Published online before print March 27, 2008, 10.1148/radiol.2472071054
(Radiology 2008;247:499-506.)
© RSNA, 2008
Radiation Dose Reduction Strategy for CT Protocols: Successful Implementation in Neuroradiology Section1
Alice B. Smith, MD,
William P. Dillon, MD,
Benison C. Lau,
Robert Gould, ScD,
Francis R. Verdun, PhD,
Edward B. Lopez, and
Max Wintermark, MD
1 From the Department of Radiology, Neuroradiology Section, University of California, San Francisco, 505 Parnassus Ave, Box 0628, San Francisco, CA 94143-0628 (A.B.S., W.P.D., B.C.L., R.G., E.B.L., M.W.); and University Institute for Radiation Physics, Centre Hospitalier Universitaire Vaudois and University of Lausanne, Switzerland (F.R.V.). Received June 16, 2007; revision requested August 22; revision received August 31; final version accepted October 4.
Address correspondence to M.W. (e-mail: Max.Wintermark{at}radiology.ucsf.edu).
 |
ABSTRACT
|
|---|
Purpose: To retrospectively quantify the effect of systematic use of tube current modulation for neuroradiology computed tomographic (CT) protocols on patient dose and image quality.
Materials and Methods: This HIPAA-compliant study had institutional review board approval, with waiver of informed consent. The authors evaluated the effect of dose modulation on four types of neuroradiologic CT studies: brain CT performed without contrast material (unenhanced CT) in adult patients, unenhanced brain CT in pediatric patients, adult cervical spine CT, and adult cervical and intracranial CT angiography. For each type of CT study, three series of 100 consecutive studies were reviewed: 100 studies performed without dose modulation, 100 studies performed with z-axis dose modulation, and 100 studies performed with x-y-z–axis dose modulation. For each examination, the weighted volume CT dose index (CTDIvol) and dose-length product (DLP) were recorded and noise was measured. Each study was also reviewed for image quality. Continuous variables (CTDIvol, DLP, noise) were compared by using t tests, and categorical variables (image quality) were compared by using Wilcoxon rank-sum tests.
Results: For unenhanced CT of adult brains, the CTDIvol and DLP, respectively, were reduced by 60.9% and 60.3%, respectively, by using z-axis dose modulation and by 50.4% and 22.4% by using x-y-z–axis dose modulation. Significant dose reductions (P < .001) were also observed for pediatric unenhanced brain CT, cervical spine CT, and adult cervical and intracranial CT angiography performed with each dose modulation technique. Image quality and noise were unaffected by the use of either dose modulation technique (P > .05).
Conclusion: Use of dose-modulation techniques for neuroradiology CT examinations affords significant dose reduction while image quality is maintained.
Supplemental material:
http://radiology.rsnajnls.org/cgi/content/full/2472071054/DC1
http://radiology.rsnajnls.org/cgi/content/full/2472071054/DC2
© RSNA, 2008
|
INTRODUCTION
|
|---|
Studies in both the United States and Europe have revealed that computed tomographic (CT) examinations account for only up to 15% of all imaging examinations but contribute up to 75% of the total radiation dose to the population (1–6). For this reason, radiation dose related to CT scans has become a public health concern. It is important for the radiologist to ensure that CT examinations are indicated and that every possible effort is made to reduce the radiation dose.
CT manufacturers have developed tools to assist radiologists in the endeavor to reduce patient dose. Dose modulation, in which tube current is adjusted separately for each CT section according to patient attenuation, is one of the available tools that permit dose reduction. To date, the effect of implementing dose modulation has been evaluated only for individual CT imaging protocols (7–10), but the overall effect of its systematic use within a radiology department or section is mostly unknown. Thus, our goal was to retrospectively quantify the effect of systematic use of tube current modulation for neuroradiology CT protocols on patient dose and image quality.
|
MATERIALS AND METHODS
|
|---|
Study Design and Imaging Protocols
The institutional review board of the University of California San Francisco approved this retrospective study, which was in compliance with the Health Insurance Portability and Accountability Act. Informed consent was waived by the institutional review board.
At our institution (the University of California San Francisco), z-axis dose modulation was introduced with our 16-section CT scanner (Lightspeed; GE Healthcare, Milwaukee, Wis) in September 2004. Prior to the introduction of dose modulation, the tube current in milliamperes was decided subjectively by a registered radiology technologist, in a range set by the radiologists, on the basis of different factors, including the patient's body habitus. When z-axis dose modulation was introduced, the same tube current range was maintained, but a computer algorithm altered the tube current applied to each CT section on the basis of a preset noise index (NI). The NI is a parameter indicative of a level of image noise considered to be acceptable to the radiologist for a given CT examination. A lower NI translates into lower noise and thus into an improved signal-to-noise ratio. However, a lower NI requires higher tube current for a given pitch and tube rotation time and therefore delivers higher patient radiation dose. Establishment of acceptable values for the NI in our institution was accomplished over a 3-month period (from September 2004 through November 2004). The NI was initially set at the lowest value (1) and was progressively increased until the image quality was determined to be insufficient. This was a consensus decision by the nine faculty members (who had 6–25 years of experience as of 2004) of our neuroradiology section for the four CT protocols described below (Tables 1–4).
In August 2006, x-y-z–axis dose modulation was introduced with the installation of our 64-section CT scanners (Lightspeed VCT; GE Healthcare). This kind of dose modulation adds one degree of complexity, in the sense that tube current not only varies from section to section but also can be modulated within a section depending on the attenuation along the x-ray beam path. The same adjustment process for the NI was employed over a 6-week period (from August 2006 through September 2006).
We evaluated our four most frequent types of neuroradiologic CT studies: adult brain CT performed without contrast material (unenhanced CT), pediatric unenhanced brain CT, cervical spine CT in adult patients, and CT angiography of the cervical and intracranial vessels in adult patients. For each type of CT study, we included three series of 100 consecutive patients each: The first series underwent CT with a 16-section CT scanner (Lightspeed; GE Healthcare) before the introduction of z-axis dose modulation, the second series underwent CT with a 16-section CT scanner after the introduction of z-axis dose modulation, and the third series underwent CT with a 64-section CT scanner after the introduction of x-y-z–axis dose modulation. Patient demographic data were recorded. The CT studies performed during the two periods of NI adjustment—September 2004 through November 2004 (3 months of practice before enrollment of patients in the second group—the group who underwent CT with z-axis dose modulation) and August 2006 through September 2006 (6 weeks of practice before enrollment of patients in the third group—the group who underwent CT with x-y-z–axis dose modulation)—were not included in the present study. Practice before enrollment of patients in the first series (the group who underwent CT with no dose modulation) was 12 months. The in-plane resolution (field of view and matrix) for each type of CT study was kept the same for the three series of patients.
For pediatric unenhanced brain CT (Table 2), we developed two categories of indications: "low-dose" indications (hydrocephalus, shunt placement) and "high-dose" indications (nonaccidental trauma, acute neurologic defect, seizure, encephalopathy, and congenital anomalies). Before implementation of dose modulation, technologists were instructed to select low tube current in the 100–150 mA range for low-dose indications and high tube current in the 200–250 mA range for high-dose indications, when image detail was crucial. After implementation of dose modulation, two NI levels were selected for the two types of indications (Table 2).
Image Analysis
For each patient group and study type, quantitative image noise and subjective image quality were evaluated. Quantitative noise was determined by using a previously reported method (11): measuring the standard deviation of the pixel values in a 100-mm2 region of interest that was drawn by a neuroradiologist (A.B.S., with 6 years of experience) in the background defined as the air surrounding the patient, as far as possible from the patient. For image quality, all studies were reviewed in random order by the same neuroradiologist (A.B.S.). The neuroradiologist was blinded as to whether or not dose modulation had been utilized, to the type of scanner with which the study was performed (16- or 64-section CT scanner), and to the tube current used for the study. The neuroradiologist graded the studies for diagnostic acceptability by using a five-point scale (in which, eg, a score of 1 indicated that the study was unacceptable; a score of 3, that the study was average but diagnostic; and a score of 5, that the study was excellent). Specific criteria were used to judge diagnostic acceptability (Appendix E1, Tables E2–E4; http://radiology.rsnajnls.org/cgi/content/full/2472071054/DC1). Grading was based on the ability to define structures and on the sharpness of tissue interfaces. This grading system is similar to those utilized in previous studies by Mulkens et al (9) and Namasivayam et al (10).
Radiation Dose
Radiation dose to the patient was monitored for each study by means of the two standard dose indicators—volume CT dose index (CTDIvol) and dose-length product (DLP)—that were calculated by the CT scanner for each CT study and were automatically saved to a dose report on our picture archiving and communication system. The CTDIvol parameter is representative of the average dose delivered within the reconstructed section. The CTDIvol represents the weighted CT dose index divided by the pitch and describes the average dose throughout a 160-mm-diameter circular Plexiglas phantom, incorporating the central dose weighted by a 1/3 factor and the peripheral dose weighted by a 2/3 factor. The DLP can be related to energy imparted to organs and can thus be used to assess the overall radiation burden of a given examination. It is equal to the product of the CTDIvol and the length of the scan in centimeters (12).
Statistical Analysis
For each type of CT study, differences between the three groups of patients in terms of demographic data, radiation dose descriptors, image quality, and noise were evaluated for statistical significance. Unpaired t tests were used to compare continuous variables between groups; Wilcoxon rank-sum (Mann-Whitney) tests were used for categorical variables such as image quality scores. The statistical significance level was set at .05. Statistical analysis was performed by using software (Stata, version 9.2; Stata, College Station, Tex).
|
RESULTS
|
|---|
Study Patients
There were no significant demographic differences among the three groups of patients (those undergoing CT with no dose modulation, z-axis dose modulation, or x-y-z–axis dose modulation) (Table E5, http://radiology.rsnajnls.org/cgi/content/full/2472071054/DC2). This was true for the four types of CT studies that were assessed.
Radiation Dose
With the 16-section scanner, both CTDIvol and DLP were lowered by up to 60.9% for unenhanced brain CT in adults performed by using dose modulation compared with these values for unenhanced brain CT performed without dose modulation (Table 5). With the 64-section scanner, a 3.75-mm section thickness could not be obtained, and thinner 2.5-mm sections were obtained instead, resulting in a greater number of sections for the 64-section scanner. Also, the tube current and CTDIvol were greater for the 64-section scanner when x-y-z–axis dose modulation was used than for the 16-section scanner when z-axis dose modulation was used, so that we could achieve the same NI for a thinner section thickness with the 64-section scanner.
Regarding pediatric unenhanced brain CT (Table 6), CTDIvol and DLP were lowered by up to 57.9% when dose modulation was used compared with CTDIvol and DLP when dose modulation was not used. The same differences in terms of numbers of sections, tube current, CTDIvol, and DLP as in the adult patients were observed between the three groups of pediatric patients.
Regarding cervical spine CT in adult patients (Table 7), CTDIvol and DLP were lowered by up to 37.4% when dose modulation was used compared with CTDIvol and DLP when dose modulation was not used.
Regarding cervical and intracranial CT angiography in adult patients (Table 8), CTDIvol and DLP were lowered by up to 38.5% when dose modulation was used compared with CTDIvol and DLP when dose modulation was not used.
Image Quality and Noise
Image quality and noise were not affected by the use of dose-modulation algorithms (Tables E6–E9, http://radiology.rsnajnls.org/cgi/content/full/2472071054/DC2).
|
DISCUSSION
|
|---|
The results of our study demonstrate that use of dose modulation as a radiation dose reduction tool for the CT examinations most frequently performed in our neuroradiology section resulted in a substantial reduction in radiation dose while image quality was maintained. Dose modulation requires selection of an acceptable noise level and a range of tube currents, after which a computer algorithm adjusts x-ray tube current to the appropriate amount within this range. This achieves the selected signal-to-noise ratio in the CT sections given the patient's attenuation and girth at each level. Identification of optimal signal-to-noise ratio for each type of CT protocol requires fine tuning to lower the tube current as much as possible while preserving image quality.
The effect of dose modulation on unenhanced brain CT was uncertain at onset given that the head is a spheroid structure with similar attenuation throughout, which differs from other body parts, such as the chest. Nevertheless, we found that dose modulation resulted in significant reductions in radiation dose to adults and children (P < .001). Indeed, for adult unenhanced brain CT, CTDIvol and DLP, respectively, were reduced by 60.9% and 60.3% by using z-axis dose modulation and by 50.4% and 22.4% by using x-y-z–axis dose modulation. Of note, z-axis dose modulation was used with a 16-section CT scanner, with which the section thickness classically used was 3.75 mm. When we transitioned to a 64-section CT scanner, x-y-z–axis dose modulation was used. With the latter scanner, the 3.75-mm section thickness was no longer available. We initially tried 5 mm, but this section thickness was judged to be insufficient by our faculty in terms of spatial resolution, and we selected 2.5 mm instead. This required an increase in tube current compared with that used with the 16-section CT scanner to maintain the same noise level. Because gantry rotation was 2 seconds for the 16-section CT scanner and 1 second for the 64-section CT scanner, 182.2 mAs (91.1 mA · 2 seconds) and 222.9 mAs (222.9 mA · 1 second) were used with these two types of CT scanners, respectively. Compared with the 16-section CT scanner with z-axis modulation, the 64-section CT scanner with x-y-z–axis modulation yielded lower CTDIvol and especially lower DLP reduction but provided images with better longitudinal resolution (thinner detector configuration and section thickness).
Lower DLP reduction was contributed to by a slightly larger spatial coverage along the z-axis. It may also have been contributed to by factors such as beam filtration and the distance of the focal spot from the system isocenter, as well as by an overranging phenomenon (also called z-overscanning) (13). Overranging happens in helical scan mode because the reconstruction algorithm requires extra rotations on both sides, outside the planned length for data interpolation during image reconstruction (14). This results in radiation exposure of tissues beyond the boundaries of the volume to be imaged (15). In our case, when the detector configuration went from 4 x 3.75 mm (beam collimation of 15 mm with the 16-section scanner) to 8 x 2.5 mm (beam collimation of 20 mm with the 64-section scanner), the CTDIvol increased from 36.2 mGy to 46.0 mGy (approximately 30%), and, for the same scanned length, the DLP should have increased in the same proportion. However, our results show that the DLP increased by 95% (from 560 to 1093 mGy · cm). The increment (95% rather than 30%) is due to the overranging phenomenon, which would have been even worse had a beam collimation of 40 mm been used with the 64-section scanner. Overranging is a major issue when scanning a small length (such as the brain) or when trying to spare some organs (lenses, thyroid gland) from radiation during a CT study (15,16). CT manufacturers are presently trying to address this issue, which will be of even greater concern for increasing detector-width CT scanners such as the 256-section CT scanner.
Radiation doses associated with pediatric unenhanced brain CT were lower than in adults. This was due to the smaller size of the head in pediatric patients and increased awareness about radiation risk in children at baseline, which resulted in lower tube current being used before implementation of dose modulation. Two protocols, high dose and low dose, were already being utilized, depending on the clinical indication for the study. However, dose modulation afforded additional dose reduction for both types of indications (CTDIvol and DLP, respectively, were reduced by 57.9% and 55.4% with z-axis dose modulation and by 52.6% and 25.8% with x-y-z–axis dose modulation).
Cervical spine CT and CT angiography were different from unenhanced brain CT in that the attenuation changed more markedly through the scan in cervical spine CT and CT angiography because of variations in anatomy, which is appropriate for the dose-modulation approach. However, a thinner section thickness (1.25 or 0.625 mm) is required to achieve better spatial resolution and enable assessment of fine anatomic details in these studies. Overall, the use of thinner sections resulted in a dose reduction that was significant (P < .001) but less pronounced than that at unenhanced brain CT. CT angiographic studies, in addition, require better contrast resolution, which is reflected by the selection of lower NIs (4–6 for CT angiography, 12 for cervical spine CT) corresponding to increased signal-to-noise ratios.
For all types of studies, dose reduction was less with the 64-section CT scanner than with the 16-section CT scanner, despite the use of x-y-z–axis dose modulation (z-axis dose modulation was used with the 16-section CT scanner). This is likely explained by a trend of obtaining more and thinner sections with the 64-section CT scanner and to the overranging phenomenon mentioned above and thus is probably unrelated to the dose-modulation approach. Dose modulation, however, was instrumental in maintaining acceptable radiation doses for 64-section CT scanner protocols despite the increase in the number of sections and the decrease in section thickness.
There were limitations to our study. Because our study was performed with CT scanners from one manufacturer only, these results should be confirmed in studies that evaluate CT scanners from other vendors. However, we do not anticipate substantial differences, as the principle and effect of tube current modulation should be similar among instruments from different vendors. Also, we did not investigate the effect of dose modulation on all our CT protocols, but rather limited our evaluation to the types of CT studies most frequently performed in our department. Our comparisons focused on the switch from 16-section to 64-section CT scanners but may have been confounded by other changes in practice that occurred over time, such as changes in radiology technology staff. Finally, we included 64-section CT scanner protocols with x-y-z–axis dose modulation in our comparison, even if they differed substantially from 16-section CT protocols (the effect of this difference is detailed in Appendix E1, http://radiology.rsnajnls.org/cgi/content/full/2472071054/DC1) and even though we did not perform any study with the 64-section CT scanner without dose modulation. After introducing dose modulation with our 16-section CT scanners and observing the significant associated dose reduction, it would have been unethical, in our opinion, not to implement dose modulation with the 64-section CT scanners when they were installed. We could have limited our report to 16-section CT scanners, but we believed that it was important to report dose findings for the 64-section CT scanner as well, considering the growing number of such scanners being installed worldwide.
In conclusion, we recommend routine use of dose modulation for neuroradiology CT examinations, because this approach affords a significant dose reduction while preserving image quality. Implementation of dose modulation requires a fine-tuning process to identify optimal signal-to-noise level for each type of CT study performed.
|
ADVANCES IN KNOWLEDGE
|
|---|
- Utilization of dose-modulation techniques resulted in a significant reduction (up to 60.9%, P < .001) in radiation dose to patients for CT studies typically utilized in our neuroradiology section.
- The dose-modulation technique can be implemented without a loss of image quality.
|
IMPLICATION FOR PATIENT CARE
|
|---|
- Dose-modulation techniques can be systematically used for neuroradiology CT examinations to reduce patient radiation dose.
|
FOOTNOTES
|
|---|
Abbreviations: CTDIvol = volume CT dose index • DLP = dose-length product • NI = noise index
Author contributions: Guarantor of integrity of entire study, M.W.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, A.B.S., M.W.; clinical studies, A.B.S., W.P.D., M.W.; experimental studies, R.G., F.R.V., M.W.; statistical analysis, B.C.L., E.B.L., M.W.; and manuscript editing, all authors
Authors stated no financial relationship to disclose.
|
References
|
|---|
- Aroua A, Burnand B, Decka I, Vader JP, Valley JF. Nation-wide survey on radiation doses in diagnostic and interventional radiology in Switzerland in 1998. Health Phys 2002;83:46–55.[CrossRef][Medline]
- Imhof H, Schibany N, Ba-Ssalamah A, et al. Spiral CT and radiation dose. Eur J Radiol 2003;47:29–37.[CrossRef][Medline]
- Kalra MK, Maher MM, Toth TL, et al. Strategies for CT radiation dose optimization. Radiology 2004;230:619–628.[Abstract/Free Full Text]
- 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]
- Mayo JR, Aldrich J, Muller NL. Radiation exposure at chest CT: a statement of the Fleischner Society. Radiology 2003;228:15–21.[Abstract/Free Full Text]
- Wiest PW, Locken JA, Heintz PH, Mettler FA Jr. CT scanning: a major source of radiation exposure. Semin Ultrasound CT MR 2002;23:402–410.[CrossRef][Medline]
- Kalra MK, Rizzo S, Maher MM, et al. Chest CT performed with z-axis modulation: scanning protocol and radiation dose. Radiology 2005;237:303–308.[Abstract/Free Full Text]
- Mastora I, Remy-Jardin M, Delannoy V, et al. Multi-detector row spiral CT angiography of the thoracic outlet: dose reduction with anatomically adapted online tube current modulation and preset dose savings. Radiology 2004;230:116–124.[Abstract/Free Full Text]
- Mulkens TH, Bellinck P, Baeyaert M, et al. Use of an automatic exposure control mechanism for dose optimization in multi-detector row CT examinations: clinical evaluation. Radiology 2005;237:213–223.[Abstract/Free Full Text]
- Namasivayam S, Kalra MK, Pottala KM, Waldrop SM, Hudgins PA. Optimization of Z-axis automatic exposure control for multidetector row CT evaluation of neck and comparison with fixed tube current technique for image quality and radiation dose. AJNR Am J Neuroradiol 2006;27:2221–2225.[Abstract/Free Full Text]
- Wintermark M, Maeder P, Verdun FR, et al. Using 80 kVp versus 120 kVp in perfusion CT measurement of regional cerebral blood flow. AJNR Am J Neuroradiol 2000;21:1881–1884.[Abstract/Free Full Text]
- Hamberg LM, Rhea JT, Hunter GJ, Thrall JH. Multi-detector row CT: radiation dose characteristics. Radiology 2003;226:762–772.[Abstract/Free Full Text]
- van der Molen AJ, Geleijns J. Overranging in multisection CT: quantification and relative contribution to dose—comparison of four 16-section CT scanners. Radiology 2007;242:208–216.[CrossRef][Medline]
- Tzedakis A, Damilakis J, Perisinakis K, Stratakis J, Gourtsoyiannis N. The effect of z overscanning on patient effective dose from multidetector helical computed tomography examinations. Med Phys 2005;32:1621–1629.[CrossRef][Medline]
- Tzedakis A, Perisinakis K, Raissaki M, Damilakis J. The effect of z overscanning on radiation burden of pediatric patients undergoing head CT with multidetector scanners: a Monte Carlo study. Med Phys 2006;33:2472–2478.[CrossRef][Medline]
- Mazonakis M, Tzedakis A, Damilakis J, Gourtsoyiannis N. Thyroid dose from common head and neck CT examinations in children: is there an excess risk for thyroid cancer induction? Eur Radiol 2007;17:1352–1357.[CrossRef][Medline]
This article has been cited by other articles:
|
|
|
|
 
C.B. Nauer, A. Eichenberger, P. Dubach, J. Gralla, and M. Caversaccio
CT Radiation Dose for Computer-Assisted Endoscopic Sinus Surgery: Dose Survey and Determination of Dose-Reduction Limits
AJNR Am. J. Neuroradiol.,
March 1, 2009;
30(3):
617 - 622.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
C.B. Nauer, F. Kellner-Weldon, G. Von Allmen, D. Schaller, and J. Gralla
Effective Doses from Scan Projection Radiographs of the Head: Impact of Different Scanning Practices and Comparison with Conventional Radiography
AJNR Am. J. Neuroradiol.,
January 1, 2009;
30(1):
155 - 159.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
J. M. Provenzale
Advances in Pediatric Neuroradiology: Highlights of the Recent Medical Literature
Am. J. Roentgenol.,
January 1, 2009;
192(1):
19 - 25.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
