Image noise and liver lesion detection with MDCT: a phantom study.

Area of Science:

• Medical Imaging Physics
• Radiology and Diagnostic Imaging

Background:

• Detecting small, low-contrast lesions in liver CT scans is challenging.
• Image noise is a critical factor affecting lesion detectability.

Purpose of the Study:

• To determine the maximum acceptable noise level for detecting small, low-contrast liver lesions using CT.
• To evaluate the impact of varying noise index on lesion detection sensitivity.

Main Methods:

• A liver phantom with simulated low-contrast lesions (2.4-10 mm) was scanned using a 64-MDCT scanner.
• Automatic tube current modulation was employed, varying the noise index.
• Three radiologists assessed lesion detection certainty using a 4-point Likert scale.
• Receiver operator characteristic (ROC) analysis and area under the curve (A(z)) were used to evaluate performance.

Main Results:

• 100% sensitivity was achieved for lesions 6.3-10.0 mm with a noise index ≤ 15 (A(z) = 0.96).
• Sensitivity decreased to 92.3% (A(z) = 0.93) with a noise index of 17-21.
• Sensitivity dropped to 81.4% (A(z) = 0.77) with a noise index of 23-27, and further to 39% at higher levels.

Conclusions:

• The optimal noise index range for >90% sensitivity in detecting small, low-contrast liver lesions is 15-21.
• Maintaining a lower noise index is crucial for accurate detection of subtle liver abnormalities.
• This study provides guidance for optimizing CT parameters in liver lesion detection.