Quality Assurance
143accuracy of the method for scatter correction, because correction techniques
cannot compensate on the noise introduced by the unwanted events and can potentially add bias to the image. The scatter fraction (SF) is defined as the ratio of the scattered events to the total events, which are measured at a sufficiently low counting rate that random coincidences, dead time effects, and pileup are negligible. Total events are the sum of the unscattered events and scattered events.
The phantom is a 20 cm diameter solid polyethylene cylinder with an overall length of 70 cm. The phantom has a hole at 4.5 cm from its centre, which goes through the whole phantom, parallel to its central axis. In the hole a Teflon LS (ID 2.3 mm), as long as the phantom, can be inserted to contain radioactivity. The LS is filled in its 70 cm central part with a solution of water and 18F. Two tests are performed, in 2D and 3D, with an initial activity of 2,664 MBq and 1,776 MBq, respectively. In both tests, the phantom is positioned at x=0 cm, y=0 cm and axially centred in the scanner FOV. In both acquisitions, random coincidences are measured by the delayed event (DE) technique. In each test (2D and 3D), 22 EM frames are acquired. Frames 1 to 4 are acquired for 900 s with no delay between consecutive frames. The remaining 18 frames are acquired for 1,500 s with a delay of 1,500s between each consecutive pair of frames.
Raw data sinograms are used in the analysis of the scatter fraction and count rates (SF & CR) test. 3D sinograms are rebinned by using a single slice-rebinning algorithm. Scatter component is calculated as for the N-94 over a fixed FOV of 40 cm diameter. Only the final frames of the SF & CR test (when the random rate was negligible, below 1%) are used to calculate the scatter fraction. Scatter fraction is calculated as the ratio between scatter component and total events. The total counts rate within a 24 cm transverse FOV is determined as a function of the radioactivity decay. The T rate (Rtrues) is determined by subtracting the random and scatter rates from the total prompts event rate (Rtotal). The Noise equivalent count (NEC) rate is calculated as follows:
Two NEC rate curves were then generated, with k=1 and k=2.
Accuracy of Corrections for Count Losses and Random The same data set as was acquired for the SF & CR test is used. Data are reconstructed with all count rate dependent corrections (dead-time losses and random coincidences) applied. A circular ROI (18 cm diameter) is drawn, centred on the reconstructed images of the phantom. 2D and 3D data sets are reconstructed as for the count rate accuracy test of the N-94.
The resulting dead-time and random corrected true image count rate (R)
144
was plotted as a function of activity concentration. The residual dead-time error Δ is calculated as follows:
% Δr = 100[1 – (R ÷ Rextrap)]
where Rextrap is the linear function of the true count rate extrapolated from the low count rate (where dead-time and random coincidences rate are negligible).
Image Quality—Attenuation and Scatter Correction Accuracy
It is desirable to compare the image quality of different imaging systems for a standardized imaging situation that stimulates a clinical imaging condition. In this, the scanners image contrast and signal noise ratios are tested under conditions that stimulate a clinical whole body study.
Overall image quality (IQ), as well as attenuation and scatter correction accuracy, is evaluated using a phantom simulating a human torso in size and shape. The IQ phantom contains six coaxial isocentre spheres with diameters of 1.0, 1.3, 1.7, 2.2, 2.8 and 3.7 cm. A cylindrical insert of 5 cm diameter, as long as the phantom, is also positioned in the centre of the phantom. The cylinder is a cold insert with a density of 0.30 g/cc to simulate the lungs. Four of the spheres, with diameters of 1.0, 1.3, 1.7 and 2.2 cm, are used to simulate hot lesions, while the other two are used to simulate cold lesions. The phantom is filled with a solution of water and 18F (5.3 kBq/cc), and the spheres with a concentration eight times higher than the background, to simulate a lesion to background (L/B) ratio of 8.
In a second experiment, radioactivity concentration in the hot spheres is such that the L/B is 4. Once filled, the phantom is positioned with the spheres both in the transverse plane and along the z-axis of the scanner FOV. To simulate body activity from outside of the scanner FOV, the phantom used for the SF & CR test is positioned at one edge of the IQ phantom. For this test, the LS of the external phantom is filled with an activity of 165.5 MBq.
A CT scan of the phantom is used for acquisition (140 kV, 90 mA). For both the experiments (L/B=8 and 4), six interleaved acquisitions (2D and 3D) are performed. The acquisition time for each 2D and 3D measurement is 8 min and 20s and 7 min and 19 s respectively. These times are derived, based on a whole body examination designed to cover a 100 cm axial FOV in 60 min, using a slice overlap for 2D and 3D mode of 5 and 11 slices, respectively.
Data were corrected for random coincidences, geometry, normalization, dead-time losses, scatter and attenuation. In order to evaluate the hot and cold sphere contrast, circular ROIs with a diameter equal to the physical size of each sphere are drawn on CT images and copied to PET images.
Twelve background ROIs (37 mm diameter) are drawn on the central slice
145 and on slices ± 10 mm and ± 20 mm from the central slice. ROIs of smaller size (10, 13, 17, 22, 28 mm) are drawn concentric to the 37 mm background ROIs. Finally, an ROI of 5 cm in diameter is drawn (in each slice of the phantom) on the central cylindrical insert to assess the accuracy of the attenuation and the scatter correction. Different parameters used to evaluate the IQ test are:
i. The hot sphere contrast recovery coefficient (HC_RC), ii. The cold sphere contrast (CC),
iii. The accuracy of attenuation and scatter correction (ΔA Clung), and iv. The background variability (BVj), are calculated as follows:
HC RC C C
where Chot and Cbkgd are the average of the counts measured in the hot spheres ROI and the average counts in all background ROIs respectively, while ahot/abkgd is the ratio of the activities in the hot sphere and background.
CC C
where Ccold is the average of the counts measured in the cold spheres ROI.
where Clung is the average counts in the lung insert ROI.
BV SD
where SDj is the standard deviation of the background ROI counts for sphere j.
Performance Evaluation of CT
Performance evaluation tests of CT includes (i) electromechanical, (ii) image quality, and (iii) radiation safety. Detailed information are also available in AAPM report No. 39 (9), AAPM-TG 66 (10) recommendations.
Electromechanical Tests
These tests include the congruence of gantry laser and imaging plane, localization of CT and pseudo CT centre, orthogonality of table top long axis to imaging plane, accuracy of table vertical and longitudinal movement, radiation and sensitivity profile widths and tests on X-ray generator.
146
Gantry and Couch
Congruence of gantry laser with centre of imaging plane and gantry tilt accuracy are verified using ready pack film(Kodak-X-V) adopting the method described in AAPM -39(9). CT center and pseudo CT center an arbitray point exactly 60 cm inferior from CT centre are localized using a commercially available laser calibration phantom. For this purpose 0.1 cm thick tranverse images are acquired at the mid plane of two parallel slabs of the phantom separated by 60 cm.
This test is also used to quantify longitudinal table motion accuracy and orthogonality of table top longitudinal axis to the image acquisition plane.
Calibrations of table linear scales are verified by moving the table both vertically and longitudinally in steps of 1 cm using an independent measuring scale. Table indexing accuracy and reproducibility are tested by irradiating a ready back film placed perpendicular to the scan plane, under scanner control longitudinal spacing of 0, 5, 10, 20 and 30 cm using 0.1 cm slice thickness.
Radiation Sensitivity and Profile Widths
A ready pack film placed horizontal to the table top and at the CT centre is exposed using all available slice thickness. The exposed films are measured using film scanner with 0.01 cm step size and the FWHM of optical density profiles corresponding to every slice thickness is obtained. These FWHM values represent the radiation profile widths. Independent verification profile width has to be done by using vendor supplied phantom.
X-ray Generator
Tests on the X-ray generator include evaluation of peak potential (kVp), timer accuracy(s),mAs linearity and repeoducibility. Non invasive measurement of kVp for different mAs are performed using suitable meters and the method adopted by AAPM-39(9). Linearity of mAs for different kVp is verified by obtaining the product of dose and time at different mA and time settings.