Multidetector computed tomographic angiography defines anatomy in complex congenital heart disease, but radiation exposure and general anesthesia requirements limit its application. The aim of this study was to compare radiation exposure, anesthesia use, and diagnostic accuracy between standard-pitch, single-source computed tomography and high-pitch, dual-source computed tomography for image quality and risk in a clinical pediatric population. Consecutive computed tomographic scans were evaluated in patients aged <2 years with complex congenital heart disease. Two groups were compared on the basis of standard- versus high-pitch scans. High-pitch scans were further divided into variable pitch (2.25 to 3.0) and highest pitch (3.4) groups. Image quality, radiation exposure, anesthesia use, and diagnostic confidence and accuracy were determined. Sixty-one scans were reviewed (29 at standard pitch, 32 at high pitch). Body surface area, scan length, and indications were similar. The median dose-length product for standard-pitch scans was 66 mGy · cm (range 29 to 372) compared to 7 mGy · cm (range 3 to 50) in all high-pitch scans. The median dose-length product was 28 mGy · cm (range 8 to 50) for variable high-pitch scans and 5 mGy · cm (range 3 to 12) for the highest fixed-pitch scans. Diagnostic confidence was similar, although high-pitch scans had higher image noise and lower contrast-to-noise ratios. All high-pitch scans were performed under sedation with free breathing, and all standard-pitch scans required general anesthesia. Diagnostic accuracy was 100% in the 2 groups, with 17 standard-pitch and 16 high-pitch patients undergoing procedural validation. In conclusion, high-pitch, dual-source computed tomography provides excellent diagnostic accuracy and markedly reduces radiation dose, although image quality is mildly reduced.
New second-generation, dual-source, high-pitch computed tomographic (CT) scanners may dramatically decrease radiation exposure risk and eliminate the requirement for breath holding in patients with complex congenital heart disease. In this study, we sought to assess the utility of such scanner technology by comparing sedation needs, image quality, diagnostic confidence and accuracy, and radiation doses between a standard-pitch, single-source, 64-slice multidetector CT angiographic (MDCTA) scanner and a second-generation, dual-source, 128-slice MDCTA scanner using a high-pitch sequence. We evaluated the youngest patients with complex congenital heart disease, those at highest risk for radiation exposure and general anesthesia.
Methods
We retrospectively reviewed 61 consecutive MDCTA scans from a single practice over a 2-year period (March 2008 to August 2010) in patients aged <2 years performed to evaluate complex congenital cardiovascular anatomy. Scans were divided into 2 groups on the basis of the scanner used. Group 1 scans (n = 29) were obtained using a Toshiba Aquilion 64-slice scanner (Toshiba Corporation, Tokyo, Japan) at a scan pitch of 0.8 and were performed at the Children’s Hospital of Minnesota from May 2008 to November 2009. Group 2 scans (n = 32) were obtained using a Siemens dual-source fast low-angle shot scanner (Siemens Medical Systems, Erlangen, Germany) at a scan pitch of 2.25 to 3.4 and were performed at the adjoining adult facility, Minneapolis Heart Institute, from November 2009 to August 2010. Group 2 high-pitch scan patients were then divided into 2 groups, those with scans done with variable pitch (2.25 to 3.0; n = 14) and those with scans obtained with the highest fixed pitch (3.4; n = 18). Institutional review board approval was obtained for retrospective review of a clinical quality improvement database. All patients underwent echocardiography and were referred by a pediatric cardiologist for clinically indicated MDCTA imaging for further definition of cardiac and vascular anatomy.
Single-source, standard-pitch, 64-slice MDCTA scans were performed under general anesthesia with suspended respiration during image acquisition lasting 7 to 15 seconds. A pediatric anesthesiologist was in attendance, and a peripheral intravenous line was placed for contrast injection. Images were acquired using a Toshiba Aquilion CT scanner (collimator 0.5 mm × 64, temporal resolution 200 ms, Z-axis coverage 32 mm/rotation for 1.0 pitch). Parameters such as tube voltage and tube current were adjusted according to patient size, with a fixed pitch of 0.8. All but 2 scans were performed without electrocardiographic gating. The scans were visually initiated without a timing bolus. Contrast injection was individualized to opacify the area of clinical interest. The scan field was adjusted to the region of interest.
Dual-source MDCTA scans were performed with free breathing. A pediatric anesthesiologist or neonatal nurse practitioner was in attendance to deliver conscious sedation through a peripheral intravenous line placed for contrast injection. Images were acquired using a high-pitch Siemens Definition dual-source fast low-angle shot scanner (collimator 0.6 mm × 128, temporal resolution 75 ms, Z-axis coverage 38.4 mm/rotation for a 1.0 pitch) with image acquisition lasting 0.3 to 0.5 seconds. Contrast injection was individualized to opacify the region of clinical interest. Localizing images were obtained in all patients, and a manual visual trigger was used during the arterial phase of contrast injection for the initiation of image acquisition. The scan range was limited to the area of interest. The initial (n = 14) dual-source patients had non–electrocardiographically gated variable-pitch (2.25 to 3.0) scans, on the basis of the reference tube voltage and tube current. For the later patients (n = 18), fixed-pitch (3.4) scans were used with electrocardiographic gating. For all high-pitch scans, automatic modulation of tube current according to patient size and anatomic region (CARE Dose4D; Siemens Medical Systems) was used.
The single- and dual-source scanner platforms use a standard 32-cm phantom for CT dose index estimates. By multiplying the scan length by the CT dose index volume, the dose-length product (DLP) was obtained and recorded from the scan protocol generated by the CT system. The effective radiation dose in millisieverts was obtained by multiplying the DLP by the organ weighting factor for the chest as the investigated anatomic region, thus estimating the lifetime attributable risk of that dose. Because of the evolving nature of the science with regard to dose and organ sensitivity in children, we calculated the effective dose using 2 different approaches. One effective dose was derived from the product of DLP and the standard organ weighting factor used for thoracic imaging (k = 0.014). This correction factor was used to determine scanner output that is comparable among all age ranges and scanner platforms (effective dose = k × DLP). A separate age-adjusted organ weighting factor was also calculated to adjust the radiation exposure to the age of the patient. To derive this effective dose, we multiplied the DLP by a conversion factor of 0.039 for patients aged ≤6 months and a factor of 0.026 for those aged 6 months to 2 years.
Two reviewers (K.H., J.L.) independently evaluated the multidetector CT angiograms for image quality according to the European Image Quality Assessment Score. Factors assessed were sharpness (1 = very sharp; 2 = questionable; 3 = noticeable blur, slice thickening), noise (1 = less than usual, 2 = optimal noise, 3 = noise affects interpretation), noise texture (0 = no noticeable change; 1 = after changing window setting, no noticeable change; 2 = perceptible change; 3 = change affects confident or blotchiness), and diagnostic confidence (1 = fully confident, 2 = probably confident, 3 = confident under limited conditions for visualization, 4 = unacceptable).
Objective indexes of image quality were determined by calculating image noise and contrast-to-noise ratio for each area of interest. Image noise was derived from the standard deviation of the density values (in Hounsfield units) within the region of imaging interest. The contrast-to-noise ratio was defined as the difference between the mean density of the contrast-filled structure and the mean density of the myocardial wall, which was divided by image noise.
Each MDCTA study was performed and tailored to answer a specific clinical question. For single-ventricle patients, the scans were performed to define arterial and venous anatomy before first- or second-stage palliation. For patients with tetralogy of Fallot and pulmonary artery atresia, the scans were performed to define pulmonary artery, aortopulmonary collaterals, and ductal anatomy for planning of the first catheter-based or surgical intervention before hospital discharge. For patients with vascular ring, aortic arch, and pulmonary venous anomalies, vascular and airway anatomy was defined. For patients who underwent subsequent catheter or surgical intervention, the intervention report was reviewed, and the pre- and postprocedural diagnosis were compared. Any discrepant or new diagnosis at the time of catheterization or surgical palliation was recorded.
All continuous variables (body surface area, DLP, effective dose, and scan range) are expressed as medians and ranges. The Kruskal-Wallis test was used to compare the median among the 3 groups; the method for comparisons between each 2 groups was the Mann-Whitney U test. All tests were 2 tailed. Statistical significance was defined as p <0.05. A third reader adjudicated any discrepant choices concerning qualitative or clinical assessment. In that unusual situation, the majority choice was used for analysis.
Results
The referral diagnoses for MDCTA imaging were similar in the 2 groups. Primary indications for angiography in group 1 versus group 2 were hypoplastic left heart syndrome or complex heterotaxy before first- or second-stage single-ventricle palliation (43% vs 47%), tetralogy of Fallot with aortopulmonary collaterals (22% vs 30%), complex aortic arch anatomy or vascular ring (30% vs 13%), and partial or total anomalous pulmonary venous return (4% vs 9%). Body surface areas were similar for the 2 groups (0.27 ± 0.1 vs 0.28 ± 0.09 m 2 ). The scan ranges in the 2 groups were similar between groups 1 and 2 (17 ± 3 vs 16 ± 3 cm, respectively). All patients in group 1 were under general anesthesia for breath-hold sequences of 7 to 15 seconds. One patient in group 2 was intubated for respiratory failure but was not paralyzed for suspended respiration, and all others were breathing freely during image acquisition of 0.3 to 0.5 seconds. There were no adverse events from anesthesia or sedation in either group.
The radiation dose for group 1 standard-pitch scans was significantly higher than for group 2 high-pitch scans (p <0.001; Table 1 ). The radiation dose for group 2 variable high-pitch (2.25 to 3.0) scans was significantly higher than for group 2 highest pitch (3.4) scans (p <0.001; Table 2 ). For all patients scanned at high pitch, this indicates a ninefold reduction in radiation dose compared to standard-pitch scanning. In addition, for all patients scanned with the highest fixed pitch (3.4), there was a fivefold decrease compared to the variable high-pitch (2.25 to 3.0) scans.
| Platform | Scans Performed | DLP (mGy · cm) | Effective Dose (mSv) | Age-Adjusted Effective Dose (mSv) | Length (cm) |
|---|---|---|---|---|---|
| Standard pitch MDCTA scan (0.8) | 29 | 66 (39–272) | 0.9 (0.4–3.8) | 2.1 (1.1–10.6) | 17 ± 3 |
| High pitch MDCTA scan (2.25–3.4) | 32 | 7 (3–50) | 0.11 (0.04–0.7) | 0.29 (0.1–1.9) | 16 ± 3 |
| Platform | Scans Performed | DLP (mGy · cm) | Effective Dose (mSv) | Age-Adjusted Effective Dose (mSv) | Length (cm) |
|---|---|---|---|---|---|
| Variable high pitch MDCTA scan (2.25–3.0) | 14 | 28 (8–50) | 0.39 (0.11–0.56) | 1.05 (0.3–1.95 | 16 ± 3 |
| Highest pitch MDCTA scan (3.4) | 18 | 4.5 (3–12) | 0.07 (0.04–0.11) | 0.20 (0.12–0.17) | 15 ± 3 |
Qualitative image quality was lower for the high-pitch scans. Sharpness was lower and noise level was higher for the high-pitch scans. Texture measurement (graininess) was also higher for the high-pitch scans. Despite this, qualitative measurement of diagnostic confidence was similar between groups, and all scans were judged as fully diagnostic ( Table 3 ).
