Deep learning and iterative image reconstruction for head CT: Impact on image quality and radiation dose...

Area of Science:

• Radiology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Computed tomography (CT) is a vital diagnostic tool.
• Image quality and radiation dose are critical considerations in CT imaging.
• Novel reconstruction algorithms aim to optimize these parameters.

Purpose of the Study:

• To objectively evaluate the Deep Learning Image Reconstruction (DLIR) algorithm.
• To compare DLIR's performance against Adaptive Statistical Iterative Reconstruction-V (ASIR-V).
• To assess image quality improvement and radiation dose reduction in unenhanced head CT scans.

Main Methods:

• Retrospective analysis of 35 unenhanced head CT scans.
• Objective image quality assessment using Signal-to-Noise Ratio (SNR) and Contrast-to-Noise Ratio (CNR).
• Evaluation of radiation dose reduction using CTDIvol and dose length product.

Main Results:

• DLIR demonstrated significant increases in SNR (up to 160%) and CNR (up to 171.5%).
• DLIR achieved radiation dose reduction of up to 44% in CTDIvol.
• Despite improvements, DLIR's dose reduction was insufficient for wide-detector scanners without gantry angulation.

Conclusions:

• DLIR offers substantial image quality enhancement and dose reduction in head CT.
• DLIR outperforms ASIR-V in objective image quality metrics.
• Further optimization is needed for dose reduction in specific scanner configurations.