Related Experiment Video
Updated: Sep 17, 2025
08:59
Morphology-Based Distinction Between Healthy and Pathological Cells Utilizing Fourier Transforms and Self-Organizing Maps
Published on: October 28, 2018
7.2K
Machine learning to detect melanoma exploiting nuclei morphology and Spatial organization.
Giulia Veronesi1,2, Nico Curti3, Aldo Gardini4
1Department of Medical and Surgical Sciences, University of Bologna, Bologna, 40126, Italy.
Scientific Reports
|July 2, 2025
Summary
This study introduces an automated tool for diagnosing cutaneous melanoma, improving accuracy and efficiency in skin cancer detection. The system analyzes cellular features from histopathology images to aid clinical decision-making.
Area of Science:
- Dermatopathology
- Computational Pathology
- Medical Image Analysis
Background:
- Cutaneous melanoma is a lethal skin cancer with increasing global incidence.
- Manual histopathological diagnosis is complex and time-consuming.
- There is a need for automated tools to support melanoma diagnosis.
Purpose of the Study:
- To develop an automated diagnostic tool for cutaneous melanoma.
- To generate interpretable results for clinical decision support.
- To refine histopathological analysis at the cellular level.
Main Methods:
- Utilized a dataset of 146 whole slide images of various skin lesions.
- Employed a multi-resolution image processing pipeline to segment nuclei and extract features.
- Applied Linear Discriminant Analysis to identify relevant diagnostic variables.
- Validated performance using Monte Carlo Cross-Validation.
Main Results:
- Identified 18 key variables demonstrating good performance in melanoma detection.
- Features extracted included nuclear geometry, morphology, and spatial organization.
- The tool's findings were interpretable within established histopathological insights.
Conclusions:
- The automated tool shows significant potential for supporting clinicians in melanoma diagnosis.
- It can aid in prioritizing critical samples and providing secondary opinions.
- The cellular-level analysis ensures verifiability for medical professionals.
