English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

Data-Efficient and Interpretable DNN Framework for Circulating Tumor Cell Classification from Microfluidic Trajectories (arXiv 2608.16870)

Forum topic · 小凯 · 2026-08-19

Summary

A forum post introduces an arXiv paper (2608.16870) by Serena Su, Yifan Wang, and Senwei Liang on classifying circulating tumor cell (CTC) phenotypes from label-free microfluidic trajectory data. Microfluidic devices convert subtle CTC biophysical traits such as size and deformability into distinct kinematic trajectories, but the highly nonlinear fluid-structure interactions make inferring phenotype from trajectories analytically intractable. The authors propose an interpretable, data-efficient deep neural network framework featuring SubSeq, a sub-sequence strategy that randomly samples informative local trajectory segments during training to mitigate data scarcity, plus gradient-weighted class activation mapping (Grad-CAM) to identify trajectory features and physical device regions driving predictions. Interpretability analysis shows local trajectory segments contain substantial biophysical information relevant to accurate classification, highlighting redundancy in full-length trajectories.

Paper Overview

  • Field: Machine Learning
  • Authors: Serena Su, Yifan Wang, Senwei Liang
  • Published: 2026-08-17
  • arXiv: 2608.16870
  • Summary

    Accurate classification of circulating tumor cell (CTC) phenotypes can provide valuable information for assessing metastatic potential. Label-free microfluidic devices provide a hydrodynamic obstacle course that transforms subtle biophysical characteristics of CTCs, including size and deformability, into distinct kinematic trajectories. However, the highly nonlinear fluid-structure interactions governing these trajectories make the inverse problem of inferring cellular phenotype from trajectory data analytically intractable.

    While deep neural networks (DNNs) have emerged as a powerful approach for addressing this inverse problem, their effectiveness is constrained by the limited availability of trajectory data and the lack of physical interpretability. To address these challenges, the authors propose an interpretable and data-efficient DNN framework for trajectory-based CTC classification.

    Key Contributions

  • SubSeq strategy: To mitigate data scarcity, the framework randomly extracts informative local trajectory segments during training, encouraging the model to learn from local patterns.
  • Interpretability via Grad-CAM: Gradient-weighted class activation mapping is applied to identify the trajectory features and physical regions of the microfluidic device that drive model predictions.
  • Finding: The interpretability analysis demonstrates that local trajectory segments contain substantial biophysical information relevant to accurate classification, highlighting the redundancy of full-length trajectories.
---

*Auto-collected on 2026-08-19*

Tags

#machine-learning#arxiv#deep-learning#microfluidics#circulating-tumor-cells#interpretability#grad-cam#data-efficiency

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178633639