CT colonography at low tube potential: using iterative reconstruction to decrease noise

Clinical Radiology 70 (2015) 981e988

Contents lists available at ScienceDirect

Clinical Radiology journal homepage: www.clinicalradiologyonline.net

CT colonography at low tube potential: using iterative reconstruction to decrease noise K.J. Chang a, *, M.A. Heisler a, M. Mahesh b, G.L. Baird a, W.W. Mayo-Smith a a

Department of Diagnostic Imaging, The Warren Alpert Medical School of Brown University, Rhode Island Hospital, 593 Eddy St., Providence, RI 02903, USA b The Russell H. Morgan Department of Radiology and Radiological Science, The Johns Hopkins School of Medicine, 601 N. Caroline St, Room 3140D, Baltimore, MD 21287, USA

art icl e i nformat ion Article history: Received 8 July 2014 Received in revised form 24 March 2015 Accepted 12 May 2015

AIM: To determine the level of iterative reconstruction required to reduce increased image noise associated with low tube potential computed tomography (CT). MATERIALS AND METHODS: Fifty patients underwent CT colonography with a supine scan at 120 kVp and a prone scan at 100 kVp with other scan parameters unchanged. Both scans were reconstructed with filtered back projection (FBP) and increasing levels of adaptive statistical iterative reconstruction (ASiR) at 30%, 60%, and 90%. Mean noise, soft tissue and tagged fluid attenuation, contrast, and contrast-to-noise ratio (CNR) were collected from reconstructions at both 120 and 100 kVp and compared using a generalised linear mixed model. RESULTS: Decreasing tube potential from 120 to 100 kVp significantly increased image noise by 30e34% and tagged fluid attenuation by 120 HU at all ASiR levels (p<0.0001, all measures). Increasing ASiR from 0% (FBP) to 30%, 60%, and 90% resulted in significant decreases in noise and increases in CNR at both tube potentials (p<0.001, all comparisons). Compared to 120 kVp FBP, ASiR greater than 30% at 100 kVp yielded similar or lower image noise. CONCLUSIONS: Iterative reconstruction adequately compensates for increased image noise associated with low tube potential imaging while improving CNR. An ASiR level of approximately 50% at 100 kVp yields similar noise to 120 kVp without ASiR. Ó 2015 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.

Introduction Computed tomography (CT) colonography (CTC) is a very useful screening examination for the detection of adenomatous polyps as well as occult colon and rectal carcinomas.1,2 A potential impediment to screening CTC is the concern regarding radiation to patients undergoing a screening examination.3 The true risk of very low levels of * Guarantor and correspondent: K.J. Chang, The Warren Alpert Medical School of Brown University, Rhode Island Hospital, Department of Diagnostic Imaging, 593 Eddy St., Providence, RI 02903, USA. Tel.: þ1 (401) 444 5184; fax: þ1 (401) 444 5017. E-mail address: [email protected] (K.J. Chang).

intermittent radiation is unknown,4e8 but patients, referring physicians, and government administrative agencies remain concerned about radiation exposure. The ability to adequately visualise the colonic mucosal surface is dependent on the attenuation difference between the colonic wall and adjacent gas and fluid or stool, a distinction that need not require a large dose of radiation. Many advances have been made in decreasing radiation dose through aggressive decreases in tube current9e11 and automatic tube current modulation.12e16 Decreasing tube potential has also led to a decrease in radiation dose, although at the expense of an increase in image noise.17 While radiation dose decreases linearly with a decrease in tube current, dose decreases by a power of 2.6 with tube

http://dx.doi.org/10.1016/j.crad.2015.05.005 0009-9260/Ó 2015 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.

982

K.J. Chang et al. / Clinical Radiology 70 (2015) 981e988

potential reduction,18,19 which makes this a powerful technique to decrease radiation exposure. In addition, at a lower tube potential, the average energy of the X-ray beam more closely approaches the K-edge of iodine leading to an increase in attenuation of iodinated contrast medium.20 The use of a lower tube potential to reduce radiation dose and increase the contrast medium attenuation of tagged colonic fluid has already been demonstrated in CTC21 as well as multiple other contrast-enhanced applications17,22e28; however, in one study, the decrease in tube potential from 120 to 100 kVp resulted in a 32% increase in image noise.21 Iterative reconstruction methods such as adaptive statistical iterative reconstruction (ASiR, GE Healthcare, Waukesha, WI, USA) use an alternative method of reconstructing CT data that reduces image noise while preserving edge detail29,30 when compared to filtered back projection. Through the use of iterative reconstruction, significant decreases in tube current have been achieved while preserving image quality. The use of ASiR has been demonstrated in CTC where tube current was decreased with ASiR resulting in a decreased level of noise and no significant change in image quality.31 Similarly, the use of other iterative reconstruction methods such as Veo (GE Healthcare), iDose4 (Philips Healthcare, Best, The Netherlands), and AIDR 3D (Toshiba Medical Systems Corporation, Otawara, Japan) have also resulted in reduced image noise and lower radiation dose at lower tube current in CTC.32e34 The complex interaction between iterative reconstruction, tube potential, and automatic tube-current modulation has been investigated in an abdominal phantom as well as in CT angiography.35e38 To the authors’ knowledge, however, the effect of iterative reconstruction on examinations using lower tube potential has not been investigated in non-contrast CT, including CTC. The use of an iterative reconstruction method should also be able to compensate for the increase in image noise that accompanies a reduction in tube potential. CTC, which inherently involves imaging patients in both the supine and prone positions without the confounding variable of different phases of intravenous contrast medium, serves as an ideal in vivo model to evaluate the effects of changes in scan variables such as tube potential and iterative reconstruction settings. The purpose of this study was 1. To assess the impact of using different levels of iterative reconstruction to mitigate the increase in image noise associated with decreasing tube potential in patients undergoing CTC, and 2. To determine what level of iterative reconstruction at 100 kVp would produce noise levels similar to filtered back projection images at 120 kVp.

Materials and methods This study was compliant with the Health Information Portability and Accountability Act and approval was obtained from the Institutional Review Board (IRB). The need for informed consent was waived by the hospital’s IRB.

Patient population and bowel preparation Between January and July 2012, 50 consecutive patients undergoing clinically indicated CTC were included in the study. The patient sample was comprised of 12 men (mean age 66 years, range 46e85 years) and 38 women (mean age 67 years, range 31e89 years) with an overall mean age of 67 years. The chief indication was incomplete colonoscopy or contraindication to colonoscopy. Patient size was not an exclusion criterion with torso anteroposterior (AP) diameters ranging from 17.6 cm to 37 cm with a mean of 25.7 cm. Patients with a history of previous incomplete colonoscopy underwent a magnesium citrate based bowel preparation the day prior to scheduled CTC (36 patients). The department’s bowel preparation consisted of a clear liquid diet for 24 hours prior to CTC and oral administration of 10 oz magnesium citrate and 20 mg bisacodyl at 6 pm on the evening prior to CTC. Patients presenting to the Radiology Department immediately following a same-day incomplete colonoscopy had taken the standard optical colonoscopy bowel preparation (polyethylene glycol-based, typically 2 l) starting at 4 pm the previous day (14 patients). Although overnight versus same-day bowel preparation could not be randomized, subsequent analyses revealed no statistically significant interaction effects for bowel preparation. Patients also underwent faecal and fluid tagging using 30 ml diatrizoate meglumine and diatrizoate sodium at a concentration of 370 g/ml (MD-Gastroview, Covidien, Mansfield, MA, USA) either undiluted or diluted in 8 oz of clear fruit juice. This was orally administered before bedtime on the night prior to the study if the study was prescheduled, or 2 hours prior to the study for those presenting immediately following incomplete colonoscopy (using a protocol described previously by Chang et al.39). Barium was not used as a tagging agent.

CT technique and image reconstruction CTC examinations were performed on one of two 64detector row multidetector CT machines with identical hardware and software (Lightspeed VCT; GE Healthcare). A balloon-tipped silicone catheter was inserted per rectum and insufflation of carbon dioxide was performed using an automated insufflation device (PROTOCO2L Colon Insufflation System; Bracco Imaging, Milan, Italy) to maximum patient tolerance or an equilibrium pressure of 25 mmHg. No spasmolytic agents or intravenous contrast materials were used. After adequate colonic insufflation (as judged on scout images), image acquisition was performed in single breath-holds to include the entire colon in both the supine and prone position. Detector acquisition parameters were 0.625 mm collimation reconstructed to 1.25 mm thin sections at 0.625 mm intervals. Automatic tube-current modulation was used with a milliampere range of 50e450 mA. All CT examinations were acquired from the diaphragm to the greater trochanters using a noise index of 25 based off 5 mm thick sections and a gantry rotation time of 0.5 seconds in both supine and prone positions with a

K.J. Chang et al. / Clinical Radiology 70 (2015) 981e988

983

pitch of 1.375:1 and a table speed of 55 mm/rotation. A peak tube potential of 120 kVp was used for supine images and 100 kVp for prone images. The automatic tube-current modulation noise index of 25 was maintained for 100 kVp images. Noise index is a user-selectable level of quantum mottle acceptable for a particular CT examination and is defined by the manufacturer to approximate the amount of image noise present in the centre of a homogeneous phantom.18 Both 120 and 100 kVp datasets were reconstructed using traditional filtered back projection as well as increasing percentage blends of ASiR at settings of 30%, 60%, and 90% yielding eight volumetric datasets per patient. In this manner, the patients acted as their own controls with the only difference in the 120 and 100 kVp datasets being the patients’ position (supine or prone).

Data collection CTC images were transferred to a picture archiving and communications system (PACS) workstation (GE Centricity, GE Healthcare) and post-processed on a commercially available 3D workstation (Advantage Workstation, version 4.3 with CTC Pro, GE Healthcare). Evaluations of regions of interest (ROI) were determined on the two-dimensional (2D) axial source images using the largest possible round or oval ROI not subject to partial volume averaging effects. On both 120 kVp axial supine images and 100 kVp axial prone images at each ASiR level, ROIs were placed in two regions of air outside the patient bilaterally to measure noise. ROIs were also placed in the bilateral psoas muscles with mean attenuation used as a surrogate for soft tissue, and in the largest dependent collection of fluid in the ascending and descending colon to measure tagged fluid (Fig 1). Care was taken to avoid including clothing material and sheets when measuring ROIs in air for noise. ROIs were also chosen to exclude any residual solid stool, as well as to avoid partial volume averaging with adjacent colonic wall, luminal gas, and haustral folds on the same section and on sections above and below. With two ROIs placed for each of the eight datasets, each patient had 16 ROIs placed for each measure; however, the attenuation of tagged colonic fluid could only be measured in 35 out of 50 patients in the study as the other patients did not show adequate amounts of tagged residual colonic fluid for ROI placement. As fluid tagging serves to increase the contrast between the soft-tissue density of a submerged polyp and the contrast in the adjacent colonic fluid, for the purposes of this analysis contrast was defined as the difference in attenuation between tagged fluid and soft tissue. Noise was defined as the mean standard deviation of air attenuation and CNR was defined as the ratio of contrast to noise, or

CNR ¼

Figure 1 Axial CTC image of a 74-year-old woman following incomplete colonoscopy. ROIs for attenuation of soft tissue drawn in the bilateral psoas muscles (white ellipses), for attenuation of tagged colonic fluid drawn in the ascending and descending colon (black ellipses), and for standard deviation representing noise drawn in two separate locations in the air outside the patient (grey circles).

A qualitative evaluation of image quality was not attempted on the axial images as readers could not be blinded to the tube voltage used to acquire the images given the supine and prone positions necessitated by the CTC protocol. Three-dimensional (3D) endoluminal views other than those acquired for clinical interpretation (one for supine and one for prone images) were also not generated for each of the ASiR levels as this would have resulted in an inordinate number of images to send to the 3D workstation (on the order of 6000e8000 images) as well as the necessity for eight separate 3D reconstructions and fly-throughs.

Statistical analysis The present study employed a within-subjects split-plot, partially factorial design. All patients received all conditions. Because supine and prone position should not influence image noise, position was not counterbalanced. All analyses were conducted using PROC GLIMMIX with an identity link function in SAS 9.3 for Windows (SAS Institute, Cary, NC, USA). The effects of tube voltage and iterative reconstruction on noise, tagged fluid attenuation, softtissue attenuation, and contrast were examined with general mixed modelling using restricted maximum likelihood

ðmean HU attenuation of tagged fluid  mean HU attenuation of psoas musclesÞ ðmean standard deviation in HU of airÞ

984

K.J. Chang et al. / Clinical Radiology 70 (2015) 981e988

and sandwich estimation assuming a Gaussian distribution. In addition, trend analysis was used to model noise and CNR across the four ASiR iterative reconstruction levels for each peak tube voltage. Mixed modelling was used for all analyses in order to include the covariance structure necessary to appropriately account for nested observations, given that subjects served as their own controls. Statistical significance was defined, a priori, at the 0.05 level. Follow-up multiple comparisons were tested using orthogonal linear contrasts with alpha maintained at 0.05 across all hypothesis tests using the Bonferroni method to adjust individual p-values. As a conservative effort, only two-tailed tests and confidence intervals (CIs) were calculated.

Results Noise Mean image noise increased with tube potential reduction and decreased with increasing levels of ASiR (Fig 2). Without ASiR, the mean image noise at 120 kVp was 22.6 HU (95% CI: 21.5, 23.7), which was lower than the mean image noise of 30.2 HU (95% CI: 28.5, 31.9) at 100 kVp. In addition, when tube voltage was reduced from 120 to 100 kVp, image noise was increased 30e34% for each level of ASiR. Follow-up comparisons indicate that the difference between tube potential was statistically significant for all levels of ASiR (p<0.0001 at each ASiR level). Moreover, mean image noise was reduced for each increase in ASiR, which follow-up comparisons revealed to be statistically significant at both levels of tube potential (all p<0.0001). Interestingly, the difference in image noise between tube

potential levels decreases with increasing levels of ASiR yielding a statistically significant interaction-effect of tube voltage and ASiR level (F(3,392)¼20.23, p<0.0001). A trend analysis of the trajectory of mean image noise for tube potential and ASiR level revealed a decreasing trend with increased levels of ASiR at both 120 kVp (p<0.0001) and 100 kVp (p<0.0001), but the steepness for the 120 kVp trend was significantly lower than for 100 kVp (p<0.0001). Lastly, follow-up comparisons indicate that when compared to a benchmark noise level of 22.6 HU at 120 kVp without ASiR (0% ASiR or filtered back projection), 100 kVp scans using an ASiR level of approximately 50% or greater (interpolated from Fig 2) should yield equivalent or lower noise levels. These differences are illustrated in Fig 2. Example axial images from a CTC study visually showing the decrease in image noise at 120 and 100 kVp at increasing ASiR settings are also illustrated in Fig 3.

Contrast The main effect for tube potential was observed revealing statistically significant differences between 120 and 100 kVp for soft-tissue attenuation, tagged colonic fluid p¼0.003; attenuation, and contrast, (F(1,392)¼8.9, F(1,272)¼80.2, p<0.0001; and F(1,272)¼69.0, p<0.0001), respectively. Specifically, soft-tissue attenuation minimally increased from 43.5 HU (95% CI: 40.8,46.2) for all ASiR levels at 120 kVp to 47.3 HU (95% CI: 45.0,49.6) for all ASiR levels at 100 kVp. Likewise, tagged colonic fluid attenuation significantly increased from 557.7 HU (95% CI: 480.6,634.9) HU for all ASiR levels at 120 kVp to 677.4 HU (95% CI: 586.8,768.1) for all ASiR levels at 100 kVp. As the increase in attenuation of tagged colonic fluid is significantly higher than that of soft tissues with a decrease in tube potential, there is also a resultant increase in tagged fluid-soft tissue contrast with a decrease in tube potential to 100 kVp (Fig 4). Specifically, contrast increased from 516 HU (95% CI: 438.7,593.3) at 120 kVp to 630.3 HU (95% CI: 538.8,721.8) at 100 kVp.

CNR A significant monotonically increasing trend was observed between ASiR level and CNR for both 120 and 100 kVp tube voltages (both p<0.0001; Fig 5); however, these trends were not significantly different from each other.

Discussion

Figure 2 Image noise (in HU) at 120 and 100 kVp at varying ASiR levels of 0% (filtered back projection), 30%, 60%, and 90%. Error bars represent 95% confidence limits. Decreasing tube potential from 120 to 100 kVp results in significantly increased image noise at all ASiR levels (p<0.001 for all measures). An ASiR level between 30 and 60% at 100 kVp is expected to yield similar noise levels to a 120 kVp scan without ASiR (dashed horizontal line). ASiR levels 60% and higher at 100 kVp yield significantly lower noise levels than the 120 kVp scan without ASiR.

The use of iterative reconstruction can decrease image noise associated with reduced tube potential imaging. The use of an iterative reconstruction technique should also be able to compensate for the increase in image noise with a reduction in tube potential. ASiR levels of 60% and 90% at 100 kVp result in significantly less image noise than filtered back projection at 120 kVp in the same patient cohort. In the present study, each patient acted as his or her own control allowing for robust statistical analysis, a novel approach compared to prior studies. When analysing the trend lines

K.J. Chang et al. / Clinical Radiology 70 (2015) 981e988

985

Figure 3 Axial CTC images of a 78-year-old woman following incomplete colonoscopy at 120 kVp at ASiR levels of 0% (a), 30% (b), 60% (c), and 90% (d), and at 100 kVp at ASiR levels of 0% (e), 30% (f), 60% (g), and 90% (h). Image noise visibly decreases with increasing ASiR level at both peak kilovoltages. A 1 cm polyp was identified in the proximal sigmoid colon (arrow). There is also a right lower quadrant urostomy with peristomal hernia containing unobstructed small bowel seen best on supine views. (eeh) Republished with permission from Chang KJ, Yee J. Dose reduction methods for CT colonography. Abdominal Imaging 2013; 38:224e232.42

in Fig 2, an ASiR level of about 50% should approximate an equivalent level of image noise present at 120 kVp using filtered back projection (dashed horizontal line). The present results also confirm that increasing levels of ASiR result in progressively decreased noise and that these noise reduction effects are linear at either 120 or 100 kVp tube voltage. Several prior studies have used iterative reconstruction as a noise reduction method with reduced tube current to reduce radiation dose29e31 but this is one of the first studies to quantify the noise reduction in reduced tube potential imaging. As expected, a reduction in tube potential also results in an increase in the attenuation of both soft tissue and tagged colonic fluid, affecting the attenuation of iodinated contrast medium more so than that of soft tissue. For example, the attenuation of tagged fluid increased by an average of 119 HU. This results in a significant increase in the contrast

between tagged fluid and soft tissue at a lower tube potential. This agrees with prior phantom studies40 as well as multiple prior studies using vascular contrast agents.17,22e27,41 Without the use of iterative reconstruction noise reduction techniques, the increase in contrast is offset by the increase in image noise resulting in no net change in CNR with tube potential reduction from 120 to 100, a result similar to that reached in a prior CTC study21 as well as in CT angiography.25,27 The addition of iterative reconstruction maintains the lower level of image noise resulting in a significant boost in CNR at either 120 or 100 kVp. This trend also appears relatively linear with increasing CNR at increasing levels of ASiR. At 120 kVp, CNR was 16%, 42%, and 79% higher at ASiR levels of 30%, 60%, and 90%, respectively, compared to 0% ASiR. At 100 kVp, CNR was 18%, 43%, and 81% higher at ASiR levels of 30%, 60%, and 90%, respectively, compared to 0% ASiR. This trend does not differ between

986

K.J. Chang et al. / Clinical Radiology 70 (2015) 981e988

Figure 3 (continued).

120 or 100 kVp. Although not formally tested in this study, increased tagged fluid attenuation at lower tube potential as well as increased CNR with iterative reconstruction should translate to improved conspicuity of submerged polyps for CTC. Of note, although the noise index setting on the CT system was kept constant for both 120 and 100 kVp acquisitions, instead of the level of image noise being maintained, noise values in patients scanned at 100 kVp were increased, a phenomenon also noted in a prior study.21 If the manufacturer has designed the noise index to result in a constant level of noise irrespective of the selected X-ray tube voltage, then image noise should have stayed the same using the same noise index. This behaviour may be explained by the use of an automatic tube-current modulation range of 50e450 mAs as a tube current greater than 450 mAs was not acquired for larger patients potentially limiting the scanner’s ability to maintain a constant noise index at lower

tube potential. A maximum tube current of 450 mAs is routinely used by the authors to limit radiation exposure for larger patients as a higher tolerance for image noise is sufficiently offset by an inherently higher amount of soft tissueefat contrast in these patients. One study limitation is that all examinations in this study were performed in humans rather than in a phantom, cadaver, or animal model. As a result, the results are subject to individual variation and anatomy. In addition, the low tube potential acquisition was not randomized to patient position in the interest of streamlining the CTC scan protocol for the CT technologists in a busy clinical setting. In addition to potentially introducing an element of systematic error, lack of randomization limits the ability to perform a subjective 2D image quality analysis due to lack of blinding of readers to patient position. A 3D endoluminal image quality analysis was also not attempted due to the impracticality of rendering eight separate 3D reconstructions for

K.J. Chang et al. / Clinical Radiology 70 (2015) 981e988

Figure 4 Contrast between tagged colonic fluid and soft tissue (psoas) in Hounsfield units at 120 and 100 kVp at varying ASiR levels of 0%, 30%, 60%, and 90%. Error bars represent 95% confidence limits. Contrast significantly increases at 100 kVp compared to 120 kVp at all ASiR levels. The amount of contrast is independent of ASiR level.

Figure 5 CNR in Hounsfield units at 120 and 100 kVp at varying ASiR levels of 0%, 30%, 60%, and 90%. Error bars represent 95% confidence limits. Increasing ASiR level significantly improves CNR at both 120 and 100 kVp. Difference in CNR between the 120 and 100 kVp scans did not reach statistical significance.

each patient to evaluate each combination of tube potential and ASiR. Another potential study limitation is that ASiR levels between 30% and 60% at 100 kVp were not specifically evaluated to determine the exact level that would generate equivalent image noise to the 120 kVp filtered back projection dataset. Images with a peak kilovoltage of 100 kVp and an ASiR level of 30% showed greater image noise than the 120 kVp benchmark filtered back projection images without ASiR, while the same datasets reconstructed with an ASiR level of 60% showed less image noise than the benchmark. From this, an ASiR level between 30% and 60% should be noise equivalent to the benchmark as illustrated in Fig 2. When designing this study the optimal level was unknown, so a wide range of ASiR values was employed. As the authors’ clinical routine already comprises imaging with an ASiR level of 30%, ASiR levels of 60% and 90% were chosen as convenient multiples to show the trend in noise

987

reduction at higher levels. Further studies could attempt to narrow the range to determine the optimal result. Another important study limitation is that neither the effect of a change in tube potential nor the effect of iterative reconstruction on the detectability of polyps was specifically addressed. Although there was a significant increase in image noise with a decrease in tube potential that appeared to be largely mitigated by ASiR levels of greater than 30%, it was not possible to gauge whether this would substantially affect diagnostic reader performance. Additional studies will be needed to evaluate the clinical utility of combining a low tube potential technique with iterative reconstruction in CTC. The present findings suggest a variety of future directions to further evaluate the effects of iterative reconstruction for noise reduction in reduced tube potential and lower radiation dose examinations. Although previous studies have shown that the increase in image noise that accompanies a decrease in tube potential can result in increased colonic mural nodularity and a decrease in 3D image quality,21 further studies are needed to determine whether iterative reconstruction techniques can offset these changes or even potentially improve 3D endoluminal image quality. The amount of increased image noise that accompanies tube potential reduction is also a function of patient size, disproportionately affecting larger patients more than smaller patients.21 The effects of iterative reconstruction will also need to be assessed as a function of patient size. The cohort of patients was not large enough to investigate this effect. In addition, the combination of iterative reconstruction and reduced tube potential may also improve the CNR for intravenous contrast-enhanced CTC as colonic polyps enhance. Finally, the effect of other vendors’ iterative reconstruction algorithms including nextgeneration model-based iterative reconstruction methods on low tube potential images should be investigated. In conclusion, iterative reconstruction can adequately compensate for the increase in image noise associated with lower dose low tube potential imaging while improving CNR. An ASiR level of approximately 50% at 100 kVp should result in similar image noise to pure filtered back projection images at 120 kVp.

References 1. Johnson CD, Chen MH, Toledano AY, et al. Accuracy of CT colonography for detection of large adenomas and cancers. N Engl J Med 2008;359:1207e17. 2. Pickhardt PJ, Hassan C, Halligan S, et al. Colorectal cancer: CT colonography and colonoscopy for detectiondsystematic review and metaanalysis. Radiology 2011;259:393e405. 3. Levin B, Lieberman DA, McFarland B, et al. Screening and surveillance for the early detection of colorectal cancer and adenomatous polyps, 2008: a joint guideline from the American Cancer Society, the US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology. CA Cancer J Clin 2008;58:130e60. 4. Brenner DJ, Georgsson MA. Mass screening with CT colonography: should the radiation exposure be of concern? Gastroenterology 2005;129:328e37. 5. Brenner DJ, Hall EJ. Computed tomographydan increasing source of radiation exposure. N Engl J Med 2007;357:2277e84.

988

K.J. Chang et al. / Clinical Radiology 70 (2015) 981e988

6. Burk RJ. Radiation risk in perspective: position statement of the Health Physics Society. Revised 2010. Available at: www.hps.org/documents/ risk_ps010-2.pdf. Accessed July 8, 2014. 7. McCollough CH, Guimaraes L, Fletcher JG. In defense of body CT. AJR Am J Roentgenol 2009;193:28e39. 8. Muirhead CR, O’Hagan JA, Haylock RG, et al. Mortality and cancer incidence following occupational radiation exposure: third analysis of the National Registry for Radiation Workers. Br J Cancer 2009;100:206e12. 9. Cohnen M, Vogt C, Aurich V, et al. Multi-slice CT-colonography in lowdose techniquedpreliminary results. Rofo 2002;174:835e8. 10. Iannaccone R, Laghi A, Catalano C, et al. Feasibility of ultra-low-dose multislice CT colonography for the detection of colorectal lesions: preliminary experience. Eur Radiol 2003;13:1297e302. 11. Ginsburg M, Obara P, Wise L, et al. BMI-based radiation dose reduction in CT colonography. Acad Radiol 2013;20:486e92. 12. Kalender WA, Wolf H, Suess C. Dose reduction in CT by anatomically adapted tube current modulation. II. Phantom measurements. Med Phys 1999;26:2248e53. 13. 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;233:241e9. 14. Rizzo S, Kalra M, Schmidt B, et al. Comparison of angular and combined automatic tube current modulation techniques with constant tube current CT of the abdomen and pelvis. AJR Am J Roentgenol 2006;186:673e9. 15. Tack D, De Maertelaer V, Gevenois PA. Dose reduction in multidetector CT using attenuation-based online tube current modulation. AJR Am J Roentgenol 2003;181:331e4. 16. Graser A, Wintersperger BJ, Suess C, et al. Dose reduction and image quality in MDCT colonography using tube current modulation. AJR Am J Roentgenol 2006;187:695e701. 17. Funama Y, Awai K, Nakayama Y, et al. Radiation dose reduction without degradation of low-contrast detectability at abdominal multisection CT with a low-tube voltage technique: phantom study. Radiology 2005;237:905e10. 18. GE. Chapter 12. In: Lightspeed VCT Technical Reference Manual. revised 9th ed. New York: General Electric Company; 2007. p. 25. 19. Elojeimy S, Tipnis S, Huda W. Relationship between radiographic techniques (kilovolt and milliampere-second) and CTDI(VOL). Radiat Prot Dosimetry 2010;141:43e9. 20. Brooks A. Quantitative theory of the Hounsfield unit and its application to dual energy scanning. J Comput Assist Tomogr 1977;1:487e93. 21. Chang KJ, Caovan DB, Grand DJ, et al. Reducing radiation dose at CT colonography: decreasing kVp to 100 kilovolts. Radiology 2013;266:801e11. 22. Nakaura T, Awai K, Oda S, et al. Low-kilovoltage, high-tube-current MDCT of liver in thin adults: pilot study evaluating radiation dose, image quality, and display settings. AJR Am J Roentgenol 2011;196:1332e8. 23. Nakaura T, Awai K, Oda S, et al. A low-kilovolt (peak) high-tube current technique improves venous enhancement and reduces the radiation dose at indirect multidetector-row CT venography: initial experience. J Comput Assist Tomogr 2011;35:141e7. 24. Nakayama Y, Awai K, Funama Y, et al. Abdominal CT with low tube voltage: preliminary observations about radiation dose, contrast enhancement, image quality, and noise. Radiology 2005;237:945e51. 25. Wintersperger B, Jakobs T, Herzog P, et al. Aorto-iliac multidetector-row CT angiography with low kV settings: improved vessel enhancement

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40. 41.

42.

and simultaneous reduction of radiation dose. Eur Radiol 2005;15:334e41. Yanaga Y, Awai K, Nakaura T, et al. hepatocellular carcinoma in patients weighing 70 kg or less: initial trial of compact-bolus dynamic CT with low-dose contrast material at 80 kVp. AJR Am J Roentgenol 2011;196:1324e31. Heyer CM, Mohr PS, Lemburg SP, et al. Image quality and radiation exposure at pulmonary CT angiography with 100- or 120-kVp protocol: prospective randomized study. Radiology 2007;245:577e83. 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;21:1881e4. Hara AK, Paden RG, Silva AC, et al. Iterative reconstruction technique for reducing body radiation dose at CT: feasibility study. AJR Am J Roentgenol 2009;193:764e71. Thibault JB, Sauer KD, Bouman CA, et al. A three-dimensional statistical approach to improved image quality for multislice helical CT. Med Phys 2007;34:4526e44. Flicek KT, Hara AK, Silva AC, et al. Reducing the radiation dose for CT colonography using adaptive statistical iterative reconstruction: a pilot study. AJR Am J Roentgenol 2010;195:126e31. Yoon MA, Kim SH, Lee JM, et al. Adaptive statistical iterative reconstruction and Veo: assessment of image quality and diagnostic performance in CT colonography at various radiation doses. J Comput Assist Tomogr 2012;36:596e601. Nagata K, Fujiwara M, Kanazawa H, et al. Evaluation of dose reduction and image quality in CT colonography: comparison of low-dose CT with iterative reconstruction and routine-dose CT with filtered back projection. Eur Radiol 2014. Lambert L, Danes J, Jahoda J, et al. Submilisievert ultralow-dose CT colonography using iterative reconstruction technique: a feasibility study. Acta Radiol 2015;56:517e25. Vardhanabhuti V, Loader R, Roobottom CA. Assessment of image quality on effects of varying tube voltage and automatic tube current modulation with hybrid and pure iterative reconstruction techniques in abdominal/pelvic CT: a phantom study. Invest Radiol 2013;48:167e74. Chen CM, Chu SY, Hsu MY, et al. Low-tube-voltage (80 kVp) CT aortography using 320-row volume CT with adaptive iterative reconstruction: lower contrast medium and radiation dose. Eur Radiol 2014;24:460e8. Zheng M, Liu Y, Wei M, et al. Low concentration contrast medium for dual-source computed tomography coronary angiography by a combination of iterative reconstruction and low-tube-voltage technique: feasibility study. Eur J Radiol 2014;83:e92e99. Zhang WL, Li M, Zhang B, et al. CT angiography of the head-and-neck vessels acquired with low tube voltage, low iodine, and iterative image reconstruction: clinical evaluation of radiation dose and image quality. PloS One 2013;8:e81486. Chang KJ, Rekhi Jr SS, Anderson SW, et al. Fluid tagging for CT colonography: effectiveness of a 2-hour iodinated oral preparation after incomplete optical colonoscopy. J Comput Assist Tomogr 2011;35:91e5. Huda W, Scalzetti EM, Levin G. Technique factors and image quality as functions of patient weight at abdominal CT. Radiology 2000;217:430e5. Nakayama Y, Awai K, Funama Y, et al. Lower tube voltage reduces contrast material and radiation doses on 16-MDCT aortography. AJR Am J Roentgenol 2006;187:W490e7. Chang KJ, Yee J. Dose reduction methods for CT colonography. Abdom Imaging 2013;38:224e32.