Multi-institution assessment of the use and risk of cardiovascular computed tomography in pediatric patients with congenital heart disease

Abstract

Background:

Cardiac computed tomography (CT) is increasingly used in pediatric patients with congenital heart disease (CHD). Variability of practice and of comprehensive diagnostic risk across institutions is not known.

Methods:

Four centers prospectively enrolled consecutive pediatric CHD patients <18 years of age undergoing cardiac CT from January 6, 2017 to 1/30/2020. Patient characteristics, cardiac CT data and comprehensive diagnostic risk were compared by age and institutions. Risk categories included sedation and anesthesia use, vascular access, contrast exposure, cardiovascular medication, adverse events (AEs), and estimated radiation dose.

Results:

Cardiac CT was performed in 1045 pediatric patients at a median (interquartile range, IQR) age of 1.7 years (0.3, 11.0). The most common indications were arterial abnormalities, suspected coronary artery anomalies, functionally single ventricle heart disease, and tetralogy of Fallot/pulmonary atresia. Sedation was used in 8% and anesthesia in 11% of patients. Peripheral vascular access was utilized for 93%. Median contrast volume was 2 ml/kg. Beta blockers were administered in 11% of cases and nitroglycerin in 2% of cases. The median (IQR) total procedural dose length product (DLP) was 20 mGy*cm (10, 50). Sedation, vascular access, contrast exposure, use of cardiovascular medications and radiation dose estimates varied significantly by institution and age (p < 0.001). Seven minor adverse events (0.7%) and no major adverse events were reported.

Conclusion:

Cardiac CT for CHD is safe in pediatric patients when appropriate CT technology and expertise are available. Scans can be acquired at relatively low radiation exposure with few minor adverse events.

1. Introduction

Congenital heart disease (CHD) is the most common congenital anomaly.¹ Patients with moderate or complex CHD require serial diagnostic evaluation throughout life, and the cumulative risks of diagnostic testing are emerging as a significant burden.^2–6 Cardiac computed tomography (CT) is increasingly used in CHD patients of all ages, but its use has been limited by concerns regarding risk including radiation dose.

Historical estimates of cardiac CT radiation dose in CHD pediatric patients are as high as 18 mSv per ECG gated scan in the literature, which is twice the estimated radiation dose for diagnostic cardiac catheterization.³ Newer generation CT scanner platforms are increasingly utilized in the assessment of cardiac disease due to rapid image acquisition of 3D and 4D information, low radiation dose compared to older technology, submillimeter isotropic spatial resolution and temporal resolution as low as 66 msec. Despite these technological improvements, incorporation of cardiac CT into CHD clinical practice in pediatric patients has been slow, with data supporting its use consisting primarily of single institution retrospective studies with relatively small patient cohorts.^7–9 No prospective, multi-institutional data exists to properly inform contemporary practice and the risk of cardiac CT in pediatric patients with CHD in the current era.

Comprehensive risk assessment for cardiac imaging in CHD includes the use of sedation/anesthesia, the need for vascular access, exposure to contrast media, cardiovascular medication administration, radiation exposure and the rate of adverse events.

This is a multi-institutional, prospective observational study assessing the variability of practice and comprehensive diagnostic risk for the use of cardiac CT in pediatric CHD patients using recent generation technology.

2. Methods

2.1. Registry description

We developed a multi-institutional, prospective, observational registry to assess the performance of cardiac CT in CHD patients of all ages.¹⁰ Herein, we present data regarding variability of practice and comprehensive diagnostic risk for the pediatric cohort of patients <18 years of age. Patient and procedural cardiac CT scan data were collected consecutively for all cardiac CT scans performed on patients with CHD from four institutions between January 6, 2017–1/30/2020. Cardiac imagers included both cardiologists and radiologists. CT scans were included in analysis if ordered for a CHD patient to evaluate cardiovascular anatomy or function. Computed tomography studies performed on a patient with CHD for assessment of non-cardiovascular indication were not included. Cardiac CT scans were performed according to local clinical care standards. De-identified clinical and procedural cardiac CT scan data were entered into a Research Electronic Data Capture (REDCap) database by practitioners at each institution. Data were analyzed at a central institution (Minneapolis Heart Institute Foundation). Each site received local Institutional Review Board approval prior to study initiation.

2.2. Patient and scan characteristics

Patient descriptors including age, weight, height, gender and inpatient vs. outpatient status were recorded. Cardiac diagnoses were divided into the following categories: arterial abnormalities; suspected coronary artery anomaly; tetralogy of Fallot/pulmonary atresia; transposition complex; valve diseases (including sub and supra valvular disease); other vascular anomalies; intracardiac shunt lesions; other left side obstruction lesions, and functionally single ventricle.

Anatomic scans were classified by type of ECG triggering/gating: a) prospective ECG triggering with high pitch spiral scan mode; b) prospective ECG triggering axial/“step + shoot”/adaptive sequential scan mode; or c) retrospective ECG gating, spiral scan mode. Anatomic scans were defined as those acquired using radiation for a single cardiac phase or a narrow acquisition window. Functional datasets were defined as those capturing data throughout the cardiac cycle (further sub-classified by either retrospective or prospective ECG gating), allowing post processing for assessment of ventricular size and function. For functional datasets, it was determined whether ECG dose modulation was used. The tube voltages available on the scanner platforms included in analysis ranged from 70 to 130 Kilovoltage peak (kVp) and were recorded for each scan.

2.3. Risk and adverse events

Comprehensive risk was categorized into the following categories: use of sedation/anesthesia, type of vascular access, iodinated contrast volume, use of cardiovascular medications, adverse events, and radiation dose estimation.

Sedation:

Categories for sedation were classified as a) no sedation; b) mild to moderate sedation or anxiolysis with patient awake and responsive; c) deep sedation without airway support with patient sedated but able to maintain airway; d) deep sedation/general anesthesia with airway support/intubation. For patients in category “d” with general anesthesia including airway support, it was determined if the patient was intubated for a clinical indication or another procedure, or if the patient was intubated specifically for a breath hold during the cardiac CT scan.

Vascular access:

Intravenous line (IV) gauge and site were identified and recorded. If a central venous catheter was used for contrast injection, it was determined whether an existing central line placed for clinical indication was used or if a central line was placed specifically for the cardiac CT scan.

Contrast:

The type of iodinated contrast and dose (reported as total volume and as milliliter per kilogram) was included.

Cardiovascular medication:

Medication given specifically for the performance of the cardiac CT scan to control heart rate or for coronary vasodilation were reported.

Adverse events (AEs):

AEs were defined as minor and major. Minor AEs included the following: vascular complication of contrast extravasation not requiring treatment; mild allergic reaction to contrast not requiring treatment, or a respiratory event requiring only an increase in oxygen support to maintain saturations during the time of the scan. Major AEs included cardiac arrest or bradycardia requiring treatment; tachycardia requiring treatment; air embolus; respiratory event requiring escalation of airway support; broviac/central line rupture; vascular compromise including contrast extravasation requiring treatment; unplanned admission post procedure; allergic reaction to contrast requiring treatment; renal injury from contrast requiring treatment; or cardiac medication side effects. For outpatient studies, AEs were included from the time the patient presented to the imaging area until the time of discharge after the cardiac CT scan was performed. For studies performed under sedation or anesthesia in patients who presented as an outpatient, AEs were included from time of initiation of sedation/anesthesia team care until the time of discharge from anesthesia care after the cardiac CT scan. For hospitalized patients, AEs were included if they occurred at the initiation of transport to the CT unit until return to the clinical hospital unit subsequent to the cardiac CT scan.

2.4. Radiation dose estimates

Radiation dose estimates generated by the CT system were recorded as volume CT dose index (CTDIvol) and scan dose length product (DLP), in mGy*cm. The estimated effective dose (ED) in milliSieverts (mSv) was calculated by multiplying the scanner derived DLP by the age specific conversion (k) factors derived from International Commission of Radiation Protection (ICRP 60)¹¹ and ICRP 103¹² (Table 1). A baseline CT modality dose was derived by using the standard k factor (0.014) based on a 32 cm phantom for historical comparison (Table 4).

Table 1.

Cardiothoracic CT Conversion Factors (k factor) for DLP to mSv Calculation.

Age group	Age interval (years)	Historic conversion factors (ICRP 60)11 (mSv*mGy−1cm−1)	Conversion factors (ICRP 103)12 (mSv*mGy−1cm−1)
0	≤0.5	0.078	0.085
1	>0.5–2.5	0.052	0.079
5	>2.5–7.5	0.036	0.065
10	>7.5–12.5	0.026	0.037
15	>12.5	0.014	0.026

CT, computed tomography; DLP, dose-length product; ED, effective dose; ICRP, International Commission on Radiological Protection (ICRP). ICRP 60 (1990),¹¹ ICRP 103 (2007).¹² Conversion factors are for DLP from a 32 cm phantom.

Table 4.

Radiation dose estimates for anatomic and functional scans.

	Entire Group	≤6 months	>6 months - 2.5 yrs	>2.5–7.5 yrs	>7.5–12.5 yrs	>12.5 - <18 yrs	p-value
Total Procedural Values, n	1045	392	163	151	129	210
DLP, mGy · cm, median (IQR)	20 (10, 50)	10 (7, 12)	16 (12, 24)	27 (19, 42)	45 (33, 75)	84 (57, 125)	<0.001
Age adjusted dose, mSv, ICRP 60† (IQR)	0.86 (0.62, 1.33)	0.78 (0.55, 0.94)	0.83 (0.62, 1.25)	0.97 (0.68, 1.51)	1.16 (0.86, 1.95)	1.18 (0.80, 1.75)	<0.001
Age adjusted dose, mSv, ICRP 103‡ (IQR)	1.24 (0.85, 2.15)	0.85 (0.60, 1.02)	1.26 (0.95, 1.90)	1.76 (1.24, 2.73)	1.65 (1.22, 2.78)	2.18 (1.48, 3.25)	<0.001
Anatomic Scan, n	1228	509	196	170	136	217
DLP, mGy · cm, median (IQR)	11 (6, 32)	5 (4, 8)	11 (8, 16)	19 (15, 29)	35 (20, 52)	61 (40, 94)	<0.001
CTDI Volume, mGy, median (IQR)	0.74 (0.45, 1.68)	0.46 (0.24, 0.66)	0.66 (0.47, 1.11)	0.85 (0.71, 1.69)	2.11 (0.90, 3.30)	3.16 (1.77, 4.90)	<0.001
Generic Modality Dose*, mSv (IQR) (for historical comparison)	0.16 (0.08, 0.45)	0.08 (0.05, 0.11)	0.15 (0.11, 0.22)	0.26 (0.21, 0.41)	0.48 (0.28, 0.73)	0.86 (0.57, 1.31)	<0.001
Age adjusted dose, mSv, ICRP 60† (IQR)	0.58 (0.38, 0.88)	0.42 (0.28, 0.62)	0.57 (0.42, 0.82)	0.68 (0.54, 1.06)	0.90 (0.52, 1.35)	0.86 (0.57, 1.31)	<0.001
Age adjusted dose, mSv, ICRP 103‡ (IQR)	0.79 (0.48, 1.36)	0.46 (0.31, 0.67)	0.86 (0.63, 1.24)	1.23 (0.98, 1.91)	1.28 (0.74, 1.93)	1.59 (1.05, 2.43)	<0.001
Functional Scan, n	101	23	14	17	15	32
DLP, mGy · cm, median (IQR)	54 (22, 82)	13 (10, 22)	22 (17, 35)	57 (41, 65)	72 (54, 107)	104 (73, 150)	<0.001
CTDI Volume, mGy, median (IQR)	3.40 (1.76, 5.63)	1.24 (0.84, 1.92)	1.65 (1.57, 2.37)	3.65 (2.79, 4.48)	4.38 (2.58, 6.89)	5.71 (3.98, 7.84)	<0.001
Generic Modality Dose*, mSv (IQR) (for historical comparison)	0.75 (0.31, 1.15)	0.19 (0.14, 0.30)	0.30 (0.24, 0.49)	0.79 (0.57, 0.90)	1.00 (0.75, 1.50)	1.45 (1.03, 2.10)	<0.001
Age adjusted dose, mSv, ICRP 60† (IQR)	1.51 (1.01, 2.16)	1.04 (0.76, 1.69)	1.13 (0.90, 1.84)	2.04 (1.46, 2.33)	1.86 (1.39, 2.79)	1.45 (1.03, 2.10)	0.003
Age adjusted dose mSv, ICRP 103‡ (IQR)	2.29 (1.50, 3.56)	1.13 (0.83, 1.84)	1.71 (1.37, 2.79)	3.68 (2.64, 4.20)	2.65 (1.98, 3.97)	2.70 (1.90, 3.89)	<0.001

yrs, years; DLP, dose length product; IQR, interquartile range; CTDI, computed tomography dose index

^*DLP x k factor of 0.014, mSv, millisievert, ICRP International Commission of Radiation Protection, Data presented here include data collected per scan, not procedural total.

^†Adapted from The 2007 Recommendations of the. ICRP publication 103 (12).

^‡1990 Recommendations of the ICRP (11).

2.5. Statistical analysis

Results are described by age group and across institutions. Continuous variables are expressed as mean ± standard deviation (SD) if distributed symmetrically, or median and interquartile range (IQR) if skewed. Categorical variables are described as frequencies and percentages. Analysis of continuous variables was performed using one-way analysis of variance or the Kruskal-Wallis test depending on the distribution. Discrete variables were compared using the chi-square test or Fisher exact test where appropriate. Monte Carlo simulations were run to obtain simulated p-values for discrete variables that were too sparsely populated (13). To assess the variability of DLP levels between practices, the differences in median doses between institutions and their corresponding 95% confidence intervals were estimated using quantile regression for each age group. Stata version 15.1 (College Station, TX) and R v 3.6.1 (R Foundation for Statistical Computing, Austria) in R studio v 1.1.463 (R Studio, Inc) were used in the analysis.¹³

3. Results

3.1. Patient and study site characteristics

Three institutions from the United States and one from England participated in this study. A total of 1045 pediatric patients underwent cardiac CT scanning for evaluation of CHD during the time of review. The median age at the time of cardiac CT scan was 1.7 years (0.3–11), 40% were female and the median body surface area (BSA) was 0.75 ± 0.59 m². Overall, 342 (33%) were inpatients and 703 (67%) were outpatients. The most common primary indications for the studies were: arterial abnormalities, suspected coronary artery anomalies, functionally single ventricle heart disease, and tetralogy of Fallot/pulmonary atresia; some patients had more than one indication. Patient characteristics by study site are presented in Table 2 and by age in Table 3. A detailed description of indications and diagnosis type can be found in Appendix 2.

Table 2.

Procedural patient characteristics, scanner platform, and risk by institution.

Characteristics	All Patients	Institution A	Institution B	Institution C	Institution D	p-value
CCT procedures, n (%)	1045	318	345	338	44
Age at procedure, yrs, median (IQR)	1.7 (0.3, 11.0)	3.3 (0.3, 12.5)	2.5 (0.2, 11.1)	0.7 (0.2, 8.0)	1.3 (0.4, 6.0)	0.001
Gender, Female, n (%)	422 (40)	143 (45)	138 (40)	119 (35)	22 (50)	0.041
Weight, kg, mean ± SD	23.2 ± 25.2	26.3 ± 26.2	25.1 ± 26.4	18.9 ± 23.1	17.9 ± 16.5	<0.001
BSA, m2, mean ± SD	0.75 ± 0.59	0.82 ± 0.60	0.64 ± 0.55	0.64 ± 0.55	0.68 ± 0.51	<0.001
Patient Status						<0.001
Outpatient, n (%)	703 (67)	216 (68)	255 (74)	212 (63)	20 (45)
Inpatient, n (%)	342 (33)	102 (32)	90 (26)	126 (37)	24 (55)
CT scanner platform						<0.001
Second gen DS-Flash, n (%)	678 (65)	318 (100)	0 (0)	316 (93)	44 (100)
Second gen DS-Drive, n (%)	17 (2)	0 (0)	1 (0)	16 (5)	0 (0)
Third gen DS-Force, n (%)	350 (33)	0 (0)	344 (100)	6 (2)	0 (0)
Sedation
Sedation administered for scan	196 (19)	72 (23)	107 (31)	16 (5)	1 (2)	<0.001
Mild to moderate, n (%)	32 (3)	2 (1)	28 (8)	1 (0)	1 (2)
Deep w/o airway support, n (%)	49 (5)	11 (3)	28 (8)	10 (3)	0 (0)
Anesthesia w/airway support, n (%)	115 (11)	59 (19)	51 (15)	5 (1)	0 (0)
Functional Scan	29 (3)	0 (0)	29 (8)	0 (0)	0 (0)
Anatomic Scan only	86 (8)	59 (19)	22 (6)	5 (3)	0 (0)
Patient clinically intubated, n (%)	92 (9)	9 (3)	39 (11)	40 (12)	4 (9)
Vascular Access
Peripheral IV, n (%)	960 (93)	283 (92)	333 (98)	302 (90)	42 (95)	<0.001
Contrast Dose ml/kg, median (IQR)	2.0 (1.7, 2.3)	2.9 (2.2, 3.0)	1.9 (1.4, 2.0)	2.0 (1.8, 2.1)	1.8 (1.5, 2.0)	<0.001
Volume, median (IQR)	25 (10, 60)	40 (16, 106)	25 (9, 55)	15 (8, 50)	20 (11, 40)	<0.001
Cardiovascular Medication
Beta blocker, n (%)	113 (11)	51 (16)	61 (18)	0 (0)	1 (2)	<0.001
Nitroglycerin, n (%)	16 (2)	0 (0)	16 (5)	0 (0)	0 (0)	<0.001
Adverse Events, n (%)	7 (1)	4 (1)	0 (0)	3 (1)	0 (0)	0.188
Radiation
Procedural DLP, mGy*cm, median (IQR)	20 (10, 50)	32 (12, 65)	22 (11, 58)	13 (7, 26)	20 (12, 36)	<0.001

CCT, cardiac computed tomography; yrs, years; IQR, interquartile range; kg, kilograms; BSA, body surface area; gen-DS, generation dual source; w/o, without; w/, with; IV, intravenous; ml, milliliters; DLP, dose length product; Continuous values are reported as mean ± standard deviation or median (IQR) if skewed. Discrete variables reported as count and percent.

Table 3.

Procedural cumulative risk profile by age.

	All Patients	≤6 months	>6 months-2.5 yrs	>2.5–7.5 yrs	>7.5–12.5 yrs	>12.5–<18 yrs	P-value
CCT procedures, n	1045	392	163	151	129	210
Sedation
Sedation for scan	196 (219)	43 (11)	69 (43)	72 (48)	8 (6)	4 (2)	<0.001
Mild to moderate, n (%)	32 (3)	14 (4)	7 (4)	9 (6)	2 (2)	0 (0)
Deep w/o airway support, n (%)	49 (5)	8 (2)	19 (12)	19 (13)	2 (2)	1 (0)
Anesthesia with airway support, n (%)	115 (11)	21 (5)	43 (27)	44 (29)	4 (3)	3 (1)
Patient clinically intubated, n (%)	92 (9)	71 (18)	15 (9)	5 (3)	1 (1)	0 (0)
Vascular Access
Peripheral IV, n (%)	960 (93)	331 (87)	147 (91)	149 (99)	127 (99)	206 (99)	<0.001
Contrast dose ml/kg, median (IQR)	2.0 (1.7, 2.3)	2.1 (1.9, 2.5)	2.1 (1.9, 2.8)	2.0 (1.7, 2.5)	1.8 (1.5, 2.3)	1.4 (1.1, 2.0)	<0.001
Total volume, ml, median (IQR)	25 (10, 60)	8 (6, 11)	18 (15, 24)	37 (30, 46)	60 (50, 85)	87 (68, 115)	<0.001
Cardiovascular Medication
Beta blocker, n (%)	113 (11)	2 (1)	2 (1)	7 (5)	41 (32)	61 (29)	<0.001
Nitroglycerin, n (%)	16 (2)	0 (0)	0 (0)	0 (0)	4 (3)	12 (6)	<0.001
Adverse Event	7 (1)	3 (1)	2 (1)	0 (0)	0 (0)	2 (1)	0.691
Radiation
Procedural DLP, mGy*cm, median (IQR)	20 (10, 50)	10 (7, 12)	16 (12, 24)	27 (19, 42)	45 (33, 75)	84 (57, 125)	<0.001

CCT, cardiac computed tomography; yrs, years; w/o, without; IV, intravenous; IQR, interquartile range; ml, milliliter; kg, kilogram;

DLP, dose length product; Continuous values are reported as median (IQR) and discrete values reported as count and percent.

3.2. Scanner platform and scan sequence

Three types of scanner platforms were used: Siemens second generation dual source (DS)-Flash (65%) (Forscheim, Germany, gantry rotation time = 280 ms, temporal resolution = 75msec, collimation of 2 × 128 × 0.6 mm), Siemens second generation DS-Drive (2%) (Forscheim, Germany, gantry rotation time = 280 ms, temporal resolution = 75msec, collimation of 2 × 128 × 0.6 mm), and Siemens third generation DS-Force (33%) (Forscheim, Germany, gantry rotation time = 250 ms, temporal resolution = 66msec, collimation of 2 × 192 × 0.6 mm) (Table 2). Median cardiac CT procedural time was 5 min (IQR 3, 8) from localizing image to scan acquisition. Two institutions used automated tube current modulation (Care kv) with adjustments as needed based on practice history, and two institutions used predetermined kv settings based on local practice history.^14,15

In the 1045 patient procedures, 1329 scans were performed, including 1228 (92%) anatomic scans and 101 (8%) functional scans. 270 patients had 2 scans in the same procedure (200 had two anatomic scans, three had two functional scans, and 67 had one functional and one anatomic scan). Seven patients (0.7%) had three anatomic scans. 96% of anatomic scans were ECG gated, including 75% prospective ECG triggering with high pitch spiral scan mode, 24% prospective ECG triggering axial/“step + shoot”/ adaptive sequential scan mode, and 1% retrospective ECG gating, spiral scan mode. All functional scans were ECG gated and used pulsed modulation with a narrow acquisition window timed to systole or diastole depending on scan indication. 68% of patients underwent cardiac CT at the lowest available tube voltage of 70 kVp kVp of 80 was used in 20%, 90 kVp in 1%, 100 kVp in 7%, 110 kVp in 0.1% and 120 kVp in 4%. kVp varied significantly (p < 0.001) by institution and age.

3.3. Risk and adverse events

Sedation:

Anesthesia and/or sedation specifically for the cardiac CT scan was used in 196 patients (19%). 115 patients (11%) underwent general anesthesia with airway support, 49 (5%) received deep sedation without airway support and 32 (3%) had scan performed with use of mild-moderate sedation. An additional 92 patients (9%) were intubated prior to the scan for clinical indications independent from the CT scan. Rates of sedation and anesthesia use varied significantly (p < 0.001) by both age and institution (Tables 2 and 3). Breath hold is needed for scans acquiring data over several cardiac cycles where motion artifact can affect scan quality. Typically this is for coronary imaging at high heart rates or for functional scans. Indication for scan in those who had general anesthesia were primarily for coronary assessment or functional assessment and the patient was too young to cooperate with breath holding.

Vascular access:

Peripheral IV access was used for contrast administration in 960 (93%) of patients. Additionally, 70 patients (7%) patients had contrast administration via a central line placed prior to cardiac CT scan for clinical reasons. In the 392 infants ≤6 months of age, hand injection was used for contrast administration in 37% and a power injection was used for contrast administration in 63%. Type of vascular access and use of power injector varied significantly by institution and age (p < 0.001) (Tables 2 and 3).

Contrast:

Omnipaque 350 (GE Healthcare Ireland, Cork, Ireland) was used in 88% of studies. Omnipaque 300 (GE Healthcare Ireland, Cork, Ireland) was used in 8%, Visipaque 320 (GE Healthcare Ireland, Cork, Ireland) in 4%. In young patients, contrast is typically calculated by weight and in older adolescents it is typically reported in total volume. The median dose of iodinated contrast was 2.1 ml/kg (IQR 1.9–2.5) for patients < 12.5 years of age. For patients 12.5–18 years of age median total contrast dose was 87 ml (IQR 68, 115). Contrast use varied by institution and age (p < 0.001) with the higher doses per kilogram given to younger patients (Tables 2 and 3).

Cardiovascular Medication:

Beta blockers were used in 113 (11%) patients and nitroglycerin was used in 16 (2%) patients. No AEs due to medication administration were reported. Cardiovascular medication varied by institution and age, respectively (p < 0.001). The most common indication for CT scan in patients who received cardiovascular medication was coronary artery imaging 102 (89%).

Adverse events:

There were no major adverse events. Seven patients (0.7%) had a minor adverse event including contrast extravasation (n = 3), mild allergic reaction to contrast (n = 2), and a temporary increase in oxygen support to maintain saturations (n = 2). Of the three patients who had contrast extravasation, none required treatment or were associated with sequelae. Two of the patients were <1 year of age and contrast was hand injected. There were no episodes of contrast extravasation in infants <1 year of age who underwent power injection with 22 or 24 gauge IV. The two patients with mild allergic reaction to contrast were 13 and 17 years old and symptoms resolved without treatment. Two patients required an increase of oxygen support to maintain saturations during the time of the scan without escalation of airway support. Rate of adverse events did not vary by institution or age.

3.4. Radiation dose estimates

The median procedural DLP for all age groups was 20 (10, 50) mGy*cm. The procedural DLP included all monitoring sequences in addition to all diagnostic scan sequences. For historical comparison of prior radiation dose estimates using the k factor of 0.014 for all ages, the estimated procedural radiation dose would be 0.28 mSv. Newer estimates of radiation dose use age adjusted k factors as published in ICRP 60 (1990)¹¹ and ICRP 103 (2007).¹² These k factors increase the radiation dose estimate in young patients for the same scanner output compared to an adult. The median procedural effective dose using updated age adjusted k factors is estimated to be 0.86 (0.62, 1.33) mSv (ICRP 60) or 1.24 (IQR 0.85, 2.15) mSv (ICRP 103). The median DLP for an anatomic scan was 11 (6, 32) mGy*cm. The median DLP for a functional cardiac CT scan was 54 (IQR 22, 82) mGy*cm. Radiation dose by procedural, anatomic and functional scans are presented in Table 4.

The variability of practice for anatomic scans and DLP are presented in greater detail in Fig. 1; this shows the distribution of radiation dose in anatomic scans by age group among the institutions. The variability among the youngest two patient groups, ≤6 months and >6 months-2.5 yrs, is highlighted in Fig. 1b. The raw data used to construct Fig. 1 is presented in Appendix 1.

Fig. 1.

Distribution of radiation dose in anatomic scans by age group among the institutions A, B, C and D. 1a depicts the variability of the entire study cohort. 1b depicts the variability in the youngest two age groups.

4. Discussion

Our data demonstrates that cardiac CT is a low risk procedure in the current era, in a real world and multi-institutional cohort of CHD patients from 1 day to less than 18 years of age. This includes a uniformly low radiation dose, use of sedation/anesthesia in a minority patients, use of peripheral IV for vascular access, and rare minor adverse events.

The median procedural DLP for this cohort of 1045 patients translates to a procedural effective dose of 1.24 mSv using the updated age adjusted conversion factor from ICRP 103.¹² Historical estimates of effective radiation dose for chest CT have previously been reported using the generic conversion factor of 0.014, which would estimate a radiation dose of 0.28 mSv for comparison to historical estimates that were not similarly age adjusted. Radiation dose reduction is dependent on scanner technology, use of dose optimization protocols and CHD trained personnel. Radiation dose estimates have been shown to vary by tube output and are lower for lower tube voltage. The most widely cited paper estimating radiation dose by tube output does not have dose estimations for 70 kVp which was used for a majority of scans in this cohort.¹⁶ Ghoshhajra and colleagues have shown up to 85% radiation dose reduction with CT technology improvements in their single institution study of 95 CHD patients.¹⁷ The importance of scanning standardization and trained personnel was illustrated in a single institution study analyzing 236 CT studies performed in patients with CHD. By implementing a quality improvement initiative that included standardized protocols and personnel training, radiation dose was reduced from a mean of 21 mSv-2 mSv per scan without change in scanner technology, and with similar accuracy compared to surgical findings.¹⁸ Surveys of non-cardiac pediatric chest CT have shown similar and wide variability of radiation exposure from 0.03 to 69.2 mSv.¹⁹ Among our cohort, there remains variability of radiation dose. Further exploration of this variation could identify potential opportunities for improvement in dose optimization across participating institutions and could serve to inform and support the wider pediatric cardiac imaging community.

Cumulative diagnostic radiation exposure for CHD patients has been shown to potentially increase lifetime risk of cancer in a dose dependent manner.^3,20 A population based study estimated an odds ratio for the development of cancer of 1.1 per 10 mSV for low dose ionizing radiation procedures in adult CHD patients. In this prior study, cardiac CT was assumed to have a radiation dose of 16 mSv, which was the highest of all modalities based on a literature review.²¹ These historical studies do not use age or size adjusted radiation dose estimates and so comparison to current estimates using updated k factors are problematic.

A linear relationship between radiation exposure and development of cancer oversimplifies a multifactorial relationship that must include intrinsic patient susceptibility. Population based data show that patients with CHD and other congenital anomalies are at significantly increased intrinsic risk for malignancy during childhood, with a hazard ratio of 2.4 for cancer in CHD patients vs healthy controls up to age 41. The hazard ratio was similar for conotruncal defects and for patients with a VSD who are unlikely to have undergone catheterization or CT angiography. This intrinsic susceptibility for malignancy in patients with congenital anomalies reinforces the need for cardiac CT radiation dose optimization.^22,23 The “Have a Heart” campaign, launched by The Image Gently Alliance, reinforces this effort across all modalities in CHD utilizing ionizing radiation.²⁴ Dose optimization strategies should be uniformly applied in cardiac CT for CHD patients, many of whom do not need the image quality required for high resolution coronary imaging recommended by coronary CT scanner protocols.

Sedation and anesthesia was used in a minority of patients unable to cooperate with breath hold, due to young age or developmental delay. Average procedural time from localizer image to complete data acquisition was <5 min. Even when anesthesia and sedation is needed, the duration is short. The risk of sedation and anesthesia in pediatric patients with CHD is twofold. There is risk of procedural adverse events and the risk of adverse long term neurodevelopmental outcome. Prolonged or repeated use of anesthesia during rapid brain development may increase the risk of learning disability and behavioral problems.^2,6,25,26

There was variability in use of sedation and anesthesia in young children among centers. The center with the highest rate of sedation/anesthesia was the only center that routinely performed functional imaging in children, which requires several heart beats for image acquisition on the scanner platform used.

Only three (0.63%) AEs occurred in our patient subset of those < 1 year of age. This contrasts with the rate of AEs associated with CMR reported in other single site studies. In a single center study exploring AEs in CMR in CHD in patients less than 1 year of age, Dorfman and colleagues reported a rate 5.2% of AEs.²⁷ In a study analyzing risk among CHD patients undergoing CMR less than 4 months of age, 9% of the study cohort had AEs.²⁸ Other risks associated with IV access and contrast allergy were very low and consistent with prior literature.²⁹ It should be noted that the two pediatric patients who had contrast extravasation with small gauge IV had hand injection of contrast and that the routine use of power injection appears safe in this setting.

Diagnostic algorithms in CHD have not yet evolved to reflect the improved capabilities and lower risk profile of modern CT technology.^7–9,30 The multi-institutional data outlined in this manuscript should help inform practice of cardiac CT in CHD patients. Our report is unique in CHD imaging because of the significantly larger sample size than previous studies, the use of prospective and multi-institutional data, and the inclusion of a comprehensive assessment of diagnostic risk.

5. Limitations

This study is limited to the assessment of practice variability and comprehensive diagnostic risk of cardiac CT in CHD, but does not include image quality assessment. Coincidentally, a single CT vendor was used at all participating institutions. Sites were chosen due to their extensive experience of using cardiac CT in patients with CHD. Centers listed in our prior publication outlining the goals of this collaborative purposely included multiple scanner platforms. Centers using alternate scanner platforms were unable to participate. These initial registry results may not be reproducible by other centers without similar scanners or professional expertise.

Estimation of radiation dose for pediatric cardiac CT is challenging since there is no uniformly accepted standard for dose calculation. Many papers that quote radiation dose do not describe the calculation method, and the recent trend is to calculate patient specific doses that vary by patient size and age and by scanner tube output. These calculations increase the radiation dose estimate 5-fold for infants compared to historical standards. For this reason we have reported a generic “modality” dose that is patient age and size independent and is based on a 32 cm phantom with the most commonly used k-factor historically of 0.014. This dose estimate may be most appropriate to compare to historical dose reports and background radiation estimates that were published in that era. We additionally used an age and size adjusted dose to reflect a patient specific dose based on the updated conversion factors. Conversion factors specific to 70 kVp imaging are not yet available. There is not consensus on conversion factors for pediatric CT imaging and the estimates of radiation dose based on current recommendations may change in the future. A standard approach for determining radiation dose estimates for pediatric cardiac CT and for comparing doses betweencardiac CT and other modalities is needed. . Radiation dose estimates in this manuscript are for a single procedural CT and do not include cumulative risk for patients undergoing repeat diagnostics.

6. Conclusions

Our study shows relatively low comprehensive diagnostic risk for the use of cardiac CT in pediatric patients with congenital heart disease in the current era. The contemporary use of cardiac CT in pediatric patients with CHD is associated with a median effective procedural radiation dose of approximately 1 mSv. Radiation dose should no longer be considered a contraindication to cardiac CT when comprehensive risk is favorable compared to alternate modalities. This is particularly relevant when anesthesia risk of adverse event is high and the scan can be performed without sedation or with limited anesthesia due to fast procedural times. According to recent multi-society CHD imaging guidelines, CT is considered an appropriate modality for assessment of most forms of CHD, particularly for evaluation of a change in clinical status.³¹ In some cases, cardiac CT should be considered a first line diagnostic modality when echocardiography is insufficient for clinical care of patients with CHD if appropriate CT technology and professional expertise are available. Having all imaging modalities available at high quality allows a tailored approach to diagnostics in CHD, choosing the modality with the best information and least risk for the clinical question at hand.

Abbreviations and acronyms

AEs

adverse events

BSA

body surface area

CHD

congenital heart disease

CT

computed tomography

CTDIvol

volume computed tomography dose index

DLP

dose length product

ECG

electrocardiogram

ED

effective dose

gen-DS

generation dual source

ICRP

International Commission of Radiation Protection

IQR

interquartile range

IV

intravenous line

k

conversion factor

kg

kilogram

mSv

milliSievert

REDCap

Research Electronic Data Capture

SD

standard deviation

W/

with

w/o

without

yrs

years

Appendix 1. DLP for Anatomic Scans by Institution & Age Group

	Institution A	Institution B	Institution C	Institution D
All Patients	N = 347 26.2 (10.0, 58.0)	N = 470 7.0 (4.0, 18.1)	N = 362 11.1 (6.4, 20.0)	N = 49 17.0 (10.0, 33.0)
≤6 months	N = 113 8.4 (7.2, 10.2)	N = 224 3.9 (3.0, 5.0)	N = 155 6.0 (4.5, 8.0)	N = 17 8.0 (7.0, 11.0)
>6 months-2.5 years	N = 56 17.1 (11.0, 22.5)	N = 67 8.0 (6.0, 10.5)	N = 61 10.9 (8.7, 12.4)	N = 12 17.0 (12.5, 21.0)
>2.5–7.5 years	N = 44 31.3 (26.3, 39.2)	N = 66 16.1 (10.3, 19.3)	N = 50 17.6 (14.9, 19.0)	N = 10 25.0 (18.0, 36.0)
>7.5–12.5 years	N = 45 54.4 (36.6, 75.2)	N = 47 21.4 (14.7, 44.2)	N = 39 24.9 (18.9, 37.0)	N = 5 38.0 (33.0, 39.0)
>12.5 - <18 years	N = 89 81.0 (58.6, 108.1)	N = 66 45.4 (29.0, 76.5)	N = 57 48.5 (39.9, 73.4)	N = 5 51.0 (34.0, 87.0)

DLP, dose length product; IQR, interquartile range; All DLP (mGy*cm) values are reported as median and IQR.

Appendix 2. Clinical Indications for Study

Diagnosis	No.
Great Arteries Abnormalities	264 (25)
Coarctation of aorta	126
Vascular ring	88
Truncus	23
Interrupted aortic arch	14
Aortic aneurysm	11
AP Window	2
Coronary artery anomaly	184 (18)
Single ventricle	165 (16)
TOF/PA	159 (15)
Pulmonary atresia	81
Tetralogy of Fallot	78
Transposition Complex	133 (13)
d-TGA	70
DORV	59
L-TGA	4
Valve Disease	74 (7)
Aortic valve, LVOT	42
Pulmonary valve, RVOT	27
Mitral valve disease	3
Tricuspid valve anomalies	2
Other Vascular anomalies	67 (6)
PAPVR	29
TAPVR	29
Pulmonary vein stenosis	8
Systemic Venous Anomalies	1
Intracardiac shunt lesions	49 (5)
AVC	20
ASD	16
VSD	13
Other	28 (3)
Other left side obstructive lesions	11 (1)
Shones syndrome	7
Cor triatriatum	4

Reported as numbers and percent AP, aortopulmonary; TOF,tetralogy of Fallot; PA, pulmonary atresia; LVOT, left ventricle outflow tract; RVOT, right ventricular outflow tract; d-TGA, D-Transposition of the great vessels; DORV, double outlet right ventricle; L-TGA, L-looped transposition of the great arteries; PAPVR, partial anomalous pulmonary venous return; TAPVR, total anomalous pulmonary venous return; AVC, atrioventricular canal; ASD, atrial septal defect; VSD, ventricular septal defect.