Implications of CT noise and artifacts for quantitative 99mTc SPECT/CT imaging K. W. Hulme and S. C. Kappadath Citation: Medical Physics 41, 042502 (2014); doi: 10.1118/1.4868511 View online: http://dx.doi.org/10.1118/1.4868511 View Table of Contents: http://scitation.aip.org/content/aapm/journal/medphys/41/4?ver=pdfcov Published by the American Association of Physicists in Medicine Articles you may be interested in Development and evaluation of convergent and accelerated penalized SPECT image reconstruction methods for improved dose–volume histogram estimation in radiopharmaceutical therapy Med. Phys. 41, 112507 (2014); 10.1118/1.4897613 An automated voxelized dosimetry tool for radionuclide therapy based on serial quantitative SPECT/CT imaging Med. Phys. 40, 112503 (2013); 10.1118/1.4824318 Simultaneous 99mTc-MDP/123I-MIBG tumor imaging using SPECT-CT: Phantom and constructed patient studies Med. Phys. 40, 102506 (2013); 10.1118/1.4820977 Quantitative simultaneous 111In/99mTc SPECT-CT of osteomyelitis Med. Phys. 40, 082501 (2013); 10.1118/1.4812421 Diminishing the impact of the partial volume effect in cardiac SPECT perfusion imaging Med. Phys. 36, 105 (2009); 10.1118/1.3031110
Implications of CT noise and artifacts for quantitative 99m Tc SPECT/CT imaging K. W. Hulme and S. C. Kappadatha) Department of Imaging Physics, The University of Texas M.D. Anderson Cancer Center, Houston, Texas 77030 and The University of Texas Graduate School of Biomedical Sciences, Houston, Texas 77030
(Received 10 October 2013; revised 26 February 2014; accepted for publication 4 March 2014; published 20 March 2014) Purpose: This paper evaluates the effects of computed tomography (CT) image noise and artifacts on quantitative single-photon emission computed-tomography (SPECT) imaging, with the aim of establishing an appropriate range of CT acquisition parameters for low-dose protocols with respect to accurate SPECT attenuation correction (AC). Methods: SPECT images of two geometric and one anthropomorphic phantom were reconstructed iteratively using CT scans acquired at a range of dose levels (CTDIvol = 0.4 to 46 mGy). Resultant SPECT image quality was evaluated by comparing mean signal, background noise, and artifacts to SPECT images reconstructed using the highest dose CT for AC. Noise injection was performed on linear-attenuation (μ) maps to determine the CT noise threshold for accurate AC. Results: High levels of CT noise (σ ∼ 200–400 HU) resulted in low μ-maps noise (σ ∼ 1%–3%). Noise levels greater than ∼10% in 140 keV μ-maps were required to produce visibly perceptible increases of ∼15% in 99m Tc SPECT images. These noise levels would be achieved at low CT dose levels (CTDIvol = 4 μGy) that are over 2 orders of magnitude lower than the minimum dose for diagnostic CT scanners. CT noise could also lower (bias) the expected μ values. The relative error in reconstructed SPECT signal trended linearly with the relative shift in μ. SPECT signal was, on average, underestimated in regions corresponding with beam-hardening artifacts in CT images. Any process that has the potential to change the CT number of a region by ∼100 HU (e.g., misregistration between CT images and SPECT images due to motion, the presence of contrast in CT images) could introduce errors in μ140 keV on the order of 10%, that in turn, could introduce errors on the order of ∼10% into the reconstructed 99m Tc SPECT image. Conclusions: The impact of CT noise on SPECT noise was demonstrated to be negligible for clinically achievable CT parameters. Because CT dose levels that affect SPECT quantification is low (CTDIvol ∼ 4 μGy), the low dose limit for the CT exam as part of SPECT/CT will be guided by CT image quality requirements for anatomical localization and artifact reduction. A CT technique with higher kVp in combination with lower mAs is recommended when low-dose CT images are used for AC to minimize beam-hardening artifacts. [http://dx.doi.org/10.1118/1.4868511] Key words: SPECT/CT, low-dose CT, attenuation correction, radiation dose 1. INTRODUCTION The fusion of single-photon emission computed tomography (SPECT) and CT provides complimentary functional and anatomical information, which has great potential to improve detection and diagnosis.1, 2 In addition to aiding in localization of uptake, CT images can be used to correct for attenuation of emission photons and improve the quantitative accuracy of SPECT reconstruction.3 The radiation burden of hybrid SPECT/CT, however, is compounded by the coupling of two ionizing radiation-based imaging modalities. Radiation dose from medical examinations has become a growing concern over the past several years as the annual effective dose per individual from medical imaging has been estimated to be six times higher than it was two decades ago.4, 5 The dramatic increase in radiation dose to the population from medical procedures is largely attributed to the growth in the number of CT exams performed each year, which approximately dou-
042502-1
Med. Phys. 41 (4), April 2014
bled from 1990 to 2000, and the increase in nuclear medicine procedures.4–6 Attenuation correction (AC) maps were previously derived from transmission scans generated by 153 Gd line-sources, which were estimated to result in an effective dose of only 1–11 μSv to the patient.7 Early generations of hybrid systems often employed low-power x-ray tubes and delivered effective doses 0 decreased as the kVp was lowered. Differences of about 0.02 cm−1 (or 7%) were observed in estimates of μ140 keV for materials with higher HU (e.g., cortical bone) between 80 and 130 kVp. The differences in μ140 keV estimated at various kVp for HU > 0 reflect the inaccuracies inherent in the manufacturer’s algorithm used for the calculation of μ140 keV from CT images. When a CT dataset is “prepared” for AC on the Symbia TruePoint SPECT/CT (T16, Siemens Medical Solutions USA, Inc.) system, it is first registered to the SPECT data reconstructed without attenuation correction. The original CT images (henceforth referred to as CT512 ) are then resampled and rebinned to match the SPECT voxel dimensions. It is the resampled CT images (henceforth referred to as CTAC,128 ) Medical Physics, Vol. 41, No. 4, April 2014
which are subsequently converted to a μ-map via the bilinear transform function. Noise in the resampled CT image (CTAC,128 ) was proportional to noise in the original CT image (CT512 ). CT images reconstructed with higher-pass filters experienced a greater reduction in noise when they were resampled compared to those images which had been more heavily smoothed prior to resampling. Noise was reduced to 58%, 15%, and 7% after resampling when the CT images were filtered using B08s, B41s, and U90s, respectively [Fig. 3(a)]. Noise in the μ-map scaled proportionally with noise in the CTAC,128 images, with a constant of proportionality equal to the slope of the bilinear transform function at that CT
F IG . 2. The bilinear transforms at 140 keV employed by the Symbia TruePoint SPECT/CT system to convert CT number [HU] acquired at three different tube potentials of 80, 110, and 130 kVp into linear-attenuation (μ) values [cm−1 ].
042502-5
K. W. Hulme and S. C. Kappadath: CT noise and artifacts for quantitative 99m Tc SPECT
TABLE II. Linear fit parameters for the bilinear transform function employed by Symbia SPECT/CT at 140 keV. HU ≤ 0
HU > 0
kVp
Slope [cm−1 HU−1 ]
Intercept [cm−1 ]
Slope [cm−1 HU−1 ]
Intercept [cm−1 ]
80 110 130
0.000154 0.000154 0.000154
0.1545 0.1545 0.1545
0.000082 0.000120 0.000137
0.1557 0.1549 0.1546
number [Fig. 3(b)]. If the distribution in CT number straddled the breakpoint of the transform function, noise in the μ-map was scaled by a weighted average of the two slopes. The use of three tube potentials and five effective-mAs settings enabled CT doses to be sampled over 2 orders of magnitude, which, in combination with the three CT reconstruction filters, enabled CT noise to be sampled over close to 3 orders of magnitude. CT512 images of the “pediatric” phantom at the scanner’s lowest dose setting (80 kVp, 12 mAs) and reconstructed with the smoothest (B08s) and sharpest (U90s) filter yielded noise levels of 6 and 168 HU, respectively, which reduced to noise levels of only 0.7% and 1.2% in the μ-map (Table I). CT512 images of the “adult” phantom had noise levels up to 393 HU and produced μ-maps with
Implications of CT noise and artifacts for quantitative 99m Tc SPECT/CT imaging K. W. Hulme and S. C. Kappadatha) Department of Imaging Physics, The University of Texas M.D. Anderson Cancer Center, Houston, Texas 77030 and The University of Texas Graduate School of Biomedical Sciences, Houston, Texas 77030
(Received 10 October 2013; revised 26 February 2014; accepted for publication 4 March 2014; published 20 March 2014) Purpose: This paper evaluates the effects of computed tomography (CT) image noise and artifacts on quantitative single-photon emission computed-tomography (SPECT) imaging, with the aim of establishing an appropriate range of CT acquisition parameters for low-dose protocols with respect to accurate SPECT attenuation correction (AC). Methods: SPECT images of two geometric and one anthropomorphic phantom were reconstructed iteratively using CT scans acquired at a range of dose levels (CTDIvol = 0.4 to 46 mGy). Resultant SPECT image quality was evaluated by comparing mean signal, background noise, and artifacts to SPECT images reconstructed using the highest dose CT for AC. Noise injection was performed on linear-attenuation (μ) maps to determine the CT noise threshold for accurate AC. Results: High levels of CT noise (σ ∼ 200–400 HU) resulted in low μ-maps noise (σ ∼ 1%–3%). Noise levels greater than ∼10% in 140 keV μ-maps were required to produce visibly perceptible increases of ∼15% in 99m Tc SPECT images. These noise levels would be achieved at low CT dose levels (CTDIvol = 4 μGy) that are over 2 orders of magnitude lower than the minimum dose for diagnostic CT scanners. CT noise could also lower (bias) the expected μ values. The relative error in reconstructed SPECT signal trended linearly with the relative shift in μ. SPECT signal was, on average, underestimated in regions corresponding with beam-hardening artifacts in CT images. Any process that has the potential to change the CT number of a region by ∼100 HU (e.g., misregistration between CT images and SPECT images due to motion, the presence of contrast in CT images) could introduce errors in μ140 keV on the order of 10%, that in turn, could introduce errors on the order of ∼10% into the reconstructed 99m Tc SPECT image. Conclusions: The impact of CT noise on SPECT noise was demonstrated to be negligible for clinically achievable CT parameters. Because CT dose levels that affect SPECT quantification is low (CTDIvol ∼ 4 μGy), the low dose limit for the CT exam as part of SPECT/CT will be guided by CT image quality requirements for anatomical localization and artifact reduction. A CT technique with higher kVp in combination with lower mAs is recommended when low-dose CT images are used for AC to minimize beam-hardening artifacts. [http://dx.doi.org/10.1118/1.4868511] Key words: SPECT/CT, low-dose CT, attenuation correction, radiation dose 1. INTRODUCTION The fusion of single-photon emission computed tomography (SPECT) and CT provides complimentary functional and anatomical information, which has great potential to improve detection and diagnosis.1, 2 In addition to aiding in localization of uptake, CT images can be used to correct for attenuation of emission photons and improve the quantitative accuracy of SPECT reconstruction.3 The radiation burden of hybrid SPECT/CT, however, is compounded by the coupling of two ionizing radiation-based imaging modalities. Radiation dose from medical examinations has become a growing concern over the past several years as the annual effective dose per individual from medical imaging has been estimated to be six times higher than it was two decades ago.4, 5 The dramatic increase in radiation dose to the population from medical procedures is largely attributed to the growth in the number of CT exams performed each year, which approximately dou-
042502-1
Med. Phys. 41 (4), April 2014
bled from 1990 to 2000, and the increase in nuclear medicine procedures.4–6 Attenuation correction (AC) maps were previously derived from transmission scans generated by 153 Gd line-sources, which were estimated to result in an effective dose of only 1–11 μSv to the patient.7 Early generations of hybrid systems often employed low-power x-ray tubes and delivered effective doses 0 decreased as the kVp was lowered. Differences of about 0.02 cm−1 (or 7%) were observed in estimates of μ140 keV for materials with higher HU (e.g., cortical bone) between 80 and 130 kVp. The differences in μ140 keV estimated at various kVp for HU > 0 reflect the inaccuracies inherent in the manufacturer’s algorithm used for the calculation of μ140 keV from CT images. When a CT dataset is “prepared” for AC on the Symbia TruePoint SPECT/CT (T16, Siemens Medical Solutions USA, Inc.) system, it is first registered to the SPECT data reconstructed without attenuation correction. The original CT images (henceforth referred to as CT512 ) are then resampled and rebinned to match the SPECT voxel dimensions. It is the resampled CT images (henceforth referred to as CTAC,128 ) Medical Physics, Vol. 41, No. 4, April 2014
which are subsequently converted to a μ-map via the bilinear transform function. Noise in the resampled CT image (CTAC,128 ) was proportional to noise in the original CT image (CT512 ). CT images reconstructed with higher-pass filters experienced a greater reduction in noise when they were resampled compared to those images which had been more heavily smoothed prior to resampling. Noise was reduced to 58%, 15%, and 7% after resampling when the CT images were filtered using B08s, B41s, and U90s, respectively [Fig. 3(a)]. Noise in the μ-map scaled proportionally with noise in the CTAC,128 images, with a constant of proportionality equal to the slope of the bilinear transform function at that CT
F IG . 2. The bilinear transforms at 140 keV employed by the Symbia TruePoint SPECT/CT system to convert CT number [HU] acquired at three different tube potentials of 80, 110, and 130 kVp into linear-attenuation (μ) values [cm−1 ].
042502-5
K. W. Hulme and S. C. Kappadath: CT noise and artifacts for quantitative 99m Tc SPECT
TABLE II. Linear fit parameters for the bilinear transform function employed by Symbia SPECT/CT at 140 keV. HU ≤ 0
HU > 0
kVp
Slope [cm−1 HU−1 ]
Intercept [cm−1 ]
Slope [cm−1 HU−1 ]
Intercept [cm−1 ]
80 110 130
0.000154 0.000154 0.000154
0.1545 0.1545 0.1545
0.000082 0.000120 0.000137
0.1557 0.1549 0.1546
number [Fig. 3(b)]. If the distribution in CT number straddled the breakpoint of the transform function, noise in the μ-map was scaled by a weighted average of the two slopes. The use of three tube potentials and five effective-mAs settings enabled CT doses to be sampled over 2 orders of magnitude, which, in combination with the three CT reconstruction filters, enabled CT noise to be sampled over close to 3 orders of magnitude. CT512 images of the “pediatric” phantom at the scanner’s lowest dose setting (80 kVp, 12 mAs) and reconstructed with the smoothest (B08s) and sharpest (U90s) filter yielded noise levels of 6 and 168 HU, respectively, which reduced to noise levels of only 0.7% and 1.2% in the μ-map (Table I). CT512 images of the “adult” phantom had noise levels up to 393 HU and produced μ-maps with
