DOI: 10.1148/rg.266065063
RadioGraphics 2006;26:1785-1794
© RSNA, 2006
Relationship between Noise, Dose, and Pitch in Cardiac Multi–Detector Row CT1
Andrew N. Primak, PhD,
Cynthia H. McCollough, PhD,
Michael R. Bruesewitz, RT(R),
Jie Zhang, PhD and
Joel G. Fletcher, MD
1 From the CT Clinical Innovation Center, Department of Radiology, Mayo Clinic College of Medicine, 200 First Street SW, Rochester, MN 55905. Recipient of a Certificate of Merit award for an education exhibit at the 2005 RSNA Annual Meeting. Received April 12, 2006; revision requested May 30 and received July 6; accepted July 10. A.N.P., C.H.M., J.Z., and J.G.F. receive research grants from Siemens Medical Solutions, Malvern, Pa; C.H.M. also receives a research grant from GE Healthcare, Waukesha, Wis; J.G.F. has an educational license from GE Healthcare and participates in a CME course sponsored by E-Z-Em, Lake Success, NY; M.R.B. has no financial relationships to disclose.
Address correspondence to A.N.P. (e-mail: primak.andrew{at}mayo.edu).
 |
Abstract
|
|---|
In spiral computed tomography (CT), dose is always inversely proportional to pitch. However, the relationship between noise and pitch (and hence noise and dose) depends on the scanner type (single vs multi–detector row) and reconstruction mode (cardiac vs noncardiac). In single detector row spiral CT, noise is independent of pitch. Conversely, in noncardiac multi–detector row CT, noise depends on pitch because the spiral interpolation algorithm makes use of redundant data from different detector rows to decrease noise for pitch values less than 1 (and increase noise for pitch values > 1). However, in cardiac spiral CT, redundant data cannot be used because such data averaging would degrade the temporal resolution. Therefore, the behavior of noise versus pitch returns to the single detector row paradigm, with noise being independent of pitch. Consequently, since faster rotation times require lower pitch values in cardiac multi–detector row CT, dose is increased without a commensurate decrease in noise. Thus, the use of faster rotation times will improve temporal resolution, not alter noise, and increase dose. For a particular application, the higher dose resulting from faster rotation speeds should be justified by the clinical benefits of the improved temporal resolution.
© RSNA, 2006
|
Introduction
|
|---|
The improved diagnostic capabilities of multi–detector row computed tomographic (CT) technology have resulted in an increasing number of CT examinations, which represent a significant portion of the radiation dose received from all medical procedures (1–3). The potential radiation risk from this increased use of CT makes it important that CT doses be kept as low as reasonably achievable. To fulfill this goal, it is important to understand the relationship between dose and image noise, as noise is a major factor in determining acceptable image quality and often dictates the dose for a particular CT protocol.
In spiral CT, both dose and noise depend on pitch, but not in the same way. While dose is always inversely proportional to pitch, the behavior of noise as a function of pitch depends on scanner type (single vs multi–detector row) and reconstruction mode (cardiac vs noncardiac). Further, in cardiac CT, the relationship between pitch and patient heart rate must also be taken into account (4,5).
In this article, we review the underlying principles and demonstrate the relationships between image noise, dose, and pitch for cardiac and non-cardiac multi–detector row CT. We also discuss the clinical implications of these relationships. The Table provides a glossary of several technical terms used in these discussions.
|
Radiation Dose
|
|---|
Radiation dose is related to the amount of energy that x-ray photons deliver during a CT scan. It depends on the total number of photons and their individual energies. The energy distribution of these photons depends on the x-ray tube potential and the spectral (bowtie) filter. The total number of photons is proportional to the tube current (in milliamperes) and the "x-ray on" time (in seconds) during a single gantry rotation, and hence it is proportional to the tube current–time product (in milliampereseconds). Thus, for a single axial scan, or the most common scenario of contiguous scans (where table increment = total nominal beam width), dose is directly proportional to the tube current–time product.
In the case of multiple axial scans or spiral scans with the table increment 
total nominal beam width, an accumulation of dose due to x-ray beam overlap, or a decrease in dose due to gaps in the radiation beams, from successive gantry rotations will occur. The quantitative measure of such overlap or gap is given by the ratio of table increment per rotation to total nominal beam width, which is known as pitch. Therefore, the general relationship exists that dose is proportional to tube current–time product/pitch (which on systems manufactured by Siemens Medical Solutions [Forchheim, Germany] is called "effective mAs" and on systems manufactured by Philips Medical Systems [Andover, Mass] is called "mAs per slice") (6–8).
For examinations where the patient table is not incremented between successive rotations of the x-ray tube (eg, dynamic CT, CT perfusion, or interventional CT examinations), dose will accumulate in the irradiated section of tissue over the multiple exposures at that location. Thus, for these examinations, dose is proportional to the tube current–time product multiplied by the number of scans in the examination (ie, the number of tube rotations where the x-ray tube is energized).
|
Noise
|
|---|
In general, noise in CT depends on the number of x-ray photons reaching the detector (quantum noise), the electronic noise of the detector system, and the reconstruction kernel (sharper kernels give noisier images). Unless images suffer from severe photon starvation (eg, in morbidly obese patients), quantum noise plays the dominant role. Since x-ray photon statistics obey the Poisson distribution, quantum noise is proportional to 
N and the corresponding image noise is approximately proportional to 1/N, where N is the number of photons that have contributed to the reconstructed image. Since the number of photons reaching the detector depends on the object attenuation, which in turn depends on photon energies, N is strongly dependent on tube potential. In addition, N is proportional to section width, tube current, and the amount of time necessary to acquire all the projection data needed for the reconstruction. In sequential mode, this time equals the "x-ray on" time per rotation, so image noise is approximately proportional to 1/mAs.
In spiral mode, however, the interpolation algorithm, which transforms the projection data acquired at various z-axis locations into projection data at one specific z-axis location, must be taken into account. Because the spiral interpolation algorithm is inherently different for multi–detector row CT compared to single detector row CT (9), the relationship between noise and pitch in spiral CT depends on the scanner type (single vs multi–detector row CT). In addition, because cardiac spiral reconstructions are optimized to decrease motion artifact (4,5) (ie, provide the best possible temporal resolution), the relationship between noise and pitch also depends on the multi–detector row CT reconstruction mode (cardiac vs noncardiac).
Single Detector Row Spiral CT
In single detector row spiral CT, the gantry always has to rotate through a certain angle (dependent on the reconstruction algorithm) in order to acquire the projection data needed for an axial image (10). Thus, N depends only on tube current–time product and not on the table speed or pitch. An intuitive way of understanding this effect is to realize that regardless of the spacing of the consecutive x-ray tube rotations (ie, pitch), the same amount of projection data is required to reconstruct an image (either 720° or 360° of data, depending on the spiral interpolation algorithm implemented on the system; 360° is the most common). Because the same amount of image data (photons) is used to reconstruct the image, the image noise remains constant even as pitch is changed.
This relationship was counterintuitive to many users, since it has been well (and correctly) publicized that increasing pitch decreases patient dose (11). Since intuition teaches that noise increases when dose decreases, the incorrect assumption that noise depended on pitch resulted. Studies have shown that indeed, noise is independent of pitch, while dose decreases with increasing pitch in single detector row CT (12).
This raises the obvious question as to how one can achieve a dose decrease at the same image noise level, as typically in imaging, one does not gain an advantage (decreased dose) without paying some price. The price of an increased pitch is a widening of the section sensitivity profile (ie, the width of the reconstructed image) (13). This widening occurs because, at higher pitch values, the 360° of required projection data are obtained from z-axis locations that are spread farther apart (relative to use of a small pitch value). Thus, data are averaged over a wider z-axis distance, widening the effective thickness of the reconstructed image (13).
Multi–Detector Row Spiral CT
Multi–detector row spiral CT reconstruction algorithms differ profoundly from single detector row spiral CT reconstruction algorithms because of the flexibility provided from having data collected simultaneously at multiple z-axis locations (multiple detector rows). The interpolation process in multi–detector row CT can select and utilize data from the detector row nearest to the requested z-axis image location, regardless of which row those data come from. Using a modified 360° interpolation approach, there are redundant data points (ie, data passing through the same z-axis position) from different detectors as they rotate about the patient. The frequency of these redundancies is determined by the pitch values, with greater redundancy at lower pitch values and less redundancy at higher pitch values.
To make use of all the photons (dose) applied to the patient, multi–detector row CT interpolation algorithms make use of the redundant data by averaging redundant projection values, which effectively decreases image noise. As the pitch is increased, there are more gaps in the data with fewer redundant points to average, so the image noise increases. Thus, the paradigm from single detector row CT is altered in multi–detector row CT, and noise becomes dependent on pitch. The relationship of dose to pitch is unchanged. However, this paradigm is valid only for routine spiral multi–detector row CT reconstruction algorithms, as the special algorithms for cardiac multi–detector row CT alter the noise versus pitch relationship for yet a third time.
Noncardiac Mode.—
As discussed earlier, in noncardiac spiral multi–detector row CT, noise depends on pitch through an interpolation algorithm used to generate a set of planar projection data (Fig 1). All spiral data within a predefined spiral interpolation window (horizontal blue dashes) are used to generate planar data for the image plane (vertical red line).
View larger version (34K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1a. Graphs of the interpolation algorithm used to generate planar data from the measured spiral data. The dotted lines show the center of every detector row, whereas the solid lines indicate the detector boundaries. (See the text for additional details.)
|
|
View larger version (48K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1b. Graphs of the interpolation algorithm used to generate planar data from the measured spiral data. The dotted lines show the center of every detector row, whereas the solid lines indicate the detector boundaries. (See the text for additional details.)
|
|
View larger version (62K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1c. Graphs of the interpolation algorithm used to generate planar data from the measured spiral data. The dotted lines show the center of every detector row, whereas the solid lines indicate the detector boundaries. (See the text for additional details.)
|
|
View larger version (33K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1d. Graphs of the interpolation algorithm used to generate planar data from the measured spiral data. The dotted lines show the center of every detector row, whereas the solid lines indicate the detector boundaries. (See the text for additional details.)
|
|
When pitch is less than 1, the measured spiral data partially overlap (shaded areas) in the z direction (perpendicular to the gantry plane) (Fig 1b), so some portions of the planar projection data for the same axial image can be generated more than once (from the spiral data acquired by different detector rows). These redundant data result in more x-ray photons contributing to the reconstruction of the axial images compared to when pitch = 1, reducing the image noise and making it pitch dependent. For example, a scan with pitch = 0.5 needs twice the number of gantry rotations to cover the same distance compared to when pitch = 1. This results in fully overlapped spiral data, which can generate two completely redundant sets of planar projection data for every axial image (Fig 1c). Combining these two redundant data sets into one allows reconstruction of axial images where noise is reduced by a factor of 2.
When pitch is more than 1, the measured spiral data have gaps in the z direction (Fig 1d). These missing data result in less x-ray photons contributing to the reconstruction of each axial image compared to when pitch = 1, increasing the image noise and making it pitch dependent.
Generally speaking, to offset the increase in noise as pitch is increased (or offset the increase in dose as pitch is decreased), CT system manufacturers provide some mechanism to adjust the tube current so that tube current–time product is increased approximately proportional to the increase in pitch (or decreased approximately proportional to the decrease in pitch). Thus, as long as effective mAs (defined as tube current–time product/pitch) is held constant, both dose and noise remain constant.
This is demonstrated in Figure 2a and 2b, where noise and dose are shown as a function of tube current–time product and as a function of effective mAs.
View larger version (49K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2a. Noise and dose versus tube current–time product and effective mAs for noncardiac (a, b) and cardiac (c, d) spiral CT modes at four different pitch (p) values and two different gantry rotation times (0.33 seconds and 0.37 seconds). The left y axis corresponds to noise curves; the right y axis corresponds to dose curves. Note that in cardiac mode, noise is dependent on tube current–time product and not on pitch (c). CTDIvol =Volume CT Dose Index.
|
|
View larger version (47K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2b. Noise and dose versus tube current–time product and effective mAs for noncardiac (a, b) and cardiac (c, d) spiral CT modes at four different pitch (p) values and two different gantry rotation times (0.33 seconds and 0.37 seconds). The left y axis corresponds to noise curves; the right y axis corresponds to dose curves. Note that in cardiac mode, noise is dependent on tube current–time product and not on pitch (c). CTDIvol =Volume CT Dose Index.
|
|
View larger version (38K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2c. Noise and dose versus tube current–time product and effective mAs for noncardiac (a, b) and cardiac (c, d) spiral CT modes at four different pitch (p) values and two different gantry rotation times (0.33 seconds and 0.37 seconds). The left y axis corresponds to noise curves; the right y axis corresponds to dose curves. Note that in cardiac mode, noise is dependent on tube current–time product and not on pitch (c). CTDIvol =Volume CT Dose Index.
|
|
View larger version (48K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2d. Noise and dose versus tube current–time product and effective mAs for noncardiac (a, b) and cardiac (c, d) spiral CT modes at four different pitch (p) values and two different gantry rotation times (0.33 seconds and 0.37 seconds). The left y axis corresponds to noise curves; the right y axis corresponds to dose curves. Note that in cardiac mode, noise is dependent on tube current–time product and not on pitch (c). CTDIvol =Volume CT Dose Index.
|
|
For Siemens Medical Solutions scanners, the manufacturer adopted the term "effective mAs" to help users realize that as long as the ratio of tube current–time product to pitch was held constant, they could reliably predict a constant noise level for a given object size. Philips Medical Systems similarly used the term "mAs per slice" to accomplish the same task. On GE Healthcare Technologies (Waukesha, Wis) systems, when the user alters the pitch values, the user interface automatically alters the prescribed tube current to a value that will maintain a constant noise. The new (suggested) tube current is highlighted with orange to alert the user that this change has occurred. If the user does not desire to hold the image noise constant at the new pitch value, the user can manually override the suggested value. Toshiba Medical Systems (Tokyo, Japan) systems have a similar mechanism of suggesting to the operator the appropriate tube current value when the pitch value is changed.
In Figure 2b, we observe that the noise is constant for a given effective mAs, completely independent of pitch. This complete "pitch independence" is a result of the chosen spiral interpolation algorithm and the manufacturer’s conscious decision to allow the user to alter pitch in a continuous fashion to prescribe the desired scan acquisition time without having to worry about the effect of the pitch on image quality. On Siemens Medical Solutions systems, both noise and section width are essentially independent of pitch for a constant effective mAs (9,14). Other manufacturers have implemented spiral interpolation and reconstruction algorithms that have preferred pitch values with regard to both image width and image noise. For example, GE Healthcare Technologies systems allow the user four discrete pitch values from which to choose, with the two lowest values providing the narrowest actual image thickness (for a selected nominal value) at the expense of higher noise. Thus, data such as presented in Figure 2b show comparable trends but do not coalesce into one curve. This makes the precise tube current value needed at a given pitch (in order to maintain constant noise) a bit more difficult to predict; thus, the user should rely on the suggested tube current values provided by the scanner interface (15).
Cardiac Mode.—
In cardiac spiral multi–detector row CT, the best possible temporal resolution is required to minimize artifacts resulting from cardiac motion. This goal is achieved by minimizing the number of projections used to reconstruct the image to those projections gathered in the shortest possible time window. The use of redundant data would not be acceptable because it would degrade the temporal resolution by averaging data over one or more rotations of the x-ray tube. The minimum amount of data required to reconstruct a CT image is 180° plus the angle (in degrees) of the x-ray beam in the plane of the image (known as the fan angle). Hence, cardiac algorithms use partial reconstruction techniques (180° + fan angle) to reconstruct an image.
These data are collected either during a single cardiac cycle (single-segment reconstruction) or during two or more consecutive heartbeats (multisegment reconstruction). In both cases, the number of photons N contributing to the cardiac reconstruction depends only on the tube current and the time it takes for the gantry to rotate through 180° plus the fan angle, and not pitch. Since this time is proportional to the rotation time, N (and hence noise) is dependent only on tube current–time product and is not affected by pitch. This is analogous to the case of single detector row CT, which uses exactly 360° of data to reconstruct an image and hence has noise that is independent of pitch. Figure 2c and 2d demonstrates the independence of noise on pitch for cardiac reconstructions, even though dose remains dependent on pitch. Hence, a constant effective mAs in cardiac multi–detector row CT does not guarantee equivalent noise, but rather only equivalent dose.
Noise versus Tube Current–Time Product Measurements
Noise versus tube current–time product data were obtained on a 64-channel CT scanner (Sensation 64; Siemens Medical Solutions) by using an anthropomorphic cardiac CT phantom (16) (QRM, Möhrendorf, Germany) (Fig 3a). The phantom was scanned by using both electrocardiographically (ECG) gated (with ECG signal provided by an ECG simulator on the CT system) and non-gated spiral modes with four different pitch values (0.18, 0.20, 0.24, and 0.27 for gated and 0.5, 0.6, 0.7, and 0.8 for nongated) and two rotation times (0.33 and 0.37 seconds). Scanning was done at 120 kVp and variable tube current–time product values. We used the 32 x 0.6-mm collimation with a z-flying focal spot technique (17), which resulted in 64 overlapping projections per rotation. Contiguous 3-mm-thick axial sections were reconstructed with a 300-mm field of view and B35 kernel. Noise was measured as the standard deviation of the pixel values within a water-equivalent cylinder embedded in the central portion of the cardiac phantom (Fig 3b). Dose was assessed by using the Volume CT Dose Index (in milligrays), per International Electrotechnical Commission publication 60601-2-44 (18).
View larger version (83K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3a. (a) Anthropomorphic cardiac phantom with small calcified cylinders used for calcium quantitation (arrows). (b) Axial image shows the water-equivalent cylindrical insert used for noise versus tube current–time product measurements (arrow).
|
|
View larger version (53K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3b. (a) Anthropomorphic cardiac phantom with small calcified cylinders used for calcium quantitation (arrows). (b) Axial image shows the water-equivalent cylindrical insert used for noise versus tube current–time product measurements (arrow).
|
|
The results of these measurements are shown in Figure 2 to demonstrate the quantitative relationships described earlier. A visual demonstration of these relationships is given in Figure 4, where for low and high pitch values, phantom images obtained at the same effective mAs are shown. In Figure 4a, the noncardiac images have the same noise texture independent of pitch because the effective mAs was held constant. In Figure 4b, the cardiac images obtained at the same effective mAs do not have the same image noise (though they do have the same dose). In Figure 4c, the cardiac images obtained at the same tube current–time product (not effective mAs) do have the same image noise (but not the same dose).
View larger version (34K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4a. Visual demonstration of the relationships between noise, dose, and pitch. Cardiac phantom images are shown for low and high pitch values. (a) Noncardiac images obtained at the same effective mAs have the same noise. (b) Cardiac images obtained at the same effective mAs do not have the same noise (but do have the same dose). (c) Conversely, cardiac images obtained at the same tube current–time product (not effective mAs) do have the same noise (but do not have the same dose). CTDIvol = Volume CT Dose Index, SD = standard deviation.
|
|
View larger version (36K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4b. Visual demonstration of the relationships between noise, dose, and pitch. Cardiac phantom images are shown for low and high pitch values. (a) Noncardiac images obtained at the same effective mAs have the same noise. (b) Cardiac images obtained at the same effective mAs do not have the same noise (but do have the same dose). (c) Conversely, cardiac images obtained at the same tube current–time product (not effective mAs) do have the same noise (but do not have the same dose). CTDIvol = Volume CT Dose Index, SD = standard deviation.
|
|
View larger version (35K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4c. Visual demonstration of the relationships between noise, dose, and pitch. Cardiac phantom images are shown for low and high pitch values. (a) Noncardiac images obtained at the same effective mAs have the same noise. (b) Cardiac images obtained at the same effective mAs do not have the same noise (but do have the same dose). (c) Conversely, cardiac images obtained at the same tube current–time product (not effective mAs) do have the same noise (but do not have the same dose). CTDIvol = Volume CT Dose Index, SD = standard deviation.
|
|
|
Pitch in Cardiac Multi–Detector Row CT
|
|---|
Temporal resolution is of fundamental importance to cardiac multi–detector row CT and was a driving force behind making gantries rotate faster and faster. However, faster gantry rotation requires a slower pitch in cardiac mode to avoid discontinuities in the anatomic coverage of the heart between images reconstructed from consecutive cardiac cycles (Fig 5) (4,5,19).
View larger version (34K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5. Diagram of an ECG-gated spiral CT scan with pitch that is too high for the heart rate. Continuous anatomic coverage of the heart is not possible because of the volume gaps between the images reconstructed from the consecutive cardiac cycles. R = R wave, Recon = reconstruction.
|
|
For example, on a CT scanner with a rotation time of 0.5 seconds (Volume Zoom, Siemens Medical Solutions), the necessary pitch was 0.375. On a CT scanner with a rotation time of 0.33 seconds (Sensation 64, Siemens Medical Solutions), the necessary pitch is 0.2.
We can derive a good approximation of the relationship between pitch, heart rate, and the rotation time by taking into account that for single-segment reconstruction (all data acquired in one cardiac cycle), the table should not move more than the total nominal beam width W during the time of one heart cycle (R-R interval time TRR). This allows all phases of the heart to be "seen" by some part of the detector at any z-axis location. This means that the table speed V should be less than or equal to W/TRR. Keeping in mind the definition of pitch (table increment/total nominal beam width), pitch = VTrot/W, where Trot is the rotation time. With some minor algebraic manipulation, we obtain the following general requirement:
 | (1) |
To derive the exact relationship, one has to consider the details of a particular cardiac reconstruction algorithm (19,20). For example, according to Ohnesorge et al (19), the maximum pitch for single-segment reconstruction is given by the following formula:
 | (2) |
where M is the number of detector rows in the cardiac mode. For modern state-of-the-art multi–detector row CT scanners with M 
32, this pitch restriction is very close to our intuitive approximation given in Equation (1).
The clinical implications of Equation (2) are quite important. At faster gantry rotation times, a lower pitch value is required. However, to achieve the same noise, one has to use the same tube current–time product value (recall that noise is independent of pitch in cardiac multi–detector row CT). The use of the same tube current–time product value (relative to slower rotation time scans), but smaller pitch values, results in higher radiation doses (dose is always inversely proportional to pitch). Thus, for single-source (ie, one x-ray tube) multi–detector row CT systems, better temporal resolution in cardiac spiral CT comes with the price of higher dose. (These conclusions do not necessarily apply to the recently introduced dual-source CT system [21], which allows pitch values to be increased as heart rate increased, thus offsetting the increased dose that accompanies improved temporal resolution in single-source cardiac multi–detector row CT.)
|
Clinical Implications
|
|---|
In noncardiac multi–detector row CT, the implications of the relationship between noise, pitch, and dose are that as long as the ratio of tube current–time product to pitch is held relatively constant, a relatively constant image noise will result. For the Siemens Medical Solutions systems, as shown earlier, the relationship is exact. For other manufacturers, the relationship holds in general, although small variations in noise (10%–20%) may occur even if the ratio of tube current–time product to pitch is held constant. This relationship can be used to the advantage of the operator when long scan ranges are required yet the total scan time is desired to be kept short (ie, a short breath hold for a thoracic-abdominal-pelvic CT angiography examination). The operator can increase the pitch (table speed) in order to keep the examination time short. Image noise will not be compromised as long as the system can provide a sufficient tube current–time product value to keep the ratio of tube current–time product/pitch constant (relative to a lower pitch examination over a short anatomic range).
In cardiac spiral multi–detector row CT, faster rotation times deliver improved temporal resolution, but require higher radiation dose to achieve equivalent image noise (compared to slower rotation times). Thus, the use of faster rotation times for a particular application should be justified by clinical benefits. We have examined in a phantom model two clinical cardiac CT applications where these issues may be relevant: coronary artery angiography (CT angiography) and quantitation of coronary artery calcium (CAC).
In our CT angiography phantom study (22), an anthropomorphic cardiac CT phantom containing an iodine-filled vessel phantom with a stenosis was scanned by using retrospective ECG-gated CT angiography protocols with 16- and 64-channel multi–detector row CT systems (Sensation 16 and 64; Siemens Medical Solutions). Data were acquired at rest (0.33 seconds rotation time only) and while the vessel phantom was moving at physiologic coronary artery velocities for multiple gantry rotation times (0.33, 0.42, and 0.5 seconds). The stenoses were measured by using a semiautomated tool (VesselView; Siemens Medical Solutions) for section widths of 0.6 mm (64-channel system only), 0.75 mm, and 1.0 mm. The accuracy of percent area stenosis measurements in a moving phantom improved significantly (P < .01) as the rotation time decreased from 0.5 to 0.33 seconds (Fig 6). Hence, for CT angiography, the benefit of faster rotation time appears to outweigh the disadvantage of the increased dose. These findings in a simple phantom model are consistent with reported clinical experiences (23–26).
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6. Percentage of area stenosis versus rotation time for a moving stenotic vessel phantom scanned on 16– and 64–detector row CT systems with coronary CT angiography protocols and multiple gantry rotation times (0.33, 0.42, and 0.5 seconds). The measurements were performed for 0.6-, 0.75-, and 1-mm section widths. (See the text for additional details.)
|
|
In our CAC phantom study (27), the central portion of an anthropomorphic cardiac phantom containing nine cylinders of varying amounts of hydroxyapatite (Fig 3a) was attached to a rotational motion device and imaged by using retrospective ECG-gated CAC protocols with Sensation 16 and 64 CT systems (Siemens Medical Solutions). The data were acquired during rotational motion corresponding to the mean velocity of coronary arteries by using multiple gantry rotation times (0.33, 0.37, 0.42, and 0.5 seconds). Calcium scores were obtained by using an automated tool (CaScoring; Siemens Medical Solutions). The accuracy of CAC measurements was not statistically different (P > .05) for calcifications greater than or equal to 3 mm and densities greater than or equal to 400 mg/cm3 (Fig 7) among the four rotation times tested. Thus, for screening purposes, slower rotation times may be more appropriate, as they appear equally accurate yet deliver lower radiation dose to the patient. These findings are also consistent with reported clinical experiences of reasonably robust CAC with multi–detector row CT systems having both slower and faster rotation times.
View larger version (9K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 7. Calcium mass score measured in a rotating insert of an anthropomorphic cardiac phantom scanned on 16– and 64–detector row CT systems with CAC protocols and multiple gantry rotation times (0.33, 0.37, 0.42, and 0.5 seconds). HA = hydroxyapatite concentration in milligrams per cubic centimeter; the diameter of the calcification is given in millimeters. (For example, 400 HA 3 mm corresponds to 400 mg/cm3 HA density and a 3-mm-diameter cylinder.)
|
|
|
Summary
|
|---|
The relationship between image noise and radiation dose is not the same for cardiac and noncardiac spiral scanning with multi–detector row CT systems. In noncardiac spiral multi–detector row CT, noise depends on pitch, which results in comparable noise when the ratio of tube current–time product to pitch is held constant. In cardiac spiral multi–detector row CT, noise depends only on tube current–time product and not on pitch, while radiation dose remains inversely proportional to pitch. Since faster rotation speeds require lower pitch values (relative to slower rotation speed), they result in higher radiation dose to achieve equivalent image noise, and their use should be justified by clinical benefits, such as in the case of coronary CT angiography.
|
Footnotes
|
|---|
Abbreviations: CAC = quantitation of coronary artery calcium, ECG = electrocardiographically
|
References
|
|---|
- Linton OW, Mettler FA Jr. National conference on dose reduction in CT, with an emphasis on pediatric patients. AJR Am J Roentgenol 2003;181: 321–329.[Free Full Text]
- Yates SJ, Pike LC, Goldstone KE. Effect of multi-slice scanners on patient dose from routine CT examinations in East Anglia. Br J Radiol 2004;77: 472–478.[Abstract/Free Full Text]
- Brix G, Nagel HD, Stamm G, et al. Radiation exposure in multi-slice versus single-slice spiral CT: results of a nationwide survey. Eur Radiol 2003; 13:1979–1991.[CrossRef][Medline]
- Ohnesorge BM, Becker CR, Flohr TG, Reiser MF. Multi-slice CT in cardiac imaging: technical principles, clinical application, and future developments. Berlin, Germany: Springer, 2002.
- Hsieh J. Computed tomography: principles, design, artifacts, and recent advances. Bellingham, Wash: SPIE Press, 2003; 348–357.
- McNitt-Gray MF. Radiation dose in CT. RadioGraphics 2002;22:1541–1553.[Abstract/Free Full Text]
- McCollough CH. Patient dose in cardiac computed tomography. Herz 2003;28:1–6.[CrossRef][Medline]
- Morin RL, Gerber TC, McCollough CH. Radiation dose in computed tomography of the heart. Circulation 2003;107:917–922.[Free Full Text]
- Flohr TG, Schaller S, Stierstorfer K, Bruder H, Ohnesorge BM, Schoepf UJ. Multi–detector row CT systems and image-reconstruction techniques. Radiology 2005;235:756–773.[Abstract/Free Full Text]
- Kalender WA, Seissler W, Klotz E, Vock P. Spiral volumetric CT with single-breath-hold technique, continuous transport, and continuous scanner rotation. Radiology 1990;176:181–183.[Abstract/Free Full Text]
- Food and Drug Administration. FDA public health notification: reducing radiation risk from computed tomography for pediatric and small adult patients. Pediatr Radiol 2002;32:314–316.[CrossRef][Medline]
- Hidajat N, Maurer J, Schroder RJ, Wolf M, Vogl T, Felix R. Radiation exposure in spiral computed tomography: dose distribution and dose reduction. Invest Radiol 1999;34:51–57.[CrossRef][Medline]
- Polacin A, Kalender WA, Marchal G. Evaluation of section sensitivity profiles and image noise in spiral CT. Radiology 1992;185:29–35.[Abstract/Free Full Text]
- Flohr T, Stierstorfer K, Bruder H, Simon J, Polacin A, Schaller S. Image reconstruction and image quality evaluation for a 16-slice CT scanner. Med Phys 2003;30:832–845.[CrossRef][Medline]
- McCollough CH, Zink FE. Performance evaluation of a multi-slice CT system. Med Phys 1999; 26:2223–2230.[CrossRef][Medline]
- Ulzheimer S, Kalender WA. Assessment of calcium scoring performance in cardiac computed tomography. Eur Radiol 2003;13:484–497.[Medline]
- Flohr TG, Stierstorfer K, Ulzheimer S, Bruder H, Primak AN, McCollough CH. Image reconstruction and image quality evaluation for a 64-slice CT scanner with z-flying focal spot. Med Phys 2005;32:2536–2547.[CrossRef][Medline]
- International Electrotechnical Commission. Medical electrical equipment, part 2-44: Particular requirements for the safety of x-ray equipment for computed tomography. IEC publication No. 60601-2-44. Ed 2.1. Geneva, Switzerland: International Electrotechnical Commission Central Office, 2002.
- Ohnesorge B, Flohr T, Becker C, et al. Cardiac imaging by means of electrocardiographically gated multisection spiral CT: initial experience. Radiology 2000;217:564–571.[Abstract/Free Full Text]
- Flohr T, Ohnesorge B. Heart rate adaptive optimization of spatial and temporal resolution for electrocardiogram-gated multislice spiral CT of the heart. J Comput Assist Tomogr 2001;25:907–923.[CrossRef][Medline]
- Flohr TG, McCollough CH, Bruder H, et al. First performance evaluation of a dual-source CT (DSCT) system. Eur Radiol 2006;16:256–268.[CrossRef][Medline]
- Primak AN, Schmidt B, Booya F, et al. The effect of temporal resolution on stenosis measurement accuracy for coronary CT angiography (CTA) using MDCT [abstr]. In: Radiological Society of North America scientific assembly and annual meeting program. Oak Brook, Ill: Radiological Society of North America, 2005; 691.
- Leber AW, Becker A, Knez A, et al. Accuracy of 64-slice computed tomography to classify and quantify plaque volumes in the proximal coronary system: a comparative study using intravascular ultrasound. J Am Coll Cardiol 2006;47:672–677.[Abstract/Free Full Text]
- Cury RC, Ferencik M, Achenbach S, et al. Accuracy of 16-slice multi-detector CT to quantify the degree of coronary artery stenosis: assessment of cross-sectional and longitudinal vessel reconstructions. Eur J Radiol 2006;57:345–350.[CrossRef][Medline]
- Hoffmann MH, Shi H, Schmitz B, et al. Noninvasive coronary angiography with multislice computed tomography. JAMA 2005;293:2471–2478.[Abstract/Free Full Text]
- Achenbach S, Ropers D, Hoffmann U, et al. Assessment of coronary remodeling in stenotic and nonstenotic coronary atherosclerotic lesions by multidetector spiral computed tomography. J Am Coll Cardiol 2004;43:842–847.[Abstract/Free Full Text]
- Primak AN, Schmidt B, Booya F, Fletcher JG, Zhang J, McCollough CH. The effect of temporal resolution on the accuracy of coronary artery calcium measurement using MDCT. Presented at the Sixth International Conference on Cardiac CT, Boston, Mass, July 22–23, 2005.
This article has been cited by other articles:
|
|
|
|
 
P. Stolzmann, S. Leschka, H. Scheffel, T. Krauss, L. Desbiolles, A. Plass, M. Genoni, T. G. Flohr, S. Wildermuth, B. Marincek, et al.
Dual-Source CT in Step-and-Shoot Mode: Noninvasive Coronary Angiography with Low Radiation Dose
Radiology,
October 1, 2008;
249(1):
71 - 80.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
B. Kim, K. H. Lee, K. J. Kim, R. Mantiuk, H.-r. Kim, and Y. H. Kim
Artifacts in Slab Average-Intensity-Projection Images Reformatted from JPEG 2000 Compressed Thin-Section Abdominal CT Data Sets
Am. J. Roentgenol.,
June 1, 2008;
190(6):
W342 - W350.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
C. H. McCollough, A. N. Primak, O. Saba, H. Bruder, K. Stierstorfer, R. Raupach, C. Suess, B. Schmidt, B. M. Ohnesorge, and T. G. Flohr
Dose Performance of a 64-Channel Dual-Source CT Scanner
Radiology,
June 1, 2007;
243(3):
775 - 784.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
