Automated procedure for slice thickness verification of computed tomography images: Variations of slice thickness, position from iso‐center, and reconstruction filter

Abstract Purpose The purpose of this study is to automate the slice thickness verification on the AAPM CT performance phantom and validate it for variations of slice thickness, position from iso‐center, and reconstruction filter. Methods An automatic procedure for slice thickness verification on AAPM CT performance phantom was developed using MATLAB R2015b. The stair object image within the phantom was segmented, and the middle stair object was located. Its angle was determined using the Hough transformation, and the image was rotated accordingly. The profile through this object was obtained, and its full‐width of half maximum (FWHM) was automatically measured. The FWHM indicated the slice thickness of the image. The automated procedure was applied with variations in three independent parameters, i.e., the slice thickness, the distance from the phantom to the iso‐center, and the reconstruction filter. The automated results were compared to manual measurements made using electronic calipers. Results The differences of the automated results from the nominal slice thicknesses were within 1.0 mm. The automated results are comparable to those from manual approach (i.e., the difference of both is within 12%). The automatic procedure accurately obtained slice thickness even when the phantom was moved from the iso‐center position by up to 4 cm above and 4 cm below the iso‐center. The automated results were similar (to within 0.1 mm) for various reconstruction filters. Conclusions We successfully developed an automated procedure of slice thickness verification and confirmed that the automated procedure provided accurate results. It provided an easy and effective method of determining slice thickness.

the decision on treatment. [7][8][9] These parameters include the reconstruction field of view (FOV), the effective mAs, the reconstruction algorithm, beam collimation, and slice thickness. 9,10 Slice thickness is one of the important parameters, and it has to be optimized as needed. Slice thickness affects the cross-plane resolution of the clinical image, which then impacts the accuracy of the size determination of the organ. 11,12 The slice thickness also directly impacts image noise. Decreasing the reconstructed slice thickness increases image noise. 13 To compensate for increased noise, the operator may choose to increase the mAs (dose) to the patient. 14 The accuracy of slice thickness determination has been investigated in previous studies using various phantoms. [15][16][17][18][19][20] In the AAPM CT performance phantom, slice thickness is measured as the thickness of a stair object using electronic calipers. [21][22][23] A more objective measurement of slice thickness can be achieved by determining the full-width at half maximum (FWHM) of the pixel profile across the stair objects. 24 However, this manual approach is tedious and time-consuming. An automated procedure would increase the measurement speed and objectivity. An automated procedure for slice thickness verification on the AAPM CT performance system utilizing MATLAB software was previously proposed by Sofiyatun et al. 21 They showed that automated procedure can produce a more accurate estimate than manually calculated results. 21 However, the study was only conducted for one slice thickness value, i.e., 5 mm. In this paper, we validated automated slice thickness results for various slice thicknesses, phantom positions from the iso-center, and reconstruction filters.

2.A | CT scanner and phantom
The study was conducted at the Radiological Installation of the Diponegoro National Hospital (RSND), using a Philips Ingenuity 128-slice CT scanner [ Fig. 1(a)] and the AAPM CT performance phantom (Model 610, CIRS, Virginia, US) [ Fig. 1(b)]. The objects for slice thickness measurement were aluminum plates each of size 0.635 mm × 25.4 mm, surrounded by water. Figure 1(c) shows an axial image of the phantom. The AAPM CT performance phantom was scanned with three different variables: nominal slice thickness, position from the iso-center, and reconstruction filter. The respective acquisition parameters are listed in Table 1.

2.B | Automated measurement
The automated procedure for slice thickness verification was carried out utilizing a program developed in MATLAB R2015b. Figure 2 shows the program workflow for processing each image to obtain the FWHM value of the slice thickness. 21 After segmenting, cropping and rotating the objects by an angle , and converted to mm using the DICOM header conversion factor. All these steps were performed automatically by tapping a single button. 21 The measurements for every variation were conducted on five frames, and the averages and standard deviations were calculated.

2.C | Manual measurement
The automated results of slice thicknesses were compared to manual measurements. Manual measurement of the slice thickness was carried out on IndoseCT software. 25 Figure 3 shows the manual measurement based on our zoomed-in view.  Table 2.

3.A | Slice thickness variation
The differences between the automated and manual results are less than 12%. The differences between the automated results and the nominal slice thicknesses are within 1.0 mm.

| DISCUSSION
This study aims to develop and validate an automated procedure for the slice thickness verification from an AAPM CT performance phantom so that easier and more effective pre-treatment measurement can be made. The automated procedure for slice thickness verification uses the thickness of the stair objects in an axial image. 21 Increase thickness of the stairs results in a wider slice thickness. 22 Previously, an automated slice thickness determination was proposed and implemented on one nominal slice thickness of 5 mm. 21 This current study further validated the algorithm for several F I G . 7. Stair object's pixels profiles and FWHM values for different phantom positions to the iso-center: (a) center of iso-center, (b) 2 cm above the iso-center, (c) 4 cm above the iso-center, (d) 2 cm below the iso-center, and (e) 4 cm below the iso-center. variations, i.e., slice thickness, position from iso-center, and image reconstruction filter. We found that all nominal slice thicknesses from 1 to 5 mm, the differences between the automated results, and the nominal slice thicknesses are within the tolerance limit, i.e., 1.5 mm. 17,26 The current study confirmed that the results from the automated procedure are independent of position of the phantom from isocenter. This suggests that users do not need to precisely locate the phantom in order to measure the slice thickness from the resulting image. This not only helps speed up image acquisition but also simplifies the image acquisition process. Another finding of the current study is that automated slice thickness results are not affected by the reconstruction filter used. This is different from the manual approach where the results may be affected by the reconstruction filter used, because user subjectivity in locating the border of the stair object may depend on the reconstruction filter.
The differences between the automated slice thickness results and manual results were within 12%. We found that the automated results were accurate, i.e., differences of less than 1 mm between them and the nominal slice thicknesses. Our automated method will be helpful in conducting a more convenient slice thickness verification. However, we need to validate its accuracy for different FOVs in a further study.
In this study, we focused on the middle stair object, assuming that its slice thickness value is no different from other two stair objects. The automated results of slice thickness may be affected by noise level, mode of acquisition (i.e., step and shot or helical modes), and pitch factor. All these parameters need to be investigated in future studies.
Apart from the slice thickness, the AAPM CT performance phantom has modules for measuring other CT performance parameters, such as noise, linearity of CT number, beam hardening, spatial inplane resolution, low contrast, and so on. 27,28 Developing an automated system for these parameters would greatly assist medical physicists in carrying out routine quality control.

| CONCLUSIONS
We have proposed and validated the algorithm for an automated procedure for slice thickness verification on an AAPM CT performance phantom. We validated it for variations of slice thickness, positions from iso-center, and reconstruction filter. The automated results are accurate, differing from the nominal thickness by less than 1.0 mm for slice thicknesses from 1 to 5 mm, for various positions, and for various reconstruction filters.