EUVIMEDEUVIMED entdecken
Übersichtsarbeit

Identification of ovarian cancer cell based on mechanical properties of regional heterogeneity of cells

Chuanzhi Liu, Luyao Tian, Fan Yang, Li Wang

Biosensors and Bioelectronics · 2027-02 · 318 · 119401

DOI: 10.1016/j.bios.2026.119401 ↗

Kurz gefasst

Übersichtsarbeit in der Zeitschrift Biosensors and Bioelectronics, 2027-02. Verfasst von Chuanzhi Liu und 3 weiteren Autor:innen. Beteiligte Einrichtung: School of Life Science and Technology, Changchun University of Science and Technology, Changchun, Jilin Province, 130022, China. Electronic address: …. Die Arbeit nennt 40 Literaturangaben.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Cellular mechanical properties are emerging as promising biomarkers for tumor diagnosis and for elucidating cancer-related mechanisms. However, most studies treat cells as mechanically homogeneous entities, overlooking regional heterogeneity and its implications for classification. Atomic force microscopy was used to characterize the mechanical properties of hepatocytes, mammary cells, ovarian cystadenoma cells, and SKOV3 ovarian cancer cells. Fluorescent staining was used to analyze the correlation between cytoskeletal topology parameters and mechanical properties. Logistic regression and support vector machine (SVM) classifiers were used to compare the performance of whole-cell, region-specific, and cross-regional models. Mechanical properties distinguished cells representing normal, benign, and malignant phenotypes. For ovarian cells, a model integrating multiple mechanical features achieved an area under the receiver operating characteristic curve (AUC) of 0.95. Region-specific models yielded AUCs of approximately 0.85, demonstrating pronounced spatial heterogeneity in cancer-cell mechanics and region-dependent remodeling of cytoskeletal branching architecture. In cross-cell-type analysis, an SVM trained on 7702/HepG2 achieved an AUC greater than 0.90 when applied to ovarian cells, indicating substantial generalizability of mechanical features across cell types. These results support the cross-cell type generalization of mechanical properties and suggest the potential feasibility of establishing a mechanical fingerprint reference across cell types. Furthermore, the difference between perinuclear and cytoplasmic regions reveals the complexity of mechanical regulation in ovarian cancer cells and provides a new perspective for understanding cellular dynamics.

Schlagwörter

Stichwörter der Autor:innen

Living cell cytologyMachine learningMechanobiologyMedical AFM

Autor:innen und Institutionen

  • School of Life Science and Technology, Changchun University of Science and Technology, Changchun, Jilin Province, 130022, China. Electronic address: duguchuanzhi@163.com.
  • School of Life Science and Technology, Changchun University of Science and Technology, Changchun, Jilin Province, 130022, China.
  • International Research Centre for Nano Handling and Manufacturing of China, Changchun University of Science and Technology, Changchun, 130022, China.
  • The 964th Hospital of the Chinese People's Liberation Army Joint Logistic Support Force, Changchun, Jilin Province, 130022, China.

Zitieren

APA

Liu, C., Tian, L., Yang, F., & Wang, L. (2027). Identification of ovarian cancer cell based on mechanical properties of regional heterogeneity of cells. Biosensors and Bioelectronics, 318, 119401. https://doi.org/10.1016/j.bios.2026.119401

Vancouver

Liu C, Tian L, Yang F, Wang L. Identification of ovarian cancer cell based on mechanical properties of regional heterogeneity of cells. Biosensors and Bioelectronics. 2027;318:119401. doi:10.1016/j.bios.2026.119401

BibTeX
@article{Liu2027,
  author = {Liu, Chuanzhi and Tian, Luyao and Yang, Fan and Wang, Li},
  title = {Identification of ovarian cancer cell based on mechanical properties of regional heterogeneity of cells},
  journal = {Biosensors and Bioelectronics},
  year = {2027},
  volume = {318},
  pages = {119401},
  doi = {10.1016/j.bios.2026.119401}
}
Kostenlos starten →