Low-dose simulation for grating-based X-ray dark-field radiography using a virtually decreased irradiation area

Area of Science:

• Medical imaging
• Radiography
• Pulmonary imaging

Background:

• X-ray dark-field radiography utilizes small-angle scattering for lung alveoli structural integrity assessment.
• Simulating low-dose images by reducing irradiated area is a method to study dose reduction effects.
• These simulations may introduce stripe artifacts, impacting image quality.

Purpose of the Study:

• Validate a low-dose simulation algorithm for dark-field radiography.
• Demonstrate that observed stripe artifacts in simulated low-dose images are algorithm-induced.
• Confirm these artifacts are absent in actual low-dose dark-field images.

Main Methods:

• Acquired dark-field radiographs of a chest phantom at varying radiation doses.
• Simulated lower-dose dark-field radiographs by virtually reducing the irradiated area from high-dose acquisitions.
• Quantitatively compared dark-field signals between simulated and measured low-dose images using regions of interest.

Main Results:

• Dark-field radiographs at one-quarter standard dose were artifact-free.
• The dark-field signal in measured low-dose images differed up to 10% from simulated ones.
• Algorithm-induced stripe artifacts degraded the image quality of simulated low-dose radiographs.

Conclusions:

• Virtual reduction of irradiation area is a feasible method for generating low-dose dark-field radiographs.
• The simulation algorithm introduces stripe artifacts at very low dose levels, unlike measured images.
• Simulated low-dose images exhibit reduced quality compared to actual low-dose dark-field radiographs.