BJR Received: 16 December 2014

© 2015 The Authors. Published by the British Institute of Radiology Revised: 2 February 2015

Accepted: 4 February 2015

doi: 10.1259/bjr.20140852

Cite this article as: Harbron RW, Pearce MS, Salotti JA, McHugh K, McLaren C, Abernethy L, et al. Radiation doses from fluoroscopically guided cardiac catheterization procedures in children and young adults in the United Kingdom: a multicentre study. Br J Radiol 2015;88:20140852.

FULL PAPER

Radiation doses from fluoroscopically guided cardiac catheterization procedures in children and young adults in the United Kingdom: a multicentre study 1

R W HARBRON, MSc, 1M S PEARCE, PhD, 1J A SALOTTI, PhD, 2K MCHUGH, FRCR, 2C MCLAREN, MSc, 3L ABERNETHY, FRCR, S REED, MSc, 5J O’SULLIVAN, MD and 6C-L CHAPPLE, PhD

4 1

The Institute of Health and Society, Newcastle University, Sir James Spence Institute, Royal Victoria Infirmary, Newcastle upon Tyne, UK Radiology Department, Great Ormond Street Hospital for Children NHS Trust, London, UK 3 Radiology Department, Alder Hey Children’s NHS Foundation Trust, Liverpool, Merseyside, UK 4 Radiology Department, University Hospitals of Leicester NHS Trust, Leicester, UK 5 Department of Paediatric Cardiology, Freeman Hospital Newcastle upon Tyne Hospitals NHS Foundation Trust, Newcastle upon Tyne, UK 6 Regional Medical Physics department, Freeman Hospital, Newcastle-upon-Tyne hospitals NHS trust, Newcastle upon Tyne, UK 2

Address correspondence to: Mr Richard W. Harbron E-mail: [email protected]

Objective: To gather data on radiation doses from fluoroscopically guided cardiac catheterization procedures in patients aged under 22 years at multiple centres and over a prolonged period in the UK. To evaluate and explain variation in doses. To estimate patient-specific organ doses and allow for possible future epidemiological analysis of associated cancer risks. Methods: Patient-specific data including kerma area product and screening times from 10,257 procedures carried out on 7726 patients at 3 UK hospitals from 1994 until 2013 were collected. Organ doses were estimated from these data using a dedicated dosimetry system based on Monte Carlo computer simulations. Results: Radiation doses from these procedures have fallen significantly over the past two decades. The organs receiving the highest doses per procedure were

the lungs (median across whole cohort, 20.5 mSv), heart (19.7 mSv) and breasts (13.1 mSv). Median cumulative doses, taking into account multiple procedures, were 23.2, 22.2 and 16.7 mSv for these organs, respectively. Bone marrow doses were relatively low (median per procedure, 3.2 mSv; cumulative, 3.6 mSv). Conclusion: Most modern cardiac catheterizations in children are moderately low-dose procedures. Technological advances appear to be the single most important factor in the fall in doses. Patients undergoing heart transplants undergo the most procedures. An epidemiological assessment of cancer risks following these procedures may be possible, especially using older data when doses were higher. Advances in knowledge: This is the first large-scale, patient-specific assessment of organ doses from these procedures in a young population.

Fluoroscopically guided cardiac catheterizations are an essential technique for the diagnosis and treatment of congenital and acquired heart conditions. The potential for relatively high radiation doses raises concerns of the long-term risk of developing cancer.1–3 The radiation doses and associated risks from cardiac catheterizations in children and young people have recently received considerable attention,4–23 although relatively fewer data on doses in the UK have been published. 7,17,22,23 Most previous assessments have been based on kerma area product (PKA, also known as dose–area product), defined as collision air kerma integrated over beam area.24 PKA is a dose indicator rather than a measure of actual patient dose. The relationship between PKA and mean

absorbed dose to a given organ can easily vary by a factor of over a hundred depending on patient size, projection angle and beam quality. PKA should therefore not be seen as an absolute measure of dose, and uncorrected comparison of PKA figures acquired using different equipment with different beam quality characteristics and different patient sizes requires considerable caution. Relatively few studies have estimated organ doses from cardiac catheterizations in young people.19,20,25 This is unfortunate as organ doses have much more relevance to assessment of the long-term risks from radiation exposure, both by direct epidemiological methods and through the application of risk models derived from other exposures.26

BJR

RW Harbron et al

The aims of the study were to (1) gather PKA data from multiple hospitals in the UK over a prolonged time period for cardiac fluoroscopy procedures in patients aged 0–22 years, (2) use these data to provide patient-specific organ dose estimates using a dosimetry system based on Monte Carlo computer simulations, and (3) characterize and explain variation in doses between hospitals and eras. The decision to focus on the under 22 years age group was largely based on the long-term goal of an epidemiological assessment of the cancer risks following these procedures. Statements concerning the higher radiosensitivity of children than that of the general population form part of the standard preamble of virtually all articles investigating paediatric radiation doses. While there is certainly some evidence to support such statements, particularly for leukaemia and cancers of the thyroid and breast,27,28 for other sites such as the lungs, oesophagus and stomach (all of particular relevance to cardiac imaging), evidence suggests that adults may be equally if not more at risk.29,30 From an epidemiological perspective, however, young subjects are more desirable owing to the relative lack of confounding variables in cancer development. To even consider an epidemiological analysis, however, information on typical organ doses and the distribution of these doses (i.e. how many patients receive a particular dose range) is required. METHODS AND MATERIALS Confidentiality Advisory Group approval for use of patient identifiable data was obtained (ECC 7-04(j)/2010), along with a favourable ethical opinion from the National Research Ethics Service Committee North East—Newcastle and North Tyneside 2 Ethics Committee (10/H0907/47). In addition, local research department approval was obtained from all hospitals supplying data. With agreement from these centres, each hospital was assigned a random identifier number that will be used from now on. Data were available between 1994 and 2010 for Hospital 1 (after 2010, the data were stored in a different manner that was difficult to access) and between 2004 and 2013 for Hospitals 2 and 3. Data from each hospital were split into a number of consecutive eras, between which equipment was replaced. Where possible, data from different eras are presented separately to allow comparison between equipment types. Details of fluoroscopic equipment used are summarized in Table 1. All equipment was subject to annual quality assurance (QA) testing by local medical physics services. Dose measuring devices were confirmed to have been calibrated to UK national standards at the National Physical Laboratory (Teddington, London) at least annually, usually via a secondary standard. All data analysis was carried out using MATLAB® v. 2013a (MathWorks®, Natick, MA) on anonymized data. Data from 10,257 procedures carried out on 7726 patients were collected. Of these, the vast majority (98%) were carried out on children aged 18 years or under. Collected examination data included patient age, mass (a further 313 procedures had no mass record), procedure type, fluoroscopic screening time and PKA. Hospitals 2 and 3 recorded PKA as a single combined figure for both bi-plane tube outputs. Hospital 1 recorded separate frontal and lateral figures, which we combined into a single figure for analysis of PKA. Some examinations had patient mass but not age, or vice versa. The units of measurement were

2 of 14

birpublications.org/bjr

confirmed with local staff or QA reports. All PKA data were normalized to the recommended unit24 of Gy cm2, assumed equal to 100 cGy cm2 or 100 mGy m2. Data on screening time and skin dose estimates (presented as air kerma, in milligray) were also collected where recorded. For presentation of data, we adopted the mass stratification used by Glatz et al31 and the age stratification used by Verghese et al16 to allow comparison with these large studies. Data were also stratified by broad procedure categories (interventional, diagnostic and “other”). These categories were further divided into individual procedure types— atrial septal defect (ASD) occlusion, patent ductus arteriosus (PDA) occlusion, pulmonary valvuloplasty, aortic valvuloplasty, pulmonary artery (PA) angioplasty, coarctation intervention, electrophysiology studies (EPSs) with or without radiofrequency ablation (RFA), pulmonary vascular resistance studies, endomyocardial biopsy, coronary angiography, atrial septostomy and valve replacements. Examinations in which more than one intervention type was carried out were not included in these latter groups but were included within the larger interventional category. The “other” category included examinations difficult to classify as being diagnostic or interventional, biopsies and EPS procedures where it was not clear if RFA was performed. Data from pacemaker insertion procedures were also collected but not included in any of the three main categories. Procedures not involving radiation exposure (i.e. ultrasound only) or with doses recorded as zero were identified and excluded from the analysis. Organ dose estimation Patient-specific equivalent doses (i.e. the mean absorbed dose to the whole organ/tissue), measured in millisieverts, to active bone marrow, lymph nodes, the stomach, the oesophagus, the liver, the thyroid, the breasts (female patients only), the heart and the lungs, along with effective dose, were estimated for each procedure using a dedicated dosimetry system developed in MATLAB. This utilized data obtained from Monte Carlo simulations using PCXMC v. 2.0 (STUK, Helsinki, Finland). Simulations were conducted for six phantom sizes corresponding to ages 0, 1, 5, 10, 15 and 30 years (3.4, 9.2, 19.0, 32.4, 56.3 and 73.2 kg, respectively) and run at 1 million photons. The source to skin distance was kept at 80 cm, while the anode angle was fixed at 12°. Field sizes were 6 3 6, 8 3 8, 9 3 9, 10 3 10, 11 3 11 and 12 3 12 cm for the respective phantom sizes. Mean simulation uncertainties were ,1% for all organs except the breasts (1.9%) and the thyroid (3.8%). The dosimetry system allowed doses to be estimated for any possible beam angle from 55° cranial to 55° caudal and in a complete 360° rotation around the patient. For two out of three hospitals, the beam projections used are not recorded for each examination. Procedure-specific combinations of beam angles were estimated based on a combination of departmental protocols, discussions with radiography staff at each hospital and angle data where it was recorded for individual examinations at Hospital 2. Doses for patient sizes between the six simulated values were calculated by linear interpolation. A correction for beam energy, represented by half-value layer (HVL) was also applied. This was estimated based on a combination of physical measurements and values recorded in structured dose reports for clinical examinations. For examinations

Br J Radiol;88:20140852

BJR

Full paper: UK cardiac doses

Table 1. Details of equipment used at participating hospitals

Hospital and era

Total procedures

Machine names

FP/II

Antiscatter grid

Acquisition frames per second

Fluoroscopy pulses per second

Hospital 1 Lab 1 1994–2000

638

Siemens Bicor

II

Fixed

30 or 60

Continuous

Lab 2 1999–2002

941

Toshiba Infinix™ CB

II

Fixed

25 or 30

25 or 30

Lab 1 2002–08

3004

Siemens Axiom artis BC

II

Not for less than approximately 10 kg

10 or 15

15

Lab 4 2007–10

651

Siemens Axiom artis dBC

FP

Not for less than approximately 10 kg

10 or 15

15

Hospital 2 2004–08

1374

Philips Integris BH3000

II

Always

Not known

Not known

2008–13

1758

Siemens Artis Zee

FP

Always

30

10 or 15

Siemens HIcor

II

Always

30

10

Siemens Axiom Artis dFC

FP

Not for less than approximately 10 kg

30

12.5

Hospital 3 2004–08

738

2008–13

1153

FP, flat panel detectors; II, image intensifier detectors; Lab, laboratory. All Siemens Artis equipment and Siemens Bicor were obtained from Siemens Healthcare, Erlangen, Germany; Siemens HIcor obtained from Siemens Healthcare, Forchheim, Germany; Philips Integris machines obtained from Philips Medical Systems, Inc., Cleveland, OH; Toshiba Infinix CB obtained from Toshiba Medical Systems Corporation, Tokyo, Japan.

carried out using Siemens Artis equipment (either Axiom or Zee®; Siemens Healthcare, Erlangen, Germany), a HVL of 6.0 mm of aluminium (Al) was used, taking into account the combination of copper filtration (between 0 and 0.9 mm) and peak tube potential ranging from around 58 to 100 kV. Inherent filtration was set at 2.5-mm Al. For older Siemens HIcor (Siemens Healthcare, Forchheim, Germany) and Philips Integris (Philips Medical Systems, Inc., Cleveland, OH) machines, a HVL of 4.8-mm Al was used, while for the older generation Siemens equipment, 3.0-mm Al was used. No adjustment was made for contrast agent administration, transoesophageal ultrasound probes or the table on which the patient lies. Although the presence of more attenuating structures should result in increased X-ray output in compensation, downstream tissues would be partially shielded and receive a lower dose per unit PKA. RESULTS A summary of median PKA, screening time, skin dose (Hospital 1 only) and PKA per kilogram are presented in Table 2 for each hospital, for all procedure types combined. A summary of median PKA for individual procedure types, without mass stratification is shown in Table 3. More comprehensive tables with age and procedure type stratification are provided as Supplementary materials. Estimated equivalent doses to individual organs are given in Table 4 for all procedure types and all hospitals combined and in Table 5 for individual hospitals and eras without mass stratification. Again, more comprehensive data including hospital and procedure stratification are presented as Supplementary material. Doses were not normally distributed and were consistently right skewed. Statistical analysis was therefore focused on median figures and non-parametric tests.

3 of 14

birpublications.org/bjr

For all hospitals and for almost all procedure types, both doses (PKA and organ doses) and screening times were seen to significantly fall with successive eras (Wilcoxon test for all procedure types combined p # 0.01 in each case). The exception was between the 2002–08 and 2007–10 eras at Hospital 1 (p 5 0.116, 0.159 and 0.234 for PKA, screening time and cardiac dose, respectively). Doses for interventional procedures were significantly higher in this latter era (PKA 5 1.67 vs 2.40 Gy cm2; p 5 0.03), although a much greater proportion of these procedures were pulmonary valve replacements. When these procedures were excluded from the analysis, there was no significant difference in PKA (p 5 0.22). The fall in organ dose was less than that for PKA as beam quality—and hence dose per unit PKA—has increased with increased usage of copper filtration. For example, the Philips Integris system tends to use 0.2- or 0.4-mm copper, while the more recent Siemens Artis equipment was found to use 0.6 or 0.9 mm for fluoroscopy, reduced to 0.1–0.3 mm for acquisitions but with a corresponding increase in tube potential. It was also noted that the fall in doses tended to be greater for smaller patient size groups than for larger patients. Organ doses for patients .25 kg were actually significantly higher in the 2008–13 era at Hospital 3 than in the 2004–08 era (cardiac dose 5 12.3 vs 16.5 mSv; p 5 0.044). PKA was positively correlated with patient mass (e.g. Spearman’s r range 0.64–0.76 for recent era interventional procedures, p # 0.01 in each case) (a table of correlations can be found in the supplementary materials for this article). For individual procedure types, there was little evidence of correlation between mass and screening time, although small but significant positive correlations were found for interventional procedures combined. Where PKA was normalized by patient mass [PKA per kilogram

Br J Radiol;88:20140852

4 of 14 birpublications.org/bjr

2.48

14.7

98

Median PKA per kilogram

Screening time

n

1.0

15.0

105

Median PKA per kilogram

Screening time

n

12

206

Median skin dose

n

43

n

9.4

Screening time

171

0.64

Median PKA per kilogram

n

2.3 (1.2–5.5)

Median PKA (IQR)

Hospital 2, 2004–08

9.2

11

Median skin dose

0.13

Median PKA per kilogram

Screening time

0.5 (0.3–1.0)

Median PKA (IQR)

Hospital 1, 2007–10

0.20

13.0

Median PKA per kilogram

Screening time

0.7 (0.4–1.3)

Median PKA (IQR)

Hospital 1, 2002–08

3.5 (1.9–7.3)

Median PKA (IQR)

Hospital 1, 1999–2001

8.5 (5–15.3)

,5

Median PKA (IQR)

Hospital 1, 1994–2000

Hospital and era

469

9.8

0.59

5.1 (3.1–9.4)

159

9

9.4

0.09

0.8 (0.5–1.4)

661

13

10.5

0.12

1.0 (0.5–2.0)

280

16.0

0.9

6.7 (3.7–12.4)

158

19.8

2.33

19.1 (12.2–30.4)

5–12.5

363

9.2

0.43

7.2 (4–14.2)

162

14

9.4

0.08

1.4 (0.9–3.5)

747

18

10.2

0.10

1.6 (0.7–3.5)

252

13.3

0.6

10.2 (6–18.2)

181

18.8

1.70

27.9 (17.9–43.4)

12.5–25

162

10.0

0.44

14.7 (7.0–27.7)

118

47

10.2

0.15

5.5 (2.1–14.3)

644

32

10.2

0.10

3.3 (1.6–7.1)

178

15.0

0.6

20.8 (9.6–36.2)

102

22.4

1.94

64.4 (38–97.1)

25–45

Patient mass range (kg)

Table 2. Dose indicators for all procedure types combined, including pacemaker insertions

153

10.4

0.33

17.4 (6.8–38.7)

117

67

10.2

0.18

9.8 (4.7–19.6)

512

69

10.5

0.13

7.1 (2.5–14.2)

107

14.7

0.9

47.2 (24.7–87.6)

65

18.7

2.00

110.9 (67.8–152.1)

45–65

56

13.0

0.51

36.9 (15.9–49.2)

52

104

12.5

0.23

15.8 (8.6–33.1)

234

131

9.4

0.17

13.5 (4.2–26.1)

27

14.7

0.8

62.4 (28.4–132.2)

34

18.5

1.52

107.9 (32.3–184.8)

.65

(Continued)

1374

9.8

0.50

6.7 (3.3–15.6)

651

20

10.0

0.12

2.1 (0.8–8.5)

3004

22

10.4

0.12

2.1 (0.8–5.9)

941

14.9

0.8

10.3 (5–25.7)

638

18.9

2.02

27.1 (13.5–65)

All

BJR RW Harbron et al

Br J Radiol;88:20140852

5 of 14

birpublications.org/bjr

7.5

99

n

11.2

117

Screening time

n

417

10.6

0.13

1.1 (0.6–1.9)

244

12.3

0.47

3.7 (2.3–6.2)

531

7.9

0.38

3.3 (2.1–5.7)

5–12.5

269

11.5

0.15

2.5 (1.3–5.0)

148

12.1

0.33

6 (3.1–9.8)

440

7.0

0.31

5.2 (3.1–8.9)

12.5–25

145

15.0

0.20

6 (2.5–11.1)

91

15.8

0.22

6.7 (2.5–16.6)

216

8.4

0.29

9.4 (4.2–17.1)

25–45

227

11.5

0.30

17 (7.9–33.4)

45–65

129

12.5

0.23

12.9 (5.1–22.9)

100

15.9

0.15

8.9 (3.1–20.4)

Patient mass range (kg)

76

10.6

0.31

22.1 (8.0–39.6)

56

18.1

0.21

16 (7.4–29.7)

79

9.7

0.23

16.6 (9.5–38.1)

.65

1153

11.3

0.16

1.9 (0.8–6.4)

738

13.40

0.36

4.8 (2.3–9.9)

1758

8.1

0.35

4.7 (2.4–10.7)

All

IQR, interquartile range; PKA, kerma area product. PKA is presented as Gy cm2. Skin dose estimates are in the form of air kerma, in units of milligray and represent the greatest of the frontal and lateral outputs. Screening time is presented as the median, in minutes.

0.14

Median PKA per kilogram

Median PKA (IQR)

0.6 (0.3–0.9)

12.7

Screening time

Hospital 3, 2008–13

0.58

2.1 (1.3–5.2)

Median PKA per kilogram

Median PKA (IQR)

Hospital 3, 2004–08

265

Screening time

n

0.42

1.6 (0.6–3.2)

,5

Median PKA per kilogram

Median PKA (IQR)

Hospital 2, 2008–13

Hospital and era

Table 2. (Continued)

Full paper: UK cardiac doses BJR

Br J Radiol;88:20140852

BJR

RW Harbron et al

Table 3. Summary of median kerma area product (Gy cm2) for each era at each participating hospitals.

Procedure type

Hospital 1

Hospital 2

Hospital 3

1994–2000

1999–2001

2002–08

2007–10

2004–08

2008–13

2004–08

2008–13

Interventional

26.93

7.64

1.65

2.28

7.15

4.28

4.68

2.25

Diagnostic

71.24

12.11

3.21

2.74

4.95

5.81

5.16

1.86

Other

25.92

14.22

1.50

1.28

6.60

5.02

0.75

0.82

7.37

0.85

1.32

7.60

4.28

5.26

3.38

Atrial septal defect occlusion Patent ductus arteriosus occlusion

27.12

6.13

0.76

0.64

6.20

3.17

3.41

1.12

Pulmonary valvuloplasty

16.87

6.05

0.94

0.83

4.50

2.68

3.44

0.69

6.77

1.24

4.86

7.90

2.07

5.24

2.03

20.16

4.65

4.67

16.95

11.51

11.05

7.24

7.78

1.73

3.68

6.50

3.41

3.47

8.24

77.30

3.28

4.30

15.70

12.00

4.27

5.89

0.72

1.20

17.11

4.84

5.42

1.32

1.40

Aortic valvuloplasty Pulmonary artery angioplasty

34.10

Coarctation angioplasty Electrophysiology studies 6 radiofrequency ablation

70.79

Heart biopsy Coronary angiography

105.07

Pulmonary vascular resistance, pressures Pacemaker procedures Atrial septostomy

22.40 5.15

2.11

8.74

1.62

5.60

2.05

2.26

1.20

0.39

Valve replacement

2.57

0.69

35.98

Data are presented for 3 broad procedure categories and for 13 individual procedure types. Data are only presented where .10 procedures were performed.

(PKA/kg)], this figure was seen to vary by patient size, although the nature of this relationship varied between hospitals and equipment types. Where antiscatter grids were used for all patient sizes, PKA/kg was the highest in the 0- to 1-year age group and tended to be negatively correlated with patient age. Where antiscatter grids were omitted for small patients, PKA/kg tended to increase with patient size or display a lop-sided U- or J-shaped pattern. This pattern was even more pronounced where PKA/kg was further normalized by the screening time (Figure 1). A breakdown of PKA originating from fluoroscopy and “cine” acquisitions was recorded at Hospital 2 between 2004 and 2008. A small but significant positive correlation was found between total examination PKA and the percentage of this figure originating from fluoroscopy rather than from cine acquisitions (Spearman’s r 5 0.17; p , 0.01). The proportion of PKA from acquisitions was higher for diagnostic (32%) than for interventional (14%) procedures. Beam angles were also recorded at

6 of 14 birpublications.org/bjr

Hospital 2 for acquisitions only. The majority were in the posteroanterior (48%) or lateral (35%) projections, with the remainder made up mainly of left/right anterior oblique (both 3%) and other unspecified oblique views. No strong correlations between usage of particular beam angles and dose were found (r ranged from 20.05 to 0.09). Where separate frontal and lateral tube PKA figures were recorded, a small but significant negative correlation was found between total dose and the percentage of this figure originating from the frontal X-ray tube (across all procedure types, Spearman’s r 5 20.12; p # 0.01). In other words, where doses are high, laterally orientated projections tend to be more responsible than frontally orientated projections. Variation in radiation dose and screening time according to the involvement of different cardiologists or radiographers was investigated using a Kruskal–Wallis test for five of the most common procedure types at Hospital 1. A Spearman’s rank test was also performed to identify any correlation between

Br J Radiol;88:20140852

7 of 14

birpublications.org/bjr

17.7 (7.6–41.1)

18.2 (7.5–46.0)

17.9 (7.3–46.5)

3.8 (1.5–9.6)

13.3 (5.5–34.0)

1.3 (0.5–3.2)

3.9 (1.6–10.1)

2.6 (1.1–6.6)

6.2 (2.5–15.7)

Breasts

Heart

Lungs

Lymph nodes

Oesophagus

Thyroid

Liver

Stomach

Effective dose

6.8 (2.8–14.1)

3.1 (1.3–6.5)

5.6 (2.4–11.6)

1.2 (0.5–2.4)

14.8 (6.1–31.0)

4.1 (1.7–8.7)

19.6 (8–42)

18.4 (7.6–37.7)

17.6 (7.4–33.8)

2.6 (1.1–5.7)

5–12.5

6.6 (2.6–14.2)

3 (1.2–6.3)

5.3 (2.0–10.9)

0.9 (0.3–1.9)

14.6 (5.8–30.5)

4.1 (1.6–8.6)

19.6 (7.7–41.9)

18.5 (7.4–38.3)

15.9 (5.4–33.2)

2.7 (1.1–6.0)

12.5–25

7.3 (2.2–16.1)

2.9 (0.9–7.0)

4.7 (1.3–10.7)

0.6 (0.2–1.6)

14.7 (4.5–33.0)

4.4 (1.3–10.4)

22.2 (6.7–52.0)

22.5 (6.9–48.6)

11.3 (3.0–37.6)

4 (1.3–10.0)

25–45

Patient mass category (kg)

These figures represent doses for individual procedures. Cumulative organ doses for cohort members are given in Table 6.

2.6 (1.1–6.8)

,5

Bone marrow

Organ

6.6 (2.2–15.7)

2.4 (0.8–6.2)

3.7 (1.1–9.3)

0.4 (0.1–1.2)

14.9 (5.4–35.3)

4.3 (1.5–10.3)

23.8 (7.8–56.3)

22 (7.6–50.8)

6.3 (1.5–21.1)

5.2 (1.9–13.2)

45–65

7.2 (2.4–17.0)

2.6 (0.9–6.7)

4.1 (1.2–11.1)

0.3 (0.1–0.7)

17.3 (5.9–39.1)

4.6 (1.6–10.7)

23.5 (8.0–55.2)

24.9 (8.7–57.4)

6.9 (1.7–19.5)

7.5 (2.5–17.8)

.65

6.7 (2.6–14.9)

2.9 (1.1–6.6)

4.9 (1.8–10.8)

0.8 (0.3–1.9)

14.7 (5.7–32.4)

4.2 (1.6–9.4)

20.5 (7.7–46.6)

19.7 (7.5–43.1)

13.1 (4.5–31.3)

3.2 (1.2–8.0)

All

Table 4. Median (interquartile range) equivalent doses and effective dose (mSv) for all procedure types and all hospitals combined, stratified by patient mass in kilograms

Full paper: UK cardiac doses BJR

Br J Radiol;88:20140852

3.1 (1.6–6.5) 6.9 (3.4–13.5) 7.8 (4.2–13.6) 3.4 (1.4–8.6) 16.2 (8.3–30.1) Effective dose

29.9 (18.4–49.3)

3.4 (1.3–8.8)

9.9 (5.5–18.3)

1.4 (0.8–2.7) 2.8 (1.5–5.4) 3.6 (1.9–6.3) 1.5 (0.7–4.2) 6.8 (3.3–13.5) Stomach

9.8 (5.8–17.5)

1.4 (0.6–3.9)

4.1 (2.2–7.7)

2.3 (1.1–5.0) 5.1 (2.1–10.4) 5.7 (2.7–10.2) 2.3 (1.0–5.8) 11.3 (5.8–20.5) Liver

18.5 (11.4–31)

2.5 (0.9–6.1)

7.3 (3.6–14)

7 (3.6–14.1)

0.4 (0.2–0.8) 0.9 (0.3–1.8) 1.1 (0.5–1.9)

14.4 (7.5–27.3) 17.2 (9.4–30.7)

0.4 (0.2–1)

1.3 (0.6–2.5)

8.1 (3.2–19.9)

2 (1.0–3.9)

0.4 (0.2–1.1)

34.6 (17.7–64.1)

2.9 (1.5–5.6) Thyroid

Oesophagus

54.4 (33.2–92.3)

7.2 (2.8–20)

20.7 (11.5–38.4)

2 (1–4.1) 4.1 (2.1–7.8) 4.9 (2.6–8.7) 2.3 (0.9–5.8) 9.7 (5–18.8) Lymph nodes

16.2 (10–27.5)

2.1 (0.8–5.8)

5.9 (3.3–11.1)

9.4 (4.9–19.4) 20.7 (10.5–39.0) 23.2 (12.7–41.2) 10.9 (4.2–27.8) 50 (25.4–94.0)

8 of 14 birpublications.org/bjr

Lungs

93.7 (57.0–157.3)

10.1 (3.8–27.6)

29.7 (16.6–55.0)

7.1 (2.6–15.4)

9.5 (5.0–19.3)

18.7 (7.1–35.9)

19.6 (10.3–36.9)

19.2 (6.9–35.3)

23.1 (12.8–41.0)

26.4 (10.8–49.8)

27.4 (15.8–50.7)

5.7 (2.4–13.0)

10.5 (4.1–27.8)

6.5 (2.1–13.6)

9.9 (3.8–26.2)

33.3 (18.2–58.7)

44.8 (22.4–84.5)

69.2 (40.0–125.1)

72.4 (44.4–125.5)

Breasts

Heart

1.5 (0.7–3.5) 3.2 (1.7–6.3) 3.4 (1.9–6.5) 1.8 (0.6–5.7) 8 (3.8–17) Bone marrow

14 (8.3–25.4)

1.7 (0.6–4.9)

4.5 (2.5–8.8)

Hospital 3, 2008–13 Hospital 2, 2008–13 Hospital 1, 2007–10 Hospital 1, 1999–2001 Organ

Hospital 1, 1994–2000

Hospital 1, 2002–08

Hospital 2, 2004–08

Hospital 3, 2004–08

RW Harbron et al

Hospital and era

Table 5. Median organ doses in millisieverts (interquartile range) for all procedure types and all patient masses combined, stratified by hospital and era

BJR

dose/screening time and the number of procedures carried out for each operator. In both cases, PKA was normalized by mass to reduce the impact of variation in patient size on the results. For most examinations, more than one staff member was listed, and it was not known who was actually at the controls. The results were characterized by mainly negative findings. A significant difference in PKA was found for cardiologists for coronary angiography (p 5 0.03) and PDA occlusions (p 5 0.02), and for ASD occlusions for radiographers (p 5 0.03). A significant negative correlation was found between the number of coronary angiography procedures carried out by each operator and the PKA (Spearman’s r 5 20.17; p # 0.01). No other significant correlations of this type were found. Estimated skin doses were available for 3640 examinations at Hospital 1. These figures are simply derived from PKA divided by beam area at a particular “reference” point representative of the patient’s entrance surface, thus representing air kerma at that distance. Skin doses were recorded for frontal and lateral tubes. The figures shown in Table 2 represent whichever figure was the highest for each examination, rather than both added together. Skin dose estimates in any one plane exceeded 1000 mGy in six examinations and 2000 mGy in one examination. Skin doses also increased with patient mass/age, although to a smaller extent than did PKA. The organs receiving the highest equivalent doses were the heart, lungs and oesophagus. Breast doses were also high for examinations involving the use of laterally orientated projections but were very low for examinations in which these projections tend to be avoided. Cumulative organ doses for cohort members are summarized in Table 6. The median cumulative effective dose and cumulative cardiac dose for this cohort was 7.6 and 22.2 mSv, respectively. Around 13% of patients received cumulative doses to the heart and lungs of .100 mSv. At Hospital 1, which carries out transplant procedures, the mean number of procedures per patient was 1.48. The majority of patients (78.2%) underwent a single procedure during the study period, while 3.9% had five or more procedures and 0.9% had eight or more procedures. Of this latter group, almost all had a history of heart transplant and typically underwent coronary angiography and biopsy procedures. Hospitals 2 and 3 did not specialize in transplants. The mean number of procedures per patient was 1.18 and 1.33, respectively, with 0.4% and 0.8% of patients at each hospital, respectively, undergoing five or more procedures. Examining only data in which clinical indications for the procedure were recorded, 3% of patients had a history of tetralogy of fallot, 4% had transposition of great arteries and 1% had Down syndrome/trisomy 21. Patients with a history consistent with heart transplant made up 5% of the cohort and were the subject of 11% of all procedures. DISCUSSION Our results show that the radiation doses from cardiac catheterizations in young people have fallen significantly, in some cases by a factor of ten or more, over the past two decades. This is in spite of the capability of increasingly complex interventions such as valve replacements. An immediate question is whether

Br J Radiol;88:20140852

Full paper: UK cardiac doses

BJR

Figure 1. Kerma air product (PKA) normalized by mass, then further normalized by screening time (ST). Data represented by open circles were acquired where antiscatter grids were removed for small patients.

the fall in doses is related to new equipment installation (thus occurring suddenly) or more owing to gradual refinement of technique. Examining patterns on a year-by-year basis is difficult, as the number of individual procedure types carried out at a single centre in a given year was generally small. For all procedures combined (Figures 2–4), the fall in median screening times between successive eras was comparable to the fall within those eras. The fall in PKA, however, tended to be much larger between successive eras than within eras. This tendency of doses to fall suddenly following the installation of new equipment suggests that the impact of technological factors is greater than increased operator experience or gradual refinement of technique. Technological explanations for reduced doses include (1) improved dose efficiency of detector materials, (2) ability to remove antiscatter grids, (3) greater control over fluoroscopic pulse rates and acquisition frame rates, (4) doseefficient post-acquisition image processing, (5) changes to beam spectrum (e.g. added copper filtration), (6) use of ultrasound guidance as an alternative to fluoroscopy and (7) other innovations such as lung shuttering and “last-frame-hold” techniques. Of all these factors, usage patterns of antiscatter grids appears to have had the greatest influence on doses over the past 10 years, for patients aged under 5 years or who weighed around 10–20 kg. The introduction of pulsed fluoroscopy, followed by a reduction in pulses per second from 25–50 down to 7.5–15.0, is a second major factor affecting doses across all age groups. Other factors such as the relative usage of different projection angles or fluoroscopy vs acquisitions are less important and have less scope for change (i.e. beam angles are to a certain extent dictated by the condition being investigated). We found no convincing evidence of lower doses associated with the use of flat panel detectors as opposed to image intensifiers. The considerable variation in median doses between procedure types ought to warn against describing doses using arbitrary

9 of 14 birpublications.org/bjr

procedure categories such as “interventional” or “diagnostic”, especially where so-called reference doses are being set. Such pigeonholing is difficult to avoid where sample sizes are small, however. The lowest doses were for endomyocardial biopsies and PDA and ASD occlusions, although the ranking varied between hospitals and eras. Many atrial septostomy procedures were carried out entirely by ultrasound guidance and involved no ionizing radiation dose, especially at Hospital 3. Transcatheter valve replacements and PA angioplasty procedures were consistently among the highest dose procedures. Assessment of interoperator variation in doses and screening times was rather inconclusive and suggests a limited impact at a given hospital. The reduction in screening times between eras does suggest some element of technique refinement, although the sudden fall in screening times at Hospital 1 between 2001 and 2002 (Figure 2) associated with new equipment installation implies the impact of technological factors, such as “last-framehold” capability. Alternatively, the definition of “screening time” may be inconsistent, that is, acquisitions may be included in the figure or not (we assume they are not). Screening time was a poor predictor of variation in median doses between different hospitals. In fact, the hospital delivering the highest doses in the most recent era had the shortest screening times, although these figures were generally much more consistent between hospitals and from one procedure to the next than were PKA or organ doses. As such, the use of screening time as the sole dose metric is not recommended. The increase in PKA with patient size is unsurprising, given the greater thickness of patient traversed by the X-ray beam (and hence greater required output) and the larger beam area required to cover the heart region. The relationship between organ doses and patient size is rather more complex, varying with organ type and equipment. Patterns for most organs are similar to that for PKA/kg, tending to decrease with patient size for older era data and increase for more recent data. The exceptions were for

Br J Radiol;88:20140852

21.4

thyroid and breast doses, which tended to fall almost monotonically with increased patient size, for all eras. Comparison with previous research There are very few studies against which to compare organ doses, and we are not aware of any previous study in which cumulative doses were estimated for a large volume of patients. Organ doses for five procedures were measured by Barnaoui et al19 by reconstructing the examinations using anthropomorphic phantoms. Lung doses were higher in proportion to breast dose by a factor of 2.7 compared with 2.3 in the present study, while lung dose was higher than thyroid dose by a factor of 13, compared with 30 in the present study. Such differences could be owing to different field sizes and the proportion of the examination in laterally orientated projections.

6.7 These data take into account multiple procedures carried out on the same patient.

112.0 27.4 37.6 Standard deviation

70.3

131.9

27.2

89.5

25.4

8.3 12.8 2.5 40.0 11.7 58.7 53.1 10.6 18.4 Mean

36.5

613 251 2277 711 Maximum

753

1264

2745

627

2246

548

1 0 13 1 3 Percentage over 100 mSv

8

14

1

8

1

2 0 27 4 8 Percentage over 50 mSv

19

28

4

20

5

9

20

1

17

4

44 14 54

55

72

44

10 of 14 birpublications.org/bjr

12 23 Percentage over 20 mSv

24 42 Percentage over 10 mSv

63

72

28

65

32

7.7 2.3 53.9 9.6 18.5 Third quartile

38.2

57.2

11.3

39.5

13.4

1.3 2.0 0.3 8.7 1.4 2.9 First quartile

5.5

8.8

1.8

6.4

5.5 0.9 22.2 3.6 7.6 Median

16.7

23.2

4.6

16.5

Liver Thyroid Oesophagus Lymph nodes Lungs Heart Breasts Bone marrow Effective dose

Organ type

Statistical measure

Table 6. Cumulative organ doses for cohort members in millisieverts

3.2

RW Harbron et al

Stomach

BJR

There have been numerous audits of PKA published, but comparison with these studies is difficult as each involves a different range of patient ages and procedure types, and measures to stratify doses by these parameters are not consistent (or absent entirely). As mentioned in the introduction, comparison of PKA figures acquired using different equipment and from different patient sizes requires a great deal of caution. The doses from the 1994 to 2000 era at Hospital 1 were compatible with those quoted by Boothroyd et al7 (presented for frontal tube output only) but much higher than those reported by Rassow et al32 for data acquired from 1994 to 1996. For both diagnostic and interventional categories, recent era PKA figures were compatible with the results of studies led by Martinez et al,33 Dragusin et al21 and McFadden et al22 when adjusted for age and with those of Barnaoui et al19 for diagnostic, ASD occlusion and PDA occlusion procedures. These investigations were relatively small, with sample sizes of no more than a few hundred procedures. Two well-reported investigations by Verghese et al16 (3365 procedures) and Glatz et al31 (2265 procedures) have been published in which stratification of PKA by age or mass for individual procedure types was sufficient to allow meaningful comparisons. Given the modern equipment and the introduction of entirely sensible radiation protection measures such as reduced frame rates and antiscatter grid removal where appropriate, one would expect doses to be comparable to those of the present study over the same period; instead, they are higher by a factor of typically between two and ten, where matched for patient size and procedure type, both for PKA and air kerma (i.e. skin dose estimates). In particular, PKA figures reported by Verghese et al16 for ASD occlusions are 30–40 times higher than those recorded at hospital 1 in the current study. The screening times were longer, although by an insufficient extent to explain the discrepancy in PKA. A further survey by Ghelani et al34 in which data from 2713 cases were acquired from 7 centres, reported median PKA and air kerma levels higher still than Verghese et al16 or Glatz et al,31 although still compatible with these studies. A recent study by Kobayashi and colleagues18 was particularly ambitious, acquiring data from 16 centres giving a sample of 8267 procedures. A range of equipment types appear to have been used, although data were combined together to produce a single huge

Br J Radiol;88:20140852

Full paper: UK cardiac doses

BJR

Figure 2. Variation in median kerma area product (PKA) and screening time (ST) for all procedure types combined with year at Hospital 1.

dataset without analysis of variation between each of the participating centres or equipment types. The reported PKA figures are high but compatible with those reported by Verghese et al,16 Glatz et al31 and Ghelani et al,34 although the majority of analysis is focussed on PKA normalized by mass. These PKA/kg figures are on average around four times higher than the recent era figures in the present study, around twice as high as those reported by Onnasch et al12 but lower than the figure quoted by Chida et al.9 The authors recommend the use of PKA/kg as a standardized measure of paediatric cardiac catheterization doses. This appears to be based on the lack of strong correlation between PKA/kg and patient size (Pearson’s r, –0.044, –0.079 and 0.014 for diagnostic, biopsy and interventional procedures, respectively), supposedly indicating that PKA/kg is similar across patient ages. The lack of a clear pattern is unsurprising considering that data from different equipment types, each of which may respond to changes in patient size differently, have been

combined together. Secondly, a lack of correlation should not imply that there is no relationship between variables—a U-shaped relationship yields a correlation coefficient of zero. While the setting of reference doses specific to individual procedure types by Kobayashi et al18 should be applauded, our data do not support the recommendation that PKA/kg should be used as the standardized measure of radiation doses, at least not in isolation. Why are the doses at some centres and in some published studies so much higher than those of others? Is the variation real or simply owing to careless recording and reporting of figures or measurement uncertainty? Confusion between units, for example, between microgray square metre and milligray square centimetre could certainly explain a ten-fold variation in doses and is the most likely explanation for the extraordinarily low PKA reported by Smith et al17 (given as mGy/cm2). It was noted

Figure 3. Variation in median kerma area product (PKA) and screening time (ST) for all procedure types combined with year at Hospital 2.

11 of 14 birpublications.org/bjr

Br J Radiol;88:20140852

BJR

RW Harbron et al

Figure 4. Variation in median kerma area product (PKA) and screening time (ST) for all procedure types combined with year at Hospital 3.

that around half of previous publications make no mention of calibration of dose measuring equipment. Recorded PKA is subject to a typical uncertainty of 615%, even when regularly calibrated to national standards. For uncalibrated dosemeters, this uncertainty may be sizably higher, although never to the extent of explaining a ten-fold variation in doses. There are also concerns over the reliability of PKA measurements using Philips Integris equipment (Hospital 2, 2004–08) in which short exposures ,0.1 Gy cm2 are not recorded, meaning dose estimates for infants may be systematically underestimated.12 Assuming dose-measuring equipment is calibrated and figures are correctly quoted, what else could explain the enormous discrepancy in doses? Wide variation in technique and protocols, including frame rates and antiscatter grid usage, has been found between UK hospitals carrying out paediatric cardiac catheterizations.35 But the variation in doses between UK hospitals in this study is relatively smaller than the international variation reported in the studies discussed above. Different emphasis on image quality is another important factor. To an extent, there is an approximate proportionality between dose and image quality, in the sense that signal-to-noise ratio increases according to the square root of photon fluence.36 Some operators may consider low-dose images to be unacceptably noisy or “grainy”. There are, however, numerous other influencing factors, including detective quantum efficiency of the imaging system, beam quality and post-acquisition image processing. It would be wrong to conclude, therefore, that high image quality must always come at the expense of increased doses. Cardiologists are discovering that selecting the “low dose” setting does not compromise the ability to carry out certain catheterization procedures.37 In particular, a combination of very low frame rates (2–3 frames per second), antiscatter grid removal and avoidance of magnification was found to produce an excellent dose reduction for EPS procedures.38 It should be noted, however, that higher doses can be justified if it can be demonstrated that patient outcome is improved accordingly.

12 of 14 birpublications.org/bjr

Errors and uncertainties Earlier, the use of PKA as an “absolute” measure of radiation dose was criticized owing to the highly variable relationship with actual patient dose. What are the potential uncertainties in organ dose estimates derived from PKA? There are two major sources of error: (1) the use of Monte Carlo simulation to estimate doses and (2) variation in beam energy and projection angle from expected values. Regarding the former point, the mathematical phantom used in PCXMC is admittedly crude and handles changes to patient mass for a given height extremely badly (all organs simply inflate like a balloon). An alternative methodology would be physical measurements in anthropomorphic phantoms, although such measurements include their own uncertainties—real people are not made from indestructible urethane. Both methodologies can only provide dose estimates and neither should be viewed as a true measure of patient dose. The impact of variation in beam angles and beam energy from expected values depends on the organ in question (Figure 5). The former was investigated by calculating the error associated with a 610° variation in beam angle from straight PA and lateral projections in both rotational (around the patient) and craniocaudal directions. The breast dose was found to be the most sensitive to small changes in beam angle, with a large increase in dose as the beam is rotated from either lateral projection to frontally orientated (i.e. anteroposterior) projections. In the lateral projection, the 610° errors for this organ were on average 650%. Such errors are an important consideration for future epidemiological analysis of cancer risks following these procedures. Doses to lymph nodes, bone marrow and the lungs are more insensitive to changes to beam angle, with 610° errors generally less than 610%. Doses to the stomach, liver and thyroid are more sensitive to variation in craniocaudal beam angle. Variation in source-to-skin distance was found to be of less importance, with variation in doses per unit PKA generally ,5% as the distance was varied from 40 to 120 cm.

Br J Radiol;88:20140852

BJR

Full paper: UK cardiac doses

Figure 5. Monte Carlo simulation of variation in dose to selected organs as a function of beam rotation around the patient (zero craniocaudal angle, 56-kg phantom size, 4.8-mm aluminium half-value layer). A straight posteroanterior projection corresponds to 90°, a straight left lateral projection is 180° and a straight anteroposterior projection is 270°.

Organ dose errors are also introduced by the use of HVL to describe beam quality as opposed to the exact spectrum as defined by tube potential and filtration. Struelens et al39 estimate these errors to be ,5%. This figure matched our calculations for HVLs .5.5 mm Al, although errors were found to increase to up to 20% as HVL was decreased to 3 mm. The HVLs used for organ dose estimation were based on limited data for older examinations in which exposure factors were not usually recorded and were not patient size dependent. It would be expected that increased patient size would be associated with increased tube potential, but we found little suggestion that this was occurring. Across the range of HVL figures seen in clinical practice for modern equipment of 4.8–6.2 mm Al, a variation of 15–25% in effective dose and 25–35% in cardiac dose could be expected, although differences across the full possible range of beam energies could be as high as a factor of two- or three-fold. Variations in beam energy thus represent the single largest source of error in organ dose calculations. Further research in this area, along with adjustments for contrast agent administration, table attenuation and fat distribution could improve dose estimates. CONCLUSION The continued fall in radiation doses from cardiac catheterizations is encouraging, and the centres involved in this study

deserve great credit. The data that we have obtained may be useful in the setting of reference doses, although caution is advised where equipment used is different to that from which our data were acquired, as beam quality characteristics may vary. We note that a wide and largely unacknowledged gulf exists in the reported doses from cardiac catheterizations, between published studies. Dialogue on this matter and evaluation of whether higher doses can be justified in terms of better patient outcomes should be encouraged. An epidemiological analysis of the cancer risks following these procedures may be feasible, particularly for older data, when doses were higher. The low proportion of patients receiving large cumulative doses and the relatively large number of patients in this cohort or among similar cohorts with confounding factors in cancer development such as Down syndrome or transplant history, however, must be taken into account in such a study. FUNDING We received funding for this study from Newcastle-upon-Tyne Healthcare Charity. ACKNOWLEDGMENTS We are grateful for the help of Katherine Kirton, Emma Thomspon, Dawn Smith, Steve Charlton, John Crompton, Richard Hardy and Mike Dunn.

REFERENCES 1.

2.

Baysson H, R´ehel JL, Boudjemline Y, Petit J, Girodon B, Aubert B, et al. Risk of cancer associated with cardiac catheterization procedures during childhood: a cohort study in France. BMC Public Health 2013; 13: 266. doi: 10.1186/1471-2458-13-266 Einstein AJ. Effects of radiation exposure from cardiac imaging: how good are the data?

13 of 14 birpublications.org/bjr

3.

4.

J Am Coll Cardiol 2012; 59: 553–65. doi: 10.1016/j.jacc.2011.08.079 Picano E, Vano E. The radiation issue in cardiology: the time for action is now. Cardiovasc Ultrasound 2011; 9: 35. doi: 10.1186/1476-7120-9-35 Ait-Ali L, Andreassi MG, Foffa I, Spadoni I, Vano E, Picano E. Cumulative patient

5.

effective dose and acute radiation-induced chromosomal DNA damage in children with congenital heart disease. Heart 2010; 96: 269–74. doi: 10.1136/hrt.2008.160309 Bacher K, Bogaert E, Lapere R, De Wolf D, Thierens H. Patient-specific dose and radiation risk estimation in paediatric cardiac catheterization. Circulation 2005; 111: 83–9.

Br J Radiol;88:20140852

BJR

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

RW Harbron et al

Beels L, Bacher K, De Wolf D, Werbrouck J, Thierens H. Gamma-H2AX foci as a biomarker for patient X-ray exposure in pediatric cardiac catheterization: are we underestimating radiation risks? Circulation 2009; 120: 1903–9. doi: 10.1161/ CIRCULATIONAHA.109.880385 Boothroyd A, McDonald E, Moores BM, Sluming V, Carty H. Radiation exposure to children during cardiac catheterization. Br J Radiol 1997; 70: 180–5. Campbell RM, Strieper MJ, Frias PA, Jeager G, Balfour G, Costello L, et al. Quantifying and minimizing radiation exposure during pediatric cardiac catheterization. Pediatr Cardiol 2005; 26: 29–33. Chida K, Ohno T, Kakizaki S, Takegawa M, Yuuki H, Nakada M, et al. Radiation dose to the pediatric cardiac catheterization and intervention patient. AJR Am J Roentgenol 2010; 195: 1175–9. doi: 10.2214/AJR.10.4466 El Sayed MH, Roushdy AM, El Farghaly H, El Sherbini A. Radiation exposure in children during the current era of pediatric cardiac intervention. Pediatr Cardiol 2012; 33: 27–35. doi: 10.1007/s00246-011-0064-z Modan B, Keinan L, Blumstein T, Sadetzki S. Cancer following cardiac catheterization in childhood. Int J Epidemiol 2000; 29: 424–8. Onnasch DG, Schr¨oder FK, Fischer G, Kramer HH. Diagnostic reference levels and effective dose in paediatric cardiac catheterization. Br J Radiol 2007; 80: 177–85. Ross RD, Joshi V, Carravallah DJ, Morrow WR. Reduced radiation during cardiac catheterization of infants using acquisition zoom technology. Am J Cardiol 1997; 79: 691–3. Schueler BA, Julsrud PR, Gray JE, Stears JG, Wu KY. Radiation exposure and efficacy of exposure-reduction techniques during cardiac catheterization in children. AJR Am J Roentgenol 1994; 162: 173–7. Watson TG, Mah E, Joseph Schoepf U, King L, Huda W, Hlavacek AM. Effective radiation dose in computed tomographic angiography of the chest and diagnostic cardiac catheterization in pediatric patients. Pediatr Cardiol 2013; 34: 518–24. doi: 10.1007/s00246-012-0486-2 Verghese GR, McElhinney DB, Strauss KJ, Bergersen L. Characterization of radiation exposure and effect of a radiation monitoring policy in a large volume pediatric cardiac catheterization lab. Catheter Cardiovasc Interv 2012; 79: 294–301. doi: 10.1002/ccd.23118 Smith BG, Tibby SM, Qureshi SA, Rosenthal E, Krasemann T. Quantification of temporal, procedural, and hardware-related factors influencing radiation exposure during pediatric cardiac catheterization. Catheter Cardiovasc Interv 2012; 80: 931–6. doi: 10.1002/ ccd.24359

14 of 14

birpublications.org/bjr

18. Kobayashi D, Meadows J, Forbes TJ, Moore P, Javois AJ, Pedra CA, et al. Standardizing radiation dose reporting in the pediatric cardiac catheterization laboratorya multicenter study by the CCISC (Congenital Cardiovascular Interventional Study Consortium). Catheter Cardiovasc Interv 2014; 84: 785–93. doi: 10.1002/ccd.25467 19. Barnaoui S, Rehel JL, Baysson H, Boudjemline Y, Girodon B, Bernier MO, et al. Local reference levels and organ doses from pediatric cardiac interventional procedures. Pediatr Cardiol 2014; 35: 1037–45. doi: 10.1007/s00246-014-0895-5 20. Axelsson B, Khalil C, Lidegran M, Schuwer P, Mortensson W. Estimating the effective dose to children undergoing heart investigations— a phantom study. Br J Radiol 1999; 72: 378–83. 21. Dragusin O, Gewillig M, Desmet W, Smans K, Struelens L, Bosmans H. Radiation dose survey in a paediatric cardiac catheterisation laboratory equipped with flat-panel detectors. Radiat Prot Dosimetry 2008; 129: 91–5. doi: 10.1093/rpd/ncn035 22. McFadden S, Hughes C, D’Helft CI, McGee A, Rainford L, Brennan PC, et al. The establishment of local diagnostic reference levels for paediatric interventional cardiology. Radiography 2013; 19: 295–301. 23. Gherardi GG, Iball GR, Darby MJ, Thomson JD. Cardiac computed tomography and conventional angiography in the diagnosis of congenital cardiac disease in children: recent trends and radiation doses. Cardiol Young 2011; 21: 616–22. doi: 10.1017/ S1047951111000485 24. International Commission on Radiation Units and Measurements. Quantities and units for measurement and calculation in medical x-ray imaging. J ICRU 2005; 5: 25–34. doi: 10.1093/jicru/ndi025 25. Yakoumakis E, Kostopoulou H, Makri T, Dimitriadis A, Georgiou E, Tsalafoutas I. Estimation of radiation dose and risk to children undergoing cardiac catheterization for the treatment of a congenital heart disease using Monte Carlo simulations. Pediatr Radiol 2013; 43: 339–46. doi: 10.1007/ s00247-012-2510-3 26. Committee to Assess Health Risks from Exposure to Low Levels of Ionizing Radiation. Health risks from exposures to low levels of ionizing radiation: BEIR VII phase 2. Washington, DC: National Research Council of the National Academies; 2006. 27. Ronckers CM, Erdmann CA, Land CE. Radiation and breast cancer: a review of current evidence. Breast Cancer Res 2005; 7: 21–32. 28. Ron E, Lubin JH, Shore RE, Mabuchi K, Modan B, Pottern LM, et al. Thyroid cancer after exposure to external radiation: a pooled

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

analysis of seven studies. Radiat Res 1995; 141: 259–77. Shuryak I, Sachs RK, Brenner DJ. Cancer risks after radiation exposure in middle age. J Natl Cancer Inst 2010; 102: 1628–36. doi: 10.1093/jnci/djq346 United Nations Scientific Committee on the Effects of Atomic Radiation. Sources, effects and risks of ionizing radiation. UNSCEAR 2013 report, volume II, scientific annex B: effects of radiation exposure of children. New York, NY: United Nations; 2013. Glatz AC, Patel A, Zhu X, Dori Y, Hanna BD, Gillespie MJ, et al. Patient radiation exposure in a modern, large-volume, pediatric cardiac catheterization laboratory. Pediatr Cardiol 2014; 35: 870–8. doi: 10.1007/s00246-0140869-7 Rassow J, Schmaltz AA, Hentrich F, Streffer C. Effective dose to patients from paediatric cardiac catheterizations. Br J Radiol 2000; 73: 172–83. Martinez LC, Vano E, Gutierrez F, Rodriguez C, Gilarranz R, Manzanas MJ. Patient doses from fluoroscopically guided cardiac procedures in pediatrics. Phys Med Biol 2007; 52: 4749–59. Ghelani SJ, Glatz AC, David S, Leahy R, Hirsch R, Armsby LB, et al. Radiation dose benchmarks during cardiac catheterization for congenital heart disease in the United States. JACC Cardiovasc Inter 2014; 7: 1060–9. doi: 10.1016/j.jcin.2014.04.013 McFadden SL, Hughes CM, Winder RJ. Variation in radiographic protocols in paediatric interventional cardiology. J Radiol Prot 2013; 33: 313–19. doi: 10.1088/0952-4746/ 33/2/313 Busburg JT, Siebert JA, Leidholdt EM Jr, Boone JM. The essential physics of medical imaging. 3rd edn. Philadelphia, PA: Lippincott Williams & Wilkins; 2011. Sutton NJ, Lamour J, Gellis LA, Pass RH. Pediatric patient radiation dosage during endomyocardial biopsies and right heart catheterization using a standard “ALARA” radiation reduction protocol in the modern fluoroscopic era. Catheter Cardiovasc Interv 2014; 83: 80–3. doi: 10.1002/ccd.25058 Gellis LA, Ceresnak SR, Gates GJ, Nappo L, Pass RH. Reducing patient radiation dosage during pediatric SVT ablations using an “ALARA” radiation reduction protocol in the modern fluoroscopic era. Pacing Clin Electrophysiol 2013; 36: 688–94. doi: 10.1111/ pace.12124 Struelens L, Vanhavere, F, Bacher, K, Thierens, H. DAP to effective dose conversion in cardiology and vascular/interventional radiology. Belgium: FCK/University of Ghent; 2012.

Br J Radiol;88:20140852