CT number

Time averaged CT numbers for each sensitometric inserts are displaced in Table 2 . HU of different materials showed a linear relationship (R² = 0.9985) with their respective attenuation coefficients over time (Fig. 2A). Data from the same scanner revealed minor drifts (≤ 8%) from baseline values for all inserts with the  exception of polystyrene, where the deviations (13% for  scanner A and -12% for scanner E) were outside the recommended limits [5] (Fig. 2B). Compared to theoretical values, there was a drift in HU of all inserts irrespective of  scanner type and the largest drifts were found for PMP (≥ -14%), LDPE (≥ -16%), and polystyrene (≥ -34%, Fig. 2C) inserts. 

Slice thickness

The change in slice thickness for all scanners differed only by a maximum of ±0.4 mm compared to the nominal value (Fig. 3A).  Thus, all slice thickness measured were within the tolerance limit (≤ ±1 mm of the nominal value). In spite of the fact that all measured slice thickness were within acceptable limit, scanners A, C and D (same vendor)  showed  5 – 7% deviation from nominal values whereas almost no deviations were found with the other scanners (Fig. 3B). 

Noise and uniformity

In general, image noise and HU uniformity measured over time were within the tolerance limits [2, 3, 5]. Consequently, only minor changes in image noise (< ±10% of baseline values) and HU uniformity (ΔHU < 3 HU) were observed for all scanner types over time (Fig. 4).  With the exception of scanner C and G, relatively large variations in HU uniformity were seen from one measurement to the next despite the fact that  they were within the acceptable limit. 

Low contrast resolution

Temporal variation of  estimated size of the smallest detectable object for 0.5% contrast is displayed in Fig. 5A whereas Fig. 5B shows the time average results for each of the three contrast levels (1%, 0.5%, and 0.3%). Apart from scanners A and D, low contrast resolution at 0.5% was stable over time for all scanners. However, compared with baselines values, low contrast resolution of all contrast levels are within recommended limits [2, 3, 5] regardless of scanner type. 

MFT

 Spatial resolution at 10% MTF for all scanners over time is shown in Fig. 6A, whereas spatial resolutions for all critical frequencies are represented in Fig. 6B. The results show almost stable MTF at all critical frequencies over time irrespective of vendor. Consequently, measured MTFs were within 0.5 lp/cm or ± 10% of the value specified by the vendors.