- Email: [email protected]
Multiphase Contrast Medium Injection For Optimization Of Computed Tomographic Coronary Angiography
Multiphase Contrast Medium Injection For Optimization Of Computed Tomographic Coronary Angiography1 Matthew Jay Budoff, Jerold S Shinbane, Janis Child, Sivi Carson, Alex Chau, Stephen H. Liu, SongShou Mao
Rationale and Objectives. Electron beam angiography is a minimally invasive imaging technique. Adequate vascular opacification throughout the study remains a critical issue for image quality. We hypothesized that vascular image opacification and uniformity of vascular enhancement between slices can be improved using multiphase contrast medium injection protocols. Materials and Methods. We enrolled 244 consecutive patients who were randomized to three different injection protocols: single-phase contrast medium injection (Group 1), dual-phase contrast medium injection with each phase at a different injection rate (Group 2), and a three-phase injection with two phases of contrast medium injection followed by a saline injection phase (Group 3). Parameters measured were aortic opacification based on Hounsfield units and uniformity of aortic enhancement at predetermined slices (locations from top [level 1] to base [level 60]). Results. In Group 1, contrast opacification differed across seven predetermined locations (scan levels: 1st versus 60th, P ⬍ .05), demonstrating significant nonuniformity. In Group 2, there was more uniform vascular enhancement, with no significant differences between the first 50 slices (P ⬎ .05). In Group 3, there was greater uniformity of vascular enhancement and higher mean Hounsfield units value across all 60 images, from the aortic root to the base of the heart (P ⬍ .05). Conclusions. The three-phase injection protocol improved vascular opacification at the base of the heart, as well as uniformity of arterial enhancement throughout the study. Key Words. Contrast medium; coronary artery imaging; CT angiography; electron beam CT; noninvasive angiography. ©
AUR, 2006
Computed tomographic (CT) coronary angiography is being increasingly used for assessment of coronary artery disease (1– 4). Electron beam angiographic (EBA) image quality is crucial to visualization of the coronary arteries and is dependent on numerous factors, including vascular opacification as measured by the number of CT Hounsfield Units (HU) and uniformity of arterial enhancement on serial image slices. Studies have been performed using single-phase injection protocols with constant injection rates (1– 4). These injection
protocols result in nonuniformity of aortic enhancement as measured by CT Hounsfield numbers, with a single peak value in the mid-portion of the study and lower values at the beginning and end of the study (5– 8). Theoretically, opacification is good if, for a given volume of injected contrast medium, the Hounsfield values of interest are high. The aim of this study was to compare three different injection protocols with respect to the degree of aortic enhancement achieved and to enhancement uniformity.
Acad Radiol 2006; 13:159 –165 1 From the Division of Cardiology, Los Angeles Biomedical Research Institute at Harbor-UCLA, 1124 W. Carson Street, RB2, Torrance, CA 90502. Received July 10, 2005; accepted September 21, 2005. Address correspondence to: M.J.B. e-mail: [email protected]
© AUR, 2006 doi:10.1016/j.acra.2005.09.087
METHODS Study Population We enrolled 244 consecutive patients (58 women; age range 27–91, mean 63 years) who were referred for EBA
159
BUDOFF ET AL
Academic Radiology, Vol 13, No 2, February 2006
Table 1 Injection Protocols
Injection Protocols Group 1 Group 2 Group 3
Injection Phase 1 Injection Rate
Injection Phase 2 Injection Rate
Injection Phase 3 Injection Rate
Contrast 3.5–4 mL/second Contrast 4–5 mL/second Contrast 4–7 mL/second
None Contrast 2.5–4 mL/second Contrast 2.5–4 mL/second
None None Saline 3 mL/second
for clinical evaluation of coronary artery disease by their cardiologist. A volume of 120 –180 mL of nonionic contrast medium was used with three different injection protocols (Table 1): a single injection phase with a constant injection rate (Group 1), a two-phase injection with two different injection rates (Group 2), and a three-phase injection composed of two contrast medium injections with two different rates, followed by a saline injection at constant rate (Group 3). All injections were made into an antecubetal vein. Contrast concentrations (CT numbers) in the aorta were measured and results were compared across the three groups. Parameters measured were aortic opacification based on Hounsfield units and uniformity of aortic enhancement at predetermined slices (locations from top [level 1] to base [level 60]). This study was approved by the Institutional Review Board of the Los Angeles Biomedical Research Institute at Harbor-UCLA. EBT angiography studies: EBA studies were performed using an Imatron C-150XLP computed tomographic scanner (San Francisco, CA). First, a non-contrast 40-slice study (3 mm slice thickness, 3 mm table increment, 100 milliseconds acquisition time) was obtained with the patients supine and no couch angulations. Subsequently, a flow study was performed to estimate the mean transit time of the contrast agent (scan delay time) to the descending aorta. These methods have been previously described (9,10). Finally, up to 60 contiguous axial EBA images were acquired to cover the entire cardiac tree with 1.5 mm slice thickness and 1.5 mm table incrementation. A volume of 120 –180 mL of contrast medium (Iopamidol 370, Bracco Diagnostics, Plainsboro, NJ) was used with injection protocols as described in the following section. Electrocardiographic triggering was used, corresponding to end systole as previously described (10). The image acquisition time was 100 milliseconds per image and total scan time was 35–55 seconds (mean 44.5 seconds). Three injection protocols were employed in this study. The injection rate is dependent on the scan time (based on heart rate and slice number). For patients in Group 1,
160
a single phase injection with a 3.5– 4 mL/seconds of fixed contrast medium injection rate was used, with a mean injection time of 44.5 seconds. The injection was completed, on average, at the 34th image. In Group 2, a dualphase contrast medium injection was employed. The first contrast medium injection rate was 4 –5 mL/seconds with 60 mL of contrast medium, followed by a second contrast medium injection rate of 2.5– 4 mL/second. The injection time was 50 seconds, with the injections completed on average at the 40th slice image, with variation from heart rate and number of images required to image the entire heart. In Group 3, two sequential contrast medium injections were followed by a saline injection. The contrast medium injection rate of the first injection was 4 –5 mL/ seconds and 2.5– 4 mL/seconds for the second contrast medium injection. The contrast medium injection time was equal to image acquisition time for 60 slices (the study acquisition time). For slower heart rates (longer acquisition times), the slower rates of contrast injection were used. The contrast injection was followed by a saline injection of 60 milliliters of dextrose 5% with halfnormal saline matching the second injection speed and continued to scan end. In Groups 1 and 2, a single injector (Envision Injector, MedRad, Inc, Indianola, PA) was used. In Group 3, a dual injector (Stellant Dual Injector, Medrad, Inc) was used, with one injector head for administration of contrast medium and the other for saline injection. Arterial opacification based on CT numbers of the descending aorta were measured at seven prespecified levels in each study: 1st, 10th, 20th, 30th, 40th, 50th, and 60th slices obtained (Insight Workstation, NeoImagery Technologies, City of Industry, CA). Recirculation To assess for the effect of contrast medium recirculation, we analyzed a subset of 30 time/density curve studies. Each contrast enhancement peak (highest CT numbers) was assessed to measure the amount and effect of
Academic Radiology, Vol 13, No 2, February 2006
OPTIMIZATION OF CT CORONARY ANGIOGRAPHY
Figure 1. Comparison of computed tomography (CT) Hounsfield units (HU) among Groups 1–3. There are significant differences in images 50 – 60 and total values, with a higher CT HU in group 3 than Groups 1 and 2 (P ⬍ .05).
recirculating contrast. Each time/density curve peak was assessed for maximum CT numbers and the timing of the peak (in seconds) in reference to the first peak and start of injection. The effect of blood laminar flow and recirculation was assessed, and the progressive drop of CT numbers was evaluated. Statistical Analyses All comparisons between the three injection protocols were achieved using the 1st (top) to 60th slice (bottom level) of each EBA study. Prespecified comparisons were set to compare the results of the slice with peak opacification as measured by CT HU to both the top slice (1st level) and bottom slice (60th level). Continuous variables with normal distributions were compared using Student’s t-test, as well as one-way analysis of variance tests. The difference between both measurements was represented as absolute difference ⫽ (abs (A ⫺ B)/(0.5 ⫻ A ⫹ 0.5 ⫻ B value) ⫻ 100%. A probability level of 5% (0.05) was used to determine statistical significance.
RESULTS There were no significant differences among the three patient groups based on factors that affect image opacification, including body surface area, contrast/kilogram, and age or gender of each group (Table 1). Additionally, the heart in its entirety was visualized within slices 1– 60
in all cases. Aortic opacification based on CT numbers for each slice decile in Groups 1–3 are listed in Table 3. Group 1 (Single-Phase Contrast Injection) The aortic opacification CT numbers were 278.8, 300.7, 321.9, 331.5, 310, 261.3, and 226.3 (overall mean 290.1) at the 1st, 10th, 20th, 30th, 40th, 50th, 60th slices, respectively. Opacification peaked halfway through the study (level 30). Opacification was significantly higher on level 30 than level 60 (P ⬍ .05; Table 3, Fig 1). There were significant differences in aortic opacification CT numbers between the 1st and 30th levels and between 30th and 60th levels of 19% and 37%, respectively, (P ⬍ .001), demonstrating nonuniformity in vascular opacification between slice levels. Group 2 (Dual-Phase Contrast Injection) The aortic opacification CT numbers were 297.5, 292.7, 302.2, 308.5, 311.4, 286.7, and 241.6 (overall mean 291.6) at the 1st, 10th, 20th, 30th, 40th, 50th, and 60th slices, respectively. There were no significant differences in CT numbers from the 1st to 50th slice (P ⬎ .05), but significant differences did exist between the 1st and 50th slice and the 60th slice (P ⬍ .05). This demonstrated that dual-phase contrast medium injection increased the uniformity of CT numbers throughout the majority of the study, but that CT HU decreases during the final slices of the study were still problematic.
161
BUDOFF ET AL
Academic Radiology, Vol 13, No 2, February 2006
Table 2 Data of the Patients Group
Number
Heart Rate
Delay Time
Weight (kg)
Height (cm)
BSA (m2)
Contrast Dose (ml/kg)
1 2 3 Analysis of variance
105 70 70
80.2 81.4 80.2 NS
20.5 19.4 21 NS
66.8 67.5 68.5 NS
179.1 180.9 176.3 NS
1.95 1.95 1.94 NS
2.62 2.60 2.55 NS
Table 3 Comparison of CT Numbers of 60 Slices in 3 Protocol Groups Slice Orders Group
1st
10th
20th
30th
40th
50th
60th
Mean
1 2 3 Analysis of variance
278.8* 397.5 314.4 P ⬍ .05
300.7 292.7 312.5 P ⬎ .05
321.9 302.2 313.2 P ⬎ .05
331.5* 308.5* 323.3 P ⬍ .05
310.0 311.4 326.4 P ⬎ .05
261.3* 286.7* 312.2 P ⬍ .05
226.3* 241.6* 274.7 P ⬍ .05
290.1* 291.6* 310.2 P ⬍ .05
*As compared with mean values in Group 3 for each level, these measures are significantly lower. Analysis of variance values shown for each level.
Group 3 (Two Sequential Contrast Medium Injection Phases Followed by a Saline Injection Phase) The aortic opacification CT numbers were 314.4, 312.5, 313.2, 323.3, 326.4, 312.2, and 274.7 (overall mean 310.2) at the 1st, 10th, 20th, 30th, 40th, 50th, and 60th slices, respectively. As in Group 2, there were no significant differences within the 1st to 50th measurements (P ⬎ .05) but a lower CT number at the 60th slice image compared with slices 1–50 (P ⬍ .05). There was though a significant larger mean CT numbers of aortic images in Group 3 as compared with Groups 1 and 2 (P ⬍ .05; Table 2, Fig 1), demonstrating that contrast opacification can be improved through the use of two phases of contrast medium injection followed by a saline injection phase. Recirculation To assess for recirculation of contrast-enhanced blood, we analyzed a subset of 30 time/density curve studies (from Group 3), with discrete second and third peak flows. The second and third peaks of contrast enhancement at the level of the superior vena cava were 5–15 seconds after primary peak (mean 7.7 seconds) and 16 –30 seconds (mean 22.3 seconds), respectively, representing recirculation from the head and neck, thorax, and upper extremity veins (Fig 2). At the insertion of the infe-
162
rior vena cava, the second and third peak times were 8 –11 seconds (mean 9.8 seconds) and 22–25 (mean 22.8 seconds) after the first peak respectively, representing recirculation from the abdomen, pelvis, and lower extremities (Fig 2).
DISCUSSION CT angiography is increasingly being used as a minimally invasive technique for visualization of the coronary arteries (1– 4). The EBA axial images requires iodinated contrast medium enhancement, administered by venous injection. The quality of three-dimensional rendered images is dependent on the quality of the original axial images. Target vessel opacification and motion artifact reduction are important factors in the acquisition of optimal axial images. The number of axial image necessary to image the heart in its entirety is 40 – 60 slices, requiring 30 – 45 second acquisition times. The prolonged acquisition time results in nonuniform opacification of the target vessels, which can also be seen in multidetector CT angiography (5– 8). Newer scanners with greater collimation width (now up to 40 mm per rotation), make this less problematic; however, these scanners are increasingly being used to cover more anatomy (entire pulmonary and coronary vasculature, aortic or peripheral imaging), resulting in similar opacification issues.
Academic Radiology, Vol 13, No 2, February 2006
OPTIMIZATION OF CT CORONARY ANGIOGRAPHY
Figure 2. This graph displays the recirculation time in 30 selected flow studies. At the SVC (superior vena cava), the timing of the second and third peak time is 5–15 seconds and 16 –30 seconds after the first peak time, respectively. At the IVC (inferior vena cava), the second and third peaks are 8 –11 seconds and 22–25 seconds after the first peak, respectively.
Several factors contribute to nonuniformity of vascular enhancement. Venous circulation blood flow is a low pressure and low velocity system, with significant variation in blood flow velocity from heart rate, right atrial pressure, left ventricular diastolic pressure, and patient postural position. An additional factor contributing to variation of vessel opacification is laminar blood flow characteristics. When blood flows through a vessel, there is a higher velocity in the center of the vessel as compared to the lumen wall (11). This nonuniformity of blood velocity from laminar flow results in increasing or decreasing opacification with different injection rates. Claussen and colleagues reported that injection in the vena cava (rather than the peripheral intravenous injections more commonly used) can decrease peak time, decrease laminar flow, and increase coronary artery enhancement (12). Another important reason for nonuniformity of vascular enhancement is recirculation of contrast-enhanced medium from different organ systems. To our knowledge, no detailed information has been reported regarding recirculation time from different venous circulations. Because of laminar flow and recirculation of blood, the CT HU increases progressively from the 1st to the 30th– 40th images (Fig 2). After the injection is completed, pressure in the injected vein falls, blood flow of the respective upper extremity slows, and the CT HU decreases, although recirculation of contrast-enhanced blood continues. Table 2, Figure 3 displays the effect of blood laminar flow and recirculation based on these study results. After the con-
Figure 3. This graph displays the time by blood laminar flow effect and recirculation from major organs.
trast medium injection is stopped, the CT number decreases smoothly and progressively because of laminar flow and recirculation. To decrease this variation and prolong the opacification time, two or more injection speed protocols have been reported for vascular CT angiograms (5–7) and other organ enhancement (13–15). These injection protocols obtained a more uniform result; however, data specifically relating to CT coronary angiography have not been reported. Dual-phase contrast medium injection protocols were used in our study, with a high injection rate (4 –5 mL/
163
BUDOFF ET AL
Academic Radiology, Vol 13, No 2, February 2006
Figure 4. After the contrast injection ends, the computed tomography Hounsfield units of aorta decreases progressively from fourth slice.
seconds, 12–15 seconds) before contrast recirculation, followed by a decreased injection rates (2.5– 4 mL/seconds) to lengthen the injection time to match the study time (40 – 45 seconds). We found greater uniformity of aortic CT numbers on the 1st to 50th slice images compared with a single-phase injection, but there was still a significant decrease after slice 50. The three-phase injection protocol, however, did demonstrate a significantly higher CT HU of the 50th– 60th slices compared with the single-phase injection or the a two-phase injection without saline injection (P ⬍ .05). Because of the loss of venous pressure after injection and the issue of wasted contrast medium present in the tubing and arm vein of the patient during imaging, we used a saline injection to maximize distal contrast opacification without increasing contrast dose to the patient. We found that the CT numbers decreased progressively starting from the third to fourth heart beat after cessation of contrast medium injection (Fig 4). Because the decrease in CT numbers was related to loss of injection pressure, a protocol for maintaining injection pressure and washing contrast medium out of the arm veins and tubing was devised using a dual injector system (Stellant, Medrad, Indianola, PA). With saline injection (Group 3), the mean CT HU of all levels was significantly higher than in Groups 1 and 2, and were significantly higher in the distal levels (images 50 and 60). Therefore, saline injection is an effective method for improving vascular opacification in CT angiography studies. Hopper and colleagues reported use of an injector with contrast agent and saline in one syringe, relying on the density difference to keep the solutions separate (16).
164
Other authors have reported that two injectors were used to deliver contrast medium and saline automatically (17,18). The aims of these CT studies (6,7,16 –18) were to save contrast medium dose and decrease the artifact of superior vena cava enhancement with conventional CT. All of the previously mentioned studies used shorter injection time (20 –25 seconds) and smaller contrast doses (⬍75 mL) than the current study. Because angiography using thin slice CT and EBA require longer injection times, the three-phase injection protocol with sequential contrast injections followed by saline injection is an important method for improving image quality. A disadvantage of the use of two separate injectors or layering saline on top of contrast is the complexity of determining injection pressures and timing delays. The new dual injector with two syringes and one control system used in our center obviates some of those difficulties.
LIMITATIONS There was a significant peak in CT HU at the 40th level and lower HU at the 60th level in all three groups, despite two phases of contrast injections followed by saline injection. For cases requiring longer injection protocols (slower heart rates or more elongated hearts), a third contrast injection phase may be necessary. Furthermore, this study addressed only the ability to increase vascular opacification density and slice uniformity, not the diagnostic accuracy of the coronary artery images obtained. This study did not explore variables such as optimal injection duration or contrast dose. These factors are impor-
Academic Radiology, Vol 13, No 2, February 2006
Figure 5. Volume rendering of a electron beam angiogram demonstrating excellent visualization of distal right coronary artery and posterior descending artery (PDA) with good opacification of lower levels (computed tomography Hounsfield units ⬎300).
tant because short injection times can lead to a lower CT HU at distal levels and produce non-diagnostic images (19). Because distal vessels are well seen with EBA given good opacification (Fig 5), there is a need to lengthen the injection time to ensure constant opacification, which needs to be balanced against total contrast dose. Nonetheless, use of dual contrast injection increased HU uniformity until the last 10 slices of the study and the addition of saline flush lead to lower contrast medium requirements (taking advantage of contrast in the arm and superior vena cava at the conclusion of injection) and higher opacification HUs at all of the levels of the CT study. ACKNOWLEDGMENT
The authors thank David Hill for his instruction, insight, and assistance, and Medrad, Inc., for providing a Stellant CT injector for evaluation. REFERENCES 1. Budoff MJ, Achenbach S, Duerinckx A. Clinical utility of computed tomography and magnetic resonance techniques for noninvasive coronary angiography. J Am Coll Cardiol 2003; 42:1867–1878.
OPTIMIZATION OF CT CORONARY ANGIOGRAPHY
2. Ha JW, Cho SY, Shim WH, et al. Noninvasive evaluation of coronary artery bypass graft patency using three-dimensional angiography obtained with contrast-enhanced electron beam CT. AJR Am J Roentgenol 1999; 172:1055–1059. 3. Budoff MJ, Lu B, Shinbane JS, et al. Methodology for improved detection of coronary stenoses with computed tomographic angiography. Am Heart J 2004; 148:1085–1090. 4. Gerber TC, Kuzo RS, Karstaedt N, et al. Current results and new developments of coronary angiography with use of contrast-enhanced computed tomography of the heart. Mayo Clin Proc 2002; 77:55–71. 5. Bae KT, Tran HQ, Heiken JP. Multiphasic injection method for uniform prolonged vascular enhancement at CT angiography: pharmacokinetic analysis and experimental porcine model. Radiology 2000; 216:872– 880. 6. Fleischmann D, Rubin GD, Bankier AA, et al. Improved uniformity of aortic enhancement with customized contrast medium injection protocols at CT angiography. Radiology 2000; 214:363–371. 7. Hittmair K, Fleischmann D. Accuracy of predicting and controlling timedependent aortic enhancement from a test bolus injection. J Comput Assist Tomogr 2001; 25:287–294. 8. Fleischmann D, Hittmair K. Mathematical analysis of arterial enhancement and optimization of bolus geometry for CT angiography using the discrete Fourier transform. J Comput Assist Tomogr 1999; 23:474 – 484. 9. Bakhsheshi H, Mao SS, Budoff MJ, et al. Preview method for electronbeam CT scanning of the coronary arteries. Acad Radiol 2000; 7:620 – 626. 10. Mao S, Budoff MJ, Lu B, et al. Optimal ECG trigger point in electronbeam CT studies: three methods for minimizing motion artifacts. Acad Radiol 2001; 8:1107–1115. 11. West JB. Best and Taylor’s physiological basis of medical practice. Baltimore: Williams & Wilkins, 1991. 12. Claussen CD, Banzer D, Pfretzschner C, et al. Bolus geometry and dynamics after intravenous contrast medium injection. Radiology 1984; 153:365–368. 13. Keogan MT, McDermott VG, Paulson EK, et al. Pancreatic malignancy: effect of dual-phase helical CT in tumor detection and vascular opacification. Radiology 1997; 205:513–518. 14. Heiken JP, Brink JA, McClennan BL, et al. Dynamic contrast-enhanced CT of the liver: comparison of contrast medium injection rates and uniphasic and biphasic injection protocols. Radiology 1993; 187:327– 31. 15. Bonaldi VM, Bret PM, Atri M, et al. A comparison of two injection protocols using helical and dynamic acquisitions in CT examinations of the pancreas. AJR Am J Roentgenol 1996; 167:49 –55. 16. Hopper KD, Mosher TJ, Kasales CJ, et al. Thoracic spiral CT: delivery of contrast material pushed with injectable saline solution in a power injector. Radiology 1997; 205:269 –271. 17. Haage P, Schmitz-Rode T, Hubner D, et al. Reduction of contrast material dose and artifacts by a saline flush using a double power injector in helical CT of the thorax. Am J Roentgenol 2000; 174:1049 –1053. 18. Irie T, Kajitani M, Yamaguchi M, et al. Contrast-enhanced CT with saline flush technique using two automated injectors: how much contrast medium does it save? J Comput Assist Tomogr 2002; 26:287–291. 19. Mao S, Takasu J, Child J, et al. Comparison of LV mass and volume measurements derived from electron beam tomography using cine imaging and angiographic imaging. Int J Cardiovasc Imaging 2003; 19: 439 – 445.
165
Rationale and Objectives. Electron beam angiography is a minimally invasive imaging technique. Adequate vascular opacification throughout the study remains a critical issue for image quality. We hypothesized that vascular image opacification and uniformity of vascular enhancement between slices can be improved using multiphase contrast medium injection protocols. Materials and Methods. We enrolled 244 consecutive patients who were randomized to three different injection protocols: single-phase contrast medium injection (Group 1), dual-phase contrast medium injection with each phase at a different injection rate (Group 2), and a three-phase injection with two phases of contrast medium injection followed by a saline injection phase (Group 3). Parameters measured were aortic opacification based on Hounsfield units and uniformity of aortic enhancement at predetermined slices (locations from top [level 1] to base [level 60]). Results. In Group 1, contrast opacification differed across seven predetermined locations (scan levels: 1st versus 60th, P ⬍ .05), demonstrating significant nonuniformity. In Group 2, there was more uniform vascular enhancement, with no significant differences between the first 50 slices (P ⬎ .05). In Group 3, there was greater uniformity of vascular enhancement and higher mean Hounsfield units value across all 60 images, from the aortic root to the base of the heart (P ⬍ .05). Conclusions. The three-phase injection protocol improved vascular opacification at the base of the heart, as well as uniformity of arterial enhancement throughout the study. Key Words. Contrast medium; coronary artery imaging; CT angiography; electron beam CT; noninvasive angiography. ©
AUR, 2006
Computed tomographic (CT) coronary angiography is being increasingly used for assessment of coronary artery disease (1– 4). Electron beam angiographic (EBA) image quality is crucial to visualization of the coronary arteries and is dependent on numerous factors, including vascular opacification as measured by the number of CT Hounsfield Units (HU) and uniformity of arterial enhancement on serial image slices. Studies have been performed using single-phase injection protocols with constant injection rates (1– 4). These injection
protocols result in nonuniformity of aortic enhancement as measured by CT Hounsfield numbers, with a single peak value in the mid-portion of the study and lower values at the beginning and end of the study (5– 8). Theoretically, opacification is good if, for a given volume of injected contrast medium, the Hounsfield values of interest are high. The aim of this study was to compare three different injection protocols with respect to the degree of aortic enhancement achieved and to enhancement uniformity.
Acad Radiol 2006; 13:159 –165 1 From the Division of Cardiology, Los Angeles Biomedical Research Institute at Harbor-UCLA, 1124 W. Carson Street, RB2, Torrance, CA 90502. Received July 10, 2005; accepted September 21, 2005. Address correspondence to: M.J.B. e-mail: [email protected]
© AUR, 2006 doi:10.1016/j.acra.2005.09.087
METHODS Study Population We enrolled 244 consecutive patients (58 women; age range 27–91, mean 63 years) who were referred for EBA
159
BUDOFF ET AL
Academic Radiology, Vol 13, No 2, February 2006
Table 1 Injection Protocols
Injection Protocols Group 1 Group 2 Group 3
Injection Phase 1 Injection Rate
Injection Phase 2 Injection Rate
Injection Phase 3 Injection Rate
Contrast 3.5–4 mL/second Contrast 4–5 mL/second Contrast 4–7 mL/second
None Contrast 2.5–4 mL/second Contrast 2.5–4 mL/second
None None Saline 3 mL/second
for clinical evaluation of coronary artery disease by their cardiologist. A volume of 120 –180 mL of nonionic contrast medium was used with three different injection protocols (Table 1): a single injection phase with a constant injection rate (Group 1), a two-phase injection with two different injection rates (Group 2), and a three-phase injection composed of two contrast medium injections with two different rates, followed by a saline injection at constant rate (Group 3). All injections were made into an antecubetal vein. Contrast concentrations (CT numbers) in the aorta were measured and results were compared across the three groups. Parameters measured were aortic opacification based on Hounsfield units and uniformity of aortic enhancement at predetermined slices (locations from top [level 1] to base [level 60]). This study was approved by the Institutional Review Board of the Los Angeles Biomedical Research Institute at Harbor-UCLA. EBT angiography studies: EBA studies were performed using an Imatron C-150XLP computed tomographic scanner (San Francisco, CA). First, a non-contrast 40-slice study (3 mm slice thickness, 3 mm table increment, 100 milliseconds acquisition time) was obtained with the patients supine and no couch angulations. Subsequently, a flow study was performed to estimate the mean transit time of the contrast agent (scan delay time) to the descending aorta. These methods have been previously described (9,10). Finally, up to 60 contiguous axial EBA images were acquired to cover the entire cardiac tree with 1.5 mm slice thickness and 1.5 mm table incrementation. A volume of 120 –180 mL of contrast medium (Iopamidol 370, Bracco Diagnostics, Plainsboro, NJ) was used with injection protocols as described in the following section. Electrocardiographic triggering was used, corresponding to end systole as previously described (10). The image acquisition time was 100 milliseconds per image and total scan time was 35–55 seconds (mean 44.5 seconds). Three injection protocols were employed in this study. The injection rate is dependent on the scan time (based on heart rate and slice number). For patients in Group 1,
160
a single phase injection with a 3.5– 4 mL/seconds of fixed contrast medium injection rate was used, with a mean injection time of 44.5 seconds. The injection was completed, on average, at the 34th image. In Group 2, a dualphase contrast medium injection was employed. The first contrast medium injection rate was 4 –5 mL/seconds with 60 mL of contrast medium, followed by a second contrast medium injection rate of 2.5– 4 mL/second. The injection time was 50 seconds, with the injections completed on average at the 40th slice image, with variation from heart rate and number of images required to image the entire heart. In Group 3, two sequential contrast medium injections were followed by a saline injection. The contrast medium injection rate of the first injection was 4 –5 mL/ seconds and 2.5– 4 mL/seconds for the second contrast medium injection. The contrast medium injection time was equal to image acquisition time for 60 slices (the study acquisition time). For slower heart rates (longer acquisition times), the slower rates of contrast injection were used. The contrast injection was followed by a saline injection of 60 milliliters of dextrose 5% with halfnormal saline matching the second injection speed and continued to scan end. In Groups 1 and 2, a single injector (Envision Injector, MedRad, Inc, Indianola, PA) was used. In Group 3, a dual injector (Stellant Dual Injector, Medrad, Inc) was used, with one injector head for administration of contrast medium and the other for saline injection. Arterial opacification based on CT numbers of the descending aorta were measured at seven prespecified levels in each study: 1st, 10th, 20th, 30th, 40th, 50th, and 60th slices obtained (Insight Workstation, NeoImagery Technologies, City of Industry, CA). Recirculation To assess for the effect of contrast medium recirculation, we analyzed a subset of 30 time/density curve studies. Each contrast enhancement peak (highest CT numbers) was assessed to measure the amount and effect of
Academic Radiology, Vol 13, No 2, February 2006
OPTIMIZATION OF CT CORONARY ANGIOGRAPHY
Figure 1. Comparison of computed tomography (CT) Hounsfield units (HU) among Groups 1–3. There are significant differences in images 50 – 60 and total values, with a higher CT HU in group 3 than Groups 1 and 2 (P ⬍ .05).
recirculating contrast. Each time/density curve peak was assessed for maximum CT numbers and the timing of the peak (in seconds) in reference to the first peak and start of injection. The effect of blood laminar flow and recirculation was assessed, and the progressive drop of CT numbers was evaluated. Statistical Analyses All comparisons between the three injection protocols were achieved using the 1st (top) to 60th slice (bottom level) of each EBA study. Prespecified comparisons were set to compare the results of the slice with peak opacification as measured by CT HU to both the top slice (1st level) and bottom slice (60th level). Continuous variables with normal distributions were compared using Student’s t-test, as well as one-way analysis of variance tests. The difference between both measurements was represented as absolute difference ⫽ (abs (A ⫺ B)/(0.5 ⫻ A ⫹ 0.5 ⫻ B value) ⫻ 100%. A probability level of 5% (0.05) was used to determine statistical significance.
RESULTS There were no significant differences among the three patient groups based on factors that affect image opacification, including body surface area, contrast/kilogram, and age or gender of each group (Table 1). Additionally, the heart in its entirety was visualized within slices 1– 60
in all cases. Aortic opacification based on CT numbers for each slice decile in Groups 1–3 are listed in Table 3. Group 1 (Single-Phase Contrast Injection) The aortic opacification CT numbers were 278.8, 300.7, 321.9, 331.5, 310, 261.3, and 226.3 (overall mean 290.1) at the 1st, 10th, 20th, 30th, 40th, 50th, 60th slices, respectively. Opacification peaked halfway through the study (level 30). Opacification was significantly higher on level 30 than level 60 (P ⬍ .05; Table 3, Fig 1). There were significant differences in aortic opacification CT numbers between the 1st and 30th levels and between 30th and 60th levels of 19% and 37%, respectively, (P ⬍ .001), demonstrating nonuniformity in vascular opacification between slice levels. Group 2 (Dual-Phase Contrast Injection) The aortic opacification CT numbers were 297.5, 292.7, 302.2, 308.5, 311.4, 286.7, and 241.6 (overall mean 291.6) at the 1st, 10th, 20th, 30th, 40th, 50th, and 60th slices, respectively. There were no significant differences in CT numbers from the 1st to 50th slice (P ⬎ .05), but significant differences did exist between the 1st and 50th slice and the 60th slice (P ⬍ .05). This demonstrated that dual-phase contrast medium injection increased the uniformity of CT numbers throughout the majority of the study, but that CT HU decreases during the final slices of the study were still problematic.
161
BUDOFF ET AL
Academic Radiology, Vol 13, No 2, February 2006
Table 2 Data of the Patients Group
Number
Heart Rate
Delay Time
Weight (kg)
Height (cm)
BSA (m2)
Contrast Dose (ml/kg)
1 2 3 Analysis of variance
105 70 70
80.2 81.4 80.2 NS
20.5 19.4 21 NS
66.8 67.5 68.5 NS
179.1 180.9 176.3 NS
1.95 1.95 1.94 NS
2.62 2.60 2.55 NS
Table 3 Comparison of CT Numbers of 60 Slices in 3 Protocol Groups Slice Orders Group
1st
10th
20th
30th
40th
50th
60th
Mean
1 2 3 Analysis of variance
278.8* 397.5 314.4 P ⬍ .05
300.7 292.7 312.5 P ⬎ .05
321.9 302.2 313.2 P ⬎ .05
331.5* 308.5* 323.3 P ⬍ .05
310.0 311.4 326.4 P ⬎ .05
261.3* 286.7* 312.2 P ⬍ .05
226.3* 241.6* 274.7 P ⬍ .05
290.1* 291.6* 310.2 P ⬍ .05
*As compared with mean values in Group 3 for each level, these measures are significantly lower. Analysis of variance values shown for each level.
Group 3 (Two Sequential Contrast Medium Injection Phases Followed by a Saline Injection Phase) The aortic opacification CT numbers were 314.4, 312.5, 313.2, 323.3, 326.4, 312.2, and 274.7 (overall mean 310.2) at the 1st, 10th, 20th, 30th, 40th, 50th, and 60th slices, respectively. As in Group 2, there were no significant differences within the 1st to 50th measurements (P ⬎ .05) but a lower CT number at the 60th slice image compared with slices 1–50 (P ⬍ .05). There was though a significant larger mean CT numbers of aortic images in Group 3 as compared with Groups 1 and 2 (P ⬍ .05; Table 2, Fig 1), demonstrating that contrast opacification can be improved through the use of two phases of contrast medium injection followed by a saline injection phase. Recirculation To assess for recirculation of contrast-enhanced blood, we analyzed a subset of 30 time/density curve studies (from Group 3), with discrete second and third peak flows. The second and third peaks of contrast enhancement at the level of the superior vena cava were 5–15 seconds after primary peak (mean 7.7 seconds) and 16 –30 seconds (mean 22.3 seconds), respectively, representing recirculation from the head and neck, thorax, and upper extremity veins (Fig 2). At the insertion of the infe-
162
rior vena cava, the second and third peak times were 8 –11 seconds (mean 9.8 seconds) and 22–25 (mean 22.8 seconds) after the first peak respectively, representing recirculation from the abdomen, pelvis, and lower extremities (Fig 2).
DISCUSSION CT angiography is increasingly being used as a minimally invasive technique for visualization of the coronary arteries (1– 4). The EBA axial images requires iodinated contrast medium enhancement, administered by venous injection. The quality of three-dimensional rendered images is dependent on the quality of the original axial images. Target vessel opacification and motion artifact reduction are important factors in the acquisition of optimal axial images. The number of axial image necessary to image the heart in its entirety is 40 – 60 slices, requiring 30 – 45 second acquisition times. The prolonged acquisition time results in nonuniform opacification of the target vessels, which can also be seen in multidetector CT angiography (5– 8). Newer scanners with greater collimation width (now up to 40 mm per rotation), make this less problematic; however, these scanners are increasingly being used to cover more anatomy (entire pulmonary and coronary vasculature, aortic or peripheral imaging), resulting in similar opacification issues.
Academic Radiology, Vol 13, No 2, February 2006
OPTIMIZATION OF CT CORONARY ANGIOGRAPHY
Figure 2. This graph displays the recirculation time in 30 selected flow studies. At the SVC (superior vena cava), the timing of the second and third peak time is 5–15 seconds and 16 –30 seconds after the first peak time, respectively. At the IVC (inferior vena cava), the second and third peaks are 8 –11 seconds and 22–25 seconds after the first peak, respectively.
Several factors contribute to nonuniformity of vascular enhancement. Venous circulation blood flow is a low pressure and low velocity system, with significant variation in blood flow velocity from heart rate, right atrial pressure, left ventricular diastolic pressure, and patient postural position. An additional factor contributing to variation of vessel opacification is laminar blood flow characteristics. When blood flows through a vessel, there is a higher velocity in the center of the vessel as compared to the lumen wall (11). This nonuniformity of blood velocity from laminar flow results in increasing or decreasing opacification with different injection rates. Claussen and colleagues reported that injection in the vena cava (rather than the peripheral intravenous injections more commonly used) can decrease peak time, decrease laminar flow, and increase coronary artery enhancement (12). Another important reason for nonuniformity of vascular enhancement is recirculation of contrast-enhanced medium from different organ systems. To our knowledge, no detailed information has been reported regarding recirculation time from different venous circulations. Because of laminar flow and recirculation of blood, the CT HU increases progressively from the 1st to the 30th– 40th images (Fig 2). After the injection is completed, pressure in the injected vein falls, blood flow of the respective upper extremity slows, and the CT HU decreases, although recirculation of contrast-enhanced blood continues. Table 2, Figure 3 displays the effect of blood laminar flow and recirculation based on these study results. After the con-
Figure 3. This graph displays the time by blood laminar flow effect and recirculation from major organs.
trast medium injection is stopped, the CT number decreases smoothly and progressively because of laminar flow and recirculation. To decrease this variation and prolong the opacification time, two or more injection speed protocols have been reported for vascular CT angiograms (5–7) and other organ enhancement (13–15). These injection protocols obtained a more uniform result; however, data specifically relating to CT coronary angiography have not been reported. Dual-phase contrast medium injection protocols were used in our study, with a high injection rate (4 –5 mL/
163
BUDOFF ET AL
Academic Radiology, Vol 13, No 2, February 2006
Figure 4. After the contrast injection ends, the computed tomography Hounsfield units of aorta decreases progressively from fourth slice.
seconds, 12–15 seconds) before contrast recirculation, followed by a decreased injection rates (2.5– 4 mL/seconds) to lengthen the injection time to match the study time (40 – 45 seconds). We found greater uniformity of aortic CT numbers on the 1st to 50th slice images compared with a single-phase injection, but there was still a significant decrease after slice 50. The three-phase injection protocol, however, did demonstrate a significantly higher CT HU of the 50th– 60th slices compared with the single-phase injection or the a two-phase injection without saline injection (P ⬍ .05). Because of the loss of venous pressure after injection and the issue of wasted contrast medium present in the tubing and arm vein of the patient during imaging, we used a saline injection to maximize distal contrast opacification without increasing contrast dose to the patient. We found that the CT numbers decreased progressively starting from the third to fourth heart beat after cessation of contrast medium injection (Fig 4). Because the decrease in CT numbers was related to loss of injection pressure, a protocol for maintaining injection pressure and washing contrast medium out of the arm veins and tubing was devised using a dual injector system (Stellant, Medrad, Indianola, PA). With saline injection (Group 3), the mean CT HU of all levels was significantly higher than in Groups 1 and 2, and were significantly higher in the distal levels (images 50 and 60). Therefore, saline injection is an effective method for improving vascular opacification in CT angiography studies. Hopper and colleagues reported use of an injector with contrast agent and saline in one syringe, relying on the density difference to keep the solutions separate (16).
164
Other authors have reported that two injectors were used to deliver contrast medium and saline automatically (17,18). The aims of these CT studies (6,7,16 –18) were to save contrast medium dose and decrease the artifact of superior vena cava enhancement with conventional CT. All of the previously mentioned studies used shorter injection time (20 –25 seconds) and smaller contrast doses (⬍75 mL) than the current study. Because angiography using thin slice CT and EBA require longer injection times, the three-phase injection protocol with sequential contrast injections followed by saline injection is an important method for improving image quality. A disadvantage of the use of two separate injectors or layering saline on top of contrast is the complexity of determining injection pressures and timing delays. The new dual injector with two syringes and one control system used in our center obviates some of those difficulties.
LIMITATIONS There was a significant peak in CT HU at the 40th level and lower HU at the 60th level in all three groups, despite two phases of contrast injections followed by saline injection. For cases requiring longer injection protocols (slower heart rates or more elongated hearts), a third contrast injection phase may be necessary. Furthermore, this study addressed only the ability to increase vascular opacification density and slice uniformity, not the diagnostic accuracy of the coronary artery images obtained. This study did not explore variables such as optimal injection duration or contrast dose. These factors are impor-
Academic Radiology, Vol 13, No 2, February 2006
Figure 5. Volume rendering of a electron beam angiogram demonstrating excellent visualization of distal right coronary artery and posterior descending artery (PDA) with good opacification of lower levels (computed tomography Hounsfield units ⬎300).
tant because short injection times can lead to a lower CT HU at distal levels and produce non-diagnostic images (19). Because distal vessels are well seen with EBA given good opacification (Fig 5), there is a need to lengthen the injection time to ensure constant opacification, which needs to be balanced against total contrast dose. Nonetheless, use of dual contrast injection increased HU uniformity until the last 10 slices of the study and the addition of saline flush lead to lower contrast medium requirements (taking advantage of contrast in the arm and superior vena cava at the conclusion of injection) and higher opacification HUs at all of the levels of the CT study. ACKNOWLEDGMENT
The authors thank David Hill for his instruction, insight, and assistance, and Medrad, Inc., for providing a Stellant CT injector for evaluation. REFERENCES 1. Budoff MJ, Achenbach S, Duerinckx A. Clinical utility of computed tomography and magnetic resonance techniques for noninvasive coronary angiography. J Am Coll Cardiol 2003; 42:1867–1878.
OPTIMIZATION OF CT CORONARY ANGIOGRAPHY
2. Ha JW, Cho SY, Shim WH, et al. Noninvasive evaluation of coronary artery bypass graft patency using three-dimensional angiography obtained with contrast-enhanced electron beam CT. AJR Am J Roentgenol 1999; 172:1055–1059. 3. Budoff MJ, Lu B, Shinbane JS, et al. Methodology for improved detection of coronary stenoses with computed tomographic angiography. Am Heart J 2004; 148:1085–1090. 4. Gerber TC, Kuzo RS, Karstaedt N, et al. Current results and new developments of coronary angiography with use of contrast-enhanced computed tomography of the heart. Mayo Clin Proc 2002; 77:55–71. 5. Bae KT, Tran HQ, Heiken JP. Multiphasic injection method for uniform prolonged vascular enhancement at CT angiography: pharmacokinetic analysis and experimental porcine model. Radiology 2000; 216:872– 880. 6. Fleischmann D, Rubin GD, Bankier AA, et al. Improved uniformity of aortic enhancement with customized contrast medium injection protocols at CT angiography. Radiology 2000; 214:363–371. 7. Hittmair K, Fleischmann D. Accuracy of predicting and controlling timedependent aortic enhancement from a test bolus injection. J Comput Assist Tomogr 2001; 25:287–294. 8. Fleischmann D, Hittmair K. Mathematical analysis of arterial enhancement and optimization of bolus geometry for CT angiography using the discrete Fourier transform. J Comput Assist Tomogr 1999; 23:474 – 484. 9. Bakhsheshi H, Mao SS, Budoff MJ, et al. Preview method for electronbeam CT scanning of the coronary arteries. Acad Radiol 2000; 7:620 – 626. 10. Mao S, Budoff MJ, Lu B, et al. Optimal ECG trigger point in electronbeam CT studies: three methods for minimizing motion artifacts. Acad Radiol 2001; 8:1107–1115. 11. West JB. Best and Taylor’s physiological basis of medical practice. Baltimore: Williams & Wilkins, 1991. 12. Claussen CD, Banzer D, Pfretzschner C, et al. Bolus geometry and dynamics after intravenous contrast medium injection. Radiology 1984; 153:365–368. 13. Keogan MT, McDermott VG, Paulson EK, et al. Pancreatic malignancy: effect of dual-phase helical CT in tumor detection and vascular opacification. Radiology 1997; 205:513–518. 14. Heiken JP, Brink JA, McClennan BL, et al. Dynamic contrast-enhanced CT of the liver: comparison of contrast medium injection rates and uniphasic and biphasic injection protocols. Radiology 1993; 187:327– 31. 15. Bonaldi VM, Bret PM, Atri M, et al. A comparison of two injection protocols using helical and dynamic acquisitions in CT examinations of the pancreas. AJR Am J Roentgenol 1996; 167:49 –55. 16. Hopper KD, Mosher TJ, Kasales CJ, et al. Thoracic spiral CT: delivery of contrast material pushed with injectable saline solution in a power injector. Radiology 1997; 205:269 –271. 17. Haage P, Schmitz-Rode T, Hubner D, et al. Reduction of contrast material dose and artifacts by a saline flush using a double power injector in helical CT of the thorax. Am J Roentgenol 2000; 174:1049 –1053. 18. Irie T, Kajitani M, Yamaguchi M, et al. Contrast-enhanced CT with saline flush technique using two automated injectors: how much contrast medium does it save? J Comput Assist Tomogr 2002; 26:287–291. 19. Mao S, Takasu J, Child J, et al. Comparison of LV mass and volume measurements derived from electron beam tomography using cine imaging and angiographic imaging. Int J Cardiovasc Imaging 2003; 19: 439 – 445.
165