Comparison of MTF Measurement Methods in CT Images for Various Reconstruction Kernels

Abstract

This study aimed to compare several methods of measuring the modulation transfer function (MTF) for different reconstruction kernels, using a point phantom, a small-circular (S-circular) Teflon object, and the automated edge of a PMMA phantom. The copper wire section of a phantom was used for the point method. The small-circular (S-circular) teflon object within the HU linearity section was used for measuring MTF using ImQuest software. The automated edge of a PMMA phantom was used to automatically measure the MTF. The three methods were implemented in images reconstructed with ten different kernels. It was found that the three methods produced comparable MTFs for all the kernels used. However, the automated edge of the PMMA phantom produced slightly smaller spatial resolutions compared with the two other methods. The differences between the automated edge of PMMA and the point method were small, i.e. 0.04 cycle/mm for both 10% MTF and 50% MTF. The differences between the automated edge of PMMA and the S-circular phantom were 0.05 cycle/mm and 0.03 cycle/mm for 10% MTF and 50% MTF. We found that the “UA” kernel produced the lowest spatial resolution values of 0.32, 0.33, and 0.31 cycle/mm of 50% MTF for point, S-circular object, and automated edge PMMA, respectively. The “YD” produced the highest spatial resolution values of 0.78, 0.76, and 0.67 cycle/mm of 10% MTF for point, S-circular object, and automated edge PMMA, respectively. We successfully compared three methods of MTF measurement. The three methods produce comparable MTFs, so that each method can be used for accurately measuring MTF depending on phantom and software available in the CT center.