MedicaMundi

J. Zhang,D.E. Roa, V. Sehgal, Q. He, M.S.A.L. Al-Ghazi -  Department of Radiation Oncology, University of  California Irvine, Orange, CA, USA.

M. Martin - Therapy Physics Inc.Gardena, CA, USA.

Computed tomography (CT) simulators have been widely installed in Radiation Oncology departments. Since CT simulators provide high-resolution organ and tissue structure images, together with the correlation of CT Hounsfield units and electron/physical densities of organs and tissues, CT images have become the standard basis for radiation treatment planning [1-7].

Conventional diagnostic CT scanners usually have a 70 cm physical bore size with a 50 cm maximum imaging field of view (FOV). For radiation oncology purposes, these scanners pose  some limitations. The delivery of fractionated radiation therapy mandates that the patient be placed in a reproducible position over the course of the treatment. To facilitate this, several treatment aids such as breast boards for breast cancer may need to be used.

The use of a big bore scanner for CT simulation facilitates the process of accurate and reproducible patient positioning. Large patients also require the Big Bore CT-simulators, in order to get undistorted images for treatment planning purposes. Some patients need specific immobilization devices during their simulation and treatment process. Consequently, Big-Bore CT simulators offer the best facilities for radiation treatment planning.
 
Recently, a Philips Brilliance Big Bore 16-slice CT-simulator (85 cm bore) (Philips Healthcare, Cleveland OH) has been installed in our department. This article presents a comprehensive study of the performance evaluation of this CT simulator through the acceptance testing and commissioning process, and outlines a quality assurance program consistent with TG-66 criteria [8]. Several phantoms were used in the acceptance testing process, commissioning and quality assurance programs for clinical use. Image quality parameters and mechanical aspects were compared to those specified by the manufacturer and published data. The results are used to demonstrate that this CT simulator can provide high-quality images for use in treatment planning.

The acceptance testing and commissioning process provide baseline values for daily and monthly comprehensive QA purposes. These exceed TG-66 recommendations and appear to be achievable in practice.

Materials and methods

Philips Brilliance Big Bore CT (85 cm bore)
The Brilliance Big Bore CT simulator is a “modified” third-generation scanner (Fan beam; multiple rotating detectors; rotation only without translation; as fast as 0.44 s per rotation) shown in Figure 1.

Figure 1. The Brilliance Big Bore 16-slice CT Simulator (85 cm bore).

The Big Bore CT simulator is a 16-slice scanner with 816 detector elements per row and 24 detector elements along the z-axis. The distance from the focal spot to the imaging isocenter is 645 mm. This is to be compared with the value of the diagnostic CT scanner of 635 mm. The heat capacity of the X-ray tube in the Big Bore simulator is 8.0 MHU. It has scanning field of view (SFOV) ranging from 50 to 600 mm, while an extended display field of view (DFOV) can be selected of up to 700 mm. This facility, designed for CT imaging for radiation treatment planning, can afford patient positioning flexibility. Unlike the diagnostic case, this aperture size is necessary for patient set-up in the treatment position, including required treatment aids.

Slice thickness, as small as 0.75 mm, is achievable for precise delineation of the tumor volume using the small focal spot. Three kVp settings are available: 90, 120, and 140 kVp. Exposure techniques range from 20 to 400 mA with six rotation times for sequential scanning (0.44, 0.5, 0.75, 1.0, 1.5, 2.0 s) and six reconstruction algorithms. The pitch setting ranges from 0.04 to 1.7. The manufacturer-specified spatial resolution is 16 line pairs/cm at 0% modulation transfer function (MTF) for both axial and spiral scanning model. Low-contrast resolution is 3.5 mm at 0.35% contrast. Three sets of coplanar laser systems are used for patient alignment in the treatment position to an accuracy of better than 1 mm.

Phantoms
In this study, several different phantoms were used for the acceptance testing and commissioning for clinical use of the Brilliance Big Bore CT simulator. Table 1 shows each phantom used in this study and its functions. In the subsections below, we will discuss the commissioning process in detail.
Table 2 shows the CT parameters used in this study in detail.

Table 1. Phantoms used for acceptance testing and commissioning for clinical use of the Brilliance Big Bore CT simulator: (a) ACR phantom; (b) CDTI phantom; (c) Philips phantom; (d) CATPHAN 504; (e) RMI phantom. 

Table 2. CT protocol parameters derived from manufacturers’ manual and published data.

CT Simulator acceptance testing and commissioning using ACR phantom
The ACR CT phantom is designed to be an integral part of the American College of Radiology CT Accreditation Program, shown in Table 1(a). This program provides an opportunity for physicians to get peer-review of their CT facility, image quality and quality assurance programs.  The ACR CT phantom was used to evaluate the image quality for CT acceptance testing and clinical commissioning. The acceptance testing includes: slice thickness, field uniformity, low contrast resolution, and high contrast resolution.

The accuracy of slice thickness is defined as the full width at half-maximum CT number, which was evaluated by setting slice thickness as 1.5 mm, 3.0 mm and 6.0 mm using module 1 in the ACR phantom. Field uniformity is used to evaluate the CT image noise and artifacts. CT numbers from Regions of Interest (ROIs) at the center of the image and four peripheral areas are measured and compared to the manufacturer-specific values. Low-contrast resolution is defined as the ability of a system to resolve adjacent objects with small density differences. Two 100 mm2 ROIs were created. The first was over the large cylinder (25 mm diameter) and the second one was over the second largest cylinder. The difference from the average CT numbers of these two ROIs was defined as the measured contrast, which is compared with the manufacturer-specified value. High-contrast resolution is defined as the minimum resolvable distance between two high-contrast objects; that is, the ability of the system to separate small objects that are placed in close proximity. 

Radiation safety and CT dose
The scanner provides shielding for primary radiation. Scatter is the main source of radiation outside the scan plane. CTDI (CT Dose Index) is a measurement of dose delivered during a scan, shown in Table 1(b). The measurement followed the protocol specified AAPM Report 31 [9] and CTDI preliminary report [10]. The Head and Body sections of the CTDI phantoms with diameters of 16 cm and 32 cm respectively were used for the measurement. The material used in these phantoms is polymethylmethacrylate (PMMA). A Radcal 10 × 6-3 CT pencil ion chamber (Radcal, Monrovia, CA) was used for measurements, together with the Keithley electrometer (Keithley Instruments, Inc. Cleveland, Ohio). Measured results were compared with the manufacturer-specified values, as well as data from published literature [4] for CT simulator dose and evaluation of the radiation safety facility.

Daily image quality and laser QA
The vendor’s phantom (Philips Healthcare; Table 1(c)) was used for daily Quality Assurance (QA), i.e. the daily image quality and CT number test.

The Philips phantom is in two parts: a head phantom and a body phantom. Due to its simple design, this phantom is convenient for quick daily image quality QA performed by the radiation therapists. The daily image quality QA includes: Quick Image Quality test (CT number, CT uniformity, CT noise and low contrast resolution), which is automatically performed by vendor software using the head phantom; detailed CT number test using both the head and body phantoms. Quick daily image quality (Quick IQ protocol) QA scan is for weekly QA in our clinic. Laser Position check is to ensure accuracy of patient position during CT-simulation.  It includes LAP laser origin and CT ISO check; Laser Alignments. Three sets of coplanar laser systems are used for patient alignment in the treatment position to an accuracy of better than 1 mm.

Comprehensive monthly QA using the CATPHAN 504
The CATPHAN 504 (The Phantom Laboratory, New York, NY) was used for image quality monthly QA, due to its comprehensive modular design, shown in Table 1(d). Evaluation of image quality parameters includes:

• field uniformity
• CT number linearity
• slice thickness
• low contrast resolution
• high spatial contrast resolution.

The parameters applied are the same as those used with ACR Phantom, except the low-contrast resolution, which was defined as the minimum resolvable diameter of an object embedded in a uniform medium with a small difference in density from the background.

CT images of the CATPHAN 504 phantom were transferred to the department image quality evaluation server, where the IRIS software (The Institute for Radiologic Image Sciences, Inc. [IRIS], Frederick, MD) is installed. IRIS software was used as an automated tool to analyze image quality parameters. In order to compare to the published data, we set similar parameters to those shown in Table 2.

CT number accuracy and linearity evaluation using RMI phantom
Due to its wide density range, the RMI Phantom (Gammex-RMI, Middleton, WI), shown in Table 1(e), was used to evaluate CT number accuracy and linearity with material physical density and electron density. There are 17 inserts of materials with a density range between 0.279 g/cm3 (lung 300) and 4.14 g/cm3 (Titanium). Three data sets were derived with different kVp settings (90, 120 and 140) for use in treatment planning systems employed in the department. The CT number vs. electron density curves were analyzed for each kVp setting, and two separate trend lines were evaluated for each kVp setting with a CT number range larger than 100 HU and smaller than 100 HU, respectively [11].

Results and Discussions

CT dose index
For the CT acceptance testing requirement, the Adult Head CTDI should be less than 75 mGy and the Adult Body CTDI should be less than 25 mGy. In our measurements, these values are 58.9 mGy and 22.8 mGy respectively, which meet the requirement. Table 3 shows the dose results from our Big Bore CT compared to published data for AcQSim [4] using the Head and Body CTDI. CTDI measurements with the Head phantom resulted in doses in the center and surface of 3.1 cGy and 3.49 cGy respectively. For AcQsim CT, these values are 4.8 cGy and 4.9-5.9 cGy respectively. Similarly for the Body phantom, the doses in the center and surface are 1.09 cGy and 2.66 cGy respectively. These are to be compared to the AcQsim values of 1.2 cGy and 2.6-3.3 cGy respectively.

Table 3. Central and surface doses of Head and Body Phantoms.  

Table 4. Quick Image Quality test: row 2 shows measured values, row 3 gives the tolerance value for each test.

Daily image quality QA
Table 4 shows the Quick Image Quality check. The tolerance values of CT number, CT uniformity, CT noise, low contrast are 0 ± 4 HU, 0 ± 4 HU, 3.7 ± 0.4 HU, 4.5 ± 1.2 HU, respectively. These values are based on the manufacturer’s recommendations. Table 5 shows the CT number evaluation for daily QA using Philips head and body phantom. Five different ROI (center and four edges) were evaluated for the body phantom using vendor provided software to obtain CT number information (ROI area, CT number mean value and standard deviation). This is shown in Figure 2.  Head and body phantoms are made of water and nylon and the tolerance value of CT number in the center (ID1) are 0±4 HU and 100±15 HU respectively. The differences of CT numbers at the edges (ID2~ID5) and that in the center should be within 5 HU. In Figure 2, there are two inserts for CT number check of water and Teflon with tolerances 0±4 HU and 920±50 HU respectively.

Figure 2. CT number check using the Philips body phantom.

Table 5. CT number evaluation for daily QA using Philips phantom. Row 1: ROI positions (ID1 –ID5) as shown in Figure 2. Row 2: measured values. Row 3: tolerance value for each test.

Table 5 is from our daily QA results. The values shown are consistent with the manufacturer’s specified reference values. Due to its convenience of use, therapists use the software supplied by the manufacturer to perform daily image quality assurance.

Laser checks follow the AAPM TG 66 recommendations [8]. Figure 3 shows the CT image of laser jig. In this image, the tolerance of laser alignment is 1 mm, and the QA reproducibility of the lasers is about 0.38 mm.

Figure 3. CT image of laser jig.

Figure 4. IRIS software used to analyze image quality from CATPHAN 504.

Monthly QA using CATPHAN 504
The CATPHAN 504 phantom was used together with IRIS software for comprehensive monthly QA purposes, as shown in Figure 4.

Slice thickness accuracy
In this study, we set the slice thickness randomly as 3.00 mm, as shown in Table 2. IRIS software reported that the average slice width is 2.82 mm, which matched the expected value well since the manufacturer’s acceptance tolerance is that the measured slice thickness should be < 1.5 mm of the  prescribed thickness.

CT number accuracy and linearity
Published data show that the acceptable criteria for CT number value for air and water are -1000 ± 3 HU and 0 ± 2 HU respectively [12]. However, due to the differences in measurement conditions, the CT number can change by up to 20 HU, even for water. Therefore, some clinics use this measured value as tolerance for low density materials [4].

Table 6. CT number accuracy and linearity.

Table 6 shows CT numbers of Air, Water, LDPE, Acrylic and Teflon obtained using CATPHAN 504. These results show that the CT scanner meets the QA tolerance. The contrast scale in this study is 0.000196 cm-1/CT # which is comparable with the published value, 0.000206 cm-1/CT #.⁴

Uniformity  
The CT acceptance testing requires that the CT number for water should be 0±5 HU and |Center-Edge|≤5 HU. Based on manufacturer supplied information (CATPHAN 504 Manual, The Phantom Laboratory, Salem, New York). This CT number can shift to about 18 HU, therefore the measured value can be used as the baseline for QA purposes. Table 7 shows that our results meet this requirement. The same conclusion can be derived when comparison is made to published data [4].

Table 7. Uniformity test.

Table 8. Low-contrast resolution (minimum resolvable diameter [mm]).

Low-contrast resolution
Low contrast resolution results from this study and from published data are shown in Table 8. A low contrast resolution of 7 mm at 0.3% for this Brilliance CT scanner /simulator is similar to data published in the literature.

High-contrast resolution
Published data shows the acceptable limit is > 5 lp/cm at the 5% MTF level [12]. The measurement in this study is 7 lp/cm.

Image quality evaluation using the ACR phantom

High contrast resolution
The highest spatial frequency for the ACR phantom was derived using abdomen technique and high resolution chest technique. 8 and 9 lp/cm were visible respectively. These exceed the manufacturer-specified value of 5 lp/cm.

Table 9. Uniformity test using the ACR phantom: adult abdomen.

Imaging field uniformity

Table 9 provides a summary of the imaging field uniformity data, which include the mean CT number from the field center and the four edges (3, 6, 9, and 12 o’clock positions). The CT acceptance testing requirement is that the water CT number should be 0 ± 5 HU and Center-Edge ≤5. The results in this study exceed these recommended values.

Low-contrast resolution

The CT acceptance testing requirement from ACR is that the 6 mm rod should be visible and the contrast should be around 6 ± 0.5 HU. The measured contrast values in this study are 5.9 HU and 6.4 HU using abdomen and head phantoms respectively, which are well within the recommended value.

Table 10. CT number and slice thickness test using module 1 of ACR phantom.

Geometric accuracy: slice thickness

Table 10 shows the CT number and slice thickness test results using module 1 of the ACR phantom. The CT acceptance testing requirement from ACR is that the measured slice width should be ≤1.5 mm of prescribed width. In our evaluation, the maximum slice thickness difference is from the 6.0 mm thickness test, which is 1.25 mm. The results in this study meet the recommended values.

Figure 5. Hounsfield Units (HU) - Electron Density (ED) curves at three different kVp settings.

CT number linearity

The RMI phantom was used for the CT number linearity study in this project (Hounsfield Units (HU) vs. Electron Density). Figure 5 shows the HU vs. Electron Density results relationship. Two different trend lines were generated using dividing point (HU=100) [11]¹. Three different kVp parameters were used in this study. For lower electron density materials (corresponding HU number less than 100), the HU numbers are the same from all three kVp settings. However, for the higher electron density materials (corresponding HU number larger than 100), the HU numbers have differences (up to 20% relative error for cortical bone) due to different kVp settings, which can be derived from the linear equations in Figure 5. The same conclusion was arrived at by the M.D. Anderson Cancer Center group [6]. These results are employed in the treatment planning system for the purpose of inhomogeneity correction.

Results

Comparative image quality analysis: Philips, CATPHAN 504 and ACR phantoms

In this study, the ACR, Philips and CATPHAN 504 phantoms were used for clinical commissioning, daily QA and comprehensive monthly QA, respectively. The results for similar measurements were compared. Table 11 shows the CT number comparison to demonstrate self-consistency of the results.

Table 11.CT number accuracy and linearity (N/A means that there is no corresponding material in that phantom).

Table 12. Slice thickness and high contrast resolution.

Table 12 shows the comparison of slice thickness and high contrast resolution. Slice thickness results were compared between ACR and CATPHAN using 3 mm as the prescribed thickness. The measured values were 2.5 mm and 2.82 mm respectively. The manufacturers’ specified tolerance is that the measured slice width should be < 1.5 mm of prescribed width. The high-contrast resolution of CATPHAN 504 and ACR Phantom were 7 lp/cm and 8 lp/cm respectively.

A comparison of these results demonstrates the self-consistency of the characteristics of the CT simulator evaluated using several different phantoms.

Conclusions

A Philips Brilliance Big Bore CT simulator has been installed in our radiation oncology department. Several phantoms were used in this study for CT simulator acceptance testing, clinical commissioning and comprehensive quality assurance. The measured data demonstrate that this Big Bore CT simulator meets the clinical requirements. Our scanner shows characteristics similar to those published in the literature for the Big Bore CT Scanner. It is well suited for use in a radiation oncology setting.

Acknowledgements

The authors acknowledge the support of the clinical faculty and staff of the Department of Radiation Oncology. Philips personnel provided valuable technical assistance during the installation, acceptance testing and commissioning process.

References

[1]  Knöös T, Nilsson M, Ahlgren L. A Method for Conversion of Hounsfield Number to Electron Density and Prediction of Macroscopic Pair Production Cross-Sections. Radiother Oncol. 1986; 5: 337–345.

[2]  Mcgee KP, Das IJ, Sims C. Evaluation of Digitally Reconstructed Radiographs (DRRs) used for Clinical Radiotherapy: A Phantom Study. Med Phys. 1995; 22: 1815–1827.

[3]  Cozzi L, Fogliata A, Buffa F, Bieri S. Dosimetric Impact of Computed Tomography Calibration on a Commercial Treatment Planning System for External Radiation Therapy. Radiother Oncol. 1998: 335–338.

[4]  Garcia-Ramirez JL, Mutic S, Dempsey JF, Low DA, Purdy JA. Performance Evaluation of an 85-cm-Bore X-Ray Computed Tomography Scanner designed for Radiation Oncology and Comparison with Current Diagnostic CT Scanners. Int J Radiat Oncol Biol Phys. 2002; 52: 1123–1131.

[5]  Saw CB, Loper A, Komanduri K, Combine T, Huq S, Scicutella C. Determination of CT-to-Density Conversion Relationship for Image-Based Treatment Planning Systems. Med. Dosimetry. 2005;  30: 145-148.

[6]  Kendall RL, Gifford KA, Kirsner SM. The Impact of Peak-Kilovoltage Settings on Heterogeneity-Corrected Photon-Beam Treatment Plans. Radiotherapy and Oncology. 2006; 81: 206-208.

[7]  Coolens C, Breen S, Purdie T, Jaffray D. Implementation and Characterization of a 320-Slice Volumetric CT Scanner for Simulation in Radiation Oncology. Med Phys. 2009; 36: 5120-5127.

[8]  Mutic S, Palta JR, Butker EK, et al. Quality Assurance for Computed-Tomography Simulators and the Computed-Tomography-Simulation Process: Report of the AAPM Radiation Therapy Committee Task Group No. 66. Med Phys. 2003; 30: 2762–2792.

[9]  Chu RYL, Fisher J, Archer BR, et al. Standardized Methods for Measuring Diagnostic X-Ray Exposures. AAPM Report 31, 1990.

[10] Computed Tomography Dose Index Preliminary Report. ACQSIM-CT Part#T55B-1440 Rev-2, Marconi Medical Systems, 2000.

[11]  Battista JJ, Bronskill MJ. Compton Scattering Imaging of Transverse Sections: An Overall Appraisal and Evaluation of Radiotherapy Planning. Phys Med Biol. 1981; 26: 81-99.

[12] Mccollough CH, Zink FE. Quality Control and Acceptance Testing of CT Systems. In: Goldman LW, Editor. Medical CT And Ultrasound: Current Technology and Applications. AAPM Summer School 1995.