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 Mueller et al. CT Dosimetry During CT Colonography
FOCUS ON:
Downloaded from www.ajronline.org by East Carolina University on 06/09/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved
Medical Physics and Informatics Original Research
JOURNAL CLUB:
JOURNA L
CLUB
Jonathon W. Mueller 1 David J. Vining2 A. Kyle Jones1 David Followill 3 Valen E. Johnson 4,5 Priya Bhosale2 John Rong1 Dianna D. Cody1 Mueller JW, Vining DJ, Jones AK, et al. Keywords: CT dose index (CTDI), CT radiation dose, CT virtual colonography, in vivo radiation dose, size-specific dose estimate DOI:10.2214/AJR.13.11092 Received April 11, 2013; accepted after revision August 18, 2013. The opinions and assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the United States Air Force or the Department of Defense. Supported by the United States Air Force Institute of Technology, Wright-Patterson AFB, OH, and by the M. D. Anderson Cancer Center Support Grant CA 016672. 1 Department of Imaging Physics, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd, Houston, TX 77030. Address correspondence to D. D. Cody ([email protected]). 2 Department of Diagnostic Radiology, The University of Texas M. D. Anderson Cancer Center, Houston, TX. 3 Department of Radiation Physics, The University of Texas M. D. Anderson Cancer Center, Houston, TX. 4 Department of Bioinformatics and Computational Biology, The University of Texas M. D. Anderson Cancer Center, Houston, TX. 5
Present address: Department of Statistics, Texas A&M University, College Station, TX.
AJR 2014; 202:703–710 0361–803X/14/2024–703 © American Roentgen Ray Society
In Vivo CT Dosimetry During CT Colonography OBJECTIVE. The purpose of this study was to develop a method of measuring rectal radiation dose in vivo during CT colonography (CTC) and assess the accuracy of size-specific dose estimates (SSDEs) relative to that of in vivo dose measurements. MATERIALS AND METHODS. Thermoluminescent dosimeter capsules were attached to a CTC rectal catheter to obtain four measurements of the CT radiation dose in 10 volunteers (five men and five women; age range, 23–87 years; mean age, 70.4 years). A fixed CT technique (supine and prone, 50 mAs and 120 kVp each) was used for CTC. SSDEs and percentile body habitus measurements were based on CT images and directly compared with in vivo dose measurements. RESULTS. The mean absorbed doses delivered to the rectum ranged from 8.8 to 23.6 mGy in the 10 patients, whose mean body habitus was in the 27th percentile among American adults 18–64 years old (range, 0.5–67th percentile). The mean SSDE error was 7.2% (range, 0.6–31.4%). CONCLUSION. This in vivo radiation dose measurement technique can be applied to patients undergoing CTC. Our measurements indicate that SSDEs are reasonable estimates of the rectal absorbed dose. The data obtained in this pilot study can be used as benchmarks for assessing dose estimates using other indirect methods (e.g., Monte Carlo simulations).
C
T colonography (CTC), also known as virtual colonoscopy, has been proposed as an effective colorectal cancer screening procedure for individuals at average risk for this disease [1]. Although the American Cancer Society endorses CTC for colorectal cancer screening [2], the U.S. Preventive Services Task Force [3] gave it a rating of 1 for insufficient evidence of the radiation risk posed by the procedure [4]. Most CTC providers currently use MDCT scanners with reduced-dose techniques [5, 6]. CT dose estimates are currently based on calculations using reference patients and uniform phantoms, raising the question of the generalizability of these estimates. Using standard phantoms to estimate radiation dose is difficult because the actual dose delivered to a patient is highly dependent on the patient’s attenuation characteristics. Therefore, the accuracy of dose estimates derived from phantom measurements is limited by how closely the phantom represents the patient’s size and attenuation.
Direct organ radiation dose measurement is better at quantifying radiation exposure than indirect estimation using phantoms or Monte Carlo simulations [7]. To our knowledge, in vivo organ dose measurements during CTC have yet to be performed. Given the absence of benchmarks for in vivo radiation dose measurements, the accuracy of indirect estimates using phantoms and Monte Carlo methods remains uncertain [8–10]. Therefore, we performed this pilot project to measure radiation doses in vivo during CTC and compare the results with dose measurements obtained indirectly. The purposes of this pilot study were to develop a method for measuring rectal radiation doses in vivo during CTC and to compare direct dose measurements with estimates obtained indirectly. Materials and Methods This pilot study was performed after institutional review board approval and patient consent. We attached thermoluminescent dosimeters (TLDs) to the interior lumen of rectal catheters
AJR:202, April 2014 703
Downloaded from www.ajronline.org by East Carolina University on 06/09/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved
Mueller et al. used for CTC to measure radiation doses in vivo without subjecting patients to unnecessary discomfort. The location of each TLD positioned within the rectum (near the isocenter of each patient) ensured these point-dose measurements would not be skewed by surface-dose variation [11]. We used a fixed technique for CTC without tube current modulation, which is currently standard at our institution for this screening examination. Thus, rectal absorbed dose was expected to vary with the size and attenuation characteristics of each patient [12]. In addition to comparing in vivo and phantom dose measurements, we compared our in vivo dose measurements with size-specific dose estimates (SSDEs), which were introduced in American Association of Physicists in Medicine (AAPM) Report 204 as a patient size–based adaptation of the volume CT dose index (CTDIvol) [13, 14]. The radiation dose metric that is typically reported (CTDIvol) represents the exposure to a standard acrylic phantom object of a fixed size (32-cm diameter) and does not directly account for differences in patient size. In a fixed scanning technique scenario such as this one, this CTDIvol value would remain constant even among patients of widely varying size and shape. The SSDE approach enables the CTDIvol value to be properly scaled to represent a test phantom that would correspond with the patient’s cross-sectional area at the location of the dimension measurements. The SSDE is calculated by multiplying the CTDIvol by a correction factor that is based on anteroposterior and lateral patient dimensions at the midpoint of the scanning extent (along the z-direction) [13, 14]. Because the SSDE scales CTDIvol on the basis of the patient’s size, the SSDE may provide a better estimate of a patient’s actual radiation dose than CTDIvol.
Indirect Dose Estimates Three indirect dose approaches were used to compute several different dose estimates that were compared with direct TLD-based measurements: Monte Carlo–based models, a uniform 32-cm-diameter CTDI phantom, and an ATOM model 701 adult male anthropomorphic phantom (Computerized Imaging Reference Systems). These indirect methods were used to obtain colonic dose estimates for the Monte Carlo–based model reported by DeMarco et al. [15], a Monte Carlo–based Image Performance Assessment of CT Scanners (ImPACT) simulation dose estimate in the lower large intestine [16], pointdose measurements with a 0.6-cm3 Farmer chamber positioned on the central axis of the 32-cm CTDI phantom, CTDIvol measurements, TLD measurements in the ATOM anthropomorphic phantom, and SSDE for each patient [13]. The Monte Carlo–based dose estimates were obtained by entering our CTC scanning parameters (Table 1) into the ImPACT Website [16]. All CTDI measurements were obtained from one axial CT scan using the same peak kilovoltage, total milliampere-seconds, and beam widths used in the clinical CTC protocol described in Table 1. Farmer chamber central axis measurements were obtained using the clinical CTC protocol extending the entire 15-cm length of the CTDI phantom according to AAPM Report 111 [17]. The ATOM phantom was modified by machining a cylindric void in one section of the pelvis to simulate an insufflated rectum. A rectal catheter was cut to the length of one section of the anthropomorphic phantom and placed in a polystyrene plug, which was then loaded with two TLDs and inserted into the simulated rectum.
Patients CTC dose measurements were made in patients already scheduled to undergo screening or diag-
nostic CTC (Table 2) between May 16, 2011, and June 15, 2011, using 64-MDCT LightSpeed VCT scanners (GE Healthcare). The patients were recruited by the attending radiologists. Patients with histories of allergic reactions to iodinated contrast agents and women of childbearing age were excluded. Written informed consent was obtained from all patients. A quality assurance process reduced the potential risks from introduction of TLDs into body cavities, including TLD-100 powder coming into direct contact with mucosal surfaces, TLDs becoming unattached from rectal catheters, and infection.
In Vivo Rectal Dose Measurement Two double-chambered TLD-100 capsules (TLD = 100, Harshaw Chemical Company) were attached to the inner lumen of Miller air tip catheters (8925, E2EM) using sutures and silicone glue (Fig. 1). Because each TLD capsule contained two separate compartments, the use of two capsules enabled the acquisition of four dose measurements per patient. Each patient was scanned in supine and prone positions with catheters and TLDs in place; thus, the measured dose was a combination of the doses from the two studies. One patient underwent a third decubitus study directed by the supervising radiologist. The means and SDs of the four measurements for each patient were calculated. The CTC procedure is summarized in Table 1. Introduction of dosimetric measurements was the only modification of the CTC technique. The TLDs were processed by the Radiologic Physics Center at The University of Texas M. D. Anderson Cancer Center using calibration and correction factors established at therapeutic doses and megavoltage energies. Specific calibration and correction factors for the beam quality and dose ranges
TABLE 1: CT Colonography Scanning Parameters Series
Scanning Mode
Scanning Direction
Rotation Time (s)
Image Scanning Length Thickness (mm)
1
Localizer
Lateral
—
Midsternum to trochanter
—
—
—
10
—
2
Localizer
Anteroposterior
—
Midsternum to trochanter
—
—
—
10
—
3
Helicala
Supine
0.5
Diaphragm to symphysis
1.25
0.984
0.8
100
Standard
4
Localizer
Lateral
—
Midsternum to trochanter
—
—
—
80
—
5
Localizer
Anteroposterior
—
Midsternum to trochanter
—
—
—
80
—
6
Helicala
Prone
0.5
Diaphragm to symphysis
1.25
0.984
0.8
100
Standard
Pitch
Image Reconstruction Interval (mm) Amperage Algorithm
Note—All used 120 kVp. Dash indicates not applicable. aDetector configuration, 64 × 0.625 mm.
704
AJR:202, April 2014
Downloaded from www.ajronline.org by East Carolina University on 06/09/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved
CT Dosimetry During CT Colonography
Fig. 1—Drawing shows thermoluminescent dosimeter (TLD) placement in rectal catheter (not drawn to scale).
expected during CTC were established by performing point-dose measurements using a 0.6cm 3 Farmer-type ionization chamber (10 × 5 = 0.6, Radcal Corporation) in a 32-cm CTDI phantom. Linear regression was applied to TLD and Farmer chamber dose measurements for a nominal 120-kVp beam for doses ranging from 0.7 to 27.0 mGy. A disinfected catheter with attached TLDs was inserted into the patient’s rectum (Fig. 1). The catheter is ordinarily used to administer an iodinated contrast enema (iohexol, Omnipaque, GE Healthcare) for stool tagging, followed by CO2 gas colon insufflation using a CO2 insufflator (PROTOCOL2, Bracco Diagnostics). At completion of CTC, the catheter was removed and placed in a double-sealed plastic bag for immediate transport to a location where it was disinfected in accordance with hospital infection control protocols. The disinfected TLDs were extracted and stored in a location away from visible light and external radiation sources for at least 14 days before being processed as described by Kirby et al. [18]. At completion of each CTC study, patient dimensions, including anteroposterior and lateral dimensions and circumference, were measured on the axial CT images at the TLD location in the rectum. The effective diameter of each patient was calculated as the square root of the product of the anteroposterior and lateral dimensions. The effective diameter was used to determine the appropriate correction factor for calculating the SSDE. Patient circumferences were entered into an anthropometric database (PeopleSize 2008 Pro, version 1.1, Open Ergonomics) to determine the percentile of American adults 18–64 years old that each patient’s measurements represented. The percentiles were assessed to determine how well the study patients’ sizes represented
the size distribution in American adults in this age group. The standard error of the TLD measurements was calculated using the pooled estimate of variance in these measurements to assess the uncertainty in the mean value of the four measurements in each patient. The mean values for each patient were used to calculate the mean square error of each indirect dose prediction, and these mean square errors were compared to assess their relative accuracy. In addition, the absolute percentage error of the SSDE relative to the mean rectal dose measurement in each patient was calculated. Finally, the patient CT images and demographic data were reviewed to identify potential causes of outlying data.
Results Five male and five female volunteers were included in this pilot study. Their demographic and dimensional data are summarized in Table 2. Their mean age was 70.4 years (range, 23–87 years). The mean body habitus in our small sample was the 27th percentile (range, 0.5–67th percentile of Americans between 18 and 64 years old). The response of TLD-100 powder to irradiation was linearly correlated with the dose measured using the ionization chamber for all dose levels measured using a 120-kVp MDCT beam (R 2 = 0.999). The TLD capsules were quite near the center of each patient’s cross-section in every case. Example images showing the TLD capsule within the enema tip in two representative patients are shown in Figure 2. The mean radiation dose (average of four measurements per patient for all exposures in each examination) to the rectum was 12.5 mGy (range, 8.8–23.6 mGy). The standard error of the mean was 0.08 mGy. TLD measurements of radiation dose delivered to the rectum, rectal SSDEs, measured CTDIvol, and dose measurements in the modified anthropomorphic phantom are plotted in Figure 3. When the CTC technique factors used in this study were applied to the Monte Carlo–based colonic radiation dose estimates published by DeMarco et al. [15], calculated radiation doses ranged from 11 to 18 mGy. The Monte Carlo–based Im-
TABLE 2: Summary of Patient Demographics, History, Effective Diameter Measurements, and Percentiles of American Adults 18–64 Years Old Represented by Study Patients Patient No.
Sex
Age (y)
History
Effective Diameter (cm)
Percentile
25.5
0.5
1
F
84
Failed optical colonoscopy
2
F
78
Failed optical colonoscopy
26.2
7
3
M
23
Diagnostic
26.9
11
4
M
82
Diagnostic
27.3
11
5
F
72
Screening
28.0
35
6
F
78
Failed optical colonoscopy
28.4
31
7
M
87
Diagnostic
29.1
67
8
F
68
Screening
30.3
27
61
Failed optical colonoscopy
30.5
31
71
Failed optical colonoscopy
31.7
61
28.4
27
9
M
10
M
Mean
50:50 (M:F)
70.4
AJR:202, April 2014 705
Downloaded from www.ajronline.org by East Carolina University on 06/09/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved
Mueller et al.
A
B
Fig. 2—Two representative subjects with thermoluminescent dosimeters (TLDs). A and B, Cross-sectional CT images in 68-year-old woman (A, patient 8) and 84-year-old woman (B, patient 1) show TLDs within enema tips and their nearly central position within overall patient section. The TLD is small dense circle inside tubular structure in air-filled colon.
PACT CT dose calculation program [16] estimated an absorbed dose of 6.3 mGy to the lower large intestine. The SSDE accurately predicted the radiation dose delivered to the rectum within 10% absolute error in eight of 10 patients (Table 3). The mean absolute percentage error of the SSDEs relative to measured doses was 7.2% (range, 0.6–31.4%). Radiation doses calculated using the SSDE were characterized by the lowest mean square error for nine of the 10 patients and the lowest overall mean square error.
Discussion We believe that this study is the first to measure radiation doses in vivo during MDCT scanning. Moreover, this technique and the data presented herein can be used as benchmarks for comparison with indirect dose estimates. We have described a method for measuring absorbed doses in the rectum in vivo in patients undergoing CTC. These measurements are important to determine whether indirect dose estimates are reliable. Our measurements suggest that the SSDE is a reasonable estimate of radiation dose delivered to the rectum dur-
ing CTC. The SSDEs were more accurate than the dose estimates obtained using a modified anthropomorphic phantom or a 32-cm CTDI phantom in nine of 10 patients. As shown in Figure 3, doses measured using TLDs in patients tended to decrease as patients’ effective diameters increased, whereas doses measured using phantoms remained constant regardless of patient size. However, patients 1 and 4 experienced higher radiation doses than the other eight patients. Review of the studies identified the causes of these differences: Patient 1 under-
TABLE 3: Measured and Estimated Radiation Doses Dose (mGy) Mean Square Error
SSDE
Modified Anthropomorphic Phantom
Farmer Chamber
CTDIvol
SSDE
SSDE Absolute Percentage Error
16.50
23.93
3.01
5.69
3.56
0.15
1.3
TLD Mean Dose
Modified Anthropomorphic Phantom
Farmer Chamber
CTDIvol
1
23.63
17.61
12.24
Patient No. 2
12.24
8.80
6.12
8.25
11.80
1.72
3.06
1.99
0.22
3.6
3
8.79
8.80
6.12
8.25
11.55
0.01
1.33
0.27
1.38
31.4
4
17.50
13.21
9.18
12.38
16.71
2.15
4.16
2.56
0.40
4.6
5
11.13
8.80
6.12
8.25
10.89
1.16
2.50
1.44
0.12
2.1
6
12.75
8.80
6.12
8.25
10.73
1.98
3.32
2.25
1.01
15.9
7
9.98
8.80
6.12
8.25
10.56
0.59
1.93
0.87
0.29
5.8
8
9.92
8.80
6.12
8.25
9.98
0.56
1.90
0.84
0.03
0.6
9
9.59
8.80
6.12
8.25
9.90
0.40
1.74
0.67
0.15
3.2
10
9.87
8.80
6.12
8.25
9.57
0.54
1.88
0.81
0.15
3.1
1.21
2.75
1.53
0.39
7.2
Average error
Note—Data show measured thermoluminescent dosimeter (TLD) and estimated modified anthropomorphic phantom, Farmer chamber, volume CT dose index (CTDIvol), and size-specific dose estimate (SSDE), radiation doses delivered to the rectum, and respective mean square errors of each prediction technique and the absolute percentage error of the SSDE. The lowest mean square error for each patient is in bold, representing the most accurate dose estimate. TLD Mean Dose represents only the helical component of the CTC exam (estimated localizer portion was subtracted).
706
AJR:202, April 2014
CT Dosimetry During CT Colonography 30
SSDE Anthropomorphic CTDIvol
20 Radiation Dose (mGy)
Farmer
15
10
25
26
27
28
29
30
Patient 10
Patient 9
Patient 8
Patient 7
Patient 6
Patient 5
Patient 4
Patient 3
0
Patient 2
5 Patient 1
Downloaded from www.ajronline.org by East Carolina University on 06/09/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved
TLD 25
31
32
Effective Patient Diameter (cm)
Fig. 3—Mean rectal radiation doses assessed using thermoluminescent dosimeters (TLDs), size-specific dose estimates (SSDEs), anthropomorphic phantom–based dose estimates, volume CT dose index (CTDIvol), and Farmer ionization chambers as function of each patient’s effective diameter. Each error bar represents one SD of mean of four TLD-based measurements for each patient.
went CTC twice at 100 rather than 50 mAs/ scan as requested by the performing radiologist and patient 4 underwent a third CTC in a decubitus position as requested by the performing radiologist. Excluding these two cases, all rectal doses were below 15 mGy (average, 10.6 mGy). In this study, the mean measured helical radiation doses delivered to the rectum ranged from 8.8 to 23.6 mGy, and the standard error in the TLD dose measurements was small (0.08 mGy). This range was similar to the range from Monte Carlo–based indirect dose estimates (6.3–18.0 mGy). Generally speaking, the accuracy of Monte Carlo–based modeling and dose calculations is unknown because historically they have been benchmarked against other forms of indirect dose measurement, such as phantoms, and not in vivo dose measurements in living patients. The measured dose in patient 3 (8.8 mGy) was accurately predicted by the modified anthropomorphic phantom measurement (8.8 mGy), most likely because the phantom was similar to this patient in both size and attenuation. Compared with the other patients in this study, patient 3 exhibited a more at-
tenuating pelvic structure, which may have shielded the TLD exposure somewhat. The measured helical radiation doses in the other patients ranged from 9.6 to 23.6 mGy. These patients were less similar in size and attenuation to the anthropomorphic phantom. Therefore, their doses were more accurately predicted by the SSDE (range, 9.6–23.9 mGy) than by any other method. These findings suggest that the SSDE is a better estimate of the radiation dose delivered to the rectum than are estimates using phantoms in patients undergoing CTC with a fixed CT technique. In this study rectal dose was accurately predicted using SSDE (within 7% on average). Accounting for the size of pediatric and small adult patients is important when estimating the absorbed dose in CT. SSDEs estimate patient doses using CTDIvol values for CT of the torso, which is most valid at the center of the scanned region (along the z-axis) in a patient. Additionally, the superficial dose distribution is highly variable owing to changing x-ray tube position with rotation speed and pitch; specific patient anatomy, size, and shape; and variable patient centering [11]. Therefore, dose
prediction using the SSDE may not be as accurate for superficial organs compared with organs located deeper in the body. For example, thyroid dose calculated using the SSDE may not be as accurate as rectal dose calculated using the SSDE or doses calculated using the SSDE for organs located near the beginning or end of a scanned region may be less accurate than doses calculated for organs at the middle of a scanned region. This screening examination was performed with a fixed technique as opposed to using tube current modulation. Tube current modulation is commonly used in routine diagnostic CT examinations of the abdominopelvic region because it adjusts the x-ray tube technique (output) to the patient’s size and shape. For screening examinations in which there is inherent excellent image contrast at the target tissues (an air-tissue boundary, for example), such as in CTC and lung cancer screening, a low-dose fixed technique has been historically applied [1, 19, 20]. However, sites have also begun to include tube current modulation for screening CT examinations to further customize the image acquisition for individual patient habitus. Incorporating tube current modula-
AJR:202, April 2014 707
Downloaded from www.ajronline.org by East Carolina University on 06/09/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved
Mueller et al. tion into CTC would result in somewhat higher exposures to larger patients and somewhat smaller exposures to smaller patients but perhaps a more consistent level of dose delivery (absorption) at the colon across patient sizes. A similar study using tube current modulation on a patient cohort with a wider distribution of body habitus would be required to confirm this speculation. Our dose measurements compare favorably with those of Berrington de González et al. [21]. They estimated rectal doses (11–12 mGy) and absorbed doses delivered to other organs to determine the risk-benefit ratio of CTC. Their analyses used a CT protocol similar to ours and data acquired from the American College of Radiology Imaging Network’s national CTC trial [22] and the National Research Council’s Biologic Effects of Ionizing Radiation VII Committee report [23]. The mean calculated benefit ratios, defined as colorectal cancer prevented by CTC to risk of radiation-induced cancer resulting from CTC, ranged from 24:1 to 35:1 for CTC screenings every 5 years for the age range of 50–80 years. Our measurements reinforce the assumptions of this study regarding CTC doses and suggest that the U.S. Preventive Services Task Force “I statement” for CTC on the basis of unknown cancer risk [4] should be reconsidered. The primary limitation of our pilot study was its small sample size of 10 patients. The study population body habitus deviated from the expected normal distribution and its median was lower than the 50th percentile for American adults. These factors may have been a consequence of the study size; the study population’s mean age, which exceeded the demographic metric on which the percentile data were based; or other physiologic factors, e.g., pathologies that may have necessitated the use of CTC rather than standard colonoscopy. As a consequence of the small body habitus, the doses presented here may represent the upper bound of rectal radiation doses for CTC using a similar technique. The standard error of these mean measurements (0.08 mGy) was quite small, indicating excellent measurement precision. The initial raw TLD dose measurements included doses delivered by localizer scanning. To more accurately compare TLD results to indirect dose precidtions, the localizer dose contributions were estimated and subtracted from the initial raw TLD values [24]. Future work in this area should involve performing TLD dose measurements in larg-
708
er populations to reduce the uncertainty in rectal doses in patients undergoing CTC. Future studies may also use our TLD approach to validate other indirect dose estimation methods. This benchmarking can be accomplished by performing Monte Carlo simulations with digital models of patients created from the CT images generated and comparing the simulated absorbed radiation doses with measured values. Future work may also include evaluating rectal doses delivered when tube current modulation is enabled during CTC. In this situation, the TLD reading would depend on the tube current modulation pattern and would represent patient dose to the region of the colon where it was located. Depending on the shape of the patient and the tube current modulation scheme used, the dose at this point in space may be less predictable using phantom-based methods. SSDE in this case might be determined by using the dimensions of the patient at the location of the TLD. The CTDIvol value reported when tube current modulation is used represents the CTDIvol based on the average of the fluctuating amperage values used during scanning. It is feasible that the CTDIvol might also be scaled for the average amperage value used at the z-location of the TLD for a more accurate result. In summary, we developed a method for measuring rectal radiation doses in vivo during CTC. To our knowledge, this is the first report of in vivo rectal doses during CTC. This study provided a novel method for measuring rectal radiation doses in vivo during CTC and assessed the accuracy of the SSDE for estimating doses. Furthermore, the present data can be used in future studies as benchmarks for indirect patient dose estimates, which are currently benchmarked against other indirect methods. Our study may inspire other researchers to develop similar techniques to measure radiation doses in other organs in vivo to compare actual and indirect dose measurements. 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; 359:1207–1217 2. 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; 58:130–160 3. Centers for Medicare and Medicaid Services (CMS). Decision memo for screening computed tomography colonography (CTC) for colorectal cancer (CAG-00396N). Baltimore, MD: U.S. Department of Health and Human Services CMS, 2009 4. Calonge N, Petitti DB, DeWitt TG, et al. Screening for colorectal cancer: U.S. Preventative Services Task Force recommendation statement. Ann Intern Med 2008; 149:627–637 5. An S, Lee KH, Kim YH, et al. Screening CT colonography in an asymptomatic average-risk Asian population: a 2-year experience in a single institution. AJR 2008; 191:[web]W100–W106 6. Cohnen M, Vogt C, Beck A, et al. Feasibility of MDCT colonography in ultra-low-dose technique in the detection of colorectal lesions: comparison with high-resolution video colonoscopy. AJR 2004; 183:1355–1359 7. Martin CJ. The application of effective dose to medical exposures. Radiat Prot Dosimetry 2008; 128:1–4 8. Brenner DJ, Georgsson MA. Mass screening with CT colonography: should the radiation exposure be of concern? Gastroenterology 2005; 129:328–337 9. Liedenbaum MH, Venema HW, Stoker J. Radiation dose in CT colonography: trends in time and differences between daily practice and screening protocols. Eur Radiol 2008; 18:2222–2230 10. Bou Serhal C, Jacobs R, Gijbels F, et al. Absorbed doses from spiral CT and conventional spiral tomography: a phantom vs. cadaver study. Clin Oral Implants Res 2001; 12:473–478 11. Zhang D, Savandi AS, Demarco JJ, et al. Variability of surface and center position radiation dose in MDCT: Monte Carlo simulations using CTDI and anthropomorphic phantoms. Med Phys 2009; 36:1025–1038 12. Turner AC, Zhang D, Khotonabadi M, et al. The feasibility of patient size-corrected, scanner-independent organ dose estimates for abdominal CT exams. Med Phys 2011; 38:820–829 13. American Association of Physicists in Medicine (AAPM). Report 204: Correction factors for patient size in CT dose estimation. New York, NY: AAPM, 2011 14. American Association of Physicists in Medicine (AAPM). Report 96: The measurement, reporting, and management of radiation dose in CT. New York, NY: AAPM, 2008 15. DeMarco JJ, Cagnon CH, Cody DD, et al. Estimating radiation doses from multidetector CT using Monte Carlo simulations: effects of different sized voxelized patient models on magnitudes of organ and effective dose. Phys Med Biol 2007; 52:2583–2597 16. ImPACT website. CTDosimetry version 1.0.3: Im-
AJR:202, April 2014
Downloaded from www.ajronline.org by East Carolina University on 06/09/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved
CT Dosimetry During CT Colonography PACT—image performance assessment of CT scanners. www.impactscan.org/ctdosimetry.htm. Published August 24, 2010. Accessed December 2010. 17. American Association of Physicists in Medicine (AAPM). Report 111: Comprehensive methodology for the evaluation of radiation dose in x-ray computed tomography. New York, NY: AAPM, 2010 18. Kirby TH, Hanson WF, Johnston DA. Uncertainty analysis of absorbed dose calculations from thermoluminescent dosimeters. Med Phys 1992; 19:1427–1433 19. Cagnon CH, Cody DD, McNitt-Gray MF, Seibert JA, Judy PF, Aberle DR. Description and imple-
mentation of a quality control program in an imaging-based clinical trial. Acad Radiol 2006; 13:1431–1441 20. McFarland EG, Fletcher JG, Pickhardt P, et al. ACR Colon Cancer Committee white paper: status of CT colonography 2009. J Am Coll Radiol 2009; 6:756–772 21. Berrington de González A, Kim KP, Knudsen AB, et al. Radiation-related cancer risks from CT colonography screening: a risk-benefit analysis. AJR 2011; 196:816–823 22. American College of Radiology (ACRIN) website. Protocol 6664: the National CT colonography trial.
www.acrin.org/TabID/151/Default.aspx. Accessed October 24, 2012 23. Committee to Assess Health Risks From Exposure to Low Levels of Ionizing Radiation, Board on Radiation Effects Research, Division of Earth and Life Studies, National Research Council of the National Academies. Health risks from exposure to low levels of ionizing radiation: BEIR VII phase 2. Washington, DC: National Academies Press, 2006 24. O’Daniel JC, Stevens DM, Cody DD. Reducing radiation exposure from survey CT scans. AJR 2005; 185:509–515
F O R YO U R I N F O R M AT I O N
For more information on Journal Clubs, see “Evidence-Based Radiology: A Primer in Reading Scientific Articles” in the July 2010 AJR at www.ajronline.org/cgi/content/full/195/1/W1.
AJR:202, April 2014 709
Mueller et al. APPENDIX 1: AJR JOURNAL CLUB
Downloaded from www.ajronline.org by East Carolina University on 06/09/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved
Study Guide In Vivo CT Dosimetry During CT Colonography Margaret Mulligan1, Joseph J. Budovec1, Alan Mautz 2 1Medical College of Wisconsin, Milwaukee, WI 2 The Aroostook Medical Center, Presque Isle, ME [email protected], [email protected], [email protected] Introduction 1. What is the research question being asked with this study? Is this study timely and relevant? Does the study attempt to answer a clinically important question? 2. What is unique about this topic? Methods 3. How were patients selected for inclusion in this study? What were the exclusion criteria? 4. Does the population excluded from the study impact the generalizability of results? 5. What are the limitations of this study? Were the limitations adequately addressed? 6. What types of bias might be introduced due to study design and patient selection? Results 7. Was the research question answered? What is the null hypothesis in this study? Physics 8. What is the size-specific dose estimate? How is this parameter related to the volume CT dose index (CTDIvol)? Which parameter may be a better estimate of the patient’s actual radiation dose? 9. What is a thermoluminescent dosimeter? How does it work? Discussion 10. What impact does this study have on clinical practice? 11. Although the authors acknowledge the study as a pilot study, does the study’s population size impact standards of practice? 12. How does the method used in this study for measuring rectal radiation doses in vivo during CT colonography vary from other accepted methods? 13. How could the study design be altered to increase the power of the results? 14. Do you agree with the authors’ assertion that the results of this study should impact the U. S. Preventive Services Task Force recommendation regarding CT colonography? Why or why not? Background Reading 1. Berrington de Gonzalez A, Kim KP, Knudsen AB, et al. Radiation-related cancer risks from CT colonography screening: a risk-benefit analysis. AJR 2011; 196:816– 823 2. Martin CJ. The application of effective dose to medical exposures. Radiat Prot Dosimetry 2008; 128:1–4
*Please note that the authors of the Study Guide are distinct from those of the companion article.
710
AJR:202, April 2014
FOCUS ON:
Downloaded from www.ajronline.org by East Carolina University on 06/09/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved
Medical Physics and Informatics Original Research
JOURNAL CLUB:
JOURNA L
CLUB
Jonathon W. Mueller 1 David J. Vining2 A. Kyle Jones1 David Followill 3 Valen E. Johnson 4,5 Priya Bhosale2 John Rong1 Dianna D. Cody1 Mueller JW, Vining DJ, Jones AK, et al. Keywords: CT dose index (CTDI), CT radiation dose, CT virtual colonography, in vivo radiation dose, size-specific dose estimate DOI:10.2214/AJR.13.11092 Received April 11, 2013; accepted after revision August 18, 2013. The opinions and assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the United States Air Force or the Department of Defense. Supported by the United States Air Force Institute of Technology, Wright-Patterson AFB, OH, and by the M. D. Anderson Cancer Center Support Grant CA 016672. 1 Department of Imaging Physics, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd, Houston, TX 77030. Address correspondence to D. D. Cody ([email protected]). 2 Department of Diagnostic Radiology, The University of Texas M. D. Anderson Cancer Center, Houston, TX. 3 Department of Radiation Physics, The University of Texas M. D. Anderson Cancer Center, Houston, TX. 4 Department of Bioinformatics and Computational Biology, The University of Texas M. D. Anderson Cancer Center, Houston, TX. 5
Present address: Department of Statistics, Texas A&M University, College Station, TX.
AJR 2014; 202:703–710 0361–803X/14/2024–703 © American Roentgen Ray Society
In Vivo CT Dosimetry During CT Colonography OBJECTIVE. The purpose of this study was to develop a method of measuring rectal radiation dose in vivo during CT colonography (CTC) and assess the accuracy of size-specific dose estimates (SSDEs) relative to that of in vivo dose measurements. MATERIALS AND METHODS. Thermoluminescent dosimeter capsules were attached to a CTC rectal catheter to obtain four measurements of the CT radiation dose in 10 volunteers (five men and five women; age range, 23–87 years; mean age, 70.4 years). A fixed CT technique (supine and prone, 50 mAs and 120 kVp each) was used for CTC. SSDEs and percentile body habitus measurements were based on CT images and directly compared with in vivo dose measurements. RESULTS. The mean absorbed doses delivered to the rectum ranged from 8.8 to 23.6 mGy in the 10 patients, whose mean body habitus was in the 27th percentile among American adults 18–64 years old (range, 0.5–67th percentile). The mean SSDE error was 7.2% (range, 0.6–31.4%). CONCLUSION. This in vivo radiation dose measurement technique can be applied to patients undergoing CTC. Our measurements indicate that SSDEs are reasonable estimates of the rectal absorbed dose. The data obtained in this pilot study can be used as benchmarks for assessing dose estimates using other indirect methods (e.g., Monte Carlo simulations).
C
T colonography (CTC), also known as virtual colonoscopy, has been proposed as an effective colorectal cancer screening procedure for individuals at average risk for this disease [1]. Although the American Cancer Society endorses CTC for colorectal cancer screening [2], the U.S. Preventive Services Task Force [3] gave it a rating of 1 for insufficient evidence of the radiation risk posed by the procedure [4]. Most CTC providers currently use MDCT scanners with reduced-dose techniques [5, 6]. CT dose estimates are currently based on calculations using reference patients and uniform phantoms, raising the question of the generalizability of these estimates. Using standard phantoms to estimate radiation dose is difficult because the actual dose delivered to a patient is highly dependent on the patient’s attenuation characteristics. Therefore, the accuracy of dose estimates derived from phantom measurements is limited by how closely the phantom represents the patient’s size and attenuation.
Direct organ radiation dose measurement is better at quantifying radiation exposure than indirect estimation using phantoms or Monte Carlo simulations [7]. To our knowledge, in vivo organ dose measurements during CTC have yet to be performed. Given the absence of benchmarks for in vivo radiation dose measurements, the accuracy of indirect estimates using phantoms and Monte Carlo methods remains uncertain [8–10]. Therefore, we performed this pilot project to measure radiation doses in vivo during CTC and compare the results with dose measurements obtained indirectly. The purposes of this pilot study were to develop a method for measuring rectal radiation doses in vivo during CTC and to compare direct dose measurements with estimates obtained indirectly. Materials and Methods This pilot study was performed after institutional review board approval and patient consent. We attached thermoluminescent dosimeters (TLDs) to the interior lumen of rectal catheters
AJR:202, April 2014 703
Downloaded from www.ajronline.org by East Carolina University on 06/09/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved
Mueller et al. used for CTC to measure radiation doses in vivo without subjecting patients to unnecessary discomfort. The location of each TLD positioned within the rectum (near the isocenter of each patient) ensured these point-dose measurements would not be skewed by surface-dose variation [11]. We used a fixed technique for CTC without tube current modulation, which is currently standard at our institution for this screening examination. Thus, rectal absorbed dose was expected to vary with the size and attenuation characteristics of each patient [12]. In addition to comparing in vivo and phantom dose measurements, we compared our in vivo dose measurements with size-specific dose estimates (SSDEs), which were introduced in American Association of Physicists in Medicine (AAPM) Report 204 as a patient size–based adaptation of the volume CT dose index (CTDIvol) [13, 14]. The radiation dose metric that is typically reported (CTDIvol) represents the exposure to a standard acrylic phantom object of a fixed size (32-cm diameter) and does not directly account for differences in patient size. In a fixed scanning technique scenario such as this one, this CTDIvol value would remain constant even among patients of widely varying size and shape. The SSDE approach enables the CTDIvol value to be properly scaled to represent a test phantom that would correspond with the patient’s cross-sectional area at the location of the dimension measurements. The SSDE is calculated by multiplying the CTDIvol by a correction factor that is based on anteroposterior and lateral patient dimensions at the midpoint of the scanning extent (along the z-direction) [13, 14]. Because the SSDE scales CTDIvol on the basis of the patient’s size, the SSDE may provide a better estimate of a patient’s actual radiation dose than CTDIvol.
Indirect Dose Estimates Three indirect dose approaches were used to compute several different dose estimates that were compared with direct TLD-based measurements: Monte Carlo–based models, a uniform 32-cm-diameter CTDI phantom, and an ATOM model 701 adult male anthropomorphic phantom (Computerized Imaging Reference Systems). These indirect methods were used to obtain colonic dose estimates for the Monte Carlo–based model reported by DeMarco et al. [15], a Monte Carlo–based Image Performance Assessment of CT Scanners (ImPACT) simulation dose estimate in the lower large intestine [16], pointdose measurements with a 0.6-cm3 Farmer chamber positioned on the central axis of the 32-cm CTDI phantom, CTDIvol measurements, TLD measurements in the ATOM anthropomorphic phantom, and SSDE for each patient [13]. The Monte Carlo–based dose estimates were obtained by entering our CTC scanning parameters (Table 1) into the ImPACT Website [16]. All CTDI measurements were obtained from one axial CT scan using the same peak kilovoltage, total milliampere-seconds, and beam widths used in the clinical CTC protocol described in Table 1. Farmer chamber central axis measurements were obtained using the clinical CTC protocol extending the entire 15-cm length of the CTDI phantom according to AAPM Report 111 [17]. The ATOM phantom was modified by machining a cylindric void in one section of the pelvis to simulate an insufflated rectum. A rectal catheter was cut to the length of one section of the anthropomorphic phantom and placed in a polystyrene plug, which was then loaded with two TLDs and inserted into the simulated rectum.
Patients CTC dose measurements were made in patients already scheduled to undergo screening or diag-
nostic CTC (Table 2) between May 16, 2011, and June 15, 2011, using 64-MDCT LightSpeed VCT scanners (GE Healthcare). The patients were recruited by the attending radiologists. Patients with histories of allergic reactions to iodinated contrast agents and women of childbearing age were excluded. Written informed consent was obtained from all patients. A quality assurance process reduced the potential risks from introduction of TLDs into body cavities, including TLD-100 powder coming into direct contact with mucosal surfaces, TLDs becoming unattached from rectal catheters, and infection.
In Vivo Rectal Dose Measurement Two double-chambered TLD-100 capsules (TLD = 100, Harshaw Chemical Company) were attached to the inner lumen of Miller air tip catheters (8925, E2EM) using sutures and silicone glue (Fig. 1). Because each TLD capsule contained two separate compartments, the use of two capsules enabled the acquisition of four dose measurements per patient. Each patient was scanned in supine and prone positions with catheters and TLDs in place; thus, the measured dose was a combination of the doses from the two studies. One patient underwent a third decubitus study directed by the supervising radiologist. The means and SDs of the four measurements for each patient were calculated. The CTC procedure is summarized in Table 1. Introduction of dosimetric measurements was the only modification of the CTC technique. The TLDs were processed by the Radiologic Physics Center at The University of Texas M. D. Anderson Cancer Center using calibration and correction factors established at therapeutic doses and megavoltage energies. Specific calibration and correction factors for the beam quality and dose ranges
TABLE 1: CT Colonography Scanning Parameters Series
Scanning Mode
Scanning Direction
Rotation Time (s)
Image Scanning Length Thickness (mm)
1
Localizer
Lateral
—
Midsternum to trochanter
—
—
—
10
—
2
Localizer
Anteroposterior
—
Midsternum to trochanter
—
—
—
10
—
3
Helicala
Supine
0.5
Diaphragm to symphysis
1.25
0.984
0.8
100
Standard
4
Localizer
Lateral
—
Midsternum to trochanter
—
—
—
80
—
5
Localizer
Anteroposterior
—
Midsternum to trochanter
—
—
—
80
—
6
Helicala
Prone
0.5
Diaphragm to symphysis
1.25
0.984
0.8
100
Standard
Pitch
Image Reconstruction Interval (mm) Amperage Algorithm
Note—All used 120 kVp. Dash indicates not applicable. aDetector configuration, 64 × 0.625 mm.
704
AJR:202, April 2014
Downloaded from www.ajronline.org by East Carolina University on 06/09/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved
CT Dosimetry During CT Colonography
Fig. 1—Drawing shows thermoluminescent dosimeter (TLD) placement in rectal catheter (not drawn to scale).
expected during CTC were established by performing point-dose measurements using a 0.6cm 3 Farmer-type ionization chamber (10 × 5 = 0.6, Radcal Corporation) in a 32-cm CTDI phantom. Linear regression was applied to TLD and Farmer chamber dose measurements for a nominal 120-kVp beam for doses ranging from 0.7 to 27.0 mGy. A disinfected catheter with attached TLDs was inserted into the patient’s rectum (Fig. 1). The catheter is ordinarily used to administer an iodinated contrast enema (iohexol, Omnipaque, GE Healthcare) for stool tagging, followed by CO2 gas colon insufflation using a CO2 insufflator (PROTOCOL2, Bracco Diagnostics). At completion of CTC, the catheter was removed and placed in a double-sealed plastic bag for immediate transport to a location where it was disinfected in accordance with hospital infection control protocols. The disinfected TLDs were extracted and stored in a location away from visible light and external radiation sources for at least 14 days before being processed as described by Kirby et al. [18]. At completion of each CTC study, patient dimensions, including anteroposterior and lateral dimensions and circumference, were measured on the axial CT images at the TLD location in the rectum. The effective diameter of each patient was calculated as the square root of the product of the anteroposterior and lateral dimensions. The effective diameter was used to determine the appropriate correction factor for calculating the SSDE. Patient circumferences were entered into an anthropometric database (PeopleSize 2008 Pro, version 1.1, Open Ergonomics) to determine the percentile of American adults 18–64 years old that each patient’s measurements represented. The percentiles were assessed to determine how well the study patients’ sizes represented
the size distribution in American adults in this age group. The standard error of the TLD measurements was calculated using the pooled estimate of variance in these measurements to assess the uncertainty in the mean value of the four measurements in each patient. The mean values for each patient were used to calculate the mean square error of each indirect dose prediction, and these mean square errors were compared to assess their relative accuracy. In addition, the absolute percentage error of the SSDE relative to the mean rectal dose measurement in each patient was calculated. Finally, the patient CT images and demographic data were reviewed to identify potential causes of outlying data.
Results Five male and five female volunteers were included in this pilot study. Their demographic and dimensional data are summarized in Table 2. Their mean age was 70.4 years (range, 23–87 years). The mean body habitus in our small sample was the 27th percentile (range, 0.5–67th percentile of Americans between 18 and 64 years old). The response of TLD-100 powder to irradiation was linearly correlated with the dose measured using the ionization chamber for all dose levels measured using a 120-kVp MDCT beam (R 2 = 0.999). The TLD capsules were quite near the center of each patient’s cross-section in every case. Example images showing the TLD capsule within the enema tip in two representative patients are shown in Figure 2. The mean radiation dose (average of four measurements per patient for all exposures in each examination) to the rectum was 12.5 mGy (range, 8.8–23.6 mGy). The standard error of the mean was 0.08 mGy. TLD measurements of radiation dose delivered to the rectum, rectal SSDEs, measured CTDIvol, and dose measurements in the modified anthropomorphic phantom are plotted in Figure 3. When the CTC technique factors used in this study were applied to the Monte Carlo–based colonic radiation dose estimates published by DeMarco et al. [15], calculated radiation doses ranged from 11 to 18 mGy. The Monte Carlo–based Im-
TABLE 2: Summary of Patient Demographics, History, Effective Diameter Measurements, and Percentiles of American Adults 18–64 Years Old Represented by Study Patients Patient No.
Sex
Age (y)
History
Effective Diameter (cm)
Percentile
25.5
0.5
1
F
84
Failed optical colonoscopy
2
F
78
Failed optical colonoscopy
26.2
7
3
M
23
Diagnostic
26.9
11
4
M
82
Diagnostic
27.3
11
5
F
72
Screening
28.0
35
6
F
78
Failed optical colonoscopy
28.4
31
7
M
87
Diagnostic
29.1
67
8
F
68
Screening
30.3
27
61
Failed optical colonoscopy
30.5
31
71
Failed optical colonoscopy
31.7
61
28.4
27
9
M
10
M
Mean
50:50 (M:F)
70.4
AJR:202, April 2014 705
Downloaded from www.ajronline.org by East Carolina University on 06/09/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved
Mueller et al.
A
B
Fig. 2—Two representative subjects with thermoluminescent dosimeters (TLDs). A and B, Cross-sectional CT images in 68-year-old woman (A, patient 8) and 84-year-old woman (B, patient 1) show TLDs within enema tips and their nearly central position within overall patient section. The TLD is small dense circle inside tubular structure in air-filled colon.
PACT CT dose calculation program [16] estimated an absorbed dose of 6.3 mGy to the lower large intestine. The SSDE accurately predicted the radiation dose delivered to the rectum within 10% absolute error in eight of 10 patients (Table 3). The mean absolute percentage error of the SSDEs relative to measured doses was 7.2% (range, 0.6–31.4%). Radiation doses calculated using the SSDE were characterized by the lowest mean square error for nine of the 10 patients and the lowest overall mean square error.
Discussion We believe that this study is the first to measure radiation doses in vivo during MDCT scanning. Moreover, this technique and the data presented herein can be used as benchmarks for comparison with indirect dose estimates. We have described a method for measuring absorbed doses in the rectum in vivo in patients undergoing CTC. These measurements are important to determine whether indirect dose estimates are reliable. Our measurements suggest that the SSDE is a reasonable estimate of radiation dose delivered to the rectum dur-
ing CTC. The SSDEs were more accurate than the dose estimates obtained using a modified anthropomorphic phantom or a 32-cm CTDI phantom in nine of 10 patients. As shown in Figure 3, doses measured using TLDs in patients tended to decrease as patients’ effective diameters increased, whereas doses measured using phantoms remained constant regardless of patient size. However, patients 1 and 4 experienced higher radiation doses than the other eight patients. Review of the studies identified the causes of these differences: Patient 1 under-
TABLE 3: Measured and Estimated Radiation Doses Dose (mGy) Mean Square Error
SSDE
Modified Anthropomorphic Phantom
Farmer Chamber
CTDIvol
SSDE
SSDE Absolute Percentage Error
16.50
23.93
3.01
5.69
3.56
0.15
1.3
TLD Mean Dose
Modified Anthropomorphic Phantom
Farmer Chamber
CTDIvol
1
23.63
17.61
12.24
Patient No. 2
12.24
8.80
6.12
8.25
11.80
1.72
3.06
1.99
0.22
3.6
3
8.79
8.80
6.12
8.25
11.55
0.01
1.33
0.27
1.38
31.4
4
17.50
13.21
9.18
12.38
16.71
2.15
4.16
2.56
0.40
4.6
5
11.13
8.80
6.12
8.25
10.89
1.16
2.50
1.44
0.12
2.1
6
12.75
8.80
6.12
8.25
10.73
1.98
3.32
2.25
1.01
15.9
7
9.98
8.80
6.12
8.25
10.56
0.59
1.93
0.87
0.29
5.8
8
9.92
8.80
6.12
8.25
9.98
0.56
1.90
0.84
0.03
0.6
9
9.59
8.80
6.12
8.25
9.90
0.40
1.74
0.67
0.15
3.2
10
9.87
8.80
6.12
8.25
9.57
0.54
1.88
0.81
0.15
3.1
1.21
2.75
1.53
0.39
7.2
Average error
Note—Data show measured thermoluminescent dosimeter (TLD) and estimated modified anthropomorphic phantom, Farmer chamber, volume CT dose index (CTDIvol), and size-specific dose estimate (SSDE), radiation doses delivered to the rectum, and respective mean square errors of each prediction technique and the absolute percentage error of the SSDE. The lowest mean square error for each patient is in bold, representing the most accurate dose estimate. TLD Mean Dose represents only the helical component of the CTC exam (estimated localizer portion was subtracted).
706
AJR:202, April 2014
CT Dosimetry During CT Colonography 30
SSDE Anthropomorphic CTDIvol
20 Radiation Dose (mGy)
Farmer
15
10
25
26
27
28
29
30
Patient 10
Patient 9
Patient 8
Patient 7
Patient 6
Patient 5
Patient 4
Patient 3
0
Patient 2
5 Patient 1
Downloaded from www.ajronline.org by East Carolina University on 06/09/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved
TLD 25
31
32
Effective Patient Diameter (cm)
Fig. 3—Mean rectal radiation doses assessed using thermoluminescent dosimeters (TLDs), size-specific dose estimates (SSDEs), anthropomorphic phantom–based dose estimates, volume CT dose index (CTDIvol), and Farmer ionization chambers as function of each patient’s effective diameter. Each error bar represents one SD of mean of four TLD-based measurements for each patient.
went CTC twice at 100 rather than 50 mAs/ scan as requested by the performing radiologist and patient 4 underwent a third CTC in a decubitus position as requested by the performing radiologist. Excluding these two cases, all rectal doses were below 15 mGy (average, 10.6 mGy). In this study, the mean measured helical radiation doses delivered to the rectum ranged from 8.8 to 23.6 mGy, and the standard error in the TLD dose measurements was small (0.08 mGy). This range was similar to the range from Monte Carlo–based indirect dose estimates (6.3–18.0 mGy). Generally speaking, the accuracy of Monte Carlo–based modeling and dose calculations is unknown because historically they have been benchmarked against other forms of indirect dose measurement, such as phantoms, and not in vivo dose measurements in living patients. The measured dose in patient 3 (8.8 mGy) was accurately predicted by the modified anthropomorphic phantom measurement (8.8 mGy), most likely because the phantom was similar to this patient in both size and attenuation. Compared with the other patients in this study, patient 3 exhibited a more at-
tenuating pelvic structure, which may have shielded the TLD exposure somewhat. The measured helical radiation doses in the other patients ranged from 9.6 to 23.6 mGy. These patients were less similar in size and attenuation to the anthropomorphic phantom. Therefore, their doses were more accurately predicted by the SSDE (range, 9.6–23.9 mGy) than by any other method. These findings suggest that the SSDE is a better estimate of the radiation dose delivered to the rectum than are estimates using phantoms in patients undergoing CTC with a fixed CT technique. In this study rectal dose was accurately predicted using SSDE (within 7% on average). Accounting for the size of pediatric and small adult patients is important when estimating the absorbed dose in CT. SSDEs estimate patient doses using CTDIvol values for CT of the torso, which is most valid at the center of the scanned region (along the z-axis) in a patient. Additionally, the superficial dose distribution is highly variable owing to changing x-ray tube position with rotation speed and pitch; specific patient anatomy, size, and shape; and variable patient centering [11]. Therefore, dose
prediction using the SSDE may not be as accurate for superficial organs compared with organs located deeper in the body. For example, thyroid dose calculated using the SSDE may not be as accurate as rectal dose calculated using the SSDE or doses calculated using the SSDE for organs located near the beginning or end of a scanned region may be less accurate than doses calculated for organs at the middle of a scanned region. This screening examination was performed with a fixed technique as opposed to using tube current modulation. Tube current modulation is commonly used in routine diagnostic CT examinations of the abdominopelvic region because it adjusts the x-ray tube technique (output) to the patient’s size and shape. For screening examinations in which there is inherent excellent image contrast at the target tissues (an air-tissue boundary, for example), such as in CTC and lung cancer screening, a low-dose fixed technique has been historically applied [1, 19, 20]. However, sites have also begun to include tube current modulation for screening CT examinations to further customize the image acquisition for individual patient habitus. Incorporating tube current modula-
AJR:202, April 2014 707
Downloaded from www.ajronline.org by East Carolina University on 06/09/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved
Mueller et al. tion into CTC would result in somewhat higher exposures to larger patients and somewhat smaller exposures to smaller patients but perhaps a more consistent level of dose delivery (absorption) at the colon across patient sizes. A similar study using tube current modulation on a patient cohort with a wider distribution of body habitus would be required to confirm this speculation. Our dose measurements compare favorably with those of Berrington de González et al. [21]. They estimated rectal doses (11–12 mGy) and absorbed doses delivered to other organs to determine the risk-benefit ratio of CTC. Their analyses used a CT protocol similar to ours and data acquired from the American College of Radiology Imaging Network’s national CTC trial [22] and the National Research Council’s Biologic Effects of Ionizing Radiation VII Committee report [23]. The mean calculated benefit ratios, defined as colorectal cancer prevented by CTC to risk of radiation-induced cancer resulting from CTC, ranged from 24:1 to 35:1 for CTC screenings every 5 years for the age range of 50–80 years. Our measurements reinforce the assumptions of this study regarding CTC doses and suggest that the U.S. Preventive Services Task Force “I statement” for CTC on the basis of unknown cancer risk [4] should be reconsidered. The primary limitation of our pilot study was its small sample size of 10 patients. The study population body habitus deviated from the expected normal distribution and its median was lower than the 50th percentile for American adults. These factors may have been a consequence of the study size; the study population’s mean age, which exceeded the demographic metric on which the percentile data were based; or other physiologic factors, e.g., pathologies that may have necessitated the use of CTC rather than standard colonoscopy. As a consequence of the small body habitus, the doses presented here may represent the upper bound of rectal radiation doses for CTC using a similar technique. The standard error of these mean measurements (0.08 mGy) was quite small, indicating excellent measurement precision. The initial raw TLD dose measurements included doses delivered by localizer scanning. To more accurately compare TLD results to indirect dose precidtions, the localizer dose contributions were estimated and subtracted from the initial raw TLD values [24]. Future work in this area should involve performing TLD dose measurements in larg-
708
er populations to reduce the uncertainty in rectal doses in patients undergoing CTC. Future studies may also use our TLD approach to validate other indirect dose estimation methods. This benchmarking can be accomplished by performing Monte Carlo simulations with digital models of patients created from the CT images generated and comparing the simulated absorbed radiation doses with measured values. Future work may also include evaluating rectal doses delivered when tube current modulation is enabled during CTC. In this situation, the TLD reading would depend on the tube current modulation pattern and would represent patient dose to the region of the colon where it was located. Depending on the shape of the patient and the tube current modulation scheme used, the dose at this point in space may be less predictable using phantom-based methods. SSDE in this case might be determined by using the dimensions of the patient at the location of the TLD. The CTDIvol value reported when tube current modulation is used represents the CTDIvol based on the average of the fluctuating amperage values used during scanning. It is feasible that the CTDIvol might also be scaled for the average amperage value used at the z-location of the TLD for a more accurate result. In summary, we developed a method for measuring rectal radiation doses in vivo during CTC. To our knowledge, this is the first report of in vivo rectal doses during CTC. This study provided a novel method for measuring rectal radiation doses in vivo during CTC and assessed the accuracy of the SSDE for estimating doses. Furthermore, the present data can be used in future studies as benchmarks for indirect patient dose estimates, which are currently benchmarked against other indirect methods. Our study may inspire other researchers to develop similar techniques to measure radiation doses in other organs in vivo to compare actual and indirect dose measurements. 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; 359:1207–1217 2. 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; 58:130–160 3. Centers for Medicare and Medicaid Services (CMS). Decision memo for screening computed tomography colonography (CTC) for colorectal cancer (CAG-00396N). Baltimore, MD: U.S. Department of Health and Human Services CMS, 2009 4. Calonge N, Petitti DB, DeWitt TG, et al. Screening for colorectal cancer: U.S. Preventative Services Task Force recommendation statement. Ann Intern Med 2008; 149:627–637 5. An S, Lee KH, Kim YH, et al. Screening CT colonography in an asymptomatic average-risk Asian population: a 2-year experience in a single institution. AJR 2008; 191:[web]W100–W106 6. Cohnen M, Vogt C, Beck A, et al. Feasibility of MDCT colonography in ultra-low-dose technique in the detection of colorectal lesions: comparison with high-resolution video colonoscopy. AJR 2004; 183:1355–1359 7. Martin CJ. The application of effective dose to medical exposures. Radiat Prot Dosimetry 2008; 128:1–4 8. Brenner DJ, Georgsson MA. Mass screening with CT colonography: should the radiation exposure be of concern? Gastroenterology 2005; 129:328–337 9. Liedenbaum MH, Venema HW, Stoker J. Radiation dose in CT colonography: trends in time and differences between daily practice and screening protocols. Eur Radiol 2008; 18:2222–2230 10. Bou Serhal C, Jacobs R, Gijbels F, et al. Absorbed doses from spiral CT and conventional spiral tomography: a phantom vs. cadaver study. Clin Oral Implants Res 2001; 12:473–478 11. Zhang D, Savandi AS, Demarco JJ, et al. Variability of surface and center position radiation dose in MDCT: Monte Carlo simulations using CTDI and anthropomorphic phantoms. Med Phys 2009; 36:1025–1038 12. Turner AC, Zhang D, Khotonabadi M, et al. The feasibility of patient size-corrected, scanner-independent organ dose estimates for abdominal CT exams. Med Phys 2011; 38:820–829 13. American Association of Physicists in Medicine (AAPM). Report 204: Correction factors for patient size in CT dose estimation. New York, NY: AAPM, 2011 14. American Association of Physicists in Medicine (AAPM). Report 96: The measurement, reporting, and management of radiation dose in CT. New York, NY: AAPM, 2008 15. DeMarco JJ, Cagnon CH, Cody DD, et al. Estimating radiation doses from multidetector CT using Monte Carlo simulations: effects of different sized voxelized patient models on magnitudes of organ and effective dose. Phys Med Biol 2007; 52:2583–2597 16. ImPACT website. CTDosimetry version 1.0.3: Im-
AJR:202, April 2014
Downloaded from www.ajronline.org by East Carolina University on 06/09/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved
CT Dosimetry During CT Colonography PACT—image performance assessment of CT scanners. www.impactscan.org/ctdosimetry.htm. Published August 24, 2010. Accessed December 2010. 17. American Association of Physicists in Medicine (AAPM). Report 111: Comprehensive methodology for the evaluation of radiation dose in x-ray computed tomography. New York, NY: AAPM, 2010 18. Kirby TH, Hanson WF, Johnston DA. Uncertainty analysis of absorbed dose calculations from thermoluminescent dosimeters. Med Phys 1992; 19:1427–1433 19. Cagnon CH, Cody DD, McNitt-Gray MF, Seibert JA, Judy PF, Aberle DR. Description and imple-
mentation of a quality control program in an imaging-based clinical trial. Acad Radiol 2006; 13:1431–1441 20. McFarland EG, Fletcher JG, Pickhardt P, et al. ACR Colon Cancer Committee white paper: status of CT colonography 2009. J Am Coll Radiol 2009; 6:756–772 21. Berrington de González A, Kim KP, Knudsen AB, et al. Radiation-related cancer risks from CT colonography screening: a risk-benefit analysis. AJR 2011; 196:816–823 22. American College of Radiology (ACRIN) website. Protocol 6664: the National CT colonography trial.
www.acrin.org/TabID/151/Default.aspx. Accessed October 24, 2012 23. Committee to Assess Health Risks From Exposure to Low Levels of Ionizing Radiation, Board on Radiation Effects Research, Division of Earth and Life Studies, National Research Council of the National Academies. Health risks from exposure to low levels of ionizing radiation: BEIR VII phase 2. Washington, DC: National Academies Press, 2006 24. O’Daniel JC, Stevens DM, Cody DD. Reducing radiation exposure from survey CT scans. AJR 2005; 185:509–515
F O R YO U R I N F O R M AT I O N
For more information on Journal Clubs, see “Evidence-Based Radiology: A Primer in Reading Scientific Articles” in the July 2010 AJR at www.ajronline.org/cgi/content/full/195/1/W1.
AJR:202, April 2014 709
Mueller et al. APPENDIX 1: AJR JOURNAL CLUB
Downloaded from www.ajronline.org by East Carolina University on 06/09/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved
Study Guide In Vivo CT Dosimetry During CT Colonography Margaret Mulligan1, Joseph J. Budovec1, Alan Mautz 2 1Medical College of Wisconsin, Milwaukee, WI 2 The Aroostook Medical Center, Presque Isle, ME [email protected], [email protected], [email protected] Introduction 1. What is the research question being asked with this study? Is this study timely and relevant? Does the study attempt to answer a clinically important question? 2. What is unique about this topic? Methods 3. How were patients selected for inclusion in this study? What were the exclusion criteria? 4. Does the population excluded from the study impact the generalizability of results? 5. What are the limitations of this study? Were the limitations adequately addressed? 6. What types of bias might be introduced due to study design and patient selection? Results 7. Was the research question answered? What is the null hypothesis in this study? Physics 8. What is the size-specific dose estimate? How is this parameter related to the volume CT dose index (CTDIvol)? Which parameter may be a better estimate of the patient’s actual radiation dose? 9. What is a thermoluminescent dosimeter? How does it work? Discussion 10. What impact does this study have on clinical practice? 11. Although the authors acknowledge the study as a pilot study, does the study’s population size impact standards of practice? 12. How does the method used in this study for measuring rectal radiation doses in vivo during CT colonography vary from other accepted methods? 13. How could the study design be altered to increase the power of the results? 14. Do you agree with the authors’ assertion that the results of this study should impact the U. S. Preventive Services Task Force recommendation regarding CT colonography? Why or why not? Background Reading 1. Berrington de Gonzalez A, Kim KP, Knudsen AB, et al. Radiation-related cancer risks from CT colonography screening: a risk-benefit analysis. AJR 2011; 196:816– 823 2. Martin CJ. The application of effective dose to medical exposures. Radiat Prot Dosimetry 2008; 128:1–4
*Please note that the authors of the Study Guide are distinct from those of the companion article.
710
AJR:202, April 2014
