COMPUTER AIDED EVALUATION OF PLEURAL EFFUSION USING CHEST CT IMAGES

Area of Science:

• Medical Imaging
• Radiology
• Computational Pathology

Background:

• Pleural effusion, characterized by abnormal fluid in the pleural space, requires accurate severity assessment for effective patient management.
• Current methods for evaluating pleural effusion severity can be subjective and time-consuming.

Purpose of the Study:

• To develop and validate an automated method for quantifying pleural effusion severity using standard chest computed tomography (CT) images.
• To compare the performance of the automated method against manual segmentation and radiologist grading.

Main Methods:

• Automated lung segmentation using region growing, mathematical morphology, and anatomical knowledge.
• Extraction of visceral and parietal pleura layers via anatomical landmarks, curve fitting, and active contour models.
• Segmentation of the pleural space and quantification of pleural effusion volume.

Main Results:

• The automated method demonstrated high accuracy in segmenting pleural effusion, with a Dice coefficient of 0.74±0.07, comparable to inter-observer variability.
• A strong Pearson correlation (0.956, P=10(-7)) was observed between the automated evaluation and radiologist's qualitative grading.
• The method was successfully tested on 15 chest CT studies.

Conclusions:

• The developed automated method provides an accurate and reliable approach for evaluating pleural effusion severity from chest CT scans.
• This automated technique has the potential to improve the efficiency and objectivity of pleural effusion assessment in clinical practice.
• The findings support the integration of automated image analysis tools in radiological workflows for pleural effusion diagnosis.