Thinner slices and broader reconstruction options improve detail and resolution
MSCT raises quality in vascular imaging
By: Hans Rigauts, M.D., And Kenneth Coenegrachts, M.D.
The advent of single-slice spiral CT in 1989 was a major leap over incremental CT scanning, offering shorter scanning time, improved contrast resolution, more seamless information, and better postprocessing results. With spiral CT, the patient is moved through the gantry, and the speed is constant during the scan, resulting in larger anatomical coverage and decreased patient exposure to x-rays.
Early spiral CT scanners were of limited use for vascular applications because of the conflicting demands of anatomical coverage and longitudinal spatial resolution. Single-slice spiral CT required compromises: Thin slices provide high spatial resolution over a relatively small anatomical coverage, whereas thick slices cover a larger area but give only low spatial resolution.1
Multislice spiral CT (MSCT) overcomes the fundamental limitations of the single-slice technique, allowing a larger anatomical volume to be scanned with thinner collimation and in a shorter time, while producing images with high spatial resolution. Images of variable thickness can be reconstructed, and they benefit from narrow collimation because partial volume artifacts are drastically suppressed. Repeated scanning is no longer required, minimizing exposure to x-rays. The contrast dose can also be reduced as a result of the shorter acquisition time.
Other advantages of MSCT include the ability to examine during optimal contrast-enhancement and to perform multiphase organ studies. Postprocessing results such as angiographic reconstructions are also improved, and three-dimensional isotropic resolution is feasible for improved volume
Slice sensitivity profile and image artifacts in single-slice spiral CT decrease monotonically as the pitch increases. The sensitivity of raw data interpolation in MSCT, however, increases in an alternating way as the pitch rises, suggesting that image quality does not decrease monotonically. Some pitches are preferred for efficient z-sampling in data collection and better artifact control. Study results show that the slice sensitivity profile, image artifacts, and signal-to-noise ratio exhibit performance peaks or valleys at certain spiral pitches in MSCT.3,4
MSCT angiography combines extended anatomical coverage with excellent spatial resolution.5-7 Overlapping thin-collimation images from MSCT refine the postprocessing techniques, improving the quality of reconstructed images such as multiplanar reconstructions (MPR), 3D, and maximum intensity projections (MIP).8,9 The combination of native and reconstructed images provides more detailed, clinically relevant information about the vascular region than is obtained with conventional angiography.
Any possible pathology of the vessel wall or pathological relationship of the vessel to adjacent structures is excluded, allowing optimized treatment planning.10 CT angiography (CTA) is more cost-effective, less invasive (requiring only the intravenous injection of the contrast), and less time- consuming, and it does not require hospitalization.7 Multislice spiral CTA still cannot replace conventional angiography for the evaluation of overall vascular status, however, and this is also important in treatment planning.11
CTA is less operator-dependent than ultrasound and not as dependent on the patient's body position. Optimization of the scan delay by use of a test bolus or a scan-triggering system is crucial. Perfect timing of the scan delay allows optimal enhancement of the vessel during image acquisition, providing an accurate appreciation of the scanned vessel.5,6,12 A dynamic review of the data with the surgeon should be performed, including source images, MPR, MIP, and multiplanar volume-rendered reconstructions, to aid presurgical planning.13
For CTA, the larger anatomical coverage during optimal contrast enhancement with thin collimation is the key to better image quality and new applications. These benefits can help in defining and optimizing clinical protocols: The spiral pitch can be selected almost freely, for example, and scan protocols can follow the diagnostic requirements without technical restrictions.1,8,14,15 Very long segment areas can be imaged, enabling CTA of the lower extremities and single acquisition of the aortoiliac or carotid system.16,17
The combination of source images and postprocessing allows comprehensive evaluation of degree of stenosis, wall abnormalities such as soft or calcified plaques, aneurysmal dilatation, the presence of collaterals, and other incidental lesions.
Three-D imaging is particularly useful in preoperative treatment planning. The main pitfall is the need for a large number of source images (600 to 1000) and the postprocessing time required. The relatively low cost, minimal invasiveness, rapid acquisition, and anticipated shortening of postprocessing time, however, should all contribute to increased use of CTA.1
Vessels that are oriented parallel to the z-axis and, therefore, to the scan direction are particularly suited for multislice spiral CTA, giving the most exact measurements. The technique offers a good detection rate of vascular stenoses, especially at 0 degrees angulation and also at 45 degrees vessel angulation to the z-axis. High pitches (6 and higher) in combination with higher tube currents allow scanning of longer distances with accuracy almost identical to that of other spiral scanning methods with lower pitches.18
It is now possible to scan eight slices per rotation, and within the next five years, 16 (or even 32) slices might become feasible. Use of a symmetrical matrix makes this technically quite simple, and greater computer power will allow simultaneous reconstruction of eight images per rotation within an acceptable period of time (one image per second).
Image quality depends not only on scan parameters such as dose and pitch, however, but also on the reconstruction algorithms. These form the mathematical "deconvolution program" that calculates images from the measured absorption data in the detector. Two reconstruction methods collect data from tube (detector) positions at either 180 degrees or 360 degrees apart. Images with a good signal-to-noise ratio are obtained when using data 360 degrees apart, but these have poorer spatial resolution. A 180 degrees reconstruction algorithm results in good spatial resolution but a worse signal-to-noise ratio. A new reconstruction algorithm combining the advantages of both interpolation algorithms should produce images with good spatial resolution and a good signal-to-noise ratio.
Another new technique involves focal spot tracking. MSCT scanners are collimated after the patient has been examined and are therefore less dose-efficient than the single-slice technique in which collimation takes place beforehand. In MSCT, the x-ray beam must be wide enough in the z-axis for the beam to remain on the activated detector rows regardless of typical movements from thermal and mechanical flexing. When a static collimator is used, the x-ray beam width has to be much wider than the thickness of the activated part of the matrix detector to prevent loss of information from a wobbling of tube anode or the x-ray beam. To reduce the beam width, and therefore the x-ray dose to the patient, a tracking system to keep the beam on the activated detector rows has been developed.
The focal spot tracking technique measures the position of the beam every few milliseconds and continuously repositions a dynamic collimator to hold the beam over all the activated detector rows at all times. The thickness of the beam can be kept very close to the thickness of the activated part of the matrix detector, allowing x-ray doses to be obtained that are comparable to the doses applied in a single-slice spiral CT.19
Manufacturers of CT equipment are also developing a dose reduction program that will automatically adjust the mA values according to the body shape while maintaining image quality. The dose will automatically be reduced in those body regions with less dense tissues. A scan of the abdomen, for example, can easily use a reduced dose in the subhepatic region and a normal dose for the liver and pelvic region. In combination with online tube current modulation, the dose can be reduced by up to 40%, without any drawbacks in image quality.
Another new technique allows scanning with submillimeter slices. This is made possible by dividing the thinnest possible collimation (1 mm) over two detector rows. Each of the detector rows will acquire information from part of one of these thin collimations, allowing submillimeter acquisitions.
Further developments concerning preoperative endovascular stent planning are also in progress. As discussed, MSCT is a useful tool for angiographic examinations, as it allows fast scanning using thin slices at optimal contrast enhancement. A semiautomated program has been developed for tracking the course of a contrast-enhanced vessel. The operator indicates four to eight points on the vessel of interest, before a bolus-tracking program obtains an optimal contrast bolus.12 The program then automatically tracks the course of the contrast-enhanced vessel, based on points entered manually by the operator. Three-D reconstructions, MPR, and axial images perpendicular to the lumen of the vessel will be obtained automatically, together with diameter and volume measurements.
The reconstruction of axial images perpendicular to the axis of the vessel is of major importance for treatment (Figure 1). Correct measurements are mandatory for choosing the stent that exactly fits in the lumen of the vessel. We are performing a study in our institution to compare the accuracy of this semiautomated program with the measurements obtained after calibrated angiography. We intend to use this program instead of calibrated angiography procedures to measure the desired graft length and diameter even more accurately when planning endovascular stent placements. Postoperatively, MSCT can also be used for the evaluation of any possible complication, such as endoleakage (Figure 2).
1. Blum A, Walter F, Ludig T, et al. Scanners multicoupes: principes et nouvelles applications scanographiques. J Radiol 2000;81:1597-1614.
2. Klingenbeck-Regn K, Schaller S, Flohr T, et al. Subsecond multi-slice computed tomography: basics and applications. Europ J Radiol 1999;31:110-124.
3. Wang G, Vannier MW. The effect of pitch in multislice spiral/spiral CT. Med Physics 1999;26:2648-2653.
4. Hu H. Multi-slice spiral CT: scan and reconstruction. Med Physics 1999;26:5-18.
5. Kalender WA, Sissler W, Klotz E, et al. Spiral volumetric CT with single-breath-hold technique, continuous transport, and continuous scanner rotation. Radiology 1990;176:181-183.
6. Brink JA, Heiken JP, Wang G, et al. Spiral CT: principles and technical considerations. Radiographics 1994;14:887-893.
7. Rankin SC. CT angiography. Europ Radiol 1999;9:297-310.
8. Schorn C, Obenauer S, Funke M, et al. Schichtempfindlichkeitsprofile und Bildpunktrauschen einer Mehrschicht Spiral-CT im Vergleich zu einer Einzelschicht Spiral-CT. Rofo 1999;171:219-225.
9. Berland LL, Smith JK. Multidetector-array CT: once again, technology creates new opportunities. Radiology 1998;209:327-329.
10. Bartolozzi C, Neri E, Caramella D. CT in vascular pathologies. Europ Radiol 1998;8:679-684.
11. Takase K, Imakita S, Kuribayashi S, et al. Popliteal artery entrapment syndrome: aberrant origin of gastrocnemius muscle shown by 3D CT. J Comput Assist Tomogr 1997;21:523-528.
12. Van Hoe L, Marchal G, Baert AL, et al. Determination of scan delay time in spiral CT-angiography: utility of a test bolus injection. J Comput Assist Tomogr 1995;19:216-220.
13. Henschke C, McCauley D, Yankelevitz D, et al. Early lung cancer action project: overall designs and findings from baseline screening. Lancet 1999;354:99-105.
14. Dawson P, Lees WR. Multi-slice technology in computed tomography. Clin Radiol 2001;56:302-309.
15. Fuchs T, Kachelriess M, Kalender WA. Technical advances in multi-slice spiral CT. Europ J Radiol 2000;36:69-73.
16. Rubin G, Shiau M, Schmidt A, et al. Computed tomographic angiography: historical perspective and new state-of-the-art using multidetector-row spiral computed tomography. J Comput Assist Tomogr 1999;23(Suppl. 1):S83-S90.
17. Cordoliani Y, Hazebrouck V, Sarrazin J, et al. Irradiation et bonnes pratiques en tomodensitometrie helicoidale. J Radiol 1999;80:903-912.
18. Funke C, Alamo L, Castillo E, et al. Mehrschicht Spiral-CT-Angiographie: Optimierung der Untersuchungsparameter an einem Gefa▀phantom. Rofo 2000;172:927-933.
19. Toth T, Bromberg N, Pan T, et al. A dose reduction x-ray beam positioning system for high-speed multislice CT scanners. Med Physics 2000;27:2659-2668.
GUIDELINES FOR BEST PRACTICE
In CTA, it is best to use a slice that is thinner than the diameter of the vessel of interest and to increase the pitch to cover the required volume, maintaining good spatial resolution.7 The resolution will degrade if the pitch is maintained and the slice thickness is increased.
In our department, we work with a LightSpeed QX/i MSCT scanner manufactured by GE Medical Systems.
-Circle of Willis: 1.25-mm slice thickness, pitch 3 (table speed of 3.75 mm per rotation), reconstruction interval of 0.6 mm, 0.8-sec rotation time, 120 kV and 200 mA. 100 mL of a nonionic contrast agent is used at an injection rate of 2 mL/sec.
-Carotid arteries: 2.5-mm slice thickness, pitch 6 (table speed of 7.5 mm per rotation), reconstruction interval of 1.25 mm, 0.8-sec rotation time, 120 kV and 200 mA. 100 mL of a nonionic contrast agent is used at an injection rate of 3 mL/sec.
-Detection of lung emboli: 2.5-mm slice thickness, pitch 6 (table speed of 7.5 mm per rotation), reconstruction interval of 1 mm, 0.8-sec rotation time, 120 kV and 220 mA. 100 mL of a nonionic contrast agent is used at an injection rate of 3 mL/sec.
-Dissection of the thoracic aorta: 2.5-mm slice thickness, pitch 6 (table speed of 7.5 mm per rotation), reconstruction interval of 1.25 mm, 0.8-sec rotation time, 120 kV and 320 mA. 100 mL of a nonionic contrast agent is used at an injection rate of 3 mL/sec.
-Abdominal aorta: First without using contrast material. 5-mm slice thickness, pitch 6 (table speed of 30 mm per rotation), reconstruction interval of 5 mm, 0.8-sec rotation time, 120 kV and 230 mA. Second with the use of contrast material. 2.5-mm slice thickness, pitch 6 (table speed of 15 mm per rotation), reconstruction interval of 1 mm, 0.8-sec rotation time, 120 kV and 200 mA. 100 mL of a nonionic contrast agent is used at an injection rate of 3 mL/sec (Figure 2).
-Renal arteries: 1.25-mm slice thickness, pitch 6 (table speed of 7.5 mm per rotation), reconstruction interval of 0.6mm, 0.8-sec rotation time, 120 kV and 200 mA. 100 mL of a nonionic contrast agent is used at an injection rate of 3 mL/sec.
-Vessels of the extremities: 1.25-mm slice thickness, pitch 6 (table speed of 7.5 mm pro rotation), reconstruction interval of 0.6 mm, 0.8-sec rotation time, 120 kV and 200 mA. 100 mL of a nonionic contrast agent is used at an injection rate of 3 mL/sec (Figure 3).