Radiation Dose Consideration in Kidney Stone CT

5 downloads 0 Views 999KB Size Report
Aug 24, 2011 - Healthcare [vendor 1], n = 12 scanners; Philips Healthcare [vendor 2], n = 2; Siemens Health- ... parentheses are numbers of examinations.
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 Andrabi et al. Radiation Dose From Kidney Stone CT

Downloaded from www.ajronline.org by Library Of Medicine on 01/25/16 from IP address 128.103.149.52. Copyright ARRS. For personal use only; all rights reserved

Medical Physics and Informatics Original Research

Radiation Dose Consideration in Kidney Stone CT Examinations: Integration of Iterative Reconstruction Algorithms With Routine Clinical Practice Yasir Andrabi1 Oleg Pianykh1 Mukta Agrawal Avinash Kambadakone Michael A. Blake Dushyant V. Sahani Andrabi Y, Pianykh O, Agrawal M, Kambadakone A, Blake MA, Sahani DV

Keywords: CT, dose monitoring, iterative reconstruction, kidney stones, radiation dose DOI:10.2214/AJR.14.13038 Received April 15, 2014; accepted after revision August 22, 2014. D. V. Sahani received research grant support from GE Healthcare. 1

All authors: Department of Radiology, Division of Abdominal Imaging and Intervention, Massachusetts General Hospital, Harvard Medical School, 55 Fruit St, White 270, Boston, MA 02114-2696. Address ­correspondence to D. V. Sahani ([email protected]).

AJR 2015; 204:1055–1063 0361–803X/15/2045–1055 © American Roentgen Ray Society

OBJECTIVE. The objective of our study was to evaluate three commercially available iterative reconstruction (IR) algorithms—ASiR, iDOSE, and SAFIRE—and conventional filtered back projection (FBP) on image quality and radiation dose in kidney stone CT examinations. MATERIALS AND METHODS. During the 6-month study period, 684 unenhanced kidney stone CT examinations of consecutive adults were performed on 17 CT scanners (GE Healthcare [vendor 1], n = 12 scanners; Philips Healthcare [vendor 2], n = 2; Siemens Healthcare [vendor 3], n = 3); these examinations were retrieved using dose-monitoring software (eXposure). A total of 347 kidney stone CT examinations were reconstructed using FBP, and 337 examinations were processed using IR (ASiR, n = 248; iDOSE, n = 50; SAFIRE, n = 39). The standard-dose scanning parameters for FBP scanners included a tube potential of 120 kVp, a tube current of 75–450 mA for vendor 1 and a Quality Reference mAs of 160–180 for vendor 3, and a slice thickness of 2.5 or 5 mm. The dose-modified protocol for the IR scanners included a higher noise index (1.4 times higher than the standard-dose FBP protocol) for vendor 1, a lower reference tube current–exposure time product for vendor 2 (150 reference mAs), and a lower Quality Reference mAs for vendor 3 (120 Quality Reference mAs). Three radiologists independently reviewed 60 randomly sampled kidney stone CT examinations for image quality, noise, and artifacts. Objective noise and attenuation were also determined. Size-specific dose estimates (SSDEs) were compared using ANOVA. RESULTS. Significantly higher subjective and objective measurements of image noise were found in FBP examinations compared with dose-modified IR examinations (p < 0.05). The radiation dose was substantially lower for the dose-modified IR examinations than the standard-dose FBP examinations (mean SSDE ± SD: 8.1 ± 3.8 vs 11.6 ± 3.6 mGy, respectively) (p < 0.0001), but the radiation dose was comparable among the three IR techniques (ASiR, 7.8 ± 3.1 mGy; iDOSE, 7.5 ± 1.9 mGy; SAFIRE, 7.6 ± 3.2 mGy) (p > 0.05). CONCLUSION. The three IRs enable 20–33% radiation dose reduction in kidney stone CT examinations compared with the FBP technique without any image quality concerns. The radiation dose and image quality were comparable among these three IR algorithms.

R

ecent data suggest that 1 in 11 Americans has a history of kidney stones and that the risk of recurrence is as high as 75% [1, 2]. CT, owing to its availability, speed, and high accuracy, has emerged as an imaging modality of choice for the diagnosis and follow-up of kidney stones in adults [3–8]. However, the relatively higher radiation exposure from CT to patients has raised concerns regarding the radiation-related potential side effects [9–15]. Efforts have already been made to apply reduced–radiation dose scanning protocols in clinical practice, which have yielded a substantial reduction in the CT dose [16]. However, excessive background noise and

beam-hardening artifacts on conventional filtered back projection (FBP) reconstructions have been some of the major limiting steps in further dose minimization efforts [17–27]. To overcome image quality concerns, various CT manufacturers introduced iterative reconstruction (IR) algorithms. In our clinical practice, three IR techniques—ASiR (GE Healthcare), iDOSE (Philips Healthcare), and SAFIRE (Siemens Healthcare)—have been in use and continue to be used. The detailed descriptions of these vendor-introduced IR algorithms have been provided in prior articles [28–30]; in brief, these IR techniques use a blending approach in image recon-

AJR:204, May 2015 1055

Downloaded from www.ajronline.org by Library Of Medicine on 01/25/16 from IP address 128.103.149.52. Copyright ARRS. For personal use only; all rights reserved

Andrabi et al. struction with contributions from FBP and IR to reconstruct a final image with reduced image noise and artifacts [28–30]. Indeed, the clinical benefits of IR algorithms in providing diagnostically acceptable CT images from dose-modified examinations have been validated in studies in the abdomen [31–43]. Monitoring an individual patient’s radiation exposure from a single examination and the cumulative dose from CT and other procedures using ionizing radiation is highly desirable. Also, in a busy practice using diverse technology from multiple manufacturers, assessing the performance of CT protocols is essential. Commercial software products have recently been introduced to automate dose tracking [35, 44–48]. One of these commercial dose-monitoring software options (eXposure, version 2, Radimetrics/Bayer) was introduced to our practice in December 2012.

ED scanners = 2 (n = 332) Total scanners = 17 (n = 684)

FBP = 1 (n = 180) ASiR (IR 1) = 1 (n = 152)

Non-ED scanners = 15 (n = 352)

FBP = 9 (n = 167)

ASiR (IR 1) = 3 (n = 96)

IR = 6 (n = 185)

iDOSE (IR 2) = 2 (n = 50) SAFIRE (IR 3) = 1 (n = 39)

Fig. 1—Flowchart summarizes kidney stone CT examinations performed at different locations on different scanners using filtered back projection (FBP) and iterative reconstruction (IR) techniques. Numbers in parentheses are numbers of examinations. ED = emergency department, IR 1 = ASiR (GE Healthcare), IR 2 = iDOSE (Philips Healthcare), IR 3 = SAFIRE (Siemens Healthcare).

We explored the feasibility of tracking radiation dose from various scanners using FBP and three IR algorithms and compared the radiation doses of standard-dose and dose-modified kidney stone CT examinations.

Materials and Methods Study Design This study was an institutional review board– approved, consent-waived retrospective analysis that followed the HIPAA guidelines for research.

Total kidney stone CT examinations = 684 Age = 51, BW = 86.6, M:F = 394:290 SSDE = 9.7, CTDIvol = 8.3

Non-ED CT examinations = 352 Age = 54, BW = 88.5 SSDE = 8.1, CTDIvol = 7.8

ED CT examinations = 332 Age = 48, BW = 84.7 SSDE = 11.4, CTDIvol = 9.9

FBP = 180

IR 1 = 152

FBP = 167

IR 1 = 96

IR 2 = 50

IR 3 = 39

Age = 48.5 BW = 86.1 SSDE = 13.6 CTDIvol = 12.2 Noise = 25.9 CNR = 13.9

Age = 47.2 BW = 83.3 SSDE = 9.2 CTDIvol = 7.2 Noise = 18.7 CNR = 19.8

Age = 54 BW = 88.8 SSDE = 9.5 CTDIvol = 8.8 Noise = 27.6 CNR = 12.9

Age = 59 BW = 88.3 SSDE = 7.8 CTDIvol = 7.2 Noise = 17 CNR = 22

Age = 54 BW = 87.6 SSDE = 7.5 CTDIvol = 6.2 Noise = 16.5 CNR = 22.4

Age = 54.5 BW = 88.3 SSDE = 7.6 CTDIvol = 6.8 Noise = 12.2 CNR = 28.4

≤ 91 kg = 117

≤ 91 kg = 124

≤ 91 kg = 108

≤ 91 kg = 70

≤ 91 kg = 29

≤ 91 kg = 24

Age = 47.4 BW = 69 SSDE = 12 CTDIvol = 9.5 Noise = 23.8 CNR = 15.4

Age = 46.9 BW = 70 SSDE = 8.8 CTDIvol = 6.9 Noise = 20 CNR = 22

Age = 53 BW = 71.7 SSDE = 7.9 CTDIvol = 6.4 Noise = 26.2 CNR = 13

Age = 55 BW = 71 SSDE = 7.2 CTDIvol = 5.9 Noise = 16.2 CNR = 23

Age = 52 BW = 74.4 SSDE = 7.1 CTDIvol = 5.8 Noise = 15 CNR = 23.5

Age = 54 BW = 74 SSDE = 7.3 CTDIvol = 6.1 Noise = 11 CNR = 29.8

> 91 kg = 63

> 91 kg = 28

> 91 kg = 59

> 91 kg = 26

> 91 kg = 21

> 91 kg = 15

Age = 49.5 BW = 103 SSDE = 15 CTDIvol = 14.9 Noise = 28 CNR = 12.4

Age = 47.5 BW = 96.7 SSDE = 9.5 CTDIvol = 8.5 Noise = 17.5 CNR = 17.9

Age = 55 BW = 105.8 SSDE = 11.1 CTDIvol = 11.2 Noise = 29.5 CNR = 12.7

Age = 63 BW = 105.7 SSDE = 8.4 CTDIvol = 8.5 Noise = 17.9 CNR = 21

Age = 56 BW = 100.7 SSDE = 7.9 CTDIvol = 6.9 Noise = 18 CNR = 21.3

Age = 55 BW = 102.6 SSDE = 7.8 CTDIvol = 7.6 Noise = 13.4 CNR = 27

Fig. 2—Flowchart shows kidney stone CT examinations stratified into groups by location and reconstruction algorithm. Age, body weight (BW), radiation dose, and objective image quality within groups are also provided. Age is listed as mean age in years; BW, as mean body weight in kilograms; male-female ratio (M:F), as number of patients; size-specific dose estimate (SSDE), as mean SSDE in milligrays; volume CT dose index (CTDIvol), as mean CTDIvol in milligrays; noise, as mean objective image noise measured as mean SD; contrast-to-noise ratio (CNR), as mean CNR measured as ratio of difference in attenuation between pelvic bone and psoas muscle to image noise. ED = emergency department, FBP = filtered back projection, IR 1 = ASiR (GE Healthcare), IR 2 = iDOSE (Philips Healthcare), IR 3 = SAFIRE (Siemens Healthcare).

1056

AJR:204, May 2015

Radiation Dose From Kidney Stone CT TABLE 1: Comparison of the Scanning Parameters Between Emergency Department (ED) and Non-ED Scanners, Various Vendors and Reconstruction Algorithms for Unenhanced Kidney Stone CT Examinations ED ASiRa

FBPa

ASiRa

iDOSEb

FBPc

SAFIREc

≤ 91 kg

120

120

120

120

120

120

80–120

> 91 kg

120

120

120

120

120

120

80–120

≤ 91 kg

350d

250d

350d

250d

125e

160 f

120 f

> 91 kg

450d

350d

450d

350d

150e

180 f

120 f

25

35

25

35







Scanning Parameters and Patient Weight Downloaded from www.ajronline.org by Library Of Medicine on 01/25/16 from IP address 128.103.149.52. Copyright ARRS. For personal use only; all rights reserved

Non-ED

FBPa

Tube potential (kVp)

Maximum tube current (mA), reference tube current–exposure time product (mAs), or Quality Reference mAs

Noise index Pitch

1.375

1.375

1.375

1.375

0.984

1.3

1.3

Slice thickness (mm)

2.5

2.5

5

5

5

5

5

Rotation time (s)

0.5

0.5

0.5

0.5

0.5

0.5

0.5

Note—Dash (—) indicates noise index is not provided because it is specific to scanners manufactured by GE Healthcare. FBP = filtered back projection. aGE Healthcare. bPhilips Healthcare. cSiemens Healthcare. dmA. eReference mAs. f Quality Reference mAs (Siemens Healthcare).

Commercially available automated dose-monitoring software (eXposure, version 2) was used to retrieve CT examinations performed for kidney stone indications. This software retrieves radiation dose information from dose reports and also estimates doses from scanning parameters in DICOM headers. After an initial validation process, this software was integrated into our clinical practice in December 2012. Most of the validation process was performed by a team of two medical physicists, an information technology technician, a CT technologist, and a few radiologists to ensure that the dose measures from the software were correct and reproducible. Some of the challenges encountered with the software included retrospective data retrieval and processing to measure dose and issues arising from incorrect examination code and incorrect patient positioning on the scanner (decubitus or lateral). These problems were appropriately resolved and fixed; for dose validation, comparison was made to scanner-generated dose reports. The size-specific dose estimate (SSDE) was validated by comparing it with manually calculated SSDE (SSDE = CT dose index [CTDI] × conversion factor). All kidney stone CT examinations performed December 20, 2012 to June 20, 2013 were retrieved. Of the total 57,280 CT examinations performed, 684 were performed for kidney stone indications (mean age, 51 years; male-female ratio, 394:290) on 17 CT machines, seven of which have IR capabilities. A total of 332 of 684 examinations (48.5%) were performed in the emergency department

(ED), whereas the other 352 examinations (51.5%) were performed on non-ED scanners. Patients were categorized on the basis of the scanner location and reconstruction algorithm used (ED: FBP, n = 180, IR, n = 152; non-ED: FBP, n = 167, IR, n = 185) (Fig. 1). The ED studies were separated from those performed in the non-ED sites because the ED scanning parameters are slightly modified. Thinner slices and a scanning length similar to that of an abdominopelvic examination were used for the ED studies because the patients were not evaluated by a urologist and therefore excluding other potential abdominal abnormalities was a concurrent objective of the kidney stone CT examination. The IR examinations were grouped as follows: IR 1 (ASiR), 152 ED and 96 non-ED examinations; IR 2 (iDOSE), 50 non-ED examinations; and IR 3 (SAFIRE), 39 non-ED examinations. To match all groups for body weight, we also divided examinations into groups by body weight. Of the ED examinations, 241 were performed of patients who weighed 91 kg or less (FBP, n = 117; IR, n = 124), and 91 were performed of patients who weighed more than 91 kg (FBP, n = 63; IR, n = 28). Of the non-ED examinations, 231 were performed of patients who weighed 91 kg or less (FBP, n = 108; IR, n = 123), and 121 were performed of patients who weighed more than 91 kg (FBP, n = 59; IR, n = 62).

Healthcare [vendor 2], n = 2; Siemens Healthcare [vendor 3], n = 3), seven of which were equipped with an IR algorithm (IR 1, n = 4; IR 2, n = 2; IR 3, n = 1). The four scanners that were excluded were two PET/CT scanners and two scanners dedicated for interventional procedures only. The standard-dose scanning parameters for FBP scanners included a tube potential of 120 kVp, a tube current of 75–450 mA for vendor 1 and a Quality Reference mAs of 160–180 for vendor 3, and a slice thickness of 2.5 mm for ED scanners and 5 mm for non-ED scanners. To reduce dose for IR examinations, the noise index was increased from 25 to 35 on vendor 1 scanners, the reference mAs was reduced from 200 to 150 reference mAs on vendor 2 scanners, and the Quality Reference mAs (Siemens Healthcare) was reduced from 180 to 120 Quality Reference mAs on vendor 3 scanners (Table 1). The term “noise index” is used exclusively on GE Healthcare scanners, whereas Philips Healthcare and Siemens Healthcare use the terms “reference mAs” or “Quality Reference mAs,” respectively, to automate tube current modulation. The noise index is inversely proportional to tube current; that is, with an increase in the noise index, images are obtained at a lower tube current, which increases image noise but also decreases radiation dose.

Image Reconstruction CT Technique Patients were studied on 17 of 21 scanners (GE Healthcare [vendor 1], n = 12 scanners; Philips

A total of 347 examinations (50.7%) were reconstructed using FBP, and the remaining 337 examinations (49.3%) were processed using IR

AJR:204, May 2015 1057

Downloaded from www.ajronline.org by Library Of Medicine on 01/25/16 from IP address 128.103.149.52. Copyright ARRS. For personal use only; all rights reserved

(ASiR, 152 ED and 96 non-ED; iDOSE, 50 nonED; SAFIRE, 39 non-ED). ASiR uses FBP-reconstructed images as a building block and models image noise by applying a few loops of iterations between raw data space and image space. The features of FBP and noise modeling are combined, and the extent of noise improvement is influenced by the strength of the applied ASiR (0–100%); for example, 50% ASiR strength has an image blended with 50% contribution from FBP and 50% from ASiR. Similarly, iDOSE and SAFIRE iterate in both raw data and image data domains, and the image blending is impacted by the selection of the IR level: There are 7 levels for iDOSE and 5 levels for SAFIRE. Because kidney stone CT examinations have high intrinsic contrast to detect stones and the relative mAs is lower than that for routine abdominal CT examinations, we used a higher strength of IR to process the kidney stone images than would be used to process routine abdominal CT examinations (ASiR, 50%; iDOSE, level 4; SAFIRE, level 5). The FBP and IR images were processed on the scanner console to generate coronal and sagittal images with a 3-mm section thickness. All image datasets were then transmitted to the PACS for image interpretation.

Andrabi et al.

A

B

C

D

Image Analysis Qualitative—The CT reports were retrieved from the institution’s electronic medical record system (ClinicalApplicationSuite, CAS version 4.6.0) to record the diagnostic interpretation of each examination. Reports were also reviewed to note any comments on image quality concerns and adequacy for making diagnostic interpretation. Qualitative and quantitative comparisons within various groups were performed by randomly sampling 10 examinations from each reconstruction technique group (FBP and IR groups in ED; FBP, IR 1, IR 2, and IR 3 groups in non-ED) (Fig. 2). Three readers with varying levels of experience (reader 1, 3 years; reader 2, 6 years; reader 3, 12 years) independently reviewed these 60 examinations on the same PACS viewing workstation (version 5.3, AGFA). The readers were blinded to the technical scanning parameters and to the reconstruction algorithm used, which was accomplished by hiding all DICOM headers from the PACS user interface. For each image analysis session, axial, coronal, and sagittal image series were presented to the reader, and the studies were rated in accordance with the European guideline on quality criteria for CT [49]. A 5-point scale was used to grade image quality (1 = unacceptable, 2 = suboptimal, 3 = acceptable, 4 = good, 5 = excellent), image noise (1 = too little noise, 2 = less than usual noise, 3 = usual noise, 4 = excessive noise not affecting diagnostic interpretation, 5 = excessive noise rendering diagnostic interpretation not possible), and artifacts

1058

Fig. 3—Coronal unenhanced kidney stone CT images of four patients of average body weight (≤ 91 kg) reconstructed using filtered back projection (FBP) and three iterative reconstruction (IR) techniques. Note image quality of IR-processed images (B–D) is similar or superior to image quality of FBP image (A), but IRprocessed images were obtained at 25–35% lower dose than FBP image. CTDIvol = CT dose index. A, Image reconstructed using FBP. Patient is 70-year-old man. B, Image reconstructed using IR 1 (ASiR, GE Healthcare). Patient is 72-year-old woman. C, Image reconstructed using IR 2 (iDOSE, Philips Healthcare). Patient is 40-year-old woman. D, Image reconstructed using IR 3 (SAFIRE, Siemens Healthcare). Patient is 35-year-old man.

AJR:204, May 2015

Downloaded from www.ajronline.org by Library Of Medicine on 01/25/16 from IP address 128.103.149.52. Copyright ARRS. For personal use only; all rights reserved

Radiation Dose From Kidney Stone CT (1 = minimum, 2 = mild, 3 = moderate, 4 = excessive, 5 = maximum). The following scores were considered diagnostically acceptable by the readers: image quality score ≥ 3, noise score ≤ 3, and artifacts score ≤ 3. Objective—One of the coauthors not involved in the qualitative analysis obtained the following quantitative measurements of these 60 randomly selected examinations by placing circular ROIs in the psoas muscle and pelvic bone: image noise (SD of the mean CT number) and attenuation (measured in Hounsfield units). The contrast-tonoise ratio (CNR) was calculated using the following formula: [(attenuation in pelvic bone) – (attenuation in psoas muscle)] / image noise.

TABLE 2: Radiation Doses in Different Weight Groups of Patients U ­ ndergoing Kidney Stone Examinations Performed on Non–Emergency ­Department Scanners Using Filtered Back Projection (FBP) or Iterative Reconstruction (IR) Technique

Radiation Dose

Image Analysis Qualitative analysis—A diagnostic interpretation was rendered for all 684 kidney stone CT examinations without any image quality disclaimers in the report. For 60 randomly sampled examinations, the mean scores for image quality, noise, and artifacts as determined by three readers were compared within FBP and IR groups. Stones were diagnosed in 42 of these 60 patients. Additional findings were also reported in 24 patients (prostate hypertrophy, n = 5; colonic diverticulosis, n = 4; renal cyst, n = 4; hepatic cyst, n = 2; pancreatic cyst, n = 2; abdominal hernia, n = 2; adrenal nodule, n = 2; fatty liver, n = 1; fibroid uterus, n = 1; prominent mesenteric lymph nodes, n = 1). Mild to moderate hydronephrosis was also reported in 11 of these selected cases. Readers reported comparable subjective image quality scores (adequate to excellent) for IR- and FBP-processed examinations (mean image quality score: ED examinations, FBP = 4.0 and IR = 4.2; non-ED, FBP = 3.7 and IR = 4.0) (p > 0.05). Image noise was also comparable between FBP and each of the IR groups in the ED (mean image noise score, FBP = 1.9 and IR = 1.6) (p > 0.05); however, significantly higher noise was found in FBP examinations than in IR examinations performed on non-ED scanners (mean image noise score: FBP = 2.4 and IR = 1.8) (p < 0.05). There was no significant difference in image artifacts as determined by the three readers (mean image artifacts score: ED examinations, FBP = 1.7 and IR = 1.4; non-ED, FBP = 2.1 and IR = 1.6) (p > 0.05) (Fig. 3). On the 5-point subjective scale, reader agreement was poor to fair because of each reader’s preference and threshold for rating images. However, when images were categorized as acceptable (image quality score ≥ 3,

The relevant radiation dose information including volume CT dose index (CTDIvol) and SSDE were obtained using dose-monitoring software. We further validated the software-generated SSDE values by comparing them with manually calculated SSDE examinations for 60 randomly selected patients.

Statistical Analysis The data were analyzed using statistical software (SPSS, version 16.0, IBM). A one-way ANOVA test was performed to compare means of quantitative measurements including age, body weight, CTDIvol, SSDE, objective image noise, and attenuation within subgroups, and a p value of ≤ 0.05 was considered significant. For statistically significant results, a follow-up Tukey-Kramer multiple-comparisons test and Bonferroni correction were performed. The subjective image quality score for individual readers was compared within groups using the Kruskal-Wallis test. A p value of < 0.05 was considered significant. For significant results, a followup Mann-Whitney test was performed. To quantify the interobserver agreement among the readers, the kappa statistic was used. The definitions of the levels of agreement on the basis of kappa values were as follows: kappa value of 0.19 or less, poor agreement; 0.20–0.39, fair agreement; 0.40–0.59, moderate agreement; 0.60–0.79, substantial agreement; and 0.80–1.0, excellent agreement.

Results Patient Groups A total of 684 kidney stone CT examinations were performed during the 6-month study period (mean age, 51 ± 1.6 years; mean body weight, 86.6 ± 1.9 kg; male-female ratio = 394:290). Patient demographics including age, sex, and body weight within various groups are summarized in Figure 2. Patient age and body weight did not show significant difference within the groups (p > 0.05).

FBP (n = 167) Radiation Dose

IR (n = 185)

≤ 91 kg

> 91 kg

≤ 91 kg

> 91 kg

CTDIvol (mGy)

6.4 ± 1.9

11.2 ± 3.4

5.9 ± 2.3

7.8 ± 2.3

SSDE (mGy)

7.9 ± 1.8

11.1 ± 2.8

7.1 ± 2.2

8.1 ± 1.9

273.7 ± 97.9

544.6 ± 286.6

318.6 ± 201.3

401.1 ± 149.5

5.4 ± 2.1

10.5 ± 5.1

6.4 ± 4.1

7.6 ± 2.6

DLP (mGy × cm) Weighted effective dosea (mSv)

Note—Data are presented as mean ± SD. CTDIvol = CT dose index, SSDE = size-specific dose estimate, DLP = dose-length product. a According to report 103 of the International Commission on Radiological Protection [64].

noise score ≤ 3, artifacts score ≤ 3) or unacceptable, there was excellent interreader agreement (κ = 1). Objective analysis—The mean objective image noise was lower for the IR examinations (mean SD: ED, 18.7; non-ED, 14.1) compared with the FBP examinations (mean SD: ED, 25.9; non-ED, 27.6) (p < 0.05). The mean CNR was also significantly higher for the IR examinations (mean SD: ED, 19.8; non-ED, 24.0) than for the FBP examinations (mean SD: ED, 13.9; non-ED, 12.9) (p < 0.05). No statistically significant differences in image noise and CNR were noted among the three IR techniques (p > 0.05) (Fig. 4). Radiation Dose Mean SSDE values were used to report and compare radiation doses within various groups. There was a significant radiation dose reduction in the IR group (ED, 9.2 mGy; non-ED, 7.6 mGy) compared with the FBP group (ED, 13.6 mGy; non-ED, 9.5 mGy) (p < 0.0001), corresponding to a 20–33% reduction in dose. However, no significant difference was found when the mean SSDEs for the three IR groups were compared with each other (non-ED: IR 1 = 7.8 mGy, IR 2 = 7.5 mGy, IR 3 = 7.6 mGy) (p > 0.05) (Figs. 2 and 5). Similarly, no significant difference in dose was found between the FBP groups from different vendors (vendor 1, 11.2 mGy; vendor 3, 9.8 mGy) (p > 0.05). Radiation dose was also significantly lower in examinations performed in non-ED settings (ED, 11.4 mGy; non-ED, 8.1 mGy) irrespective of reconstruction algorithm used (p < 0.05) (29% lower dose) (Table 2 and Fig. 5). When dose variability was tracked for patients grouped by body weight, statistically significant lower radiation doses were observed for patients weighing 91 kg or less

AJR:204, May 2015 1059

Discussion The choice of scanning parameters for a diagnostic CT examination is a moving target, and currently there is no defined dose level that is accepted as optimal for a body part and clinical indication. Therefore, the efforts to lower CT dose with the scanning parameter selection have been somewhat arbitrary. These challenges are amplified when diverse CT technology is being applied in a high-output practice [50]. The reported cumulative radiation doses associated with repeated kidney stone CT examinations range from 8.5 mSv to as high as 154 mSv [51]. Using data from the Dose Index Registry (DIR), Lukasiewicz et al. [52] reported a mean dose-length product (DLP) of 746 mGy × cm (range, 307– 1497 mGy × cm) and median CTDIvol of 14.3

Fig. 4—Box plots depict objective image noise in kidney stone CT examinations performed on non–emergency department scanners. FBP = filtered back projection, IR 1 = ASiR (GE Healthcare), IR 2 = iDOSE (Philips Healthcare), IR 3 = SAFIRE (Siemens Healthcare). The upper and lower limits of the box represent the 3rd and the 1st HYPERLINK http://en.wikipedia.org/ wiki/Quartile\o “Quartile” quartile (Q3 and Q1), and the horizontal line inside the box is median (Q2). Upper whisker represents data within 1.5 times interquartile range (IQR) above Q3, and lower whisker is data within 1.5 times IQR of Q1.  = outlier.

60

Image Noise

(ED: FBP = 12.1 mGy, IR = 8.8 mGy; nonED: FBP = 7.8 mGy, IR = 7.2 mGy) compared with patients who weighed more than 91 kg (ED: FBP = 15.1 mGy, IR = 9.6 mGy; non-ED: FBP = 11.1 mGy, IR = 8.1 mGy) (p < 0.001). When SSDE values were plotted for different weight groups using box plots, we found four outliers with radiation doses greater than 2 SDs above the median (Fig. 6). Radiation dose values (SSDE) generated using dose-monitoring software showed a strong correlation with manually calculated SSDE values (R2 = 0.95).

40

20

0 FBP

mGy (range, 4.9–32.3 mGy) for a single unenhanced kidney stone CT examination performed in 93 U.S. institutions from 2011 to 2013. The mean CTDI at our institution (8.3 mGy) is nearly 45% lower than the mean national and academic averages provided in the DIR feedback report for the period from January 2013 through June 2013 [53] (Fig. 7). This lower CTDI has been possible through continuous protocol optimizations based on patient body weight, clinical indications, image quality expectations, and scanner specifications and technology. Modifications in the scan-

IR 1

IR 2

IR 3

ning acquisition parameters such as lowering the tube voltage (kVp), decreasing the tube current–exposure time product (mAs), using thicker slices and an increased pitch, and limiting the scanning region have been effective in lowering dose from each CT examination [16, 20, 26]. The data from 10 CT scanners using FBP (347 patients) showed a 30% lower mean CTDI (10.5 mGy) than reported in the DIR [53]. The IR algorithms enabled a nearly 50% dose reduction (mean CTDI = 6.8 mGy) compared with the dose reported in the DIR. Similar dose-lowering benefits from

25.0

15.0

SSDE (mGy)

20.0

SSDE (mGy)

Downloaded from www.ajronline.org by Library Of Medicine on 01/25/16 from IP address 128.103.149.52. Copyright ARRS. For personal use only; all rights reserved

Andrabi et al.

15.0

10.0

5.0

10.0

5.0

0

0 FBP

IR ED

FBP

FBP

IR

ASiR

iDOSE

SAFIRE

Non-ED

A

B

Fig. 5—Box plots depict radiation doses for different reconstruction algorithms as size-specific dose estimates (SSDEs).  = outliers. A, Radiation doses are shown for kidney stone CT examinations performed on emergency department (ED) and non-ED scanners using filtered back projection (FBP) and iterative reconstruction (IR) techniques. B, Radiation doses are shown for kidney stone CT examinations performed on non-ED scanners using FBP and three IR techniques: ASiR (GE Healthcare), iDOSE (Philips Healthcare), and SAFIRE (Siemens Healthcare). The upper and lower limits of the box represent the 3rd and the 1st HYPERLINK http://en.wikipedia.org/wiki/Quartile\o “Quartile” quartile (Q3 and Q1), and the horizontal line inside the box is median (Q2). Upper whisker represents data within 1.5 times interquartile range (IQR) above Q3, and lower whisker is data within 1.5 times IQR of Q1.

1060

AJR:204, May 2015

Radiation Dose From Kidney Stone CT 25.0

12.0 ≤ 91 kg

> 91 kg

20.0 SSDE (mGy)

9.0

15.0 10.0

** *

5.0 8.0

IR 1

IR 2

IR 3

A

IR, > 91 kg

IR, ≤ 91 kg

FPB, > 91 kg

FPB

FPB, ≤ 91 kg

6.0

IR, ED, > 91 kg

7.0

IR, ED, ≤ 91 kg

FBP, ED, ≤ 91 kg

0 FBP, ED, > 91 kg

SSDE (mGy)

10.0

Non-ED

B

Fig. 6—Graphics present radiation doses for different reconstruction algorithms as size-specific dose estimates (SSDEs) stratified by patient weight and scanner location. A, Bar graph shows SSDE for kidney stone CT examinations performed using filtered back projection (FBP) and three iterative reconstruction (IR) techniques stratified by patient weight. IR 1 = ASiR (GE Healthcare), IR 2 = iDOSE (Philips Healthcare), IR 3 = SAFIRE (Siemens Healthcare). B, Box plot shows SSDE for kidney stone CT examinations performed using FBP and IR techniques stratified by patient weight and scanner location. ED = emergency department. The upper and lower limits of the box represent the 3rd and the 1st HYPERLINK http://en.wikipedia.org/wiki/Quartile\o “Quartile” quartile (Q3 and Q1), and the horizontal line inside the box is median (Q2). Upper whisker represents data within 1.5 times interquartile range (IQR) above Q3, and lower whisker is data within 1.5 times IQR of Q1.  = outliers.

IR have been reported by other investigators for examinations performed for various indications in the abdomen [31–43, 54]. K ­ ulkarni et al. [33] have reported even higher (80%) dose reduction with the use of ASiR over FBP examinations in follow-up kidney stone CT examinations. Although IR techniques vary in technical specifications, the dose reduction benefits and image quality improvements achieved using our applied scanning protocols were comparable among the three algorithms. These benefits were also evident for patients who were larger than average size (> 91 kg) in whom image quality concerns are amplified in low-dose examinations [36, 42]. The scanners equipped with IR are often relatively expensive; therefore, IR is not available on all scanners. These develop-

SSDE (mGy)

Downloaded from www.ajronline.org by Library Of Medicine on 01/25/16 from IP address 128.103.149.52. Copyright ARRS. For personal use only; all rights reserved

11.0

45.0 40.0 35.0 30.0 25.0 20.0 15.0 10.0 5.0 0

FBP

IR

ments along with other technologic advancements such as multienergy CT introduce a few challenges in maintaining and monitoring an individual patient’s radiation exposure in a busy practice. Furthermore, efforts have been made by many groups, including the American College of Radiology (ACR), U.S. Food and Drug Administration, National Institute of Biomedical Imaging and Bioengineering, The Joint Commission, and National Institutes of Health, to record doses and also to encourage centers to report doses [50, 53, 55–58]. In its efforts to help sites report doses and compare patient dose exposure with national averages and other institutions, ACR opened a DIR in May 2011 and provided a feedback report for January 2013 through June 2013 to all participating insti-

All DIR DIR Participants Participants at Academic Centers

tutions [53]. States such as California have moved further forward by mandating dose reporting from July 2012, and other states are expected to follow this trend as well [59]. This demand for dose reporting has necessitated the introduction of a validated software-based approach that is accurate, reproducible, quantitative, vendor-neutral, and easy to apply [60]. Because the scanner-generated CTDI is an estimate of dose received by standard-sized homogeneous phantoms, the SSDE, which incorporates an individual patient’s size, is considered more appropriate for dose estimation and reporting [50, 61–63]. An automated SSDE calculation is desirable and is made feasible by dose-monitoring software. The software estimates SSDE after determining the effective

Fig. 7—Box plot compares radiation doses at Massachusetts General Hospital using filtered back projection (FBP) and iterative reconstruction (IR) for kidney stone examinations during study period with national and academic center averages for period from January 2013 through June 2013 (center horizontal lines) as provided by American College of Radiology (ACR) in feedback report to Dose Index Registry (DIR) participants [53]. The upper and lower limits of the box represent the 3rd and the 1st HYPERLINK http://en.wikipedia.org/wiki/Quartile\o “Quartile” quartile (Q3 and Q1), and the horizontal line inside the box is median (Q2). Upper whisker represents data within 1.5 times interquartile range (IQR) above Q3, and lower whisker is data within 1.5 times IQR of Q1. SSDE = size-specific dose estimate.

AJR:204, May 2015 1061

Downloaded from www.ajronline.org by Library Of Medicine on 01/25/16 from IP address 128.103.149.52. Copyright ARRS. For personal use only; all rights reserved

Andrabi et al. patient diameter and water-equivalent diameter. This software also estimates effective organ dose by applying an advanced Monte Carlo simulation. Monitoring cumulative doses is also feasible, but we did not calculate the cumulative doses for our patient cohort because it was outside the scope of our study objectives. Despite diverse scanner locations and technologies, the dose-monitoring software enabled us to retrieve all kidney stone CT examinations performed on various scanners along with patient information, CT operator information, radiation dose information, and scanning parameters. Moreover, we identified four outlier examinations of patients who received doses that were higher than expected, but these doses were justified because of the large patient body size (body mass index [weight in kilograms divided by the square of height in meters] > 45). Realtime dose monitoring and setting up dose alerts that are based on accepted dose thresholds for each examination and for cumulative dose for individual patients are some other features available on the dose-monitoring software to augment effective dose monitoring and reporting in the future. Limitations Our retrospective analysis has some limitations. The diversity in scanner and IR technology makes optimal comparison of scanner performance difficult because the dose estimates are based on applied scanning protocols for each examination. Because of logistics, we matched patients by their demographic characteristics and compared various reconstruction techniques within these matched groups. Further, we have not studied the diagnostic performance of all the kidney stone CT examinations to assess any impact of the dose-modified examinations on the overall diagnostic yield and radiologist confidence. Additionally, subjective analysis of image quality was not performed for all patients; instead, we relied on the diagnostic acceptability of all examinations by reviewing the radiology reports for image quality disclaimers. In conclusion, the use of IR enabled a 20–33% reduction in radiation dose for kidney stone CT examinations over standarddose protocols without compromising image quality. The CT dose and image quality were comparable for the applied three IR algorithms (ASiR, iDOSE, and SAFIRE). Dosemonitoring software reliably facilitated radiation dose monitoring for kidney stone CT

1062

examinations in a busy practice with diverse CT technology. It is conceivable that the routine use of IR algorithms and incorporating automated dose-tracking solutions will improve both the image quality of low-dose CT and the radiation safety profile for patients undergoing CT. References 1. Bartoletti R, Cai T, Mondaini N, et al. Epidemiology and risk factors in urolithiasis. Urol Int 2007; 79(suppl 1):3–7 2. Scales CD, Smith AC, Hanley JM, Saigal CS. Project UDiA: prevalence of kidney stones in the United States. Eur Urol 2012; 62:160–165 3. Chu G, Rosenfield AT, Anderson K, Scout L, Smith RC. Sensitivity and value of digital CT scout radiography for detecting ureteral stones in patients with ureterolithiasis diagnosed on unenhanced CT. AJR 1999; 173:417–423 4. Gottlieb RH, La TC, Erturk EN, et al. CT in detecting urinary tract calculi: influence on patient imaging and clinical outcomes. Radiology 2002; 225:441–449 5. Ahmad NA, Ather MH, Rees J. Unenhanced helical computed tomography in the evaluation of acute flank pain. Int J Urol 2003; 10:287–292 6. Cullen IM, Cafferty F, Oon SF, et al. Evaluation of suspected renal colic with noncontrast CT in the emergency department: a single institution study. J Endourol 2008; 22:2441–2445 7. Dhar M, Denstedt JD. Imaging in diagnosis, treatment, and follow-up of stone patients. Adv Chronic Kidney Dis 2009; 16:39–47 8. Kambadakone AR, Eisner BH, Catalano OA, Sahani DV. New and evolving concepts in the imaging and management of urolithiasis: urologists’ perspective. RadioGraphics 2010; 30:603–623 9. Brenner DJ, Hall EJ. Computed tomography: an increasing source of radiation exposure. N Engl J Med 2007; 357:2277–2284 10. Mancini JG, Ferrandino MN. The impact of new methods of imaging on radiation dosage delivered to patients. Curr Opin Urol 2010; 20:163–168 11. Brenner DJ, Hall EJ. Cancer risks from CT scans: now we have data, what next? Radiology 2012; 265:330–331 12. Astroza GM, Neisius A, Wang AJ, et al. Radiation exposure in the follow-up of patients with urolithiasis comparing digital tomosynthesis, noncontrast CT, standard KUB, and IVU. J Endourol 2013; 27:1187–1191 13. Neisius A, Wang AJ, Wang C, et al. Radiation exposure in urology: a genitourinary catalogue for diagnostic imaging. J Urol 2013; 190:2117–2123 14. Shuryak I, Sachs RK, Brenner DJ. Cancer risks after radiation exposure in middle age. J Natl Cancer Inst 2010; 102:1628–1636

15. Fahmy NM, Elkoushy MA, Andonian S. Effective radiation exposure in evaluation and follow-up of patients with urolithiasis. Urology 2012; 79:43–47 16. Sodickson A. Strategies for reducing radiation exposure from multidetector computed tomography in the acute care setting. Can Assoc Radiol J 2013; 64:119–129 17. Heneghan JP, McGuire KA, Leder RA, DeLong DM, Yoshizumi T, Nelson RC. Helical CT for nephrolithiasis and ureterolithiasis: comparison of conventional and reduced radiation-dose techniques. Radiology 2003; 229:575–580 18. Hyams ES, Shah O. Evaluation and follow-up of patients with urinary lithiasis: minimizing radiation exposure. Curr Urol Rep 2010; 11:80–86 19. Kalra MK, Maher MM, D’Souza RV, et al. Detection of urinary tract stones at low-radiation-dose CT with z-axis automatic tube current modulation: phantom and clinical studies. Radiology 2005; 235:523–529 20. Karmazyn B, Frush DP, Applegate KE, Maxfield C, Cohen MD, Jones RP. CT with a computer-simulated dose reduction technique for detection of pediatric nephroureterolithiasis: comparison of standard and reduced radiation doses. AJR 2009; 192:143–149 21. Katz DS, Venkataramanan N, Napel S, Sommer FG. Can low-dose unenhanced multidetector CT be used for routine evaluation of suspected renal colic? AJR 2003; 180:313–315 22. Kim BS, Hwang IK, Choi YW, et al. Low-dose and standard-dose unenhanced helical computed tomography for the assessment of acute renal colic: prospective comparative study. Acta Radiol 2005; 46:756–763 23. Liu W, Esler SJ, Kenny BJ, Goh RH, Rainbow AJ, Stevenson GW. Low-dose nonenhanced helical CT of renal colic: assessment of ureteric stone detection and measurement of effective dose equivalent. Radiology 2000; 215:51–54 24. Poletti PA, Platon A, Rutschmann OT, Schmidlin FR, Iselin CE, Becker CD. Low-dose versus standard-dose CT protocol in patients with clinically suspected renal colic. AJR 2007; 188:927–933 25. Tack D, Sourtzis S, Delpierre I, de Maertelaer V, Gevenois PA. Low-dose unenhanced multidetector CT of patients with suspected renal colic. AJR 2003; 180:305–311 26. Kluner C, Hein PA, Gralla O, et al. Does ultralow-dose CT with a radiation dose equivalent to that of KUB suffice to detect renal and ureteral calculi? J Comput Assist Tomogr 2006; 30:44–50 27. Sohn W, Clayman RV, Lee JY, Cohen A, Mucksavage P. Low-dose and standard computed tomography scans yield equivalent stone measurements. Urology 2013; 81:231–234 28. Beister M, Kolditz D, Kalender WA. Iterative reconstruction methods in X-ray CT. Phys Med 2012; 28:94–108

AJR:204, May 2015

Downloaded from www.ajronline.org by Library Of Medicine on 01/25/16 from IP address 128.103.149.52. Copyright ARRS. For personal use only; all rights reserved

Radiation Dose From Kidney Stone CT 29. Willemink MJ, de Jong PA, Leiner T, et al. Iterative reconstruction techniques for computed tomography. Part 1. Technical principles. Eur Radiol 2013; 23:1623–1631 30. Willemink MJ, Leiner T, de Jong PA, et al. Iterative reconstruction techniques for computed tomography. Part 2. Initial results in dose reduction and image quality. Eur Radiol 2013; 23:1632–1642 31. Kalra MK, Woisetschläger M, Dahlström N, et al. Radiation dose reduction with Sinogram Affirmed Iterative Reconstruction technique for abdominal computed tomography. J Comput Assist Tomogr 2012; 36:339–346 32. Sagara Y, Hara AK, Pavlicek W, Silva AC, Paden RG, Wu Q. Abdominal CT: comparison of lowdose CT with adaptive statistical iterative reconstruction and routine-dose CT with filtered back projection in 53 patients. AJR 2010; 195:713–719 33. Kulkarni NM, Uppot RN, Eisner BH, Sahani DV. Radiation dose reduction at multidetector CT with adaptive statistical iterative reconstruction for evaluation of urolithiasis: how low can we go? Radiology 2012; 265:158–166 34. Singh S, Kalra MK, Do S, et al. Comparison of hybrid and pure iterative reconstruction techniques with conventional filtered back projection: dose reduction potential in the abdomen. J Comput Assist Tomogr 2012; 36:347–353 [Erratum in J Comput Assist Tomogr 2012; 36:627] 35. Noël PB, Renger B, Fiebich M, et al. Does iterative reconstruction lower CT radiation dose: evaluation of 15,000 examinations. PLoS ONE 2013; 8:e81141 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 2012; 22:1631–1640 37. Singh S, Kalra MK, Hsieh J, et al. Abdominal CT: comparison of adaptive statistical iterative and filtered back projection reconstruction techniques. Radiology 2010; 257:373–383 38. Silva AC, Lawder HJ, Hara A, Kujak J, Pavlicek W. Innovations in CT dose reduction strategy: application of the adaptive statistical iterative reconstruction algorithm. AJR 2010; 194:191–199 39. Fuentes-Orrego JM, Hayano K, Kambadakone AR, Hahn PF, Sahani DV. Dose-modified 256MDCT of the abdomen using low tube current and hybrid iterative reconstruction. Acad Radiol 2013; 20:1405–1412 40. Desai GS, Fuentes Orrego JM, Kambadakone AR, Sahani DV. Performance of iterative reconstruction and automated tube voltage selection on the

image quality and radiation dose in abdominal CT scans. J Comput Assist Tomogr 2013; 37:897–903 41. Martinsen AC, Sæther HK, Hol PK, Olsen DR, Skaane P. Iterative reconstruction reduces abdominal CT dose. Eur J Radiol 2012; 81:1483–1487 42. Schindera ST, Odedra D, Mercer D, et al. Hybrid iterative reconstruction technique for abdominal CT protocols in obese patients: assessment of image quality, radiation dose, and low-contrast detectability in a phantom. AJR 2014; 202:[web]W146–W152 43. Desai GS, Thabet A, Elias AY, Sahani DV. Comparative assessment of three image reconstruction techniques for image quality and radiation dose in patients undergoing abdominopelvic multidetector CT examinations. Br J Radiol 2013; 86:20120161 44. Cook TS, Zimmerman S, Maidment AD, Kim W, Boonn WW. Automated extraction of radiation dose information for CT examinations. J Am Coll Radiol 2010; 7:871–877 45. Rehani MM, Berris T. Radiation exposure tracking: survey of unique patient identification number in 40 countries. AJR 2013; 200:776–779 46. Cook TS, Sundaram A, Boonn WW, Kim W. An interactive RADIANCE toolkit for customizable CT dose monitoring and reporting. J Digit Imaging 2013; 26:663–667 47. Rehani MM, Berris T. Templates and existing elements and models for implementation of patient exposure tracking. Radiat Prot Dosimetry 2014; 158:36–42 48. Sodickson A, Warden GI, Farkas CE, et al. Exposing exposure: automated anatomy-specific CT radiation exposure extraction for quality assurance and radiation monitoring. Radiology 2012; 264:397–405 49. Bongartz G, Golding SJ, Jurik AG, et al. European Guidelines on quality criteria for computed tomography. Luxembourg, Luxembourg: European Commission, 2000:EUR 16262 50. Boone JM, Hendee WR, McNitt-Gray MF, Seltzer SE. Radiation exposure from CT scans: how to close our knowledge gaps, monitor and safeguard exposure—proceedings and recommendations of the Radiation Dose Summit, sponsored by NIBIB, February 24–25, 2011. Radiology 2012; 265:544–554 51. Katz SI, Saluja S, Brink JA, Forman HP. Radiation dose associated with unenhanced CT for suspected renal colic: impact of repetitive studies. AJR 2006; 186:1120–1124 52. Lukasiewicz A, Bhargavan-Chatfield M, Coombs L, et al. Radiation dose index of renal colic protocol CT studies in the United States: a report from the American College of Radiology National Radiology Data Registry. Radiology 2014; 271:445–451

53. American College of Radiology website. National Radiology Data Registry. www.acr.org/QualitySafety/National-Radiology-Data-Registry/DoseIndex-Registry. Accessed October 25, 2013 54. Schindera ST, Odedra D, Raza SA, et al. Iterative reconstruction algorithm for CT: can radiation dose be decreased while low-contrast detectability is preserved? Radiology 2013; 269:511–518 55. U.S. Food and Drug Administration website. Initiative to reduce unnecessary radiation exposure from medical imaging. www.fda.gov/Radiation emitting Products / RadiationSafety/RadiationDoseReduction/default.htm. Published 2010. Accessed January 7, 2014 56. Amis ES Jr, Butler PF; American College of Radiology. ACR white paper on radiation dose in medicine: three years later. J Am Coll Radiol 2010; 7:865–870 57. Neumann RD, Bluemke DA. Tracking radiation exposure from diagnostic imaging devices at the NIH. J Am Coll Radiol 2010; 7:87–89 58. The Joint Commission website. Sentinel event alert: radiation risks of diagnostic imaging. www. jointcommission.org/assets/1/18/sea_471.pdf. Published August 24, 2011. Accessed December 17, 2013 59. California State Senate website. SB 1237. http:// www.legalinfo.ca.gov/pub/09-10/bill/sen/sb_12011250/sb_1237_bill_20100929_chaptered.html. Accessed August 21, 2013 60. Brink JA. Dose tracking and rational examination selection for the medically-exposed population. Health Phys 2014; 106:225–228 61. Brink JA, Morin RL. Size-specific dose estimation for CT: how should it be used and what does it mean? Radiology 2012; 265:666–668 62. Ikuta I, Warden GI, Andriole KP, Khorasani R, Sodickson A. Estimating patient dose from x-ray tube output metrics: automated measurement of patient size from CT images enables large-scale size-specific dose estimates. Radiology 2014; 270:472–480 63. American Association of Physicists in Medicine. Size-specific dose estimates (SSDE) in pediatric and adult body CT examinations: report 204 of the AAPM Task Group 204 of the Diagnostic Imaging Council CT Committee. College Park, MD: American Association of Physicists in Medicine, 2011: Report 204 64. [No authors listed]. The 2007 recommendations of the International Commission on Radiological Protection: ICRP publication 103. Ann ICRP 2007; 37:1–332

AJR:204, May 2015 1063