The Best Parameters for Imaging Agent Injection and Scanning Methods in Computed Tomography Angiography

Patients All cases with suspicion of PE or AD referred for CTA in our hospital between March 2018 and February 2020 had been reviewed. Those cases with known allergic history, poor renal function (creatinine level>1.5 mg/dl), pregnant women or age less than twenty years old were excluded from the study. The investigators finally enrolled and analyzed 101 cases of CTA. The collected data included age, sex, blood pressure, body mass index (BMI), Hounsfield unit (HU) at bifurcation of PA, HU at the descending aorta with the same level of PA bifurcation, time to peak enhancement, and radiation dose. This study had been approved by the institutional review board in our hospital. All patients signed informed consent.

In addition, forty cases with suspicion of PE or AD performing traditional two-time CT scans were reviewed retrospectively. For controlling the dose of contrast media, the two-time CT scans in our hospital were performed with single 80ml contrast media injection at 3ml/sec, followed by 20ml normal saline injection at 3ml/sec. The CT PA angiography was performed when the density of the PA achieved baseline PA density plus 150 HU during tracking scans, followed immediately by the CT aortogram. The investigators compared the experimental groups and traditional groups.

Experimental designs of contrast media injection After the contrast injection, the contrast media goes to the right heart via peripheral veins and then the contrast is pumped out to the PA. The interval from contrast injection to the peak enhancement of PA is defined as P seconds. After pulmonary circulation, the contrast media travels through the pulmonary veins, the left atrium, and the left ventricle and contracts into the aorta. The interval from contrast injection to the peak enhancement of aorta is defined as A seconds. Afterward, the contrast media starts a new cycle of systemic circulation.

Based on the published data, for the pulmonary artery, adequate attenuation is known to be greater than 180 HU, and for the aorta, sufficient attenuation is known to be greater than 200 HU. However, the requirement of vascular enhancement for diagnosing PE is high because the size of the PA is significantly tapering and the thrombus in the distal PA branches will be vague if the enhancement of the PA is not vivid. Therefore, insufficient enhancement of the PA restricts the ability to diagnose PE. The investigators design a biphasic contrast media injection protocol, including a low rate(2ml/s) first phase injection and high rate(3ml/s) second phase injection, followed by high rate(3ml/s) normal saline injection. The first phase injection is kept for (A - P) seconds and the second phase injection is kept until the rest contrast media is completely injected. Theoretically, when the start time of the scan is at the A seconds, the first phase contrast media is about to significantly enhance the aorta, though the peak time and peak enhancement will not be optimal because the time intervals are calculated on the basis of high rate (3ml/s) injection. On the other hand, the second phase contrast media will maximally enhance the PA at the same time because the second phase contrast media has been injected for P seconds [A-(A-P)=P].

The pre-diagnostic axial CT images were set at the level of the PA bifurcation using 10ml test bolus contrast media injection at the 3ml/s rate. Radiology technologists in our team placed circular region of interest (ROI) within the areas of PA bifurcation and the descending aorta at the same level. The CT densities were collected and demonstrated the time-enhancement data. The peak enhancement time for the PA was chosen as P seconds and that for the aorta as A seconds.

The investigators designed four groups, including two groups using the test bolus (TB) method and two groups using the bolus-tracking (BT) method. The two TB groups were different at the total amount of contrast media with 80ml and 70ml respectively. The P and A seconds were calculated from the pre-diagnostic images and the start time of scan was fixed at A seconds. Based on the results of initial 22 cases in the two TB groups, the mean interval between P and A seconds was approximately 9.5 seconds. Therefore, we roughly assumed the A-P interval in most cases was about 10 seconds and designed the two BT groups with a fixed 10-second first phase contrast media injection time. The start time of scan was based on the tracking method when the descending aorta at the PA bifurcation achieved the baseline density plus 150 HU. The tracking scan started after contrast injection was initiated for 15 seconds. The two BT groups were also different at the total amount of contrast media with 80ml and 70ml respectively, including the 10ml for the pre-diagnostic images in order to have the same condition before the diagnostic CT images.

Details of the four experimental groups:

1. TB I: The investigators use the TB method with biphasic injection, followed by the saline flush. Initially, the investigators inject the 10 ml of test contrast media with a velocity of 3ml/s and apply the ROI at the bifurcation of PA and descending aorta at the same level. The dynamic curve demonstrates the time to peak enhancement of P second and A second. The first phase of contrast media injection uses a velocity of 2ml/s and the volume of contrast is measured as 2ml/s multiply (A-P) second. The second phase of contrast media injection uses a velocity of 3ml/s and the volume of contrast media is 70ml minus the amount of first phase injection. The total volume of the contrast media is 80ml, including the 10ml for pre-diagnostic test bolus images. The investigators performed the saline flush following the administration of the contrast media with 20 ml normal saline. The start time of the diagnostic scan is at the A second.
2. TB II: This group is similar to the TB I. However, the investigators use the lesser contrast media in the second phase of injection. The second phase of contrast media injection is applied with a velocity of 3ml/s and the volume of the contrast media is 60ml minus the amount of first phase injection. The total volume of the contrast media is 70 ml, also including the 10ml for pre-diagnostic test bolus images. The following saline flush uses 30 ml of normal saline. The start time of the diagnostic scan is at the A second as well.
3. BT I: Initially, 10ml of test contrast media was administrated but no calculation was performed for these pre-diagnostic test bolus images in this group. Unlike the TB method, the investigators use 10 seconds as the fixed interval of (A-P) in this group. The first phase of contrast media injection uses a velocity of 2ml/s and thus the volume of contrast is 20ml. The second phase of contrast media injection is administered with a velocity of 3ml/s and the volume of the contrast media is 70ml minus 20ml. The total volume of the contrast media is 80 ml. The following saline flush uses 20 ml of normal saline. The tracking scan started after contrast injection was initiated for 15 seconds. The ROI is placed in the descending aorta at the same level of PA bifurcation and the diagnostic CT scan is triggered when the density in the ROI achieves the baseline density plus 150HU.
4. BT II: This group is similar to the BT I. The first phase of injection is the same as the BT I group. The investigators use lesser contrast media in the second phase of contrast injection. The second phase of contrast media injection is applied with a velocity of 3ml/s and the volume of the contrast media is 60ml minus 20ml. The total volume of the contrast media is 70 ml, including the 10ml for pre-diagnostic test bolus images. The following saline flush uses 30 ml of normal saline. The protocol for triggering diagnostic CT scan is the same as that in BT I.

CT scan technique and contrast medium administration CT was performed using a 320-MDCT (Canon Aquilion PRIME CT scanner, Toshiba Medical Systems, Otawara, Japan). Prior to the initiation of the diagnostic CT scan, a series of low-radiation sequential axial pre-diagnostic CT images were acquired at a fixed level of the PA bifurcation while the test bolus of contrast media was being injected at 3ml/s with 10ml amount. This pre-diagnostic CT scan was applied for all patients. The scan parameters were as the following: 5 mm section thickness, 100kVp, automatic exposure control of current (mA), and 0.5 second rotation time. Following the completion of the pre-diagnostic CT scan, the patient was given a breath-hold instruction for the standard diagnostic CT using either TB or BT method. Scanning was performed in the cranial-caudal direction.

All cases were administrated contrast medium with 300 mg iodine (I)/ml Omnipaque using a dual-power injector (MEDRAD Stellant CT injector). All cases were administrated 10ml of contrast medium for establishment of pre-diagnostic CT images. The total volume of contrast media was fixed at 70 or 80 ml in accordance with the designed injection protocols. For all cases, a right upper limb antecubital venous access was achieved using an 18 or 20 Gauge venous catheter. The biphasic contrast injection protocol was conducted with two connected phases of contrast medium injection immediately followed by saline flush. The investigators performed the saline flush at the rate of 3ml/s following the administration of the contrast medium with different amounts (20 or 30 ml) to achieve 100ml fluid injection in total.

Estimation of radiation dose The dose length product (DLP) of all cases was recorded. Effective radiation dose was estimated by multiplying the DLP by a conversion factor. The effective radiation dose correlates with different parts of the body which varies the radiation sensitivity of the tissue. Our study applied conversion factor of 0.014 mSV/(mGy x cm) based on the region of chest and abdomen.

Statistical analysis:

A one-way analysis of variance (ANOVA) test was used for comparing the data of different groups. If the test of homogeneity of variances was not statistically different, Scheffe method was used for post hoc analysis. On the contrary, Games-Howell method was used for post hoc analysis. A p value of ≤0.05 was considered statistically significant. All statistical analysis was performed using software (SPSS version 25.0, IBM Corp.).