Seesaw Balancing Radiation Dose and IV Contrast Dose - AJR

M e d i c a l P hy s i c s a n d I n f o r m a t i c s • O r i g i n a l R e s e a r c h Fält et al. CT Risk Reduction Medical Physics and Informatics Original Research

Seesaw Balancing Radiation Dose and IV Contrast Dose: Evaluation of a New Abdominal CT Protocol for Reducing Age-Specific Risk Tobias Fält 1 Marcus Söderberg2 Lisa Hörberg1 Ingela Carlgren1 Peter Leander 1 Fält T, Söderberg M, Hörberg L, Carlgren I, Leander P

OBJECTIVE. The purpose of this study was to evaluate an abdominal CT protocol in which radiation dose was reduced and IV contrast dose increased for young patients and radiation dose was increased and IV medium dose decreased for elderly patients. The hypothesis was that these adjustments would result in constant image quality and a reduction in age-specific risk. MATERIALS AND METHODS. Patients were divided into four age groups of 25 patients each: group 1, 16–25 years; group 2, 26–50 years; group 3, 51–75 years; and group 4, older than 75 years. The quality reference tube load ranged from 100 to 300 mAs, and the IV contrast dose ranged from 600 to 350 mg I/kg. Group 3 was the reference group. Signal-tonoise and contrast-to-noise ratios for a hypothetical hypovascular liver metastatic lesion were calculated. Subjective image quality was evaluated by visual grading characteristic analysis in which four readers assessed the reproduction of seven image-quality criteria. RESULTS. Radiation dose was reduced 57% in the youngest group, and the IV contrast dose was reduced 18% in elderly patients. There were no statistically significant differences between the groups with respect to signal-to-noise and contrast-to-noise ratios. Subjective image quality was graded significantly lower for four criteria in group 1 compared with group 3. No significant difference was found in comparisons of groups 2 (except for one criterion) and 4 with group 3. CONCLUSION. It is possible to balance radiation dose and contrast dose against each other and maintain signal-to-noise and contrast-to-noise ratios. Subjective image quality was affected by increased noise level on the images but was judged acceptable in all groups except the one with the lowest radiation dose.

C

Keywords: contrast medium, CT, image quality, radiation dose DOI:10.2214/AJR.12.8534 Received January 8, 2012; accepted after revision April 18, 2012. 1 Department of Radiology, Skåne University Hospital, Lund University, Malmö, Södra Förstadsgatan 101, SE-205 02 Malmö, Sweden. Address correspondence to T. Fält ([email protected]). 2 Department of Clinical Sciences, Medical Radiation Physics Malmö, Lund University, Skåne University Hospital, Malmö, Sweden.

AJR 2013; 200:383–388 0361–803X/13/2002–383 © American Roentgen Ray Society

T is commonly used, and its use is steadily increasing [1]. Contrastenhanced CT (CECT) of the abdomen is an important part of the workup for a great number of conditions in patients of all ages. It is fast, robust, and readily available. As with most other procedures, though, CECT also has possible adverse effects for the patient undergoing the examination. The most important risks are radiation-induced injuries and contrast-induced nephropathy (CIN). There is evidence that exposure to high doses of ionizing radiation is a risk factor for cancer development [2]. It has also been found that the lifetime risk of cancer is greater the younger a patient is at exposure [3]. Even though the risks imposed by lower doses of ionizing radiation associated with radiologic examinations such as CT are still under debate, the consensus is to adhere to the as low as reasonably achievable principle and keep a cautious approach to the medical use of ionizing radiation, especially for younger patients.

CIN is defined as a greater than 25% or 44-µmol/L increase in serum creatinine within 3 days after administration of IV contrast medium [4]. The risk of CIN is related to the dose of contrast medium and the number of risk factors. Important risk factors are reduced renal function, congestive heart failure, diabetes, and age older than 70 years [4, 5]. CIN is uncommon in patients with normal renal function [6]. There have been reports [6, 7] that the risk of CIN after CECT may be overstated. It remains important, however, to keep to a minimum the dose of IV contrast medium administered to patients with risk factors and to elderly patients. Quantum noise is inversely related to the square root of the tube load (in milliampere seconds), and tube load is directly related to radiation absorbed dose. Reduction of tube load in a CT protocol thus leads to an increase in image noise. The signal-tonoise ratio (SNR) is a description of the relation between attenuation and image noise in

AJR:200, February 2013 383

Fält et al. TABLE 1: CT Parameters and Sex Distribution in Four Age Groups Age (y)

Female-to-Male Ratio

Tube Load (mAs)a

IV Contrast Dose (mg I/kg)

Maximum Iodine Dose (g)

1

16–25

15:10

100

600

45.0

2

26–50

18:7

150

500

37.5

3

51–75

13:12

200

420

31.5

4

> 75

16:9

300

350

26.3

Group

aQuality reference.

2.0

350 Quantum Noise (Arbitrary)

Quality Reference Tube Load (mAs)

400

300 250 200 150 100

1.5

1.0

0.5

50 0

0

10

20

30

40

50 60 Age (y)

70

80

90

0.0

100

0

10

20

30

40

50 60 Age (y)

70

80

90

100

A

B 2

800

600 CNR (Arbitrary)

IV Contrast Dose (mg IL/kg)

700

500 400 300

1

200 100 0

0

10

20

30

40

50 60 Age (y)

70

80

90

100

0

0

10

20

30

40

C

50 60 Age (y)

70

80

90

100

D

Fig. 1—Theoretic model behind selection of examination parameters for protocol studied. A, Graph shows quality reference tube load plotted in linear relation to patient age. Dot represents formerly used parameter. B, Graph shows resulting arbitrary noise, which is inversely proportional to square root of quality reference tube load. C, Graph shows IV contrast dose plotted to match noise curve. D, Graph shows resulting theoretic constant contrast-to-noise (CNR) ratio.

a specified area. Of greater implication for diagnostic imaging is the attenuation difference between adjacent structures, that is, the contrast. The lower the contrast between two structures, the greater their conspicuity is reduced by increased noise. This relation is described by the contrast-to-noise ratio (CNR).

384

In abdominal CT, diagnosis of most pathologic entities relies on use of contrast medium. A higher contrast dose increases image contrast by a higher iodine concentration in enhancing tissues. Thus it is theoretically possible to balance noise and image contrast against each other by varying the radiation

and contrast doses and maintaining a constant CNR. A similar idea was evaluated by Watanabe et al. [8]. The purpose of this study was to evaluate the idea behind an abdominal CECT protocol introduced in our department in May 2009 to reduce age-specific risk while theoretically

AJR:200, February 2013

CT Risk Reduction

A

B

C

D

Fig. 2—Examples of CT images of abdomen in four age groups. A, 23-year-old woman with minor liver laceration, in group 1. B, 39-year-old man with suspected incarcerated inguinal hernia in group 2. C, 57-year-old woman with abdominal trauma, in group 3. D, 78-year-old woman with suspected gynecologic cancer, in group 4.

maintaining constant image quality. According to this protocol, younger patients receive a reduced radiation absorbed dose and an increased contrast dose, and elderly patients receive a reduced contrast dose compensated by an increased radiation dose. Materials and Methods The institutional review board approved this retrospective study, and the requirement for informed consent was waived. Patients 16 years or older undergoing CECT of the abdomen because of acute symptoms were divided into four age groups: group 1, 16–25 years; group 2, 26–50 years; group 3, 51– 75 years; and group 4, older than 75 years. One hundred patients (63 women, 37 men) examined between May and December 2009 were consecutively selected as 25 patients per group (Table 1). Because adipose tissue is poorly perfused, it may be helpful to adapt the IV contrast dose to the

lean body weight [9]. For this reason, the IV contrast dose in the evaluated protocol was linearly proportional to body weight only up to 75 kg. To keep the number of variables at a minimum in the evaluation of the theoretic model behind the protocol, patients with a body weight exceeding 75 kg were excluded from the study. All patients were examined with a 16-MDCT scanner (Somatom Sensation 16, Siemens Healthcare). The reconstruction kernel was B30f; collimation, 1.5 mm; pitch, 0.65; and tube voltage, 120 kV. By use of an automatic exposure control system (CARE Dose 4D, Siemens Healthcare) the tube load was set to a level-denominated quality reference. The IV contrast medium used was iohexol (Omnipaque 300 mg I/mL, GE Healthcare).

theoretic model shown in Figure 1, which we have likened to a seesaw balancing radiation dose and IV contrast dose. Radiation dose (quality reference tube load in milliampere seconds) was plotted in a linear relation to age. Quantum noise is inversely related to the square root of the radiation dose) (quality reference tube load), and arbitrary noise values were calculated with this relation. To compensate for variation in image noise, the IV contrast dose curve was plotted to match the noise curve. Theoretically, this plot would yield a constant CNR for all ages. The resulting parameters used for the different groups are shown in Table 1. The parameters used for group 3 were the same as in the formerly used abdominal CECT protocol in the department.

Radiation Dose Theoretic Model The quality reference tube load and contrast dose for the different groups were determined with the

Volume CT dose index and dose-length product (DLP) were recorded for each patient. An estimation of the effective dose (E) for each group


Fält et al. TABLE 2: Mean Radiation Dose and Noise Group

CT Dose Index (mGy)

Effective Dose (mSv)

Noise (HU)

1

4.6 ± 0.62

3.6 ± 0.6

16.4 ± 1.7

10

2

8.1 ± 1.3

6.6 ± 1.3

14.9 ± 2.3

3

10.7 ± 1.4

8.5 ± 1.3

12.8 ± 1.9

4

15.9 ± 3.0

12.4 ± 3.2

10.7 ± 1.5

SNR

Fig. 3—Graph shows mean liver signal-tonoise ratio (SNR) and contrast-to-noise ratio (CNR) for hypothetical hypovascular liver metastatic lesion in different groups. Error bars represent 1 SD of measurements.

CNR

8 6 4 2 0

Group 1

Group 2

Group 3

was derived from the dose-length product and the conversion factor EDLP as follows: E = EDLP × DLP, where EDLP is region specific and DLP is the normalized effective dose (mSv / mGy · cm) determined by Huda et al. [10] (abdominal region 0.016).

IV Contrast Dose At administered dose of IV contrast was recorded for each patient.

Objective Evaluation of Image Quality With the PACS (Sectra PACS, Sectra, Linöping, Sweden), the attenuation (signal) in HU was measured by placement of a circular region of interest with a diameter of approximately 3 cm in a homogeneous part of the right liver lobe, avoiding vascular and biliary structures (Fig. 2). Standard deviation (SD) was used as a measurement of image noise, and the liver SNR was calculated. For calculation of the CNR, the difference in attenuation between the measured value and a hypothetical hypovascular liver metastatic lesion with attenuation of 40 HU was used for contrast and then divided by the noise value. Liver metastatic lesions often have attenuation slightly lower than the liver parenchyma, are typically hypovascular, and are not enhancing in the portal venous phase [11, 12].

Subjective Evaluation of Image Quality Image quality was evaluated by visual grading characteristic (VGC) analysis [13]. In VGC analysis, the task of the observers is to rate their opinion about reproduction or visibility of a certain structure. In a previous study [14], the use of VGC analysis made it possible to discriminate small differences in noise levels between abdominal CT images. In the current

386

Group 4

study, the following seven criteria from the European guidelines on quality criteria for CT (EUR 16262) [15] were used: visually sharp reproduction of the liver parenchyma and intrahepatic portal veins; visually sharp reproduction of the pancreatic contours; visually sharp reproduction of the kidneys and proximal ureters; reproduction of the gallbladder wall; visually sharp reproduction of the right adrenal gland from adjacent structures; visually sharp reproduction of the structures of the liver hilum; and reproduction of the choledochal ductus in the pancreatic parenchyma. All criteria were judged absolutely on a 5-grade scale on which grade 1 was unacceptable; 2, substandard; 3, acceptable; 4, above average; and 5, superior. A sixth alternative, not applicable, was allowed for reproduction of the gallbladder wall if the gallbladder had been removed, and the data were excluded from further analysis. The images were presented and evaluated with the free software ViewDEX [16]. The software and all images were loaded on a USB memory stick and inserted for the evaluation into one of the PACS workstations in our department. All images were viewed individually by four readers: two gastrointestinal radiologists with 15 and 20 years of experience and two residents in radiology with 3 and 4 years of experience. The information on patient identity and scanning parameters was removed, and the images were presented at random to each radiologist. The readers could scroll through the stacks, magnify the images, and alter the window width and window level settings. First an adaption series with 12 patients, three from each group, was presented to each reader to accustom the readers to the method and the range of image quality. The results of the test series were not used for evaluation. Thereafter

images from the full set of 100 patients were presented individually to each reader. The rating data were analyzed with methods developed in receiver operating characteristic analysis. A VGC curve was obtained by plotting the cumulative distributions of rating data for two compared systems against each other [13]. The area under the VGC curve (AUC) was used as a measure of the difference in image quality between the two systems. An AUC of 0.5 corresponded to equal image quality of the two systems, an AUC < 0.5 indicated that the image quality was higher for the reference system, and an AUC > 0.5 indicated that the image quality was higher for the evaluated system. Because the parameters used for group 3 were the same as in the formerly used protocol, this group was selected as the reference for the VGC analyses. Software for multiple observers (VGC Analyzer, Båth M, Hansson J, Sahlgrenska Academy at University of Gothenburg) was used to perform calculations comparing groups 1, 2, and 4 with group 3 individually for all seven criteria and all readers combined. If the 95% confidence interval (CI) for the estimation of AUC did not cover the value 0.5, a statistically significant difference at the 95% level between the two evaluated systems was established. Receiver operating characteristic software (Rockit, CE Metz, University of Chicago) was used to calculate interobserver agreement comparing groups 1, 2, and 4 with group 3 individually for all four readers. In a separate analysis, data on 20 patients (age range, 16–25 years) examined before May 2009 with the old protocol (200 mAs quality reference, 420 mg I/kg) were analyzed by three readers and compared with data on 20 weight-matched patients from group 1. This step was taken to study the effect of radiation dose reduction on this age group. The young patients examined with the formerly used protocol were also compared with group 3 to analyze whether there was a difference in image quality between age groups when the same dose parameters were used. VGC analysis gives information only about image quality in relation to another group and not about absolute image quality. It is statistically incorrect to calculate mean values of the grades because they are ordinal data. Therefore, the percentage of grades in each group rated acceptable or better was used as a measure of the absolute level of image quality.

Results Radiation Dose The mean volume CT dose index and mean effective doses for the four groups are shown in Table 2. Compared with the values in group 3, representing the formerly used protocol,


CT Risk Reduction the mean effective dose was reduced 57% in group 1 and 22% in group 2 but increased 46% in group 4.

Objective Evaluation of Image Quality The mean liver SNR in groups 1–4 ranged from 7.3 for group 2 to 8.4 for group 4, and the mean CNR values in groups 1–4 ranged from 4.3 for group 3 to 5.1 for group 1 (Fig. 3). Analysis of variance and Tukey test results did not establish a significant difference in SNR or CNR between the groups. Subjective Evaluation of Image Quality The VGC AUCs for all readers combined are presented in Figure 4. In groups 1 and 2 the AUCs showed significantly lower subjective image quality for four and one of the seven criteria compared with the results in group 3. No significant difference between groups 3 and 4 was found. The percentages of grades acceptable or better (grades 3–5) were 71% for group 1, 80% for group 2, 85% for group 3, and 83% for group 4. The comparison between 20 patients from group 1 and 20 patients 16–25 years old examined with the formerly used protocol (200 mAs, 420 mg I/kg), showed significantly lower subjective image quality in group 1 for four criteria (visually sharp reproduction of the pancreatic contours, visually sharp reproduction of the kidneys and proximal ureters, visually sharp reproduction of the right adrenal gland from adjacent structures, and visually sharp reproduction of the structures of the liver hilum). No significant difference was seen between group 3 and the young patients examined with the formerly used protocol. Interobserver Agreement In a direct comparison of all grades, that is, seven criteria in each of the 100 patients, the agreement between the four readers ranged from 29% to 50%. This interobserver agreement was affected by baseline variation in grades between the readers. Analyzing the agreement of significant differences shown by the AUCs for each criterion instead avoids this baseline variation, and the agreement is better, between 69% and 95% (Table 3).

AUC

IV Contrast Dose The mean administered IV contrast dose in each group was: group 1, 594 mg I = kg; group 2, 492 mg I = kg; group 3, 418 mg I = kg; group 4, 344 mg I = kg. This equals a reduction in IV contrast dose by 18% in group 4 and an increase in IV contrast dose by 42% in group 1 and by 18% in group 2.

1.0

Group 1

Group 2

Group 4

0.5

0.0

1

2

3

4 Criteria

5

6

7

Fig. 4—Graph shows values of area under the visual grading characteristic curve per criterion for all readers combined. Error bars represent the 95% CI of AUC. If the error bars do not include 0.5, subjective image quality is statistically different from reference group (group 3). AUC represents difference in image quality between groups and is not an absolute value of subjective image quality.

TABLE 3: Interobserver Agreement (%) Reader

Reader 1

Reader 2

Reader 3

4

89

69

83

3

95

76

2

82

Discussion According to the generally accepted view, the radiation doses associated with CT examinations may increase the lifetime risk of cancer, and IV contrast medium may induce CIN. The protocol evaluated in this study reduced the radiation effective dose 57% in the youngest patient group with maintained CNR, a reduction made possible by a concurrent increase in IV contrast dose. At the other end of the age spectrum, the IV contrast dose given to elderly patients was reduced 18% at the cost of an increased radiation dose. In this way, the most important risk for each age group was reduced. In support of our theory that radiation dose and contrast dose can be balanced against each other is the lack of difference in SNR and CNR between the groups and the lack of significant difference in subjective image quality between groups 3 and 2 (except for one criterion) and between groups 3 and 4. The lack of significant differences in SNR between the groups shows that the variation in contrast enhancement of the liver can match the variation in noise level caused by as much as a threefold difference in radiation dose. Subjective image quality was significantly lower for four criteria in group 1, which was examined with a considerably reduced radiation dose compared with the reference group. Analysis of the percentage of grades acceptable or better within the groups shows similar levels in groups 2–4, whereas the level was slightly lower for group 1. From these findings

we conclude that even though the subjective image quality was rated lower for one criterion in group 2, it was still acceptable according to the general opinion of the readers, but it was slightly lower than acceptable for group 1. The discrepancy between objective and subjective image quality in this study is of interest and needs to be analyzed further. One explanation may be that the noise level simply was too high in the low radiation groups, thereby negatively affecting image quality. It may also reflect a weakness in the method in that reproduction of normal structures may not be affected by image noise in a way that is directly comparable to the conspicuity of low-contrast lesions. The theory behind the protocol is that increased noise is compensated by increased contrast of a lesion, and this may not be applicable to the normal structures rated. Yet another explanation is that the readers, from everyday practice, were not accustomed to reading high-noise-level images and rated these images lower. An assumption with VGC analysis is that there is a correlation between the reproduction of anatomic and pathologic features, but further studies are needed to determine the diagnostic performance of this protocol. Another weakness of this study was that CNR was calculated with the difference in attenuation between the measured CT number in the liver and a hypothetical hypovascular lesion with the CT number 40 HU. This simulation was used because liver lesions are


Fält et al. rare in patients in groups 1 and 2 and it would not have been possible to accrue enough patients with liver lesions for the study. The four readers had different baselines in grading subjective image quality. Our method included an adaption series of 12 patients viewed individually by each reader to calibrate the grading. The study showed that 12 was not enough. A joint adaptation session may be a more effective way of avoiding baseline differences between readers in future studies. It my also be wise to standardize the individual reading sessions regarding, for example. session length and time of day, thereby further reducing variations in the study. Only patients with a body weight up to 75 kg were included in the study, whereas in clinical practice the protocol is used for all patients independently of body weight. The assumption is that in most individuals, the additional body weight over 75 kg consists mainly of adipose tissue with limited vascular supply and small interstitial space. The amount of adipose tissue therefore has only a minor effect on the concentration of contrast medium in blood and enhancing organs [9]. A more exact way to determine the contrast dose would be to calculate the lean body weight for all patients, but this is not suitable in emergency care. The effect of body size on the noise level of the images is compensated by the automatic exposure control system. Therefore, we believe the findings of this study are applicable to patients with body weights over, as well as under, 75 kg. To our knowledge, there have been no previous studies exploring the concept of balancing radiation dose against contrast dose for abdominal CT. Watanabe et al. [8] evaluated the idea of increasing the IV contrast dose to allow a reduced radiation dose. Those authors, however, used a slightly different method of rating the depiction of vessels when different radiation and contrast doses were used. They found preserved qualitative image quality at 30% reduced radiation dose compensated by a 15% increase in contrast dose. They used a noise index 12 HU as the definition for a standard radiation dose and 15 HU as a low dose, values similar to the noise levels in groups 2 and 3 in our study (Table 2). Previous studies by Kalra et al. [17] and Wessling et al. [18] showed possible radiation dose reduction up to approximately 40% for abdominal CT without a negative effect on subjective image quality or diagnostic acceptability, this without compensation by an increased contrast dose. Hence there seems to be support for a reduction in radiation dose for abdominal CT in general. The finding that it

388

is possible to compensate a reduction in radiation dose by an increase in IV contrast dose may represent a further step in the pursuit of minimizing radiation dose to young patients. Increasing radiation dose to be able to lower the IV contrast dose administered to older patients, leaving other parameters unchanged, has not, to our knowledge, been proposed before. Our results support this concept, and this may be an important finding for reducing the risk of CIN in examinations of older patients with reduced renal function. Future development of the protocol may be to lower the kilovoltage for smaller patients to achieve a higher CNR for the same amount of iodine and to add iterative reconstruction to reduce the noise level on low-radiation-dose images. Another finding of this study is that the noise level affects subjective image quality more than age-related anatomic differences between patients, such as the amount of intraabdominal fat. This knowledge may be useful for future studies comparing image quality between patients of different ages. Conclusion This study showed that increasing the dose of IV contrast medium can compensate for a reduced radiation dose and vice versa while SNR and CNR are maintained. Subjective image quality was affected by increased noise level on the images but was judged acceptable in all groups except the one with the lowest radiation dose. The discrepancy between objective and subjective image quality may, at least in part, be caused by weaknesses in the method of rating normal anatomic structures. In this protocol the radiation dose was reduced 57% in young patients and the dose of IV contrast medium was reduced 18% in elderly patients. References 1. National Council on Radiation Protection and Measurements (NCRP). Ionizing radiation exposure of the population of the United States. Report no. 160. Bethesda, MD: National Council on Radiation Protection, 2009 2. Brenner DJ, Doll R, Goodhead DT, et al. Cancer risks attributable to low doses of ionizing radiation: assessing what we really know. Proc Natl Acad Sci USA 2003; 100:13761–13766 3. Brenner DJ, Hall EJ. Computed tomography: an increasing source of radiation exposure. N Engl J Med 2007; 357:2277–2284 4. Morcos SK, Thomsen HS, Webb JA. Contrastmedia-induced nephrotoxicity: a consensus report. Contrast Media Safety Committee, European Society of Urogenital Radiology (ESUR). Eur

Radiol 1999; 9:1602–1613 5. Stacul F, van der Molen AJ, Reimer P, et al. Contrast induced nephropathy: updated ESUR contrast media safety committee guidelines. Eur Radiol 2011; 21:2527–2541 6. Katzberg RW, Newhouse JH. Intravenous contrast medium-induced nephrotoxicity: is the medical risk really as great as we have come to believe? Radiology 2010; 256:21–28 7. Bruce RJ, Djamali A, Shinki K, Michel SJ, Fine JP, Pozniak MA. Background fluctuation of kidney function versus contrast-induced nephrotoxicity. AJR 2009; 192:711–718 8. Watanabe H, Kanematsu M, Miyoshi T, et al. Improvement of image quality of low radiation dose abdominal CT by increasing contrast enhancement. AJR 2010; 195:986–992 9. Kondo H, Kanematsu M, Goshima S, et al. Body size indexes for optimizing iodine dose for aortic and hepatic enhancement at multidetector CT: comparison of total body weight, lean body weight, and blood volume. Radiology 2010; 254:163–169 10. Huda W, Ogden KM, Khorasani MR. Converting dose-length product to effective dose at CT. Radiology 2008; 248:995–1003 11. Kanematsu M, Kondo H, Goshima S, et al. Imaging liver metastases: review and update. Eur J Radiol 2006; 58:217–228 12. Heiken JP, Menias CO, Elsayes K. Liver. In: Lee JK, Sagel SS, Stanley RJ, Heiken JP, eds. Computed body tomography with MRI correlation, 4th ed., vol. 1. Philadelphia, PA: Lippincott Williams & Wilkins, 2006:829–930 13. Båth M, Månsson LG. Visual grading characteristics (VGC) analysis: a non-parametric rank-invariant statistical method for image quality evaluation. Br J Radiol 2007; 80:169–176 14. Leander P, Söderberg M, Fält T, Gunnarsson M, Albertsson I. Post-processing image filtration enabling dose reduction in standard abdominal CT. Radiat Prot Dosimetry 2010; 139:180–185 15. Bongartz G, Golding SJ, Jurik AG, et al. European guidelines on quality criteria for computed tomography. Report EUR 16262. Luxembourg, Luxembourg: European Commission, 1999 16. Börjesson S, Håkansson M, Båth M, et al. A software tool for increased efficiency in observer performance studies in radiology. Radiat Prot Dosimetry 2005; 114:45–52 17. Kalra MK, Maher MM, Kamath RS, et al. Sixteen-detector row CT of abdomen and pelvis: study for optimization of Z-axis modulation technique performed in 153 patients. Radiology 2004; 233:241–249 18. Wessling J, Esseling R, Raupach R, et al. The effect of dose reduction and feasibility of edge-preserving noise reduction on the detection of liver lesions using MSCT. Eur Radiol 2007; 17:1885–1891
