Learning effects in visual grading assessment of model-based reconstruction algorithms in abdominal Computed...

Area of Science:

• Radiology
• Medical Imaging
• Image Reconstruction

Background:

• Iterative reconstruction algorithms in CT imaging can alter spatial frequencies, potentially affecting diagnostic performance.
• Higher strengths of Advanced modeled iterative reconstruction algorithm (ADMIRE) may present unusual image characteristics.
• Radiologist perception and adaptation to novel image reconstruction techniques require investigation.

Purpose of the Study:

• To evaluate if radiologists can adapt to images generated by higher strengths of ADMIRE.
• To assess the impact of prolonged exposure on radiologist perception of ADMIRE-reconstructed images.
• To determine if a learning effect occurs with increased review of ADMIRE images.

Main Methods:

• Analysis of two prior studies on ADMIRE performance in abdominal CT (non-contrast and contrast-enhanced).
• Images reconstructed using ADMIRE strengths 3 and 5, and filtered back projection (FBP).
• Radiologists assessed image quality using European guidelines; time variable introduced to assess learning effect via logistic regression.

Main Results:

• A significant negative attitude towards ADMIRE 5 images intensified over time for liver parenchyma and overall image quality.
• ADMIRE 3 initially received a positive attitude, but a negative trend in overall image quality was observed in one dataset over time.
• Increased dislike for ADMIRE 5 images was evident with continued review, suggesting no adaptation.

Conclusions:

• Radiologists did not demonstrate a learning effect to adapt to or accept ADMIRE 5 images over time.
• Prolonged review of ADMIRE 5 images resulted in a worsening perception of image quality.
• Further research may be needed to optimize iterative reconstruction algorithms for clinical acceptance.