Room: Exhibit Hall | Forum 8
Purpose: To implement an MTF measurement method in radiography that can be integrated into routine clinical QC practice without the use of slit or edge test objects.
Methods: Our department uses the RaySafe P Fluoro phantoms in our digital radiography routine weekly QC. The phantom contains a set of line-pair test objects at 20 different spatial frequencies, ranging from 0.6 lp/mm to 5.0 lp/mm, based on which MTF can be directly measured and calculated from the change of contrast at different spatial frequencies. A square region of the image containing the line-pair test objects is first automatically identified and extracted. The segmented image is then interpolated to double its original matrix size, and re-oriented so that the line-pair test objects align orthogonally with the square image. Line profiles across the line-pairs, averaged over approximately 70% of the length of the test objects, are extracted, from which â€œpeaksâ€? and â€œvalleysâ€? are then identified. The mean values of these signal amplitudes are then used to calculate the contrast at different spatial frequencies. Contrast at 0 lp/mm is estimated by evaluating the histogram of the entire resolution test objects, from which all contrast values are normalized to.
Results: We implemented this method in a Matlab script and applied it to our weekly QC images. The program was able to produce consistent values of MTF for all QC images (including AP abdomen, PA chest and AP foot protocols). Using MTF = 0.1 as the cutoff threshold, the spatial resolution of the tested systems were found to be generally between 3.4 lp/mm (abdomen) to 4.0 lp/mm (foot and lung).
Conclusion: MTF can be conveniently calculated using only routine QC phantom images with line-pair patterns, without resorting to the use of slit or edge test objects.
Spatial Resolution, MTF, Quality Control