Abu Saleh,
  • Ariel Shemesh,
  • Polina Stein,
  • Ido Livneh,
  • Peleg Hasson

摘要

Accurate fibrosis quantification is essential for understanding muscle and cardiac disease, yet current manual and semi‑automated methods remain slow, subjective, and poorly reproducible. We introduce FibroTrack, a standalone deep learning platform with a graphical user interface (GUI) that streamlines fibrosis analysis across Sirius Red (SR), Masson’s Trichrome (MT), and immunohistochemistry (IHC) stainings. FibroTrack uniquely integrates LAB (lightness, green–red, blue–yellow) color space normalization with a You Only Look Once version 11 (YOLOv11) segmentation model trained on 2,034 histological images. This approach achieved 99.5% mask precision for muscle segmentation and demonstrated excellent concordance with blinded pathologists (Spearman correlation, r = 0.87–0.96). Automated outputs include segmented images and structured spreadsheets, ensuring high reproducibility and scalability. By combining advanced color analysis with state‑of‑the‑art segmentation in an accessible tool, FibroTrack provides a novel, accurate, and clinically relevant solution for high‑throughput fibrosis quantification in both preclinical research and pathology practice.