Original Research
Medical Physics and Informatics
April 2010

Estimating Effective Dose for CT Using Dose–Length Product Compared With Using Organ Doses: Consequences of Adopting International Commission on Radiological Protection Publication 103 or Dual-Energy Scanning

Abstract

OBJECTIVE. The objective of our study was to compare dose–length product (DLP)–based estimates of effective dose with organ dose–based calculations using tissue-weighting factors from publication 103 of the International Commission on Radiological Protection (ICRP) or dual-energy CT protocols.
MATERIALS AND METHODS. Using scanner- and energy-dependent organ dose coefficients, we calculated effective doses for CT examinations of the head, chest, coronary arteries, liver, and abdomen and pelvis using routine clinical single- or dual-energy protocols and tissue-weighting factors published in 1991 in ICRP publication 60 and in 2007 in ICRP publication 103. Effective doses were also generated from the respective DLPs using published conversion coefficients that depend only on body region. For each examination type, the same volume CT dose index was used for single- and dual-energy scans.
RESULTS. Effective doses calculated for CT examinations using organ dose estimates and ICRP 103 tissue-weighting factors differed relative to ICRP 60 values by –39% (–0.5 mSv, head), 14% (1 mSv, chest), 36% (4 mSv, coronary artery), 4% (0.6 mSv, liver), and –7% (–1 mSv, abdomen and pelvis). DLP-based estimates of effective dose, which were derived using ICRP 60–based conversion coefficients, were less than organ dose–based estimates for ICRP 60 by 4% (head), 23% (chest), 37% (coronary artery), 12% (liver), and 19% (abdomen and pelvis) and for ICRP 103 by –34% (head), 37% (chest), 74% (coronary artery), 16% (liver), and 12% (abdomen and pelvis). All results were energy independent.
CONCLUSION. These differences in estimates of effective dose suggest the need to reassess DLP to E conversion coefficients when adopting ICRP 103, particularly for scans over the breast. For the evaluated scanner, DLP to E conversion coefficients were energy independent, but ICRP 60–based conversion coefficients underestimated effective dose relative to organ dose–based calculations.

Introduction

Effective dose (E) is a single parameter meant to reflect the relative risk from exposure to ionizing radiation. It reflects the risk of detrimental biologic effects from a nonuniform, partial-body exposure in terms of a whole-body exposure [1, 2]. The risk coefficients used in calculating effective dose were derived from a cohort that included both sexes and all ages and depended primarily on the excess risk observed in survivors of the Japanese atomic bombings. The values are a broad estimate of risk for an average (thin by today's standards) adult hermaphrodite phantom, which is a fairly unrealistic description of the human body (see Fig. 1A, 1B, 1C, 1D, 1E). Hence, effective dose is not applicable to any single individual. Nonetheless, it is useful for comparing and optimizing imaging procedures that use ionizing radiation, particularly when comparing examinations from different techniques, such as radiography, CT, and nuclear medicine [35].

Background

Methods for Calculating Effective Dose

In this study, two common methods used to estimate effective dose for a CT examination were compared: first, the gold standard method based on organ dose estimates [6, 7] that explicitly uses tissue-weighting coefficients as specified by the International Commission on Radiological Protection (ICRP) [2, 8, 9]; and, second, the computationally more simple method based on the dose–length product (DLP) and a DLP to E conversion coefficient, referred to as “k,” that depends on only the anatomic region examined [1015].
Fig. 1A Scan ranges for studied examinations as shown on mathematic model used for National Radiological Protection Board–Monte Carlo simulations. (Figure of adult hermaphrodite mathematical phantom reproduced with kind permission from Shrimpton PC, Jones DG, Hillier MC, Wall BF, Le Heron JC, Faulkner K; U.K. Health Protection Agency. Report NRPB-R249 (1991): Survey of CT practice in the U.K. Part II. Dosimetric aspects. Chilton, United Kingdom: National Radiological Protection Board) Drawings show scan ranges.
Fig. 1B Scan ranges for studied examinations as shown on mathematic model used for National Radiological Protection Board–Monte Carlo simulations. (Figure of adult hermaphrodite mathematical phantom reproduced with kind permission from Shrimpton PC, Jones DG, Hillier MC, Wall BF, Le Heron JC, Faulkner K; U.K. Health Protection Agency. Report NRPB-R249 (1991): Survey of CT practice in the U.K. Part II. Dosimetric aspects. Chilton, United Kingdom: National Radiological Protection Board) Drawings show scan ranges.
Fig. 1C Scan ranges for studied examinations as shown on mathematic model used for National Radiological Protection Board–Monte Carlo simulations. (Figure of adult hermaphrodite mathematical phantom reproduced with kind permission from Shrimpton PC, Jones DG, Hillier MC, Wall BF, Le Heron JC, Faulkner K; U.K. Health Protection Agency. Report NRPB-R249 (1991): Survey of CT practice in the U.K. Part II. Dosimetric aspects. Chilton, United Kingdom: National Radiological Protection Board) Drawings show scan ranges.
Fig. 1D Scan ranges for studied examinations as shown on mathematic model used for National Radiological Protection Board–Monte Carlo simulations. (Figure of adult hermaphrodite mathematical phantom reproduced with kind permission from Shrimpton PC, Jones DG, Hillier MC, Wall BF, Le Heron JC, Faulkner K; U.K. Health Protection Agency. Report NRPB-R249 (1991): Survey of CT practice in the U.K. Part II. Dosimetric aspects. Chilton, United Kingdom: National Radiological Protection Board) Drawings show scan ranges.
Fig. 1E Scan ranges for studied examinations as shown on mathematic model used for National Radiological Protection Board–Monte Carlo simulations. (Figure of adult hermaphrodite mathematical phantom reproduced with kind permission from Shrimpton PC, Jones DG, Hillier MC, Wall BF, Le Heron JC, Faulkner K; U.K. Health Protection Agency. Report NRPB-R249 (1991): Survey of CT practice in the U.K. Part II. Dosimetric aspects. Chilton, United Kingdom: National Radiological Protection Board) Drawings show scan ranges.

Calculation of Organ Doses

Monte Carlo simulations—The most complete computational method for estimating organ and tissue doses is based on Monte Carlo simulations [6, 7]. The simulations account for many scanner and technique specifics, including scanner geometry, bow-tie filtration, beam collimation, tube potential, and current as well as the CT dose index (CTDI) [5, 1618] and the scan length for a given CT examination.
The Monte Carlo–based organ dose coefficient data used for this study were published in 1991 by Jones and Shrimpton [6, 19]. (Similar data are available from the Institute for Radiation Protection [7].) Jones and Shrimpton used a simulated hermaphroditic patient (MIRD-5 phantom) [2022] having mathematically modeled organs and tissues (Fig. 1A, 1B, 1C, 1D, 1E). The mathematic phantom was divided from head to mid thigh into 208 axial slabs of 5 mm thickness. Then, accounting for tube voltage and using CT scanner–specific data for geometry and beam shaping, they simulated a CT scan and calculated absorbed doses to all organs of the body for the irradiation of each axial slab. Summing contributions from all slabs exposed during a particular CT examination yielded the total organ doses. These doses were normalized by the CTDI100 measured in air at the gantry isocenter, CTDIair [3, 5, 12]. The resultant data tables were published in 1991 in paper [6] and in 1993 in electronic [19] formats by the National Radiological Protection Board (NRPB) in the United Kingdom. (Starting in 2005, the NRPB became the Radiation Protection Division of the Health Protection Agency, UK.)
Assessment of CT scanners (ImPACT) spreadsheet and scanner match—By 2000, many new scanners were in use. The ImPACT Group (UK National Health Service CT Evaluation Centre, part of the Medical Physics Department at St. George's Hospital, London, UK) developed an Excel (Microsoft) spreadsheet to provide a convenient user interface for determining organ doses using the NRPB Monte Carlo–generated data sets. (The spreadsheet is available free of charge at www.ImPACTscan.org [23]) They further developed a method [24] to map results from the original 23 scanner data sets to new CT scanners through the use of so-called “ImPACT factors”; these factors are based on tube voltage–dependent CTDI measurements using a standard 100-mm pencil dosimeter in air and either a standard head or standard body CTDI dose phantom [12]. The ImPACT Group's spreadsheet can be used to determine organ doses for a wide range of relevant examination parameters [5, 17, 18]: scanner type, tube voltage, tube current, rotation time, head or body scan, and either scan width and increment for sequential (nonhelical) scans or pitch for helical scans.

Calculation of Effective Dose

Method 1: using organ dose estimates and ICRP 26, 60, or 103 tissue-weighting factors— Tissue-weighting factors are meant to represent the relative radiation sensitivity of each type of body tissue as determined from population averages over age and sex and are derived primarily from the atomic bomb survivors cohort [35, 8]. For partial-body irradiation, effective dose is the weighted summation of the absorbed dose to each specified organ and tissue multiplied by the ICRP-defined tissue-weighting factor for that same organ or tissue [8]:
\[ \[E={{\sum}_{z}}\{{{\sum}_{T}}w_{T}{\times}H_{T}\},\] \]
where T is all ICRP-specified tissues and organs, wT is the ICRP-specified tissue-weighting factor, HT is the dose to a particular organ or tissue, the inside summation \( \({{\sum}_{T}}\) \) is over all tissues, and the outside summation \( \({{\sum}_{Z}}\) \) is over all irradiated slabs.
Since 1977, three different sets of tissue-weighting factors (Table 1 and Fig. 2) have been defined in publications by the ICRP: ICRP 26, published in 1977 [2]; ICRP 60, in 1991 [8]; and ICRP 103, in 2007 [9]. These revisions were intended to reflect advances in knowledge about the radiation sensitivity of various organs and tissues. In ICRP 26, the term “effective dose equivalent” was used. In ICRP 60, the name of the summed quantity was changed to “effective dose” in addition to changes to the tissue-weighting factors. (Throughout this article, we use the term “effective dose,” or E, to mean effective dose equivalent when referring to calculations based on ICRP 26 tissue-weighting factors.) Although ICRP 103 assigns different tissue-weighting factors for several primary organs, it retains the name, “effective dose.” In addition, the three ICRP recommendations differ somewhat in calculation methodology. For example, in ICRP 26, organ doses are defined by a single-point dose in the organ of interest, whereas in ICRP 60, the mean organ dose is to be used. With each publication, the trend has been to specify weighting factors for an increasing number of organs and tissues, which decreased the weighting of “remainder tissues.” An example is the brain, which was treated as one of the “remainder” organs in ICRP 26 and ICRP 60. The brain was first listed as a primary organ in ICRP 103. Over time, weighting of specific tissues has also changed. For example, the weighting of the gonads has decreased in each subsequent publication. However, for the breast, weighting was decreased in ICRP 60 but then increased in ICRP 103. As a result of the changes, the estimates of effective dose for the exact same CT examination can differ substantially depending on which ICRP report was used.
TABLE 1: Tissue-Weighting Factors for International Commission on Radiological Protection (ICRP) Publications 26, 60, and 103
Publication
Tissue or OrganICRP 26ICRP 60ICRP 103
Gonads0.250.200.08
Red bone marrow0.120.120.12
Lung0.120.120.12
Colon 0.120.12
Stomach 0.120.12
Breast0.150.050.12
Bladder 0.050.04
Liver 0.050.04
Esophagus 0.050.04
Thyroid0.030.050.04
Skin 0.010.01
Bone surface0.030.010.01
Brain  0.01
Salivary glands  0.01
Remainder
0.30
0.05
0.12
Total
1.00
1.00
1.00
Fig. 2 Bar graph shows tissue-weighting factors specified by International Commission on Radiological Protection (ICRP) publications 26, 60, and 103.

Method 2: Using DLP and k Coefficients From the European Guidelines

DLP is defined as the product of the volume CTDI and the irradicted scan length.
\[ \[DLP=CTDI_{vol}{\times}\mathrm{irradiated\ length},\] \]
where CTDIvol is the volume CTDI [11, 12]. Comparing effective dose values estimated from DLP for a wide range of scanner models with effective dose values derived from NRPB organ dose calculations and ICRP 60 tissue-weighting coefficients, a linear relationship was found [11] when data sets were restricted to the same anatomic region (e.g., head, neck, chest, and abdomen and pelvis). This led the European Commission to present in 2000 [12] a generic method to quickly estimate effective dose from CT examinations, with updates published in 2004 and 2005 [10, 13, 14]. By this widely used shortcut method, effective dose is calculated as follows:
\[ \[E=k{\times}DLP,\] \]
where the k coefficient (Table 2) is specific only to the anatomic region scanned.
TABLE 2: Published DLP to E “k” Conversion Coefficients a
DLP to E “k” Conversion Coefficients [mSv / (mGy × cm)]
Anatomic RegionJessen et al., [11] (1999)EC [12] (2000)EC Appendix B [10] (2004)EC Appendix C [13] (2004) and NRPB-W67 [14] (2005)Phantom (cm)
Head0.00210.00230.00230.002116
Head and neck   0.003116
Neck0.00480.0054 0.005932
Chest0.0140.0170.0180.01432
Abdomen0.0120.0150.0170.01532
Pelvis0.0190.0190.0170.01532
Chest, abdomen, and pelvis



0.015
32
Note—EC = European Commission, NRPB = National Radiological Protection Board.
a
E = k× DLP, where DLP = dose—length product. The phantom size is specified for the volume CT dose index measurements on which DLP is based.
Deviations in estimates of effective dose of ± 15% have been reported using this method relative to the gold standard organ dose–based technique for CT scans obtained at 120 kV [25]. In helical CT, this calculation method is apt to underestimate E when DLP is calculated with only the CTDIvol and the prescribed scan range because the irradiated length typically exceeds the prescribed scan length [26, 27]. Because of the widespread use of this method, most manufacturers of CT scanners now compute and display DLP, taking into account the entire irradiated length rather than the lesser prescribed scan length.
In spite of these sources of variation in the calculation of effective dose, effective dose is widely used by the academic, clinical, and manufacturing communities. Therefore, the purpose of this investigation was to determine how well estimates of E calculated using DLP agree with calculations based on organ dose estimates after adopting the revised tissue-weighting factors of ICRP 103 or when using tube potential values other than 120 kV, such as for dual-energy CT protocols.

Materials and Methods

Evaluated CT Examinations

Clinical protocols (Table 3) were investigated for head, chest, coronary artery, liver, and abdomen and pelvis CT examinations using typical scan ranges scaled to the NRPB mathematic patient model (Fig. 1A, 1B, 1C, 1D, 1E). All single-energy (120-kV) protocols had a counterpart in which two tube energies were used, neither being 120 kV. Organ doses and DLP were calculated using our clinical technique parameters, where the CTDIvol of the single-energy examination matched that of the corresponding dual-energy examination.
TABLE 3: Clinical Scan Parameters for the CT Examinations Studied
Tube Current
Anatomic RegionScan TypeTube Potential (kV)Effective mAsamAs/rotationCollimation (mm)PitchRotation Time (s)
HeadSE120400 2 × 32 × 0.60.71
 DE140150 2 × 32 × 0.60.50.5
  80500 2 × 32 × 0.60.50.5
ChestSE120250 24 × 1.20.50.5
 DE14080 14 × 1.20.70.5
  80340 14 × 1.20.70.5
Coronary arteriesSE120 2402 × 32 × 0.60.220.33
 DE140 902 × 32 × 0.60.220.33
  100 1802 × 32 × 0.60.220.33
LiverSE120450 2 × 32 × 0.60.70.5
 DE140190 14 × 1.20.50.5
  80556 14 × 1.20.50.5
Abdomen and pelvisSE120240 2 × 32 × 0.61.20.5
 DE14090 14 × 1.20.60.5


80
382

14 × 1.2
0.6
0.5
Note—SE = single energy, DE = dual energy.
a
Effective mAs = mAs / pitch.
For helical coronary artery examinations, ECG gating is required [28, 29]. Because data are typically not reconstructed at all parts of the cardiac cycle, patient dose can be reduced by decreasing the x-ray tube current during unused parts of the cycle. For this study, we assumed a heart rate of 70 beats per minute and an image reconstruction window that resulted in a 30% reduction in irradiation.

Evaluated Scanner

The CT protocols investigated were in clinical use on our dual-source CT scanners (SOMATOM Definition DS, Siemens Healthcare) at the time of the study. The scanner's so-called “z-flying focal spot” technique [30] was used for collimations described with the prefix “2 ×”; for example, 2 × 32 × 0.6 mm describes a scanner mode that uses the z-flying focal spot technique to collect two interleaved data sets for a detector having a physical size of 32 rows, each 0.6 mm in length.
Measured CTDI values used for scanner matching, through the use of ImPACT factors, were an average over all available quality control data for three identical scanners at our institution. Data were collected over more than 2 years and included, on average, 14 measurements for each set of scan parameters. Specifically, quality control data for CTDIair, CTDIcenter, and CTDIedge were used for 2 × 32 × 0.6 mm collimation at all energies studied (80, 100, 120, and 140 kV) and for collimations of 14 × 1.2 and 24 × 1.2 mm at 120 kV. The remaining requisite dose data for collimations of 14 × 1.2 and 24 × 1.2 mm at 80, 100, and 140 kV were measured using our standard technique [17, 31] on one scanner.

Effective Dose Calculation Based on Organ Doses

The ImPACT Group's Excel spreadsheet (version 0.99x [23]) was used to calculate organ doses based on the NRPB Monte Carlo data sets [19]. Because data for the studied scanner were not reported by the ImPACT Group, we performed the needed scanner matching according to their method [24] using measurements from our scanners.
Data entered into the dose calculator spreadsheet for each examination (Table 4) included clinical CT scan parameters and CTDIair. Additionally, the weighted CTDI100—defined as nCTDIw = 1/3 × CTDIcenter + 2/3 × CTDIedge [12, 32] and normalized to 100 mAs—and tube current–time product, which is the tube current (in milliamperes) × the gantry rotation time (in seconds), were entered to allow the spreadsheet to calculate the CTDIvol for the described examination. Comparison of the spreadsheet-calculated CTDIvol with the scanner-reported CTDIvol served as a verification of correct data entry. After the organ dose calculations, a modified version of the ImPACT Group's spreadsheet was used, along with the tissue-weighting coefficients and calculation methodology specified in ICRP publications 26, 60, or 103 to yield an estimate of effective dose (or effective dose equivalent): E26, E60, or E103, respectively. For a dual-energy examination, effective dose was the sum of E values for the low- and high-kV scans. Because use of the dose calculator method results in identical organ doses for any given protocol (e.g., kV, scan region, mAs), any differences in the calculated values of effective dose for E26, E60, and E103 resulted from the use of the different sets of tissue-weighting factors and calculation rules from the three ICRP publications.
TABLE 4: Input Parameters for the ImPACT Dose Calculator [ 23 ] Spreadsheet
Anatomic RegionTube Potential (kV)Monte Carlo Set No.aScan Startb (cm)Scan Endb (cm)Scan Lengthb (cm)CTDIair (mGy)nCTDIwc (mGy)
Head1202080.594.013.519.112.5
 1402080.594.013.527.618.5
 80980.594.013.56.43.8
Chest1201542.069.527.519.06.2
 1402342.069.527.527.09.6
 80542.069.527.57.02.0
Coronary arteries1202043.055.512.519.15.4
 1402243.055.512.527.09.6
 100543.055.512.57.02.0
Liver1201919.546.026.518.66.4
 1402319.546.026.527.09.6
 80519.546.026.57.02.0
Abdomen and pelvis120190.546.045.518.66.4
 140230.546.045.527.09.6

80
5
0.5
46.0
45.5
7.0
2.0
a
Resulting from scanner matching, as described in the text.
b
Scan start, scan end, and scan length refer to the mathematic phantom used for National Radiological Protection Board—Monte Carlo calculations as shown in Figure 1A, 1B, 1C, 1D, 1E.
c
nCTDIw is the weighted CTDI100, normalized to 100 mAs, as defined in the text.
Fig. 3A Comparison of effective dose (E) estimates by anatomic region according to recommendations of International Commission on Radiological Protection (ICRP) publications 26, 60, or 103 or calculated using dose–length product (DLP) and k coefficients. SE = single energy, DE = dual energy. Bar graphs show data for head (A), chest (B), coronary CT angiogram (C), liver (D), and abdomen and pelvis (E) examinations.
Fig. 3B Comparison of effective dose (E) estimates by anatomic region according to recommendations of International Commission on Radiological Protection (ICRP) publications 26, 60, or 103 or calculated using dose–length product (DLP) and k coefficients. SE = single energy, DE = dual energy. Bar graphs show data for head (A), chest (B), coronary CT angiogram (C), liver (D), and abdomen and pelvis (E) examinations.
Fig. 3C Comparison of effective dose (E) estimates by anatomic region according to recommendations of International Commission on Radiological Protection (ICRP) publications 26, 60, or 103 or calculated using dose–length product (DLP) and k coefficients. SE = single energy, DE = dual energy. Bar graphs show data for head (A), chest (B), coronary CT angiogram (C), liver (D), and abdomen and pelvis (E) examinations.
Fig. 3D Comparison of effective dose (E) estimates by anatomic region according to recommendations of International Commission on Radiological Protection (ICRP) publications 26, 60, or 103 or calculated using dose–length product (DLP) and k coefficients. SE = single energy, DE = dual energy. Bar graphs show data for head (A), chest (B), coronary CT angiogram (C), liver (D), and abdomen and pelvis (E) examinations.
Fig. 3E Comparison of effective dose (E) estimates by anatomic region according to recommendations of International Commission on Radiological Protection (ICRP) publications 26, 60, or 103 or calculated using dose–length product (DLP) and k coefficients. SE = single energy, DE = dual energy. Bar graphs show data for head (A), chest (B), coronary CT angiogram (C), liver (D), and abdomen and pelvis (E) examinations.

Effective Dose Calculated From DLP

DLP for each CT examination was also calculated using the ImPACT spreadsheet. From this result, we determined an estimate for EDLP as the product of DLP and the body region–appropriate DLP to E conversion coefficient, k (Table 2). The values used for k were the most recently reported values, published in 2004 [13]. They were based on ICRP 60 and are 0.0021 mSv/mGy × cm (head), 0.014 mSv/mGy × cm (chest), and 0.015 mSv/mGy × cm (abdomen and pelvis). The estimate of effective dose for a dual-energy examination was the sum of E values for the two energies.

Energy Dependence of Effective Dose

To assess the influence of CT tube potential (80, 100, 120, or 140 kV) on estimations of effective dose, the values for EDLP, E60, and E103 at each tube potential were normalized to CTDIvol to obtain EDLP / CTDIvol, E60 / CTDIvol, and E103 / CTDIvol, respectively. For each examination type, we computed the coefficient of variation as a function of energy for the normalized effective dose values (E60 / CTDIvol or E103 / CTDIvol), where [coefficient of variation = SD / mean]. These coefficients of variation were used to quantitate any energy dependence of the organ dose calculations.

Comparative Assessments

Changes from E26 to E60 were normalized to E26 and changes from E60 to E103 were normalized to E60 to show the step-wise changes in risk estimates due to changes in the definition of effective dose. Having no knowledge of the “true” value of E, because it is a nonphysical mathematic construct that cannot be measured, all other normalizations were performed relative to E60, because ICRP 60's definitions were used to determine the current k coefficients.

Results

Comparison of Effective Dose Based on Organ Doses Between ICRP 26, 60, or 103

The relative values of effective dose calculated using ICRP 26, 60, or 103 tissue-weighting factors and organ dose estimates based on Monte Carlo simulations depended on the body region examined (Table 5 and Fig. 3A, 3B, 3C, 3D, 3E). The range—that is, the absolute value (maximum – minimum)—of the three E values (E26, E60, and E103) was ∼ 2 mSv (head), 2 mSv (chest), 6–7 mSv (coronary arteries), 3 mSv (liver), and 1 mSv (abdomen and pelvis) corresponding, respectively, to 155% (head), 29% (chest), 53% (coronary arteries), 25% (liver), and 7% (abdomen and pelvis) of the E60 values. For four of the five investigated examinations, E26 was the largest. The exception was the abdomen and pelvis examination, where there was a relatively small (7%) difference between effective dose estimates. E60 was the smallest of the three estimates of effective dose in three of the five investigated examinations.
TABLE 5: Organ Dose–Based Effective Dose (E) Estimates and Comparisons for International Commission on Radiological Protection (ICRP) Publications 26, 60, and 103
Anatomic RegionScan TypeTube Potential (kV)E26(mSv)E60(mSv)E103(mSv)Range of Ea(mSv)Range of Ea / E60(%)E60—E26 (mSv)(E60—E26) / E26 (%)E103—E60 (mSv)(E103—E60)/E60 (%)
HeadSE1203.01.40.92.2155—1.6—54—0.5—39
 DE1401.70.80.51.2155—0.9—54—0.3—39
  801.30.60.40.9153—0.7—53—0.2—39
  DE total3.01.40.92.1154—1.6—54—0.5—39
ChestSE1209.37.28.22.129—2.1—231.014
 DE1404.43.53.91.028—1.0—220.514
  804.33.23.81.136—1.1—260.618
  DE total8.86.67.72.132—2.1—241.016
Coronary arteriesSE12018.012.016.06.253—6.2—354.236
 DE14010.06.58.93.656—3.6—362.538
  1008.35.17.33.364—3.3—392.243
  DE total18.012.016.06.959—6.9—374.640
LiverSE12016.613.313.83.325—3.3—200.64
 DE14010.78.68.92.225—2.2—200.44
  805.64.54.61.125—1.1—200.24
  DE total16.313.013.63.325—3.3—200.64
Abdomen and pelvisSE12013.313.712.71.070.43—1.0—7
 DE1407.27.56.90.570.23—0.5—7
  805.35.55.10.470.23—0.4—7


DE total
12.5
12.9
12.0
0.9
7
0.4
3
—0.9
—7
Note—SE = single energy, DE = dual energy.
a
Range of E = maximum E minus minimum E for the set E26, E60, E103.
In changing from ICRP 60 to ICRP 103 tissue-weighting factors, effective dose estimates were least affected for liver (0.6 mSv, 4% increase) and abdomen and pelvis (–1 mSv, 7% decrease) examinations. For CT of the head, effective dose decreased by 0.5 mSv, or 39%, because the weighting of the brain was reduced in ICRP 103 to 0.01, compared with that for ICRP 60, where half of the remainder weighting factor (0.025) was applied. The largest changes were an increase of 14% for chest (1-mSv increase) and 36% for coronary artery (4-mSv increase) examinations due to the increase in the breast tissue-weighting factors from 0.05 to 0.12.

Comparison of Effective Dose Based on Organ Doses or DLP

Results were compared for effective dose calculated from organ dose estimates and tissue-weighting factors versus DLP and k coefficients (Table 6). EDLP underestimated E for all investigated examinations relative to the organ-based calculation of E60. The percentage differences [100% × (EDLPE60) / E60] for the 120-kV single-energy examination were –4% (head), –23% (chest), –37% (coronary arteries), –12% (liver), and –19% (abdomen and pelvis).
TABLE 6: Dose–Length Product (DLP)–Based Effective Dose (E) Estimates and Comparisons for International Commission on Radiological Protection (ICRP) Publications 26, 60, and 103
Anatomic RegionTube Potential (kV)CDTIvol (mGy)DLP (mGy × cm)k valuea (mSv) / (mGy × cm)EDLP= k× DLP(mSv)EDLP — E60 (mSv)(EDLPE60) / E60 (%)EDLP — E103 (mSv)(EDLP — E103) / E60 (%)(EDLP — E103) / E103 (%)E103 = E103 / DLP (mSv) / (mGy × cm)
Head12046.96330.00211.3—0.01—40.534560.0013
 14027.83750.00210.8—0.00—20.337610.0013
 8019.12580.00210.5—0.01—80.231510.0014
 DE total46.96330.00211.3—0.01—40.534560.0013
 DE — SE0.00        
Chest12014.53990.0145.6—1.6—23—2.6—37—320.021
 1407.72110.0143.0—0.5—14—1.0—28—250.019
 806.81870.0142.6—0.6—18—1.1—36—300.020
 DE total14.53980.0145.6—1.1—16—2.1—32—270.019
 DE — SE0.0—1        
Coronary arteries12041.55190.0147.3—4.3—37—8.6—74—540.031
 14023.32910.0144.1—2.4—37—4.8—75—540.031
 10018.22270.0143.2—1.9—38—4.1—81—560.032
 DE total41.55180.0147.3—4.3—37—8.9—77—550.031
 DE — SE0.0—2        
Liver12029.37770.01512.0—1.6—12—2.2—16—160.018
 14018.34840.0157.3—1.3—15—1.7—20—190.018
 8011.02950.0154.40.0—1—0.2—4—40.016
 DE total29.37790.01512.0—1.3—10—1.9—14—140.017
 DE — SE0.02        
Abdomen and pelvis12016.27370.01511.0—2.6—19—1.7—12—130.017
 1408.63940.0155.9—1.6—21—1.0—14—150.018
 807.63440.0155.2—0.3—50.1120.015
 DE total16.27380.01511.0—1.8—14—0.9—7—80.016

DE — SE
0.0
1








Note—SE = single energy, DE = dual energy.
a
Based on ICRP 60 [8] as reported by Shrimpton [13, 14].
Likewise, EDLP underestimated E compared with the organ-based calculations of E103 except for the head examination, where it was nearly equivalent (absolute increase of 0.5 mSv, relative increase of 34%). The largest differences (EDLPE103) were found for the chest (–2.6 mSv, –37%) and coronary artery (–8.6 mSv, –74%) examinations. The E103 per DLP ratio—that is, k values for ICRP 103—were computed for each examination (Table 6, see Discussion).

Comparison of Effective Dose for Single- or Dual-Energy Protocols

Organ dose–based estimates of effective dose for single-energy and dual-energy examinations were virtually the same with an observed difference of no more than 0.8 mSv, or 8% (Table 5), thus confirming that the experimental design goal was achieved. For dual-energy examinations, effective dose of the high-energy and low-energy components of the examination were split almost equally for head and chest regions, differed by about 20% for coronary artery and abdomen and pelvis regions, and differed nearly 50% for the liver examination.

Energy Dependance of Effective Dose

The k values for ICRP 103 (Table 6, see Discussion) computed for each examination were energy independent. Normalized values for EDLP per CTDIvol (Table 7) were 0.028 mSv/mGy (head), 0.39 mSv/mGy (chest), 0.18 mSv/mGy (coronary arteries), 0.40 mSv/mGy (liver), and 0.68 mSv/mGy (abdomen and pelvis). These values must be energy independent because each of the input parameters used to compute EDLP (CTDIvol, DLP, and the DLP to E (k) conversion coefficients) were fixed, independent of energy (tube potential). For each examination type, the coefficients of variation for E60 and E103 values, which were based on energy-dependent organ-dose calculations, were within 1% of each other and were approximately 3–4% (head), 5–6% (chest), 1–3% (coronary arteries), 7–8% (liver), and 10% (abdomen and pelvis). Thus, for the same total CTDIvol, the tube potential has a minimal effect on estimates of effective dose.
TABLE 7: Energy Dependence of EDLP, E60, and E103
Normalized to CTDIvolCoefficient of Variationb
Anatomic RegionTube Potential (kV)CDTIvol(mGy)EDLP/ CTDIvol (mSv/mGy)E60/ CTDIvol (mSv/mGy)E103/ CTDIvol (mSv/mGy)E60/ CTDIvol(%)E103/ CTDIvol(%)
Head12046.90.0280.0300.01843
 14027.80.0280.0290.018  
 8019.00.0280.0310.019  
Chest12014.50.390.500.5756
 1407.70.380.450.51  
 806.80.390.470.55  
Coronary arteries12041.50.180.280.3813
 14023.30.170.280.38  
 10018.20.170.280.30  
Liver12029.30.400.450.4778
 14018.30.400.470.49  
 8011.00.400.400.42  
Abdomen and pelvis12016.20.680.850.791010
 1408.60.690.870.81  

80
7.6
0.68
0.72
0.67


ak based on ICRP 60 [8].
b
Coefficient of variation = 100% × (SD) / mean. The coefficient of variation for a given examination indicates any energy dependence.

Discussion

The results of this study reinforce the fact that effective dose is a derived parameter. It is always computed through multiple steps and approximations. Depending on which set of tissue-weighting coefficients are used, E values may vary substantially. Effective dose is meant as an estimate of relative biologic risk [9] and is not a physical parameter that can be measured.
EDLP, although based on ICRP 60, underestimates E60 (based on organ doses) for the studied scanner. This finding is not surprising because k is based on data averaged over many scanner makes and models and is therefore not specific to this scanner. However, generating many specialized k values is not consistent with the definition of effective dose and its intended use.
Adopting ICRP 103 but retaining the current k coefficients would further increase the underestimations of EDLP for chest, coronary artery, and abdomen examinations compared with E values derived from organ doses. Hence, to use the DLP-based method of estimating E for ICRP 103 weighting factors requires a comprehensive analysis of a broad variety of scanners to estimate the mean values for DLP to E conversion coefficients. To account for the increased sensitivity assigned to breast tissue by ICRP 103 over ICRP 60, it may be useful to introduce distinct coefficients for cardiac examinations.
Martin [4] has reported the inherent relative uncertainties in estimating effective dose (using organ doses) to a reference patient to be about ± 40%. He further reminds readers that because E has been defined by the ICRP [9] as a single parameter to reflect overall risk averaged over all ages and both sexes for a reference patient, neither the Monte Carlo–based organ dose coefficients nor the DLP-based k values [13] should be used to calculate E estimates for individual patients [33].
In general, the accuracy of E values derived from Monte Carlo calculations cannot be substantially improved by measuring organ doses using an anthropomorphic phantom and radiation measurement devices because experimental uncertainties will exist, including variations in the physical anthropomorphic phantom used. Further, Monte Carlo methods to estimate absorbed dose have been shown to be extremely accurate (e.g., they are used in radiation therapy treatment planning). Martin's [4] analysis of ± 40% uncertainties in E estimates applies in either case, with Monte Carlo methods being more standardized and reproducible. The results of this study show that estimating effective dose from DLP works about equally well for dual-energy CT as it does for single-energy CT.
A limitation of this work is that only one scanner was studied. Additionally, a circa-1990 scanner with similar dose characteristics was used as a surrogate for the studied scanner. However, until Monte Carlo organ dose coefficient values for newer scanners are available, scanner matching to an older scanner model is the only available option. Finally, use of the simplified and relatively small anthropomorphic model makes these estimates of effective dose applicable to scans of a small adult. With today's continued increase in the number of overweight and obese patients, the calculated values of E should be used with caution. They apply to the CT examination generally and cannot be used for any one individual, particularly persons who have a marked difference in body habitus compared with the MIRD phantom.
In conclusion, the use of ICRP 103 tissue-weighting factors in place of ICRP 60 factors decreased organ dose–based estimates of effective dose for CT examinations of the abdomen and pelvis by 7% and head by 39%, but increased estimates of effective dose for scans of the liver by 4%, chest by 14%, and coronary arteries by 36%. The absolute value of the largest change was 4 mSv for the coronary artery examination. These changes primarily reflect the increased tissue-weighting factor for breast tissue together with decreased factors for the gonads and brain.
For the evaluated CT scanner and examinations, EDLP, although based on ICRP 60, underestimated E relative to E60 by 4–37% and relative to E103 by up to 74%. These results were essentially independent of tube potential, suggesting that estimates of E based on DLP work equally well for single-energy and dual-energy CT examinations.

Footnotes

Address correspondence to C. H. McCollough ([email protected]).
CME
This article is available for CME credit.
See www.arrs.org for more information.

References

1.
Jacobi W. The concept of effective dose: a proposal for the combination of organ doses. Radiat Environ Biophys 1975; 12:101 –109
2.
International Commission on Radiological Protection. Recommendations of the International Commission on Radiological Protection. Oxford, UK: Pergamon Press, 1977:ICRP publication no. 26
3.
McCollough CH, Schueler BA. Calculation of effective dose. Med Phys 2000; 27:828 –837
4.
Martin CJ. Effective dose: how should it be applied to medical exposures? Br J Radiol 2007; 80:639–647
5.
American Association of Physicists in Medicine (AAPM). The measurement, reporting and management of radiation dose in CT: report of AAPM Task Group 23 of the Diagnostic Imaging Council CT Committee. College Park, MD: AAPM, 2007: report no. 96
6.
Jones DG, Shrimpton PC. Survey of CT practice in the UK. Part 3. Normalised organ doses calculated using Monte Carlo techniques. Chilton, UK: National Radiological Protection Board,1991 :report no. NRPB-R250
7.
Zankl M, Panzer W, Drexler G. The calculation of dose from external photon exposures using reference human phantoms and Monte Carlo methods. Part VI. Organ doses from computed tomographic examinations. Neuherberg, Germany: GSF - Forschungszentrum fur Umwelt und Gesundtheit, Institut fur Strahlenschutz,1991
8.
International Commission on Radiological Protection. 1990 recommendations of the International Commission on Radiological Protection. Ann ICRP 1991; 21 :publication no. 60
9.
International Commission on Radiological Protection. 2007 recommendations of the International Commission on Radiological Protection. Ann ICRP 2007; 37 :publication no. 103
10.
Bongartz G, Golding SJ, Jurik AG, et al. European guidelines for multislice computed tomography. Brussels, Belgium: European Commission, 2004:FIGM-CT2000-20078-CT-TIP
11.
Jessen KA, Shrimpton PC, Geleijns J, Panzer W, Tosi G. Dosimetry for optimisation of patient protection in computed tomography. Appl Radiat Isot 1999; 50:165 –172
12.
Jessen KA, Panzer W, Shrimpton PC, et al. European Guidelines on quality criteria for computed tomography. Brussels, Belgium: European Commission, 2000:EUR 16262
13.
Shrimpton P. Assessment of patient dose in CT. In: EUR. European guidelines for multislice computed tomography funded by the European Commission 2004: contract number FIGMCT2000-20078-CT-TIP. Luxembourg, Luxembourg: European Commission, 2004:Appendix C
14.
Shrimpton PC, Hillier MC, Lewis MA, Dunn M. Doses from computed tomography (CT) examinations in the UK: 2003 review. Chilton, UK: National Radiological Protection Board, 2005:report NRPB-W67
15.
Shrimpton P. Reference doses for computed tomography. Radiological Protection Bulletin 1997; 193:16–19
16.
Shope TB, Gagne RM, Johnson GC. A method for describing the doses delivered by transmission x-ray computed tomography. Med Phys 1981; 8:488 –495
17.
McNitt-Gray MF. AAPM/RSNA physics tutorial for residents: topics in CT—radiation dose in CT. RadioGraphics 2002; 22:1541 –1553
18.
McCollough C. CT dose: how to measure, how to reduce. Health Phys 2008; 95:508–517
19.
Jones DG, Shrimpton PC. Normalised organ doses for x-ray computed tomography calculated using Monte Carlo techniques. Chilton, UK: National Radiological Protection Board, 1993:report NRPB-SR250
20.
Cristy M. Mathematical phantoms representing children of various ages for use in estimates of internal dose. Oak Ridge, TN: Oak Ridge National Laboratory, 1980
21.
Jones DG, Wall BF. Organ doses from medical x-ray examinations calculated using Monte Carlo Techniques. Chilton, UK: National Radiological Protection Board, 1985:report NRPB-R186
22.
Snyder WS, Fisher HL Jr, Ford MR, Warner GG. Estimates of absorbed fractions for monoenergetic photon sources uniformly distributed in various organs of a heterogeneous phantom. J Nucl Med 1969 ; [suppl 3]:7 –52
23.
ImPACT Group Website. www.impactscan.org. Accessed January 20, 2010
24.
Lewis M, Edyvean S, Sassi S, Kiremidjian H, Keat N, Britten A. Estimating patient dose on current CT scanners: results of the ImPACT CT dose survey. RAD Magazine 2000; 26:17–18
25.
McCollough CH. Patient dose in cardiac computed tomography. Herz 2003; 28:1 –6
26.
Crawford CR, King KF. Computed tomography scanning with simultaneous patient translation. Med Phys 1990; 17:967 –982
27.
Nicholson R, Fetherston S. Primary radiation outside the imaged volume of a multislice helical CT scan. Br J Radiol 2002; 75:518 –522
28.
Jakobs TF, Becker CR, Ohnesorge B, et al. Multislice helical CT of the heart with retrospective ECG gating: reduction of radiation exposure by ECG-controlled tube current modulation. Eur Radiol 2002; 12:1081 –1086
29.
McCollough CH, Bruesewitz MR, Kofler JM Jr. CT dose reduction and dose management tools: overview of available options. RadioGraphics 2006; 26:503–512
30.
Flohr T, Stierstorfer K, Raupach R, Ulzheimer S, Bruder H. Performance evaluation of a 64-slice CT system with z-flying focal spot. Rofo 2004; 176:1803 –1810
31.
International Electrotechnical Commission. Medical electrical equipment. Part 2-44. Particular requirements for the safety of x-ray equipment for computed tomography, ed. 2.1 Geneva, Switzerland: International Electrotechnical Commission (IEC) Central Office,2002 :IEC publication no. 60601-2-44
32.
Leitz W, Axelsson B, Szendro G. Computed tomography dose assessment: a practical approach. Radiat Prot Dosimetry 1995; 57:377 –380
33.
Shrimpton PC, Wall BF. Effective dose and dose–length product in CT. Radiology 2009; 250:604–605

Information & Authors

Information

Published In

American Journal of Roentgenology
Pages: 881 - 889
PubMed: 20308486

History

Submitted: August 10, 2009
Accepted: October 16, 2009

Keywords

  1. CT
  2. DLP
  3. dose–length product
  4. dual-energy CT
  5. effective dose
  6. effective dose equivalent
  7. ICRP publication 26
  8. ICRP publication 60
  9. ICRP publication 103
  10. organ dose
  11. tissue-weighting factors

Authors

Affiliations

Jodie A. Christner
All authors: Mayo Clinic Rochester, 200 First St., SW, Rochester, MN 55905.
James M. Kofler
All authors: Mayo Clinic Rochester, 200 First St., SW, Rochester, MN 55905.
Cynthia H. McCollough
All authors: Mayo Clinic Rochester, 200 First St., SW, Rochester, MN 55905.

Metrics & Citations

Metrics

Citations

Export Citations

To download the citation to this article, select your reference manager software.

Articles citing this article

View Options

View options

PDF

View PDF

PDF Download

Download PDF

Media

Figures

Other

Tables

Share

Share

Copy the content Link

Share on social media