CT Scan Parameters and Radiation Dose: Practical Advice for Radiologists Siva P

Home , CT scan

CT Scan Parameters and Radiation Dose: Practical Advice for Radiologists Siva P. Raman, MD, Mahadevappa Mahesh, MS, PhD, Robert V. Blasko, BS, RT(R)(CT), Elliot K. Fishman, MD

Although there has been increasing recognition of the importance of reducing radiation dose when performing multidetector CT examinations, the increasing complexity of CT scanner technology, as well as confusion about the importance of many different CT scan parameters, has served as an impediment to radiologists seeking to create lower dose protocols. The authors seek to guide radiologists through the manipulation of 8 fundamental CT scan parameters that can be altered or optimized to reduce patient radiation dose, including detector conﬁguration, tube current, tube potential, reconstruction algorithm, patient positioning, scan range, reconstructed slice thickness, and pitch. Although there is always an inherent trade-off between image quality or noise and patient radiation dose, in many cases, a reasoned manipulation of these 8 parameters can allow the safer imaging of patients (with lower dose) while preserving diagnostic image quality. Key Words: Radiation dose, automated tube current modulation, low kVp, pitch, iterative reconstruction J Am Coll Radiol 2013;10:840-846. Copyright © 2013 American College of Radiology

INTRODUCTION can affect radiation dose and image quality and examine Over the past several years, there has been increasing how these can be altered to reduce dose. recognition of the importance of reducing radiation dose In this review, we seek to provide radiologists with a in radiologic studies, particularly with regard to multide- simple approach to the manipulation of 8 CT parameters tector CT. The growth in the volume of multidetector that can be changed by the radiologist, while designing or CT studies performed in the United States, the increas- altering scan protocols, to reduce patient radiation dose: ing use of CT in susceptible populations (including chil- (1) detector configuration, (2) tube current, (3) tube dren), and growing concerns on the part of the general potential (kVp), (4) reconstruction algorithm, (5) pa- public with regard to radiation exposure have provided tient positioning, (6) scan range, (7) reconstructed slice an impetus for performing these studies with the least thickness, and (8) pitch. possible radiation dose [1]. It is now believed that as many as 0.4% of all malignancies in the United States can be traced back to radiation from CT studies per- DETECTOR CONFIGURATION formed between 1991 and 1996 and that radiation from Detector configuration is a term encompassing the num- CT studies currently being performed may ultimately ber of data channels being used in the z axis and the account for 1.5% to 2% of all cancers in the future [2]. “effective detector thickness” of each data channel. For Unfortunately, despite this growing awareness among example, a detector configuration of 64 ϫ 0.5 mm would radiologists, the introduction of each new generation of suggest the use of 64 data channels in the z-axis, each of CT scanners has resulted in scanning protocols that are which has an effective thickness of 0.5 mm. Notably, the increasingly complex and difficult to manipulate for a effective detector thickness represents the smallest possi- radiologist whose primary expertise lies in clinical imag- ble reconstructed slice thickness: if a study is acquired ing rather than medical physics. When facing increasing criticism regarding CT dose, the best way to tackle the with an effective detector thickness of 0.5 mm, images issue is to understand all the factors and parameters that cannot be reconstructed to any smaller interval (such as 0.25 mm). Both the number of channels used and the effective detector thickness can be varied depending on how many channels in a detector array are used, which Department of Radiology, Johns Hopkins University, Baltimore, Maryland. Corresponding author and reprints: Siva P. Raman, MD, Johns Hopkins channels in a detector array are used, and the manner in University, Department of Radiology, JHOC 3251, 601 N Caroline Street, which different channels are combined [3]. Different scan- Baltimore, MD 21287; e-mail: [email protected]. ners from different manufacturers have different included

detector channels, and the detector configurations available most applications. This computer software, which has on different equipment can vary widely [4]. various names depending on the specific vendor (ie, The product of the detector configuration (the num- CARE Dose 4D on Siemens, Dose-Right on Phillips, ber of data channels multiplied by the effective detector Auto mA/Smart mA on GE, and SUREExposure 3D on row thickness) is equivalent to the beam collimation. The Toshiba), automatically increases the mAs in those parts physical manifestation of the beam collimation is the of the body with the greatest attenuation and decreases x-ray beam size as defined at the gantry isocenter [5].In the mAs in those parts of the body with lower attenua- the example given previously, with a detector configuration. As a result, the software usually increases the mAs in tion of 64 ϫ 0.5 mm, the beam collimation would be 32 the shoulder and hips (which have relatively more atten- mm. This has a critical impact on patient dose because uation) and decreases the mAs in the abdomen and tho- the x-ray beam is actually slightly wider than the beam rax (which have relatively less attenuation) [1,7]. In most collimation as a result of a penumbra effect (also referred cases, automated tube current modulation reduces pa- to as “overbeaming”), resulting in a small amount of tient dose and minimizes photon starvation artifacts. “wasted” radiation dose with each rotation [6].Asare- Automated tube current modulation works on differ- sult, if a smaller beam collimation (4 ϫ 0.5 mm) is used, ent principles depending on the specific vendor, al- the relative wasted dose is relatively high compared with though a combination of 4 different strategies is generally the primary beam size. On the other hand, with a larger used: Patient size modulation varies the mAs on the basis 64 ϫ 0.5 mm collimation, the wasted dose is relatively of a global evaluation of the overall size of the patient as low compared with the primary beam size. However, it seen on the scout radiograph. Z-axis modulation changes should be noted that with the newest CT scanners, the the mAs constantly along the z-axis of the patient de- penumbra effect is much less apparent compared with pending on the patient attenuation at each point, as earlier generations of spiral CT technology because the determined using the scout image. Angular (x, y) modu- extent of the overbeaming or penumbra effect is less lation changes the mAs as the x-ray tube rotates 360° pronounced (relative to the beam width). Therefore, the around the patient to account for different attenuations need to obtain a smaller effective detector thickness to in different projections of the x-ray beam [1]. Combined improve spatial resolution in the z-axis must be balanced x, y, and z modulation adjusts the mAs in all 3 axes on the against the increased radiation dose associated with such basis of the patient’s attenuation [8]. The user must de- a detector configuration, as well as longer scan times fine a minimum acceptable image quality, the definition [3,5]. of which will vary depending on the specific vendor and The detector configuration should be determined on system. For example, GE users must specify a noise index the basis of the type of study performed, the necessary value that is maintained as a constant on every image in a slice thickness for multiplanar reformations (MPRs), and series, while Siemens users must specify an image quality the need for 3-D images. For routine acquisitions, with- reference mAs, the mAs that would be used for an aver- out the need for thin-section MPR images or 3-D imag- age-sized patient [1]. ing, an extremely thin effective detector thickness is In almost all cases, automated tube current modulation unnecessary. In such cases, if thick-section 5-mm images should be enabled and used by all users, although some in the axial, coronal, and sagittal planes are sufficient for significant caveats and exceptions must be kept in mind: radiologist review, the effective detector thickness can be ● Small or pediatric patients, when incorrectly positioned widened to 1.25 or 1.5 mm (rather than 0.75 or 0.5 or improperly centered within the CT gantry, can result mm), without a marked impact on image quality (either in excessively noisy images with artificially low mAs. Not for the axial images or MPR images). However, if thin- only must technologists be properly taught to correctly section MPRs or high-resolution 3-D images (with in- position patients, but it may also be helpful to modify the creased spatial resolution in the z-axis) are required, a noise index or increase the reference mAs for smaller thin effective detector thickness (0.5-0.75 mm) may be patients (while still capping the maximum mAs to avoid necessary, although there is an associated dose penalty. inadvertently high radiation doses) [8,9]. ● Many institutions do not use automated tube current TUBE CURRENT modulation when performing scans of the head, particu- Increases in tube current or the product of tube current larly in pediatric patients, because of the difficulty in and scan time (mAs) result in improved image quality, correctly placing the head in the isocenter of the gantry. A decreased image noise, and increased patient dose. In small deviation of the head from the isocenter can result general, the relationship between tube current and pa- in a significant increase in image noise. tient dose is essentially linear, with increases in mAs ● Patients’ arms should be out of the field of view for resulting in a comparable percentage increase in patient both scout-image and axial-image acquisition. Any dose [3]. Although tube current can be manually con- discrepancy in the position of the arms between the trolled, most operators use automated tube current mod- scout and axial images can adversely affect the algo- ulation (also known as automated exposure control) for rithm’s function and theoretically increase dose. If 842 Journal of the American College of Radiology/Vol. 10 No. 11 November 2013

lips), which are relevant to any discussion of tube current. These two essentially synonymous terms, defined as tube current ϫ rotation time/pitch, are used to convey the concept that as the pitch increases on these two vendors’ scanners (with resultant increases in noise), the scanner software will automatically increase the tube current proportionally to compensate for the increased pitch. As long as the effective mAs or mAs per slice remains constant, the noise level (and dose) will remain unchanged [11]. Kvp Reducing kVp can be an effective means of reducing the radiation dose imparted during an examination. As a general rule of thumb, the radiation dose changes with the square of kVp, and a reduction in kVp from 120 to Fig 1. CT colonoscopic prone image acquired at a fixed 100 reduces radiation dose by 33%, while a further re- low 50-mAs protocol. duction to 80 kVp can reduce dose by 65% [12,13].Asa further advantage, particularly in angiographic studies, in which vascular contrast is critical, reductions in kVp result neck and chest CT studies are to be performed at the in an increased attenuation of iodine (even with a constant same time, either two separate acquisitions should be dose of contrast) as a result of the photoelectric effect and performed (with the arms out of the field of view for approaching the k-edge of iodine, potentially improving each), or a single acquisition with the arms up by the contrast-to-noise ratios. Nevertheless, unlike reductions in neck should be used [8]. mAs, which have a linear and relatively predictable effect on ● Only some vendors’ software can correctly account for image noise and contrast-to-noise ratios, decreases in kVp metallic orthopedic hardware. Users should check can result in nonlinear, exponential increases in image noise, with their vendors as to whether their specific software often necessitating a concomitant increase in mAs to pre- is able take this into account. If not, the modulation serve image quality. Regardless, several studies evaluating software could theoretically increase patient dose [8]. the use of low kVp in a number of different protocols (an- In such cases, automatic tube current modulation giography, venous phase studies, cardiac and coronary stud- should be turned off [10]. ies, etc) have shown that such reductions can markedly ● The same noise levels may not be necessary for every reduce radiation doses, while still preserving acceptable im- scan. Each type of study should be individually ap- age quality [12,14,15]. praised, and the appropriate noise level should be de- It is now widely accepted that low-kVp protocols termined. Moreover, for extremely low dose studies (100, 80, and even 70 kVp) are most useful in thin, (such as screening chest CT), it may be easier to use a nonobese patients, particularly in vascular or angio- low-fixed-mAs protocol (Fig. 1) [8]. graphic studies, in which improvements in the attenua- ● In extremely large or obese patients, there may be a tion or conspicuity of iodine can be helpful in terms of conflict between the specified noise index (or reference image contrast-to-noise ratios (Fig. 2) [16]. Most scan- mAs) and the maximum allowed mAs because the ners offer no tools to correctly select an appropriate kVp, maximum mAs may not be sufficient to provide diag- and the decision to decrease kVp must be made at the nostic images. In such cases, the maximum mAs may discretion of the radiologist, taking into account the type need to be increased or a higher kVp used [9]. More- of study being performed (ie, arterial or venous phase over, extreme care should be taken when using tube acquisition), the portions of the body being imaged, and current modulation with obese patients, particularly a global assessment of patient body habitus. Although on the latest generation of scanners that can produce some sites have attempted to use algorithms that base the extremely high tube currents. In these cases, the tube decision on a patient’s body mass index, this may often be current must be appropriately limited or “capped,” or unsatisfactory because it is the distribution of the pa- else patient doses could increase markedly [10]. tient’s body fat that has the ultimate impact on image ● Pediatric modulation settings should not be used for quality. Data regarding the correct kVp selection in any adults, and vice versa, because some scanner software given patient are still evolving, and each practice must packages use separate modulation curves for children determine its comfort level with low-kVp techniques by and adults. Such use may increase doses for children or slowly instituting such changes over time. Although the result in inadequate doses for adults. Society of Cardiovascular Computed Tomography has In addition, users should be aware of the terms effective published recommendations regarding kVp levels in pa- mAs (used by Siemens) and mAs per slice (used by Phil- tients undergoing cardiac CT, including a suggested kVp Raman et al/CT Scan Parameters and Radiation Dose 843

quality, even in cases in which the imparted radiation dose and contrast-to-noise ratios are low. A number of studies using each of the different vendors’ software packages have concluded that iterative reconstruction is a viable option in terms of both reducing radiation dose and improving image quality [19-22]. For example, a study by Kaza et al [23] found that iterative reconstruction techniques could lower radiation dose by up to 30% in CT enterographic studies, while still maintaining acceptable image quality. Overall, when iterative reconstruction is available on a group’s CT scanner, there is little downside to its imple- mentation over more traditional FBP techniques in terms of either diagnostic quality or radiation dose. However, regardless of which vendor’s iterative reconstruction algorithm is used, there is no simple way to determine the extent to which dose can subsequently be reduced. In general, once iterative reconstruction is insti- tuted, dose should gradually be decreased over time (via changes in the acceptable noise index or reference mAs Fig 2. Pediatric chest CT study acquired at 80 kVp. A on the system’s automatic exposure control software) low–tube voltage technique may be most useful in thin or until reaching a point at which the radiologist no longer pediatric patients. feels comfortable with the image quality. Importantly, the images acquired with iterative reconstruction can have a smoothed, “artificial” appearance compared to of 100 in patients weighing Յ90 kg, no similar guide- FBP images, necessitating some acclimatization on the lines have been published for conventional body and part of the interpreting radiologist. neurologic imaging applications [17]. Notably, the increase in image noise associated with lower kVp techniques must be balanced by an increase in PATIENT POSITIONING the tube current. The reference mAs (on Siemens scan- In a study by Li et al [24] in 2006, 95% of patients in the ners) or noise index (on GE scanners) must be appropri- series were found to be improperly centered within the ately altered to increase tube current, but the tube current scanner gantry by the CT technologist. Although this should not be raised too much because the dose savings may seem to be of only minimal importance, improper from a lower kVp would be canceled out. Instead, the positioning can have a significant impact on both image tube current must be mildly increased, preserving image noise and patient surface dose. Habibzadeh et al [25] quality (with perhaps minimal degradation), while still found improper centering of phantoms in the scanner ensuring some dose savings [16]. gantry to be associated with an average surface dose increase of 23% and an image noise increase of 7%. Abnor- RECONSTRUCTION ALGORITHM mal positioning in both the horizontal and vertical axes Traditionally, CT reconstruction algorithms have used can have a similar impact, as a study by Kaasalainen et al filtered back projection (FBP). The mathematical algo- [26] using child-sized phantoms found an increase in the rithms underlying FBP assume a “nonexact” relationship breast surface dose of 16% and in the thyroid gland of between the projection data acquired at the CT scanner 24% for phantoms improperly positioned in the vertical and the data displayed in the final image. As a result, FBP axis. Just as important, small variations in patient posi- removes very little noise from the final image, is limited tioning can have a marked effect on image noise, as in the use of data with lower contrast-to-noise ratios, and vertical displacement of a patient by just 6 cm can result generally requires studies to be performed with larger in an increase in noise of 22%. Needless to say, this can radiation doses to be of diagnostic quality [1,18]. have a dramatic secondary effect on radiation dose if As a result, each of the major CT vendors now offers its automatic exposure control is used because radiation own proprietary iterative reconstruction algorithms as a re- doses can be markedly increased by the software to com- placement for FBP. Iterative reconstruction, now increasingly pensate for this increased noise [26-28]. Although pa- feasible because of better algorithms and improvements in tient positioning must be carefully monitored in all hardware, involves the use of computationally intensive itera- patients, extra attention should be paid to this issue in tive “correction loops,” which repeatedly compare the recon- pediatric patients and patients with smaller body habitus structed data with an “ideal” image, until an acceptable final because small errors in positioning such patients are more image is created [1]. Accordingly, these iterative techniques likely to have a significant impact on both noise and allow the reconstruction of images with improved image dose. Moreover, although correct patient centering is 844 Journal of the American College of Radiology/Vol. 10 No. 11 November 2013

an important factor regardless of the type of scanner being used, it is particularly important when using dual- source CT scanners. Given the scanner geometry of dual- source mode (with the second source typically having a smaller field of view), the dose penalty from centering the patient away from the scanner isocenter can be greater than with conventional single-source acquisitions. The primary cause for this association between improper patient centering and increased patient radiation dose is the use of bowtie filters in modern CT scanners. Fig 3. CT coronary artery study with a minimized scan These filters normally compensate for patient attenua- range encompassing only the necessary aspect of the tion during tube rotation, resulting in increased x-ray thorax in the z-axis. intensity to the thickest parts of the patient (ie, the center of the patient) and decreased intensity to the thinnest parts of a patient (ie, the patient surface). The bowtie minimum for any examination, particularly when imag- filter functions under the assumption that the patient is ing structures such as the heart, for which an increased correctly centered within the gantry [26]. In optimal scan range is unnecessary (Fig. 3). Although cardiac stud- circumstances, the bowtie filter can reduce patient dose ies are certainly the most obvious applications in which a by up to 50% [24]. However, when the patient is im- limited scan length can be used, a small scan range may properly positioned, the mathematical assumptions un- be possible in many traditional body imaging applica- derlying this filter break down, and doses to the surface of tions as well: Patel et al [29] found that using a small scan the body increase. range (from the top of the aortic arch to the bottom of the In general, incorrect positioning in the horizontal axis heart) in CT pulmonary angiographic studies can allow should not be a problem in the vast majority of patients. diagnosis of pulmonary embolism without any loss of Virtually every recent scanner has a laser guidance system sensitivity but with a reduction in radiation dose of 48%. that should allow the patient to be correctly placed at the Unfortunately, even if the scan range is kept to the isocenter of the scanner, and all technologists should be absolute minimum, helical CT requires the acquisition trained to correctly use this guidance mechanism. How- of raw data on either end of the scan length, outside of the ever, correct positioning in the vertical axis can be more intended scan range, to acquire sufficient projection data. tricky: in patients who are markedly scoliotic or unable to This can particularly be a problem with higher pitches lie completely flat in the gantry, portions of the patient, and can result in additional dose as a result of wasted dose by necessity, will inevitably be positioned outside of the at the margins of the scan range. However, this effect can isocenter with vertical displacement. In addition, tech- be mitigated with the use of dynamic z-collimation. Un- nologists must be very careful when placing pillows unlike a fixed collimator, which is fully open throughout the der the patient, as excessive elevation of the patient’s head scan range, the dynamic z-collimator opens only that with pillows may cause portions of the patient’s upper portion of the collimator entering the scan range (while body and head to be above the scanner’s isocenter. Un- at the beginning of the scan) and closes the proximal edge fortunately, an error in positioning may not be immedi- of the collimator while leaving the scan range, thus min- ately apparent to either the radiologist or the technologist imizing dose outside the intended scan range and result- because there may not be a dramatic change in the vol- ing in a more rectangular dose profile. Dose savings with ume CT dose index: increased surface dose to portions of the use of dynamic z-collimation can be incremental, the body above the scanner’s isocenter may be balanced although the exact extent of savings will depend on the by decreased surface dose to portions of the body below CT pitch [30,31]. the isocenter, and the volume CT dose index may seem unchanged, although portions of the body have received RECONSTRUCTED SLICE THICKNESS markedly increased skin dose. As a result, this error must As the reconstructed slice thickness decreases, the num- be prevented through active training of technologists, ber of photons within each voxel also decreases, resulting who must be instructed not only to correctly position the in increased image noise. To maintain constant noise patients but also to turn off automatic tube current mod- levels within an image with a smaller slice thickness, the ulation in any case in which there is difficulty in properly radiation dose must be consequently increased [3]. With positioning the patient at the isocenter of the scanner. the introduction of automatic tube current modulation, the interaction between the user-specified level of accept- SCAN RANGE able noise, reconstructed slice thickness, and radiation For many CT applications, significant reductions in scan dose can be quite complex and will undoubtedly vary range may be neither possible nor desirable. Neverthe- depending on the CT scanner and the specific automatic less, the scan range should be reduced to the needed tube current modulation software being used [32].In Raman et al/CT Scan Parameters and Radiation Dose 845

general, the larger the reconstructed slice thickness used, true dose savings in nonvascular studies in the chest or the lower the patient dose. It is worth repeating, however, abdomen. In particular, in such cases, some systems will that the reconstructed slice thickness cannot be smaller attempt to maintain a constant effective mAs (Siemens) than the acquired slice thickness. or mAs per slice (GE) (tube current ϫ rotation time/ pitch) and will consequently increase the tube current to PITCH compensate for the increased pitch. In addition, in patients who are large, the system may attempt to increase Pitch in the multidetector, spiral CT era is defined as the tube current but may be precluded from doing so as a table travel per rotation divided by beam collimation. result of a capped or limited mAs, and image noise may Pitch Ͻ1 suggests overlap between adjacent acquisi- actually rise. Accordingly, in nonvascular studies in tions, pitch Ͼ1 implies gaps between adjacent acquisi- which scan speed is not as important, using a high-pitch tions, and pitch of 1 suggests that acquisitions are technique may offer little dose savings and potentially contiguous, with neither overlap nor gaps [7]. A smaller worse image quality. That said, high-pitch techniques on pitch, with increased overlap of anatomy and increased these scanners still offer significant value in nonvascular sampling at each location, results in an increased radia- studies in children, in whom scan speed is critical. tion dose. Alternatively, a larger pitch implies gaps in the anatomy and hence lower radiation dose. As a result, if all other parameters are unchanged, increasing pitch reduces CONCLUSIONS radiation dose in a linear fashion [7,33]. Low-pitch tech- As both the general public and clinicians have become nique is associated with less image noise, fewer artifacts, increasingly aware of the concerns associated with mul- and improved signal-to-noise and contrast-to-noise ra- tidetector CT–related radiation dose, it has become crit- tios [34]. For routine body CT protocols, pitch Ͼ1is ical for radiologists to better understand the CT scanner generally acceptable with no compromise to image qual- technology they work with. In particular, a detailed under- ity. However, using higher pitches (pitch Ͼ1.5) can re- standing of a few basic CT scan parameters is essential, and sult in interpolation artifacts and image noise that is knowledge of how to manipulate these parameters to pro- typically unacceptable. For most cardiac CT applica- duce diagnostic images at lower doses is critical for safe tions, pitch values Ͻ0.5 are routinely used, resulting in imaging. higher doses in such studies. These lower pitch values are used to overcome motion artifacts and are also stipulated TAKE-HOME POINTS by the reconstruction algorithms in use. As mentioned earlier, one major caveat to this discussion of the normal ● Several multidetector CT parameters can be changed relationship between pitch and radiation dose is the fact to reduce radiation dose, including detector configu- that certain scanners use the concept of effective mAs or ration, tube current, kVp, reconstruction algorithm, mAs per slice [33]. In these cases, the effect of pitch on patient positioning, scan range, reconstructed slice dose is negated by proportional increases in tube current thickness, and pitch. to maintain similar image noise. ● Although all users should generally use automated Although higher pitch values have traditionally been tube current modulation, users must be careful about thought to be associated with poor image quality, this is using this software in obese patients, children, and no longer true with the latest dual-source CT technology. patients with metallic hardware. A study by Apfaltrer et al [35] using pitch values Ͼ3ona ● Low-kVp protocols may be most useful in thin, non- dual-source scanner for vascular studies suggested that obese patients, particularly in vascular or angiographic image quality was more than acceptable, with radiation studies. dose savings as high as 45% to 50%. The newest dual- ● Iterative reconstruction techniques allow the recon- source scanners offer high-pitch or FLASH mode acqui- struction of images with improved image quality, even sitions with pitch values Ͼ3, consequently decreasing in cases in which the imparted radiation dose and scan times precipitously and potentially reducing radia- contrast-to-noise ratios are low. tion doses significantly. These high-pitch modes mark- ● The scan range should be reduced to the needed min- edly reduce scan times and eliminate many motion- imum for any examination. related artifacts, particularly those previously seen with ● Incorrect positioning of patients in the scanner gantry can cardiac imaging [36]. However, it should be noted that be associated with significant radiation dose penalties. there are significant caveats that should induce some caution before using high-pitch or FLASH acquisitions REFERENCES on every case: Although the data suggest that FLASH 1. Raman SP, Johnson PT, Deshmukh S, et al. CT dose reduction applica- acquisitions offer a significant dose reduction for vascular tions: available tools on the latest generation of CT scanners. J Am Coll studies in the chest, for which retrospective or prospec- Radiol 2013;10:37-41. tive gating would otherwise be used to prevent motion- 2. Brenner DJ, Hall EJ. Computed tomography—an increasing source of related artifacts, it is unclear whether there is actually any radiation exposure. N Engl J Med 2007;357:2277-84. 846 Journal of the American College of Radiology/Vol. 10 No. 11 November 2013

3. Saini S. Multi-detector row CT: principles and practice for abdominal 21. Martinsen AC, Saether HK, Hol PK, et al. Iterative reconstruction re- applications. Radiology 2004;233:323-7. duces abdominal CT dose. Eur J Radiol 2012;81:1483-7. 4. Mahesh M. CT physics: the basics of multi-detector physics. Philadel- 22. Leipsic J, Labounty TM, Heilbron B, et al. Estimated radiation dose phia: Lippincott Williams & Wilkins; 2009. reduction using adaptive statistical iterative reconstruction in coronary CT angiography: the ERASIR study. AJR Am J Roentgenol 2010;195: 5. Verdun FR, Gutierrez D, Schnyder P, et al. CT dose optimization when 655-60. changing to CT multi-detector row technology. Curr Probl Diagn Radiol 2007;36:176-84. 23. Kaza RK, Platt JF, Al-Hawary MM, et al. CT enterography at 80 kVp with adaptive statistical iterative reconstruction versus at 120 kVp with 6. Hamberg LM, Rhea JT, Hunter GJ, et al. Multi-detector row CT: radi- standard reconstruction: image quality, diagnostic adequacy, and dose ation dose characteristics. Radiology 2003;226:762-72. reduction. AJR Am J Roentgenol 2012;198:1084-92. 7. Goldman L. Principles of CT: multislice CT. J Nucl Med Technol 2008; 24. Li J, Udayasankar UK, Toth TL, et al. Automatic patient centering for 36:57-68. MDCT: effect on radiation dose. AJR Am J Roentgenol 2007;188:547-52. 8. Singh S, Kalra MK, Thrall JH, et al. Automatic exposure control in CT: 25. Habibzadeh MA, Ay MR, Asl AR, et al. Impact of miscentering on patient applications and limitations. J Am Coll Radiol 2011;8:446-9. dose and image noise in x-ray CT imaging: phantom and clinical studies. 9. Kalra M, Maher M, Toth T, et al. Techniques and applications of auto- Phys Med 2012;28:191-9. matic tube current modulation for CT. Radiology 2004;233:649-57. 26. Kaasalainen T, Palmu K, Lampinen A, et al. Effect of vertical positioning 10. Mahesh M, Fishman EK. CT dose reduction strategy: to modulate dose or on organ dose, image noise and contrast in pediatric chest CT-phantom not in certain patients? J Am Coll Radiol 2012;9:931-2. study. Pediatr Radiol 2013;43:673-84. 11. Primak AN, McCollough CH, Bruesewitz MR, et al. Relationship be- 27. Toth T, Ge Z, Daly MP. The influence of patient centering on CT dose tween noise, dose, and pitch in cardiac multi-detector row CT. Radio- and image noise. Med Phys 2007;34:3093-101. graphics 2006;26:1785-94. 28. Gudjonsdottir J, Svensson JR, Campling S, et al. Efficient use of auto- 12. Gnannt R, Winklehner A, Eberli D, et al. Automated tube potential matic exposure control systems in computed tomography requires correct selection for standard chest and abdominal CT in follow-up patients with patient positioning. Acta Radiol 2009;50:1035-41. testicular cancer: comparison with fixed tube potential. Eur Radiol 2012; 29. Patel H, Coughlin B, LaFrance T, et al, eds.Comparison of full chest CTA 22:1937-45. with limited CTA and triple rule-out CTA for PE detection and effective 13. Paul JF. Individually adapted coronary 64-slice CT angiography based on dose implications. Chicago, Illinois: Radiological Society of North Amer- precontrast attenuation values, using different kVp and tube current ica; 2007. settings: evaluation of image quality. Int J Cardiovasc Imaging 2011; 30. Shirasaka T, Funama Y, Hayashi M, et al. Reduction of the unnecessary 27(suppl):53-9. dose from the over-range area with a spiral dynamic z-collimator: com- 14. Feuchtner GM, Jodocy D, Klauser A, et al. Radiation dose reduction by parison of beam pitch and detector coverage with 128-detector row CT. using 100-kV tube voltage in cardiac 64-slice computed tomography: a Radiol Phys Technol 2012;5:53-8. comparative study. Eur J Radiol 2010;75:e51-6. 31. Christner JA, Zavaletta VA, Eusemann CD, et al. Dose reduction in 15. Dong F, Davros W, Pozzuto J, et al. Optimization of kilovoltage and tube helical CT: dynamically adjustable z-axis X-ray beam collimation. AJR current-exposure time product based on abdominal circumference: an Am J Roentgenol 2010;194:W49-55. oval phantom study for pediatric abdominal CT. AJR Am J Roentgenol 32. Kanal K, Stewart B, Kolokythas O, et al. Impact of operator-selected 2012;199:670-6. image noise index and reconstruction slice thickness on patient radiation 16. McCollough CH, Chen GH, Kalender W, et al. Achieving routine sub- dose in 64-MDCT. AJR Am J Roentgenol 2007;189:219-25. millisievert CT scanning: report from the summit on management of 33. Mahesh M, Scatarige J, Cooper J, et al. Dose and pitch relationship for a radiation dose in CT. Radiology 2012;264:567-80. particular multislice CT scanner. AJR Am J Roentgenol 2001;177: 17. Halliburton SS, Abbara S, Chen MY, et al. SCCT guidelines on radiation 1273-5. dose and dose-optimization strategies in cardiovascular CT. J Cardiovasc 34. Hu H, Fox SH. The effect of helical pitch and beam collimation on the Comput Tomogr 2011;5:198-224. lesion contrast and slice profile in helical CT imaging. Med Phys 1996; 18. Fleischmann D, Boas FE. Computed tomography—old ideas and new 23:1943-54. technology. Eur Radiol 2011;21:510-7. 35. Apfaltrer P, Hanna EL, Schoepf UJ, et al. Radiation dose and image 19. Desai GS, Uppot RN, Yu EW, et al. Impact of iterative reconstruction on quality at high-pitch CT angiography of the aorta: intraindividual and image quality and radiation dose in multidetector CT of large body size interindividual comparisons with conventional CT angiography. AJR adults. Eur Radiol 2012;22:1631-40. Am J Roentgenol 2012;199:1402-9. 20. Korn A, Fenchel M, Bender B, et al. Iterative reconstruction in head CT: 36. Korn A, Fenchel M, Bender B, et al. High-pitch dual-source CT angiog- image quality of routine and low-dose protocols in comparison with standard raphy of supra-aortic arteries: assessment of image quality and radiation filtered back-projection. AJNR Am J Neuroradiol 2012;33:218-24. dose. Neuroradiology 2013;55:423-30.