Scholarlink

Purpose <p>Microscopy-based cancer cell classification traditionally relies on cell-based morphological features, while subcellular organelle organization remains underutilized. Existing machine learning methods often require manual preprocessing and handcrafted feature extraction, limiting scalability and introducing user bias. This study proposes an automated, interpretable, and organelle-focused deep learning framework for classifying breast cancer cell lines from high-resolution fluorescence microscopy images.</p> Methods <p>We developed an end-to-end framework that incorporates patch-based sampling, sparsity filtering, and a channel-wise intermediate fusion strategy to independently extract and integrate organelle-specific features. Model interpretability was assessed using Grad-CAM visualizations and single-organelle classifier analyses. The framework was evaluated on fluorescence microscopy images from six breast cancer cell lines using 5-fold cross-validation.</p> Results <p>The proposed framework achieved a classification accuracy of <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(97.1 \pm 1.1\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mn>97.1</mn> <mo>±</mo> <mn>1.1</mn> </mrow> </math></EquationSource> </InlineEquation>%, performing comparably to or exceeding conventional handcrafted feature-based approaches while eliminating the need for manual segmentation and 3D rendering steps. Interpretability and classifier analyses revealed inter-organelle dependencies and mitochondria as the most informative contributors to classification decisions.</p> Conclusion <p>Organelle morphology and spatial organization provide strong discriminative signals for cancer cell classification. The proposed framework offers a scalable, automated, and interpretable deep learning solution that advances microscopy-based phenotyping and supports broader applications in computational pathology and cellular informatics.</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

A data fusion deep learning approach for accurate organelle-based classification of cancer cells

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年2月6日

查看全文
  • Harrison Yee,
  • Megan Bouyea,
  • Joshua Goldwag,
  • John M. Lamar,
  • Xavier Intes,
  • Uwe Kruger,
  • Margarida Barroso

摘要

Purpose

Microscopy-based cancer cell classification traditionally relies on cell-based morphological features, while subcellular organelle organization remains underutilized. Existing machine learning methods often require manual preprocessing and handcrafted feature extraction, limiting scalability and introducing user bias. This study proposes an automated, interpretable, and organelle-focused deep learning framework for classifying breast cancer cell lines from high-resolution fluorescence microscopy images.

Methods

We developed an end-to-end framework that incorporates patch-based sampling, sparsity filtering, and a channel-wise intermediate fusion strategy to independently extract and integrate organelle-specific features. Model interpretability was assessed using Grad-CAM visualizations and single-organelle classifier analyses. The framework was evaluated on fluorescence microscopy images from six breast cancer cell lines using 5-fold cross-validation.

Results

The proposed framework achieved a classification accuracy of \(97.1 \pm 1.1\) 97.1 ± 1.1 %, performing comparably to or exceeding conventional handcrafted feature-based approaches while eliminating the need for manual segmentation and 3D rendering steps. Interpretability and classifier analyses revealed inter-organelle dependencies and mitochondria as the most informative contributors to classification decisions.

Conclusion

Organelle morphology and spatial organization provide strong discriminative signals for cancer cell classification. The proposed framework offers a scalable, automated, and interpretable deep learning solution that advances microscopy-based phenotyping and supports broader applications in computational pathology and cellular informatics.

Graphical abstract