Related Experiment Video
Updated: Sep 14, 2025
07:05
Area-based Image Analysis Algorithm for Quantification of Macrophage-fibroblast Cocultures
Published on: February 15, 2022
2.7K
Relationship between Artificial Intelligence-Based Cell Detection and Cytomorphological Variations Induced by Cell
Nanako Sakabe1, Yuma Yoshizaki1, Kenta Fukuda1
1Pathophysiology Sciences, Department of Integrated Health Sciences, Nagoya University Graduate School of Medicine, Nagoya, Japan.
Acta Cytologica
|July 21, 2025
Summary
Cell processing solutions significantly alter cell appearance, impacting artificial intelligence (AI) accuracy in cytology. Data augmentation effectively improves AI cell detection by enhancing cytomorphological features for AI recognition.
Area of Science:
- Computational pathology
- Digital cytology
- Artificial intelligence in medicine
Background:
- Cytomorphological variations from specimen preparation impede artificial intelligence (AI) implementation in cytology.
- Challenges include small-scale research and insufficient datasets for AI development in this field.
Purpose of the Study:
- To investigate the link between cell-processing solutions, cytomorphological changes, and AI cell detection accuracy.
- To evaluate the efficacy of data augmentation in improving AI performance for cytology.
Main Methods:
- MKN45 human gastric cancer cells were treated with different solutions and processed.
- Papanicolaou staining was performed, followed by mathematical analysis of nuclear and cytoplasmic features (area, HSB values).
- Deep learning models were trained with and without data augmentation to assess AI cell detection rates.
Main Results:
- Heparin sodium solution treatment induced significant cytomorphological differences, affecting nuclear and cytoplasmic features.
- AI cell detection rates were notably lower for heparin-treated cells compared to controls.
- Data augmentation enhanced AI cell detection accuracy across all tested cell samples.
Conclusions:
- Key cytomorphological features influencing AI cell recognition were identified.
- Data augmentation was confirmed as a valuable strategy to boost AI-based cell detection accuracy in cytology.
