Program Information
Avoiding a Common Pitfall in Performing MTF Measurements with High Resolution Kernels
N Rubert*, T Szczykutowicz , F Ranallo , University of Wisconsin, Madison, WI
Presentations
SU-E-I-14 (Sunday, July 12, 2015) 3:00 PM - 6:00 PM Room: Exhibit Hall
Purpose: The modulation transfer function (MTF) is frequently used to assess high contrast spatial resolution for routine quality assurance in CT imaging. High resolution reconstruction kernels provide new challenges for MTF estimation. The purpose of this work is to demonstrate the effect of CT number saturation on MTF estimation.
Methods: A small bead contained in a CATPHAN 600 phantom was imaged at two different slice thicknesses, 2.5 mm and 5.0 mm, on a GE CT 750 HD scanner using an HD Bone Plus reconstruction algorithm in a high resolution scan mode. The resulting images were 2-D point spread functions (PSF's), which were integrated to create line spread functions, and Fourier transformed to yield MTF's.
Results: Using a high resolution reconstruction algorithm and the 2.5 mm slice thickness, it was found that the CT number for several pixels in the PSF reached 3071, the maximum value allowed for this system. Using the same algorithm with the 5.0 mm slice, no pixel values reached the CT number limit. Comparing the two MTF's it was found that the MTF affected by CT number saturation had a 10% cut-off that was shifted downwards, from 15.6 lp/cm to 14.9 lp/cm. Additionally, the MTF affected by CT number saturation exhibited a spurious lobe beyond the true limiting resolution of the CT scanner.
Conclusion: The downwards shift in MTF cut-off is a subtle but important effect to consider when attempting to measure MTF for routine quality assurance. While a downwards shift in MTF cut-off is not easily identified, CT number saturation is readily identified in by examining the entire MTF curve and identifying the spurious lobe produced. CT number saturation is an important consideration when examining test objects to be used for MTF estimation when the algorithm being used is high resolution.
Funding Support, Disclosures, and Conflict of Interest: This work was supported by the AAPM grant for Diagnostic Imaging Residencies
Contact Email: