Image analysis of benign and malignant neck masses

Area of Science:

• Radiology and Medical Imaging
• Computational Pathology
• Oncology

Background:

• Current computed tomography (CT) scan analysis for neck masses relies on basic features like size and density.
• Digital imaging advances enable sophisticated computerized image processing for enhanced diagnostic capabilities.
• Distinguishing benign from malignant neck masses is crucial for appropriate patient management.

Purpose of the Study:

• To evaluate the effectiveness of four different image processing algorithms in differentiating benign and malignant neck masses.
• To compare the diagnostic performance of edge analysis, Fourier analysis, texture parameters, and histogram analysis.

Main Methods:

• Retrospective analysis of CT scans from 10 patients with squamous cell carcinoma neck masses and 11 with benign neck nodes.
• Application of four image processing algorithms: edge analysis, Fourier analysis, texture analysis, and histogram analysis.
• Statistical comparison of algorithm outputs between malignant and benign mass groups.

Main Results:

• Statistically significant differences were observed between malignant and benign masses using edge analysis and Fourier analysis.
• Several texture parameters showed significant differences, but were deemed less reliable.
• Histogram analysis did not reveal significant differences between the two groups.

Conclusions:

• Edge analysis and Fourier analysis demonstrate potential as valuable tools for improving the distinction between benign and malignant neck masses on CT scans.
• These advanced image processing techniques can provide supplementary information beyond traditional assessment methods.
• Further validation is warranted to integrate these methods into routine clinical practice for neck mass evaluation.