Thin-slice brain CT with iterative model reconstruction algorithm for small lacunar lesions detection: Image quality...

Area of Science:

• Radiology
• Medical Imaging
• Neuroradiology

Background:

• Brain computed tomography (CT) is crucial for diagnosing neurological conditions.
• Evaluating image quality and lesion detection is vital for accurate diagnosis.
• Thin-slice CT reconstruction algorithms offer potential improvements over conventional methods.

Purpose of the Study:

• To assess and compare the image quality and lacunar lesion detection capabilities of thin-slice brain CT images reconstructed with different algorithms.
• To evaluate filtered back projection (FBP), hybrid iterative reconstruction (HIR), and iterative model reconstruction (IMR) algorithms.

Main Methods:

• Sixty-one patients underwent noncontrast brain CT.
• Images were reconstructed using 5.0 μm routine slice with FBP and 1.0 μm thin-slice with IMR, HIR, and FBP.
• Objective (CT attenuation, noise, artifacts, CNR) and subjective image quality assessments were performed.
• Lacunar lesion detection was compared against MRI reference in 26 patients.

Main Results:

• Thin-slice IMR demonstrated the lowest noise, artifact index, and best contrast-to-noise ratio (CNR).
• Both IMR and HIR thin-slice images achieved superior subjective image quality scores compared to routine slice FBP.
• Thin-slice IMR and HIR showed higher sensitivity and positive predictive value (PPV) for detecting 3.5 mm lacunar lesions.

Conclusions:

• Thin-slice iterative model reconstruction (IMR) significantly enhances brain CT image quality.
• Iterative reconstruction algorithms (IMR and HIR) improve the detection of small lacunar lesions compared to filtered back projection (FBP).
• Thin-slice CT with advanced reconstruction techniques offers improved diagnostic performance in brain imaging.