Noise-reducing algorithms do not necessarily provide superior dose optimisation for hepatic lesion detection with...

Area of Science:

• Medical Imaging
• Radiology
• Image Reconstruction

Background:

• Multidetector computed tomography (MDCT) is crucial for hepatic lesion detection.
• Image reconstruction algorithms play a vital role in optimizing image quality and diagnostic performance.
• Standard filtered backprojection (FBP) is a widely used algorithm, but iterative methods offer potential for noise reduction.

Purpose of the Study:

• To evaluate the dose-optimization potential of smoothing FBP and hybrid FBP/iterative algorithms compared to standard FBP.
• To assess the impact of these algorithms on hepatic lesion detection at varying slice thicknesses.
• To determine if noise reduction translates to improved detection of low-contrast lesions.

Main Methods:

• A liver phantom with a simulated low-contrast lesion was scanned using MDCT at various milliampere-second (mAs) settings and slice thicknesses (5, 3, and 1 mm).
• Data were reconstructed using standard FBP, smoothing FBP (A), and hybrid FBP/iterative (iDose(4)) algorithms.
• Image noise, sensitivity, and diagnostic performance (Figure of Merit - FOM) were assessed by 10 observers using a free-response receiver operating characteristic (FROC) analysis.

Main Results:

• Significant reductions in FOM and sensitivity were observed at 1-mm slice thickness and at 25 mAs across all algorithms, indicating decreased detection performance.
• The smoothing FBP (A) algorithm showed significantly reduced sensitivity at 3-mm thickness.
• Noise levels were reduced by approximately 13% for smoothing FBP (A) and 21% for iDose(4) compared to standard FBP.

Conclusions:

• Neither the smoothing FBP nor the hybrid FBP/iterative algorithm demonstrated superior performance for hepatic lesion detection compared to standard FBP.
• Despite achieving noise reduction, particularly with thinner slices, these advanced algorithms did not improve the detection of low-contrast hepatic lesions.
• Reductions in image noise do not automatically lead to enhanced detection of low-contrast objects in CT imaging.