Assessment of Acute Pulmonary Embolism by Computer-Aided Technique: A Reliability Study

Area of Science:

• Cardiovascular Imaging
• Medical Informatics
• Radiology

Background:

• Acute pulmonary embolism (PE) is a common cardiovascular disease.
• Computer-aided detection (CAD) is crucial in chest imaging for PE assessment.
• Reliability and quantitative analysis of CAD techniques are essential for clinical application.

Purpose of the Study:

• To evaluate the reliability of a geometry-based CAD system for detecting and quantifying emboli morphology and severity in acute PE.
• To compare the accuracy and efficiency of CAD versus manual interpretation for PE assessment.

Main Methods:

• Thirty patients with suspected acute PE underwent both manual and CAD interpretation.
• Vascular obstruction index, Qanadli score, and Mastora score were assessed.
• CAD measurements of emboli quantitative parameters were obtained.
• Reliability was analyzed using intraclass correlation coefficient (ICC).

Main Results:

• CAD significantly reduced interpretation time (121.07±51.76s) compared to manual methods (374.90±150.16s) (P<0.001).
• High agreement was found between CAD and manual interpretations for Qanadli score (ICC=0.998) and Mastora score (ICC=0.997).
• Emboli quantitative parameters showed moderate correlation with established scoring systems (P<0.001).

Conclusions:

• Geometry-based CAD is a reliable tool for assessing acute PE.
• CAD reduces interpretation time and improves the reliability of vascular obstruction index assessment.
• CAD provides additional quantitative parameters, enhancing disease assessment for acute pulmonary embolism.