A novel method for pulmonary embolism detection in CTA images

Area of Science:

• Medical Imaging
• Radiology
• Computer-Aided Diagnosis

Background:

• Pulmonary embolism (PE) detection in computed tomography angiography images (CTAI) is crucial.
• Accurate lung vessel segmentation is essential for high sensitivity PE detection.

Purpose of the Study:

• To develop a novel computer-aided detection (CAD) method for detecting pulmonary embolism (PE).
• To improve the accuracy and efficiency of PE diagnosis using CTAI.

Main Methods:

• Accurate lung vessel segmentation using a novel CAD approach.
• Identification of PE starting regions and reference points based on anatomical structures.
• Distinguishing PEs using region, intensity, and size characteristics.

Main Results:

• The developed CAD system detected 428 out of 450 PEs.
• Achieved a sensitivity of 95.1% with 14.4 false positives per dataset (FP/ds).
• Tested on datasets with co-existing heart disease or lung disease (excluding PE).

Conclusions:

• The proposed CAD system demonstrates high performance for PE detection in CTAI.
• The system is a valuable tool for radiologists as a second reader.
• The method shows promise for improving diagnostic accuracy in clinical practice.