Room: Room 202
Purpose: To develop an automated method for quantifying the visibility of the anti-scatter grid in 2D digital mammography images and assess its utility as an indicator of system performance.
Methods: Four digital mammography systems (Hologic Dimensions 3D 5000) were configured to automatically transfer 2D views to a server where the normalized power spectral density at the grid frequency, S(kgrid), was automatically calculated using Fourier analysis and stored in an SQL database. S(kgrid) was analyzed retrospectively and compared with image quality and service records for weekly ACR phantom images and patient images to establish preliminary action limits for physicist intervention and/or notification. Systems were then monitored in real-time for 3 months to evaluate the effectiveness of the proposed action limits.
Results: All systems exhibited a gradual increase in S(kgrid) in ACR phantom images prior to having identical major components of the generator subsystem repaired despite not having any visible grid lines. Retrospective analysis suggests that that the physicist should consider additional testing when S(kgrid) exceeds a threshold value of 3x10â?»â?· and that clinical image quality may be impacted above 4x10â?»â?·. Real-time monitoring of the weekly ACR phantom images successfully led to a preemptive service call before clinical image quality was impacted in one system. Analysis of patient images suggested grid lines are visible above a threshold of 1.45x10â?»â?¶ and was successfully used to notify physics of intermittent artifacts which ultimately led to major service on the grid assembly.
Conclusion: Routine monitoring of the grid frequency can be a useful indicator of system performance in digital mammography systems. ACR phantom images are suitable for longitudinal tracking of long term system degradation and can identify artifacts before they become visible whereas patient images are more suited for high frequency monitoring of gross errors in system performance.