LETTER TO EDITOR
|Year : 2016 | Volume
| Issue : 2 | Page : 153-156
Assessment of some image quality tests on a 128 slice computed tomography scanner using a Catphan700 phantom
Eric Naab Manson1, John Justice Fletcher2, Vivian Della Atuwo-Ampoh3, Eric K.T. Addison4, Cyril Schandorf5, Luc Bambara1
1 Department of Medical Physics, School of Nuclear and Allied Science, University of Ghana-Atomic Campus, Navrongo, Ghana, West Africa
2 Department of Applied Physics, Faculty of Applied Sciences, University for Development Studies, Navrongo, Ghana, West Africa
3 Department of Oncology, Komfo Anokye Teaching Hospital, Kumasi, Ghana, West Africa
4 Department of Physics, Kwame Nkrumah University Science and Technology, Kumasi, Ghana, West Africa
5 Department of Radiation Protection, School of Nuclear and Allied Science, University of Ghana-Atomic Campus, Navrongo, Ghana, West Africa
|Date of Web Publication||3-May-2016|
Eric Naab Manson
School of Nuclear and Allied Sciences, University of Ghana – Atomic Campus, P. O. Box LG80, Legon, Ghana
Source of Support: None, Conflict of Interest: None
|How to cite this article:|
Manson EN, Fletcher JJ, Atuwo-Ampoh VD, Addison EK, Schandorf C, Bambara L. Assessment of some image quality tests on a 128 slice computed tomography scanner using a Catphan700 phantom. J Med Phys 2016;41:153-6
|How to cite this URL:|
Manson EN, Fletcher JJ, Atuwo-Ampoh VD, Addison EK, Schandorf C, Bambara L. Assessment of some image quality tests on a 128 slice computed tomography scanner using a Catphan700 phantom. J Med Phys [serial online] 2016 [cited 2021 Dec 1];41:153-6. Available from: https://www.jmp.org.in/text.asp?2016/41/2/153/181637
Routine quality control (QC) procedures in computed tomography (CT) centers in most African countries are not given much attention due to lack of access to newer technology and no/limited certified professional medical physicists trained to carry out this task. This has rendered the services of medical physicist generally limited especially in West Africa than the rest of the world. The objective of this letter is to adequately provide information on some image quality tests on CT systems to medical physicist/radiographers, in order for them to perform QC tests with good confidence. This letter describes some essential image QC procedures performed at CT diagnostic facilities using Catphan700 phantom. The need for this is pronounced due to the increase in design capabilities and extended applications of CT scanners  and the increasing number of CT facilities coupled with a shortage of qualified professionals in West Africa.
These advancements in CT technology have raised many concerns about the differences in the quality of images produced by these scanners. An image that contains all the information needed for correct diagnosis of a patient disease condition is said to be a quality image. The quality of an image is affected by composite factors such as contrast, spatial resolution, noise, blur, artifacts, and distortion.,
Quality assurance for single slice CT (SSCT) and multi-slice CT (MSCT) scanners usually consists of some basic required elements of testing such as contrast scale, CT number, high-contrast resolution, low contrast resolution, image noise, uniformity, and artifacts; and other tests such as laser light alignment and accuracy, slice thickness and localization, and patient dose., Basically, the primary difference between the SSCT and MSCT lies with the design of the detector arrays.
Garayoa and Castro  performed a study to evaluate the image quality on a cone beam CT scanner by determining various physical parameters that characterize a system's performance, using Catphan phantom. Catphan700 phantom is a diagnostic imaging tool specially designed for comprehensive evaluation of axial, spiral, multi-slice, cone beam, and volume CT scanners from the point of view of maximum performance.
The scanner investigated is an Aquilion 128 slice CT scanner manufactured by Toshiba Company (Toshiba Aquilion 128 CXL Edition, manufactured in the city of Otawara-shi, located in Tochigi state in the country of Japan). In this letter, some parameters that characterize a system's image quality performance have been assessed and their effects on image quality discussed. These include spatial linearity, pixel size, CT number, spatial resolution, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR).
The initial test carried out was to verify the accuracy of distance measurement and pixel size on the image. The aim of the distance measurement is to ensure that the displayed image represents the true phantom image. The CTP682 module of the phantom was used to evaluate the accuracy of distance measurement by measuring the horizontal (x) and the vertical (y) dimensions using the distance tool in the DICOM viewer. Furthermore, the pixel size was evaluated to verify whether the image pixel size agrees with the nominal value. The number of pixels on the image display was determined by counting the pixels along both horizontal and vertical measured lines. By knowing the distance and the number of pixels along the distance, the pixel size was calculated using Equation 1.
The nominal value of the phantom dimensions in the horizontal (x) and vertical (y) direction is 150.5 mm. The x and y value for the acquired image of the CT system was measured to be 150.2 mm each. This presents an error percentage of 0.19 from the nominal value. It can be concluded that the displayed image properly represents the phantom dimension with insignificant variation observed. Furthermore, the expected pixel size is estimated to be 0.49 mm, while the measured pixel size value calculated on the image display from the study was found to be 0.44 mm ± 0.1 mm (variation range: 0.08 and 0.12mm).
The second test was on CT number accuracy. The CT number measurement accuracy was obtained from the image of module CTP682 for the different sensitometry targets located in the phantom. These targets include Teflon, bone 50%, Delrin, bone 20%, acrylic, polystyrene, low-density polyethylene, polymethylpentene, lung foam #7112, and air. A 1 cm 2 region of interest (ROI) was selected on each target using DICOM viewer and the mean CT number value determined on each of the targets. The mean CT number of each material was then compared to the actual CT number range from the phantom specifications.
The measured CT numbers in the study for the various targets were all within the minimum and maximum range as specified by the manufacturer except for the Delrin and bone 20% which were found to be slightly outside the range. The deviations may be due to the variation in the density and composition of the materials, the type of scanner as well as the scanning parameters used. On average, the estimated CT values for all the other targets are within the range of phantom specification as shown in [Table 1].
|Table 1: Measured computed tomography numbers of various materials and their computed tomography number range as specified by the manufacturer|
Click here to view
To establish a constancy of contrast scale over the range of CT numbers which is of clinical interest, the CT linearity was verified. This was performed by checking whether the CT numbers measured vary in a linear fashion with their linear attenuation coefficient values. Clinically, CT number relevant linearity was observed with r2 = 0.973 close to unity, despite the slight deviation of CT mean number values from the line of best fit, especially as observed with bone 50% in [Figure 1].
|Figure 1: The measured computed tomography numbers of the target materials plotted against linear attenuation coefficients for energy of 120 kV|
Click here to view
The third test was on measurement of spatial resolution using modulation transfer function (MTF). There are several methods that have been described to calculate the MTF of imaging systems, based on the slit, edge, or bar pattern images. In this letter, the MTF was calculated based on the method described in the Catphan700 manual. By obtaining the pixel values surrounding the image of a 0.18 mm tungsten carbide bead in the CTP682 module, the line spread function (LSF) along the x-and y-axes were determined by measuring the average pixel values of point spread function in the horizontal and vertical directions respectively, for four routine scanning protocols [Table 2]. The MTF curve was then computed by taking the Fourier transform of the LSF data using Matrix Laboratory (MATLAB) software (MATLAB software version R2012b, RadiAnt DICOM Viewer, Poznan, Poland).
The 10% MTF results of the study in [Table 2] reflect the characteristics of the manufacturer specifications (6.5 ± 0.68) of CT system. The CNR clearly depends on both the SNR and MTF. An increase in SNR results in an increase in CNR and spatial resolution. In addition, since the MTF increases with an increase in the type of filter from FC8 to FC23, it can be concluded from the study that, the spatial resolution of CT images can be increased through further enhancement of reconstructed images by filtering. The spatial resolutions as expressed in terms of MTF for the four scan protocols are within the specifications set by the manufacturer.
The fourth test was SNR calculated from reconstructed images of module CTP712 of the Catphan700 phantom. By selecting four different ROIs on each image, the SNR was calculated as the ratio of average pixel value, avep to the standard deviation of the pixel values, σp between the ROIs [Equation 2].
The SNR was calculated from reconstructed images using three different convolution filters. These filters which are used in smoothing or enhancing images of high frequencies are described as FC8 (sharp), FC18 (medium), and FC23 (smooth). Noise depends on the filter function F (and can be reduce with smooth filter kernel). It was found that, as the filter sharpness is increased from FC8 to FC23, the SNR increases. Increase in mAs decreases noise which increases the spatial resolution element as shown in [Table 2]. The high value of SNR as observed from analysis of the image obtained using the head scan protocol is due to the high mAs and smooth filter used. This is contrary to the thorax protocol with a low value of SNR due to low mAs and sharp filter. Although not obvious from the table, it may be added here that there would be excessive noise if size of the matrix and FOV used are not appropriate; also, noise varies with slice thickness (h) as (h)−½. Further work is planned to determine the variation of range of values of noise with mA, spatial resolution, and slice thickness.
The final test on the assessment was CNR. The CNR was determined from scan images of module CTP515 of the phantom. By taking the mean pixel values of ROIs on eight different contrast targets and backgrounds, the CNR was calculated from Equation 3 below;
Where, xT, xB, and σN are mean pixel values of the contrast targets, mean pixel values of backgrounds and standard deviation of the noise, respectively. The results indicate the visibility of all the contrast targets relative to the background at 1% nominal contrast level. The thorax routine scan protocol demonstrated a low CNR with a mean deviation of ± 0.43 in comparison to the routine head, abdomen, and pelvis protocol. This difference may be attributed to the increase of noise produced at low mAs. However, it is difficult for one to tell exactly which parameters directly have an influence on the CNR. Results of CNR vary at different scan routine times with inconsistent values observed especially between the thorax and abdomen scan protocol. Despite the inconsistency, it was realized that the CNR increases with the reconstruction filter [Table 2]. On an average, all values of CNR calculated for the four scan protocols are within tolerance of ± 0.5 as specified by manufacturer.
In this study, some image quality parameters of a CT scanner have been analyzed from acquired images of Catphan700 phantom using four default scan techniques. The analysis was performed using DICOM and MATLAB software. In general, results of the CT images obtained from analysis reveal that the CT system is adequate for accurate diagnostic purposes. It must be noted that these are preliminary tests from image quality test only. For quality assurance which includes type approval and acceptance/constancy testing, more comprehensive tests are needed. It is planned to carry out further work and publish the complete results. The methods described in this study can be used for future routine image QC assessment of CT systems.
This study was supported jointly by the Department of Medical Physics from the Graduate School of Nuclear and Allied Sciences - University of Ghana and Ghana Atomic Energy Commission.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Rae WI. Status of Education and Training in Africa: Focus on South Africa. World Congress on Medical Physics and Biomedical Engineering, 7-12 September, 2009, Munich, Germany. Berlin, Heidelberg: Springer; 2010.
Hendee WR, Chaney EL, Rossi RP. Radiologic Physics, Equipment and Quality Control. Chicago: University of Michigan, Digitized 2008, Year Book Medical Publishers; Inc.; 1977. p. 189-221.
ACR Technical Standard for Diagnostic Medical Physics Performance Monitoring of Computed Tomography (CT) Equipment. ACR-American College of Radiology Web Site Revised 2002 (Res. 21) Effective 1/1/03. Available from: http://www.acr.org
. [Last accessed on 2004 Jun 08].
Ballinger PW, Frank ED. Merrill's Atlas of Radiographic Positions and Radiologic Procedures. 10th
ed., Vol. 3. Missouri: Mosby; 2003. p. 330-71.
Goldman LW. Principles of CT: Multislice CT. J Nucl Med Technol 2008;36:57-68.
Garayoa J, Castro P. A study on image quality provided by a kilovoltage cone-beam computed tomography. J Appl Clin Med Phys 2013;14:3888.
American Association of Physicists in Medicine, Diagnostic Radiology Committee, Judy PF. Phantoms for Performance Evaluation and Quality Assurance of CT Scanners. Chicago, Illinois: American Association of Physicists in Medicine; 1977.
Buhr E, Günther-Kohfahl S, Neitzel U. Simple method for modulation transfer function determination of digital imaging detectors from edge images. In: Medical Imaging. San Diego, CA: International Society for Optics and Photonics; 2003. p. 877-84.
[Table 1], [Table 2]
|This article has been cited by|
||Deep Learning Versus Iterative Reconstruction for CT Pulmonary Angiography in the Emergency Setting: Improved Image Quality and Reduced Radiation Dose
| ||Marc Lenfant, Olivier Chevallier, Pierre-Olivier Comby, Grégory Secco, Karim Haioun, Frédéric Ricolfi, Brivaël Lemogne, Romaric Loffroy |
| ||Diagnostics. 2020; 10(8): 558 |
|[Pubmed] | [DOI]|