Increasing efficiency in the ACR Digital Mammography QC Manual's DBT Z-resolution evaluation

Area of Science:

• Medical Imaging Physics
• Quality Control in Healthcare
• Radiologic Technology

Background:

• The ACR Digital Mammography (DM) QC manual specifies Z-resolution evaluation using pseudo full width half max (pFWHM) derived from speck group signals.
• This evaluation involves measuring up to 63 regions of interest (ROIs) across multiple phantom slices, comparing pFWHM to a baseline.

Purpose of the Study:

• To assess if reducing the number of ROIs in the Z-resolution evaluation impacts the reliability of results.
• To determine if a simplified measurement protocol can maintain diagnostic accuracy in digital mammography QC.

Main Methods:

• Retrospective analysis of Z-resolution data from 468 surveys across 147 mammography units.
• Calculation of pFWHM using the standard 6-speck mean (pFWHM₆), single speck (pFWHM₁), and 2-speck mean (pFWHM₂).
• Evaluation of measurement deviation and failure detection rates for each pFWHM method using longitudinal data.

Main Results:

• The 2-speck method (pFWHM₂) showed significantly lower average absolute error (3.9%) compared to the single-speck method (pFWHM₁) at 6.0% relative to pFWHM₆.
• Longitudinal testing revealed pFWHM₂ had a mean absolute deviation of 5.6% versus 7.8% for pFWHM₁, closely matching pFWHM₆ at 5.1%.
• Both pFWHM₁ and pFWHM₂ detected component-related failures with 0% false negatives, but pFWHM₁ had a higher false positive rate (12.9%) than pFWHM₂ (2.8%).

Conclusions:

• A single-speck pFWHM measurement leads to high variability and unreliable Z-resolution evaluations in ACR DM QC.
• The 2-speck pFWHM method provides clinically equivalent results to the 6-speck method while reducing measurements by 57%.
• Implementing the 2-speck method enhances testing efficiency without compromising the accuracy of Z-resolution evaluation.