Introduction of an Effective Method for the Optimization of CT Protocols Using Iterative Reconstruction Algorithms: Comparison With Patient Data

Transcription

1 Medical Physics and Informatics Original Research Kordolaimi et al. Medical Physics and Informatics Original Research Sofia D. Kordolaimi 1 Ioannis Saradeas 1 Agapi Ploussi 1 Ioannis Pantos 2 Stylianos Argentos 1 Efstathios P. Efstathopoulos 1 Kordolaimi SD, Saradeas I, Ploussi A, Pantos I, Argentos S, Efstathopoulos EP Keywords: contrast-to-noise ratio, CT optimization, dose, idose4, iterative reconstruction DOI:.2214/AJR Received September 13, 13; accepted after revision December 22, 13. idose4 software and hardware were provided by Philips Healthcare within the auspices of Philips CT Publication of the year award won by our department in 9. 1 Second Department of Radiology, Medical School, University of Athens, 1 Rimini Str, Haidari, Athens, Greece. Address correspondence to E. P. Efstathopoulos (stathise@med.uoa.gr). 2 Department of Cardiology, Athens Euroclinic, Athens, Greece. WEB This is a web exclusive article. AJR 14; 3:W434 W X/14/34 W434 American Roentgen Ray Society Introduction of an Effective Method for the Optimization of CT Protocols Using Iterative Reconstruction Algorithms: Comparison With Patient Data OBJECTIVE. The purpose of this study is to introduce an efficient method for the optimization of iterative reconstruction CT protocols based on phantom image analysis and the comparison of obtained results with actual patient data. MATERIALS AND METHODS. We considered chest, abdomen, and pelvis CT examinations before the installation of an iterative reconstruction algorithm (idose4) to define the exposure parameters used in clinical routine with filtered back projection (FBP). The body area of a CT phantom was subsequently scanned with various tube voltages and tube currents exposure time products, and acquired data were reconstructed with FBP and different levels of idose4. The contrast-to-noise ratio () for FBP with the original exposure parameters was calculated to define the minimum acceptable value for each tube voltage. Then, an optimum tube current exposure time products for each tube voltage and level of iterative reconstruction was estimated. We also compared findings derived by the phantom with real patient data by assessing dosimetric and image quality indexes from a patient cohort scanned with exposure parameters gradually adjusted during 1 year of adoption of idose4. RESULTS. By use of the proposed phantom method, dose reduction up to 7% was achievable, whereas for an intermediate level of iteration (level 4), the dose reduction ranged between % and 6%, depending on the tube voltage. For comparison, with the gradual adjustment of exposure settings, the corresponding dose reduction for the same level of iteration was about %. CONCLUSION. The proposed method provides rapid and efficient optimization of CT protocols and could be used as the first step in the optimization process. V arious techniques have been reported for the optimization of CT protocols [1 6]; however, the recent introduction of iterative reconstruction algorithms in clinical practice has raised further need for optimization. The goal of optimization is to acquire a diagnostically acceptable outcome with the minimum radiation dose [7]. An optimization study must provide a selection of identical images with different exposure settings, an adequate number of observers to avoid bias, and a definition of the threshold of image quality that is regarded as diagnostically acceptable [8]. Moreover, an optimization method should be as simple as possible and time effective, because if it requires complicated workflows and exhaustive data analysis, its applicability is limited. The most common optimization method currently adopted is the gradual modification of exposure parameters to re- duce radiation exposure while maintaining imaging quality [9]. Imaging quality is evaluated by both subjective assessment of CT images based on predefined imaging quality criteria and by objective measurements based on ROI analysis [ 13]. An acceptable diagnostic level of imaging quality is defined, and the optimized scanning protocols with reduced radiation dose are adjusted accordingly. This procedure must be repeated for all available CT protocols. Although this method is straightforward and widely accepted, it requires a large patient cohort for each protocol; thus, it is a lengthy procedure, especially for CT examinations that are scarcely performed and for centers with limited numbers of patients. Alternatively, CT optimization can be based on CT phantoms. Similarly to clinical imaging quality evaluation, the CT phantom images are evaluated both objectively (by ROI analysis) and subjectively (by, for example, spatial resolution, low contrast resolution, W434 AJR:3, October 14

2 Fig. 1 Images used to calculate effective diameter. A, 68-year-old woman. Anterior-posterior and lateral dimensions of patient in abdomen area are shown. B, Effective diameter of patient is equal to diameter of phantom. and visual evaluation of noise) to determine the relationship between optimum dose and acceptable imaging quality [14]. The purpose of the current study is to introduce an efficient method for optimizing the CT examination protocols to the requirements of iterative reconstruction algorithms based on CT phantom image analysis. This method was applied to the CT examination protocol for the chest, abdomen, and pelvis. To compare the results derived from the phantom study with real patient data, we assessed dosimetric and image quality indexes from a patient cohort scanned with exposure parameters gradually adjusted during 1 year of adoption of idose4 (Philips Healthcare). Materials and Methods Optimization Process We considered chest, abdomen, and pelvis CT examinations of patients of medium weight (61 9 kg) whose effective diameter in the abdomen area was similar (mean ± SD,. ± 1. cm) to that of the body area of the phantom. According to Boone et al. [], the effective diameter represents the diameter of a patient, assuming that the patient has a circular cross-section, and is calculated from Equation 1: effective diameter = AP LAT (1), where AP is the anterior-posterior dimension and LAT is the lateral dimension of the patient (Fig. 1). Examinations were performed with a chest, abdomen, and pelvis CT protocol that had been thoroughly optimized before the installation of the iterative reconstruction algorithm. The chest, abdomen, and pelvis CT protocol used tube current exposure time product modulation based on the patient s attenuation along the z-axis. The exposure parameters (kilovoltage and tube current exposure A time product) were recorded, and the effective diameter was defined at the same slice at the level of the upper abdomen area at each patient. The abdomen was selected because it contains limited air cavities and is filled mostly with tissue rather than bony structures; thus, it is closely simulated by the body area of the phantom. All reconstructed images were transferred for quantitative and qualitative analysis to a DICOM-compatible PACS workstation (Evorad RID-PACS, version 2.2. build 81, Evorad SA), with a 3-megapixel display monitor certified for medical use (EISO, Eizo Nanao Technologies). To verify that all acquired images were clinically acceptable, subjective evaluation analysis was performed by two experienced radiologists (with 23 and years of experience in chest, abdomen, and pelvis CT). All image files were randomized and made anonymous by a resident radiologist (fourth year of residency) so that the reviewers were unaware of the scanning protocol and patient demographics. The resident radiologist was also present during the whole evaluation process. The readers TABLE 1: Scanning Parameters Parameter Voltage (kv) 8,, 1, 1 Tube current exposure time product (mas) Acquisition mode Reconstruction algorithm Slice thickness (mm) 1 Matrix size Detector configuration (rows mm) 64.6 Dose modulation Off Filtration Standard FOV (cm) Scanning length (cm) 17.1 were given a training session that included a set of written guidelines concerning the scoring system based on relevant published articles. In addition, both radiologists were trained on four image datasets for the grading of different aspects of subjective image quality so that they would understand the evaluation system. These four datasets belonged to four patients who underwent chest, abdomen, and pelvis CT examinations and who were not included in the patient cohort of our study. The resident radiologist loaded the CT images separately on the workstation at a preset window width of 36 HU and level of 6 HU; however, reviewers had the option to vary the window settings for optimal evaluation. Each reader independently evaluated the axial datasets for image noise, sharpness, contrast, and diagnostic confidence on the following -point scale: 1 is excellent, 2 is very good, 3 is acceptable or fair, 4 is poor, and is unacceptable. The degree of interobserver agreement between the two readers was assessed using a kappa statistical test, with κ <. signifying poor agreement, κ =.. indicating fair agreement, Value 6 (for 8 kvp), (for kvp), (for 1 kvp), (for 1 kvp) Spiral FBP, idose4 (levels 1 6; Philips Healthcare) B AJR:3, October 14 W4

3 Kordolaimi et al. κ =.41.6 indicating moderate agreement, κ =.61.8 indicating good agreement, and κ = indicating almost perfect agreement. The statistical analysis was performed using SPSS software (version 21., IBM). The exposure settings for each patient were recorded. These exposure settings correspond to a specific contrast-to-noise ratio () value at the phantom images reconstructed with the filtered back projection (FBP) algorithm. Thus, a correlation between the subjective and objective imaging quality can be set through the. By defining a value as the minimum threshold of clinically acceptable imaging quality (threshold ), one can adjust accordingly the exposure parameters for every level of the iterative reconstruction algorithm to acquire images of acceptable imaging quality. In this study,, an image quality metric frequently used for the optimization of the CT protocols [4, 16, 17], was considered. However, other imaging quality indexes, such as signal-to-noise ratio and image noise, are also suitable. To investigate the effect of the imaging quality index in the optimization procedure, noise measurements were also considered for particular exposure settings. Acquisition and Reconstruction Protocol For the definition of threshold and the measurement of the at various exposure settings, a CT quality assurance phantom (Philips 7736 System Phantom Kit, Philips Healthcare) was used. This phantom consists of a waterfilled cylinder with an external diameter of cm, which simulates the adult head (head area), in contact with a nylon cylinder of -cm diameter that simulates the adult body (body area), which also contains a Teflon (DuPont) strip and a water hole (Fig. 1). The phantom was scanned on a 64-MDCT scanner (Brilliance 64, Philips Healthcare) with the acquisition parameters defined from the patient cohort to define the threshold. It was also scanned at 8,, 1, and 1 kvp for six tube currents exposure time product per voltage to define the values at various exposure settings and to derive the optimum tube current exposure time product for each tube voltage (Table 1). All images were reconstructed with FBP and idose4 (Philips Healthcare). idose4 is an iterative reconstruction algorithm that provides seven levels of image noise reduction (from level 1, the lowest noise reduction, to level 7, the highest noise reduction). The proportion of idose4 blending with FBP as well as the noise removal for each level as provided by the vendor are presented in Table 2. Level 7 is available for only particular examinations (i.e., CT angiography); thus, reconstructions in the current study were conducted for levels 1 6 only. Reconstruction parameters are presented in Table 1. TABLE 2: idose4 (Philips Healthcare) Blending With FBP and Noise Removal for Each Level of idose4 Level of idose4 Percentage of idose4 Blending With FBP (%) Noise Removal (%) Note FBP = filtered back projection. Evaluation of Image Quality To assess the imaging quality of the reconstructed images, ROI analysis was conducted in the body area of the phantom. Mean attenuation and SD measurements were recorded using two ROIs (Fig. 1B). A 3.-cm 2 ROI was positioned at the Teflon hole and a -cm 2 ROI was placed at the background area (nylon cylinder). ROIs shape and position were kept constant for all images. was used as an image quality index and was calculated by Equation 2 [18]: = Mean Teflon Mean Nylon (SD 2 Teflon + SD 2 Nylon)/2 (2), where Mean Teflon, SD Teflon, Mean Nylon, and SD Nylon, are the mean CT numbers and SDs in the Teflon and nylon areas, respectively. Patient-Based Optimization Process After the installation of idose4 (April 12), a conservative reduction of exposure parameters for the various CT examination protocols with a gradual increase in the level of the iteration algorithm was followed in our institution. A patient cohort of patients (Table 3) of medium weight (61 9 kg) who underwent chest, abdomen, and pelvis CT examination on the same 64-MDCT scanner using reduced tube current exposure time product settings, as adjusted a year after the installation of idose4, was reviewed. The patients had.18 ± 1.22 cm effective diameter in the abdomen. The % level (level 4) of idose4 algorithm was selected as the default level for the current chest, abdomen, and pelvis protocol based on radiologists experience (changes in image texture were noticed in higher levels of the idose4 algorithm) and vendor s recommendations. Thus, all image quality analysis and radiation exposure adjustment correspond to the level 4 of the idose4 algorithm. Subjective imaging quality analysis was conducted by the same radiologists using the same scale as in the first step of the optimization process. Interobserver agreement was also assessed with the use of kappa test in SPPS software. TABLE 3: Radiation Exposure Quantities and Subjective Image Quality Evaluation for the Two Patient Groups Before and 1 Year After the Installation of idose4 (Philips Healthcare) Characteristic Before idose4 1 Year After idose4 Demographic characteristics No. of patients Age (y) ± ± 16.9 Effective diameter (cm). ± ± 1.22 Radiation exposure parameters Voltage (kvp) 1 1 Tube current exposure time product (mas) ± ± 18 CT dose index (mgy) 9.2 ± ± 1.1 Subjective image quality (score) Noise 2.8 ±. 2.1 ±.4 Sharpness 2.4 ± ±.4 Contrast 2.1 ± ±.4 Diagnostic confidence 2.3 ±. 2. ±. Note Except where noted otherwise, data are mean ± SD. W436 AJR:3, October 14

4 Exposure Indexes For phantom scans, the volume CT dose index (CTDI vol ) was recorded from the console display of the CT scanner. For the patient cohort, the CTDI vol was recorded from the DICOM information report, at the same slice where the effective diameter was measured. Results The considered patients were all scanned with 1 kvp at the abdomen area, and the corresponding average tube current exposure time product was ± 29 mas. Dosimetric as well as subjective imaging quality results for the patients are presented in Table 3. The interobserver agreement was perfect (κ = 1.). Subjective image analysis revealed that the chest, abdomen, and pelvis CT protocol was well optimized; thus, these settings provided an acceptable imaging quality. The mean effective diameter of all patients at the level of the origin of the superior mesenteric artery was.4 ± 1. cm, almost identical to the phantom s diameter. At the body area of the phantom, ROI analysis was conducted in 168 images (4 tube voltages 6 tube currents exposure time products 7 reconstructions). values were estimated from Equation 1, and results for all reconstructions are presented in Figure 2. The threshold, which was derived with the original exposure parameters (1 kv and mas), was This value is shown as a brown line in the graphs of Figure 2. In Figure 3, CTDI vol, as displayed at the CT scanner console, is presented in relationship to the considered tube current exposure time product for all kilovoltage settings. The intersecting points between the threshold and lines in Figure 2 represent the tube current exposure time product for each idose4 level, which maintains the imaging quality at the desirable level. In Figure 4, the optimum tube current exposure time product (calculated with linear interpolation from the acquired data) for maintaining the predetermined level of image quality (threshold, 23.9) in relationship to the tube voltage for each reconstruction level is presented. Tube current exposure time product differences for the various levels of idose4 are greater for low kilovoltage compared with high kilovoltage in absolute values; however, the corresponding percentage differences do not follow the same trend. TABLE 4: Image Noise (SD) at kvp for Various Tube Current Exposure Time Products, Optimum Tube Current Exposure Time Products, and Percentage Difference in Optimum Tube Currents Exposure Time Products Using the Contrast-to-Noise Ratio () and Various Levels of idose4 (Philips Healthcare) Variable 6 Image Noise (HU) FBP idose4 Level 1 idose4 Level 2 idose4 Level 3 idose4 Level 4 idose4 Level idose4 Level 6 Tube current exposure time product (mas) Optimum tube current exposure time product SD difference (%) Note FBP = filtered back projection. 8 kvp Tube Current Exposure Time Product (mas) Tube Current Exposure Time Product (mas) 1 kvp FBP idose4 Level 1 idose4 Level 2 idose4 Level 3 kvp 6 8 Tube Current Exposure Time Product (mas) Tube Current Exposure Time Product (mas) idose4 Level 4 idose4 Level idose4 Level 6 Threshold 1 kvp Fig. 2 Graphs show threshold contrast-to-noise ratio () (gray horizontal lines) and values for various kilovoltage and tube current exposure time product settings in all available reconstruction algorithms. FBP = filtered back projection. AJR:3, October 14 W437

5 Kordolaimi et al. Dose Reduction in Phantom The original exposure settings (1 kv and mas) were associated with a CTDI vol value of 9.8 mgy. The adaptation of the scanning protocol to idose4 allowed substantial dose reductions up to 7% of the original CTDI vol value. Figure illustrates the CTDI vol of the adapted protocol in relationship to the iteration level of idose4 for all tube voltages. It is evident that CTDI vol decreases linearly with the level of idose4. For an intermediate level of iterations (idose4 level 4), CTDI vol reduction between % and 6% is achievable for tube voltages between 8 and 1 kvp. As expected, the highest level of idose4 (level 6) is associated with greater dose reduction between 6% and 7% for the various tube voltages. Optimization Based on Image Noise To investigate the effect of the imaging quality index on the optimization procedure, image noise measurements, as expressed by the SD, were also considered for a particular tube voltage ( kv). For the original exposure parameters, the SD that provided acceptable image quality and was considered as the threshold value for image noise was a threshold SD of 26.7 HU. In Table 4, image noise at various tube currents exposure time products for kvp and the interpolated tube currents exposure time products for achieving the predefined noise level for each reconstruction level are tabulated. The percentage differences between the optimum tube current exposure time products values, as estimated with SD, compared with are small and range between.3% and 6.3%. Patient-Based Optimization Process One year after the installation of idose4, chest, abdomen, and pelvis CT examinations Tube Current Exposure Time Product (mas) 6 8 kvp kvp 1 kvp 1 kvp CTDI vol (mgy) were performed with 1 kvp and a mean tube current exposure time product of 92 ± 18 mas. Scanning exposure parameters and subjective imaging quality results for the patients scanned with reduced tube current exposure time products are presented in Table 3. There was excellent agreement between the two radiologists in subjective imaging quality (κ = 1.). One year after the installation of idose4, the corresponding dose reduction expressed by CTDI vol, compared with the situation before the installation of idose4, was %. The resulting imaging quality with the reduced exposure parameters and level 4 of idose4 was excellent (Table 3), indicating that further dose reduction would be feasible without compromising diagnostic imaging quality. Discussion The adaptation of exposure parameters of CT protocols at the transition from FBP to iterative reconstruction algorithms is challenging, especially because different levels of blending of the two techniques usually are available from CT vendors. In the current study, we present a straightforward method of transition that requires that the existing FBP CT protocols are optimized, and, thus, the clinical FBP images are of acceptable imaging quality while patient radiation exposure is also optimized. The method is applied to a CT scanner equipped with the idose4 iterative reconstruction algorithm; however, it can be applied to any CT scanner and iterative reconstruction algorithm. Our results indicate that, with the proposed method, radiation dose reduction between % and 6% compared with the corresponding FBP protocol is achievable with an intermediate level of iteration (level 4), whereas a dose reduction up to 7% is achievable with the use of the higher Tube Current Exposure Time Product (mas) Tube Voltage (kvp) FBP idose4 Level 1 idose4 Level 2 idose4 Level 3 idose4 Level 4 idose4 Level idose4 Level 6 level (level 6) of idose4. These findings are in general agreement with previous phantom studies that reported dose reduction of 76% compared with FBP [19, ], and patient studies reporting dose reduction up to % with level of idose4 [12, 21 23]. Concerning the time efficiency of the idose4 algorithm, although we have not recorded the reconstruction time in our study, no delays were experienced using idose4 compared with FBP. Both methods took less than seconds to reconstruct 171 slices of the phantom CT images. As a consequence, idose4 can be applied in all clinical CT protocols and especially in high-dose examinations (i.e., coronary CT angiography) because it did not affect the clinical efficiency. To compare the results of the proposed phantom-based method with dose reduction achievable with conservative reduction of exposure parameters, we considered a patient cohort scanned with the current chest, abdomen, and pelvis protocol, which was optimized during the 1 year of application of idose4 in our department. Conservative reduction of exposure parameters resulted in a % dose reduction compared with the situation before the installation of idose4. However, image analysis showed that imaging quality was very good or excellent; thus, reduction of exposure parameters is still feasible and there is potential to further reduce dose to the levels indicated by the phantom study ( 6% compared with FBP). The fact that the optimization of the protocol of a frequently used examination requires considerable time emphasizes the need for an effective and prompt optimization procedure like the one introduced in the current study. This method allows an initial rough optimization based on objective imaging quality analysis that does not CTDI vol (mgy) 8 8 kvp 7 kvp 1 kvp 6 1 kvp Percentage idose 4 Blending with FBP Fig. 3 Volume CT dose index (CTDI vol ), in relationship to acquired tube current exposure time product, for all kilovoltage settings in CT phantom is shown. Fig. 4 Tube current exposure time product for each tube voltage value for all iterative reconstruction levels using idose4 (Philips Healthcare) is shown. FBP = filtered back projection. Fig. Graph shows volume CT dose index (CTDI vol ) required for achieving least acceptable image quality in relationship to level of idose4 (Philips Healthcare) for four tube voltages. FBP = filtered back projection. W438 AJR:3, October 14

6 require the participation of busy radiologists, sure and image quality is required. Patientbased optimization is a safe method; however, and low-contrast material dose abdominal CT in provement of image quality at low-radiation dose and it is quick and straightforward. Then patient-based optimization could be applied for fine adjustments in the exposure settings. In this study, was considered as the imaging quality index to determine image quality; however, the use of image noise has led to similar results regarding protocol optimization. Quantum noise is also an easily measured index on both phantom studies and patients; however, it has been shown that noise index alone is a poor representative of image quality [24]. For the optimization of abdomen-pelvis CT protocols in particular, noise derivative (derivative of the function of noise with respect to dose) and relatively low-contrast detectability are considered suitable indexes for managing radiation dose []. This it is time consuming and also requires a large number of examinations. The proposed method provides rapid and efficient optimization of existing CT protocols based on actual patient data, and it could be the first step in the optimization process before patient image analysis. References 1. Kalra MK, Maher MM, Toth TL, et al. Strategies for CT radiation dose optimization. Radiology 4; 2: Halliburton SS, Abbara S, Chen MY, et al. SCCT guidelines on radiation dose and dose-optimization strategies in cardiovascular CT. J Cardiovasc Comput Tomogr 11; : Strauss KJ, Goske MJ, Kaste SC, et al. Image gently: patients with cirrhosis: intraindividual comparison of low tube voltage with iterative reconstruction algorithm and standard tube voltage. J Comput Assist Tomogr 12; 36: Hou Y, Xu S, Guo W, Vembar M, Guo Q. The optimal dose reduction level using iterative reconstruction with prospective ECG-triggered coronary CTA using 6-slice MDCT. Eur J Radiol 12; 81: Zarb F, Rainford L, McEntee M. Developing optimized CT scan protocols: phantom measurements of image quality. Radiography 11; 17: Boone JM, Strauss KJ, Cody DD, et al. Size-specific dose estimates (SSDE) in pediatric and adult body CT examinations: AAPM report no. 4. American Association of Physicists in Medicine optimization method can be applied to CT ten steps you can take to optimize image qual- website. protocols other than chest, abdomen, and pelvis. In that case, and especially in inhomogeneous areas such as the chest, the effective diameter of the scanning area should be estimated by taking into account not only the spatial dimensions of the patient but also the x-ray attenuating equivalence of the body area compared with that of the CT phantom used for optimization. Alternatively, a dedicated phantom (i.e., with air cavities for simulating the lungs) for the particular body area could be used. This study has several limitations. The CT examinations used as a reference were all acquired with 1 kv; thus, the threshold was defined only for that tube voltage. If thresholds for were defined for different tube voltages, the optimum tube current exposure time products for each voltage probably would be different. The proposed method is applied for a specific patient size that matches that of the CT phantom; thus, the findings are strictly applicable only for this patient size. The same procedure could be followed by using the head part of the CT phantom to optimize pediatric protocols. For optimizing protocols of different patient sizes, the findings of this study can be used as a reference. The relative exposure differences already exist among the various FBP protocols based on a patient s weight, body mass index, or abdominal circumference could be applied in the phantom-based optimized patient category. As a result, a proportional reduction in the exposure settings of all patient categories will be feasible. Finally, this procedure must be repeated for the optimization of all the available CT protocols. Conclusion Iterative reconstruction algorithms allow significant dose reduction, but careful assessment of the tradeoff between radiation expoity and lower CT dose for pediatric patients. AJR ; 194: Huda W, Ogden KM, Khorasani MR. Effect of dose metrics and radiation risk models when optimizing CT x-ray tube voltage. Phys Med Biol 8; 3: Dougeni E, Faulkner K, Panayiotakis G. A review of patient dose and optimisation methods in adult and paediatric CT scanning. Eur J Radiol 12; 81:e66 e Yerramasu A, Venuraju S, Atwal S, Goodman D, Lipkin D, Lahiri A. Radiation dose of CT coronary angiography in clinical practice: objective evaluation of strategies for dose optimization. Eur J Radiol 12; 81: The Council of the European Union. Council directive 97/43/Euratom of June 1997 on health protection of individuals against dangers of ionizing radiation in relation to medical exposure, and repealing Directive 84/466/Euratom. European Union website. ec.europa.eu/energy/nuclear/ radioprotection/doc/legislation/9743_en.pdf. Published June, Accessed June 11, Nagel HD, ed. Guideline for optimization of the radiation exposure of CT examinations, 2nd ed. Science & Technology for Radiology website. com/attachments/file/ct_guideline_%282nd_ Ed%29.pdf. Published. Accessed June 11, Wallace AB, Goergen SK, Schick D, Soblusky T, Jolley D. Multidetector CT dose: clinical practice improvement strategies from a successful optimization program. J Am Coll Radiol ; 7: Kalra MK, Woisetschlager M, Dahlstrom N, et al. Radiation dose reduction with Sinogram Affirmed Iterative Reconstruction technique for abdominal computed tomography. J Comput Assist Tomogr 12; 36: Desai GS, Uppot RN, Yu EW, Kambadakone AR, Sahani DV. Impact of iterative reconstruction on image quality and radiation dose in multidetector CT of large body size adults. Eur Radiol 12; 22: Namimoto T, Oda S, Utsunomiya D, et al. Im- pdf. Published 11. Accessed June 11, Boone JM, Geraghty EM, Seibert JA, Wootton- Gorges SL. Dose reduction in pediatric CT: a rational approach. Radiology 3; 228: Dong F, Davros W, Pozzuto J, Reid J. Optimization of kilovoltage and tube current-exposure time product based on abdominal circumference: an oval phantom study for pediatric abdominal CT. AJR 12; 199: Gagne RM, Boswell JS, Myers KJ. Signal detectability in digital radiography: spatial domain figures of merit. Med Phys 3; : Funama Y, Taguchi K, Utsunomiya D, et al. Combination of a low-tube-voltage technique with hybrid iterative reconstruction (idose) algorithm at coronary computed tomographic angiography. J Comput Assist Tomogr 11; : Singh S, Kalra MK, Gilman MD, et al. Adaptive statistical iterative reconstruction technique for radiation dose reduction in chest CT: a pilot study. Radiology 11; 9: Niu YT, Mehta D, Zhang ZR, et al. Radiation dose reduction in temporal bone CT with iterative reconstruction technique. AJNR 12; 33: Hosch W, Stiller W, Mueller D, et al. Reduction of radiation exposure and improvement of image quality with BMI-adapted prospective cardiac computed tomography and iterative reconstruction. Eur J Radiol 12; 81: Gervaise A, Osemont B, Lecocq S, et al. CT image quality improvement using Adaptive Iterative Dose Reduction with wide-volume acquisition on 3-detector CT. Eur Radiol 12; 22: Wilting JE, Zwartkruis A, van Leeuwen MS, Timmer J, Kamphuis AG, Feldberg M. A rational approach to dose reduction in CT: individualized scan protocols. Eur Radiol 1; 11: Brisse HJ, Brenot J, Pierrat N, et al. The relevance of image quality indices for dose optimization in abdominal multi-detector row CT in children: experimental assessment with pediatric phantoms. Phys Med Biol 9; 4: AJR:3, October 14 W439