Tumor segmentation via enhanced area growth algorithm for lung CT images

Area of Science:

• Medical Imaging
• Computational Biology
• Radiology

Background:

• Accurate lung tumor segmentation is crucial for diagnosis due to the dynamic nature of tumors.
• Understanding tumor growth and changes aids in primary diagnosis.

Purpose of the Study:

• To introduce and evaluate the Enhanced Area Growth (EAG) algorithm for lung tumor segmentation.
• To improve the accuracy and reliability of lung tumor detection in CT images.

Main Methods:

• The Enhanced Area Growth (EAG) algorithm was developed using MATLAB on 60 patient CT images from diverse databases.
• The algorithm involves contrast augmentation, intensity analysis, thresholding, and iterative boundary determination for precise tumor delineation.
• Edge correction and region annexation refine the segmented tumor surface.

Main Results:

• The EAG algorithm demonstrated a significant enhancement in tumor identification, improving accuracy by over 16%.
• The algorithm achieved high accuracy with dice coefficients of 0.92 ± 0.03, indicating precise tumor delineation.
• The multipoint-growth-starting-point approach ensures independence from the initial seed point, reducing human error.

Conclusions:

• The Enhanced Area Growth (EAG) algorithm substantially improves lung tumor detection accuracy by over 18%.
• The algorithm's independence from matrix size and image thickness suggests broad applicability to other tumor imaging datasets.
• This method has the potential to reduce errors in radiologists' interpretation and segmentation of lung tumors.