Performance evaluation of the General Electric eXplore CT 120 micro-CT using the vmCT phantom.

The eXplore CT 120 is the latest generation micro-CT from General Electric. It is equipped with a high power tube and a flat panel detector. It allows high resolution and high contrast fast CT scanning of small animals. The aim of this study was to compare the performance of the eXplore CT 120 with the one of the eXplore Ultra, its predecessor for which the methodology using the vmCT phantom was already described [1].
The phantom was imaged using typical rat (fast scan or F) or mouse (in vivo bone scan or H) scanning protocols. With the slanted edge method, a 10% modulation transfer function (MTF) was observed at 4.4 (F) and 3.9-4.4 (H) mm-1 corresponding to 114 μm resolution. A fairly larger MTF was obtained with the coil method with the MTF for the thinnest coil (3.3 mm-1 ) equal to 0.32 (F) and 0.34 (H). The geometric accuracy was better than 0.3%. There was a highly linear (R2 > 0.999) relationship between measured and expected CT numbers for both the CT number accuracy and linearity sections of the phantom. A cupping effect was clearly seen on the uniform slices and the uniformity-to-noise ratio ranged from 0.52 (F) to 0.89 (H). The air CT number depended on the amount of polycarbonate surrounding the area where it was measured: a difference as high as approximately 200 HU was observed. This hindered the calibration of this scanner in HU. This is likely due to the absence of corrections for beam hardening and scatter in the reconstruction software. However in view of the high linearity of the system, the implementation of these corrections would allow a good quality calibration of the scanner in HU. In conclusion, the eXplore CT 120 achieved a better spatial resolution than the eXplore Ultra (based on previously reported specifications) and future software developments to include beam hardening and scatter corrections will make the new generation CT scanner even more promising.