Dedicated infant and pediatric CT protocols typically incorporate low mA techniques, optimized to provide the best image quality, which is a critical part of reducing radiation exposure in children. There is a linear relationship between radiation dose and the tube current. In small children, the mA can be reduced to 80% of the adult value without compromise of image quality, especially for evaluation of fairly large targets such as dimensions of the dilated aortic root, coarctation, and branch pulmonary artery stenosis (1). But, reducing the tube current results in increased image noise. While it is tempting to use low doses for CT angiography, it should be noted that image noise mainly affects low-contrast resolution and evaluation of small structures such as coronary arteries, aortopulmonary collaterals and pulmonary veins, and could lead to a nondiagnostic study for such indications. Therefore, a cautious reduction of the mA must be undertaken, based on the indication for the study and the size of the patient. Numerous other factors, besides tube current, determine the amount of noise—the reconstruction method (360 degrees or 180 degrees), sharpness of kernels and filters, slice thickness, kilovoltage, beam filtering, sensitivity of the x-ray detectors, and quality of the amplifiers. In general, there is a trade-off between mAs, slice thickness, image noise, and spatial resolution. A decrease in diagnostic efficacy due to increased image noise may be offset to some extent by optimizing the contrast injection protocol and reduced respiratory and pulsation motion artifact.

Gantry cycle time (14) is closely related to mA, and a combination of both is mAs (milli-Ampere second). Again, there is a linear relationship between gantry rotation time and radiation dose. In situations where the contrast resolution is very good, as in first-pass angiographic studies, reducing the gantry rotation time to <0.5 seconds is an excellent means of reducing radiation dose as well as scan time.

The choice of slice collimation and table speed (pitch) is an important determinant of image quality, spatial resolution, and radiation exposure. A narrow x-ray beam collimation has the advantage of better spatial resolution along the z-axis, and reduced partial volume effects. A wider beam collimation has the advantage of less radiation dose and/or less image noise, resulting in better contrast resolution, and shorter scan duration. The choice of collimation setting depends on the clinical indication. Larger volumes are imaged with a large section thickness in order
to reduce the scan duration. If there is a need to depict fine anatomic detail, then collimation should be reduced. MDCT has a thinner collimation than single-slice CT, along with the advantage of shorter scan duration. In multislice volumetric imaging, thin slices (0.5 to 1.5 mm) may also be reconstructed retrospectively from the raw data.

Pitch (15,16) is a unitless ratio, originally defined for single-slice spiral CT, as table movement per 360-degree gantry rotation (mm) divided by x-ray beam collimation (mm). Scanning at a pitch less than 1 produces an overlapping scan pattern that increases the radiation dose to the patient, but offers slight advantages for the 3D reconstruction of contours that are roughly parallel to the scanning plane. Increasing the pitch from 1:1 to 2:1 reduces radiation exposure and scan duration by half, but results in a broadening of the slice sensitivity profile of approximately 30%, with the consequence of a higher degree of partial voluming. Higher pitch also has a consequence of reducing z-axis resolution.

In MDCT, there are two definitions of pitch (17), depending on whether a single section collimation (pitch P) or the total collimation of the detector array (volume pitch P*) is chosen as the reference. While the physicists prefer the former, most manufacturers use the latter definition. While P is independent of the number of detector rows, P* increases as the number of detector rows grows. Pitch is a useful marker of image quality. High pitch values degrade image quality due to increased noise, slice broadening and artifacts, but also result in faster coverage and reduced radiation.

The relationship between pitch and slice width is simple and easy to understand in single-slice CT. As pitch increases, so does effective slice width. But, the relationship is complex in multislice CT. By using different slice reconstruction algorithms and an optimal choice of pitch, a thinner effective slice width may be obtained. Thinner slices can also be traded for a thicker slice with less image noise.

While it is tempting to default to a pitch of less than or equal to 1 in order to obtain the best image resolution, there are instances when a shorter scan time is desirable to avoid patient motion, or when the indication does not require a very high spatial resolution. In such cases, pitches of 1.5, 1.7, and 2.0 may be used with resultant savings in scan time and radiation exposure (16). In larger children, changes in image quality between a pitch of 1 and 1.5 are barely discernible (15). Decreased collimation and table increments are reserved for detailed examinations.

This refers to the extent of the body that is covered by the scan. The volume of coverage is directly proportional to radiation dose. This is a powerful tool to reduce radiation exposure in children. Restricting the CT examination to the volume of interest is only possible in cases in which the pathology is either known or likely to be present only in one specific part of the chest. For indications like anomalous coronary artery origin, branch pulmonary artery stenosis, pulmonary vein stenosis, and evaluation of stent patency, the scanning volume can be restricted to the structure of interest. Thus, only one-half to one-third of the chest is exposed to radiation, with corresponding reduction in scan time. Figure 15.1 is a good example of a restricted scan range in a 3-year-old patient with tetralogy of Fallot
and RV-PA conduit placement, bilateral branch pulmonary artery stenosis status post stenting, and severe tracheobronchomalacia, where the field of view was limited to evaluation of the pulmonary arteries and the proximal airway. On the other hand, for indications like heterotaxy (evaluation of the systemic and pulmonary veins, as well as cardiac morphology) (Fig. 15.2), scimitar syndrome, lung anomalies, intracardiac anomalies, aortopulmonary collaterals, or hereditary hemorrhagic telangiectasia, the entire thorax, as well as parts of the neck and abdomen will have to be covered by the scan. Increasing the pitch, and reducing the rotation time can offset the increased scan time and radiation exposure associated with the large volume of coverage.

A high temporal resolution is needed to freeze cardiac motion and avoid artifacts, especially with 64-detector scanners. The 0.3- to 0.4-second gantry rotation time in most modern scanners, combined with half scan reconstruction, and optional multisector reconstruction result in excellent temporal resolution in multislice CT. ECG synchronization of data is necessary to capture the heart in a motionless phase. With single sector reconstruction, only data from the prescribed time range during one cardiac cycle are used for partial scan reconstruction of images. This yields a temporal resolution of about 200 ms for a 0.4-second gantry rotation time. Multisector reconstruction (18) (segmented reconstruction) can increase the temporal resolution by using scanned data from more than one heart cycle for image reconstruction. A temporal resolution of 50 to 65 ms can be reached with up to four sectors. The more sectors that are used, the higher the overlap during data acquisition has to be. This requires lower pitches and results in an increased patient dose. Multisector reconstruction yields the shortest effective scan time, but also needs the highest radiation dose to accomplish this goal.

Non-ECG–gated scanning can provide diagnostic images of the extracardiac vasculature in most patients with CHD. For evaluation of the aorta, pulmonary artery and pulmonary veins, ECG synchronization is usually not necessary. ECG gating is the preferred technique for evaluation of the morphology of the heart chambers, including assessment of ventricle aneurysms, cardiac thrombi, cardiac tumors, evaluation of small aortopulmonary collaterals in pulmonary atresia, and for coronary artery assessment (19). In prospective ECG gating, a prospective trigger is derived from the ECG signal to initiate scanning during a phase of little heart movement. The advantage of prospective ECG gating over nongated scanning is the lower radiation dose, because all data that are acquired are also used for image reconstruction. Prospective ECG gating uses information from the preceding RR interval in order to calculate the trigger delay. The trigger can be selected by the user as a percentage of the RR interval (usually between 40% and 80%) or as a fixed time (absolute delay) after the preceding R peak or before the next R peak. For lower heart rates, 70% to 80% percent of the RR interval yields the best results. But, with higher heart rates (>90 beats per minute); 30% to 40 of the RR interval may yield better results (20). There is one caveat with the use of prospective ECG gating, as in children with high heart rate variability, the starting phase of data acquisition does not precisely match with the preselected starting phase. Temporal padding overcomes this limitation by increasing acquisition time before and after the predicted acquisition window. This technique permits retrospective modification of the reconstruction window to ensure image reconstruction within low motion and identical cardiac phases from one cardiac cycle to another. However, this technique will increase the radiation burden to the patient. In addition, this technique could not be utilized in helical scanning. The acquisition of temporally uniform volumetric data sets with 320-detector scanner enables retrospective half scan reconstruction within different cardiac phases in raw data. This post-acquisition data processing technique, called Target Mode, finds and reconstructs the best motion-free cardiac phase in raw data without increasing the radiation dose to the patient (1). In retrospective ECG gating technique, the tube current is on during the whole cardiac cycle, for approximately five heartbeats using a 64-multislice system, so that data can be reconstructed during the best motion-free phases. This boosts the radiation dose three to four times compared to a nongated examination. The radiation dose can be lessened by using ECG-based tube current modulation that reduces the beam strength during nonoptimal cardiac phases. Retrospective ECG-gated scanning is the technique of choice for evaluation of cardiac function that needs multiphasic images, and for coronary artery assessment in patients with irregular or high hear rates.