Improved image quality and dose reduction in liver CT using deep learning-based reconstruction: A comparative study

Area of Science:

• Radiology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Computed Tomography (CT) is crucial for diagnosing focal liver lesions.
• Traditional image reconstruction methods like Hybrid Iterative Reconstruction (HIR) have limitations in image quality and radiation dose.
• Deep Learning-based Image Reconstruction (DLIR) offers a potential advancement in CT image processing.

Purpose of the Study:

• To compare the image quality and radiation dose of CT scans reconstructed using DLIR versus HIR in patients with focal liver lesions.
• To evaluate the quantitative and qualitative performance of DLIR in hepatic CT imaging.

Main Methods:

• 153 patients with focal liver lesions underwent two CT scans using scanners with DLIR and HIR algorithms.
• Image quality was assessed using a Likert scale, with measurements of CT attenuation, noise (SD), SNR, and CNR.
• Radiation dose was quantified using CTDI and DLP values.

Main Results:

• DLIR demonstrated significantly higher image quality scores compared to HIR (p < 0.001).
• DLIR yielded higher CT attenuation, lower noise (SD), and improved SNR and CNR across all regions (p < 0.001).
• Radiation dose was significantly lower with DLIR compared to HIR (p < 0.001).

Conclusions:

• DLIR significantly enhances both qualitative and quantitative image quality in liver CT.
• DLIR achieves a substantial reduction in radiation dose for patients undergoing hepatic CT.
• DLIR represents a superior reconstruction algorithm for focal liver lesion detection and patient safety.