An automated method for comparing motion artifacts in cine four-dimensional computed tomography images

Area of Science:

• Medical Imaging
• Radiology
• Image Processing

Background:

• Motion artifacts in four-dimensional computed tomography (4D CT) can significantly degrade image quality.
• Objective assessment of these artifacts is crucial for accurate diagnosis and treatment planning.
• Current methods for artifact evaluation often rely on subjective human observation.

Purpose of the Study:

• To develop an automated, objective method for comparing motion artifacts in 4D CT image sets.
• To identify 4D CT image sets with fewer or smaller motion artifacts compared to human observer perception.
• To validate the proposed automated method against human assessments.

Main Methods:

• The study proposes a novel method based on the difference of normalized correlation coefficients between edge slices at couch transitions.
• Ten pairs of 4D CT image sets with subtle, human-identifiable artifacts were used for evaluation.
• Image sets were sorted using breathing traces with both miscalculated and corrected respiratory phases.

Main Results:

• The automated method successfully identified the 4D CT image sets with corrected respiratory phases as having fewer or smaller artifacts in nine out of ten pairs.
• The method's accuracy in identifying reduced artifacts mirrored the consensus of two independent human observers.
• One image pair showed no discernible difference in artifacts between the two sorting methods.

Conclusions:

• The developed automated method provides a reliable and objective approach to quantify and compare motion artifacts in 4D CT imaging.
• This technique has the potential to assist in selecting optimal 4D CT datasets, improving diagnostic confidence.
• The method demonstrates proof of principle in replicating human observer judgment for artifact reduction.