Image Contrast Dependent Spatial Resolution Differences for Varying Levels of ASiRô Implementation
S Brady*, R Kaufman, St. Jude Children's Research Hospital, Memphis, TNTH-E-217BCD-3 Thursday 1:00:00 PM - 2:50:00 PM Room: 217BCD
Purpose: Introduce MTF calculation technique for contrast dependent spatial resolution measurements of ASiR™ using the first module of the ACR CT phantom. Compare three approaches to account for spatially variant MTF for CT.
Methods: Each material insert within module one of the ACR phantom was imaged with a pixel pitch of 0.188. An MTF methodology described by Boone was used to validate the use of the ACR phantom. To account for spatially variant MTF of CT, three MTF calculation methodologies were investigated: an ensemble of edge response functions (ERFs) were aligned and averaged to produce one MTF; MTF curves were calculated for each ERF and averaged into one; and a cumulative distribution function was created by integrating a 300 bin histogram for one MTF calculation. Three reconstruction filters were investigated: standard, bone, and lung.
Results: ACR phantom and Boone methodology MTF curves agreed to within 12%. Three methodologies investigated to account for spatially variant CT agreed to within 1%; histogram methodology was susceptible to number of bins used to form the histogram, and was not reliable. High-contrast resolution improved by 10% when 100% ASiR™ was compared to 0% ASiR™; this trend agreed with previous reports and with Boone methodology calculations. Lung and bone reconstruction filters measured increased spatial resolution with increased levels of ASiR™: 10% and 17% for 0 and 100% ASiR™, respectively. Intermediary contrast resolution measured at acrylic-water and at polyethylene-water interfaces did not change with increasing ASiR™ reconstruction.
Conclusions: ACR phantom MTF calculations agreed with the Boone methodology. Of the three methodologies investigated to account for spatially variant MTF in CT, method three varied depending on sampling interval. MTF values for high-contrast objects were found to improve between 10-17% with 100% ASiR™ reconstruction. ASiR™ did not affect intermediary contrast object spatial resolution.