Introduction
In clinical practice, non-contrast CT, CT perfusion (CTP), and CT angiography (CTA) are usually combined to obtain comprehensive information about the site of arterial occlusion and the extent of ischemic damage in acute stroke patients [
1-
4]. In addition, combination of these three modalities could be helpful in the assessment of the cerebral tumor grading and the tumor blood supply as well [
5-
7]. Usually the CTA and CTP images of the brain could be obtained by two separate scan protocols with two contrast medium injections on 16 multidetector CT (MDCT). Moreover, the coverage of the CTP is limited to 2 cm on 16 MDCT [
8-
12]. Although the development of 64 MDCT allows the simultaneous acquisition of CTP and CTA images in a single scan, the coverage of CTP is limited to 4 cm in thickness [
13,
14]. Thus, we developed a novel technique — multiphase dynamic helical scan to expand the anatomic coverage of CT perfusion acquisition to the whole-brain and to obtain CTA source images simultaneously in a single scan on 16 MDCT. The new protocol was tested in various brain abnormality examinations.
Materials and methods
Patients
From October, 2006 to January, 2008, a total of 90 patients were included in our study. This study had been approved by the Ethics Committee of the Huazhong University of Science and Technology. Informed consent was obtained from all the patients before participation in the study. They were randomly assigned into 3 groups. Thirty patients (18 men and 12 women, age range: 11-84 years; mean age: 43.6±18.8 years) undergoing the new protocol were included in group 1, which consisted of 25 patients with brain tumors and 5 patients with chronic headaches.
Thirty patients (14 men and 16 women, age range: 10-74 years; mean age: 42.8±16.9 years) undergoing CTA of the brain with a standard protocol were chosen for group 2. They were composed of patients with brain tumor (n=11), stroke (n=2), aneurysm (n=9), arteriovenous malformation (n=5) and chronic headache (n=3).
Group 3 included 30 patients (11 men and 19 women, age range: 11-76 years; mean age: 45.9±16.6 years) who accepted standard CTP examination. This group consisted of patients with brain tumors (n=14), stroke (n=3), epilepsy (n=7) and chronic headache (n=6).
Imaging protocol
All CT scans were performed at our institution by a 16 MDCT scanner (Lightspeed; GE Medical Systems, Milwaukee, Wis, USA). All the data was post-processed on the Sun Ultra 80 workstation. For all the groups, a scout image and a standard non-contrast CT scan (120 kV, 150 mA, FOV: 25 cm, thickness: 10 mm) were obtained prior to the contrast medium injection to determine the scan range. Then, an 18-gauge plastic intravenous catheter was placed in the antecubital vein for injection of non-ionic contrast material [iopamidol (Iopamiron, Bracco), 50 mL, 3.5 mL/s] with a power injector.
In group 1, the new protocol was performed. After standard non-contrast CT scan, each patient acquired dynamic helical CT scans from skull base to vertex for 14 times. Detailed technical parameters are given in Table 1. Then, the CTA source images were obtained by reconstruction from the raw data of the phase in which the enhancement of main cerebral arteries came to peak with a section thickness of 1.25 mm and an overlap of 1.25 mm.
The patients in group 2 underwent routine CTA scans. The delay time of data acquisition was determined by the bolus testing method. The patients in group 3 received standard CTP examination. Because of the limited coverage, for each patient, four locations (0.5 cm thick for each) were chosen at the level of lesions on the noncontrast CT or at the level of the insula/basal ganglia. Detailed technical parameters of routine CTA and CTP protocols were provided in Table 1.
Data analysis
CT angiography analysis
The CTA images were generated on the workstation Sun Ultra 80 by maximum intensity projection (MIP), multi-planar reformation (MPR) and volume rendering (VR) techniques. The diagnostic image quality of the main cerebral arteries (internal carotid artery and basilar artery) and their first-order to third-order branches was scored from 0 to 2 (0 for no visualization, 1 for only visualization, and 2 for diagnostic image quality) by two experienced radiologists without knowing the image acquisition details.
CT perfusion imaging analysis
Four perfusion parameters were calculated and corresponding functional color maps were generated by the CT perfusion software: blood volume (BV) map, blood flow (BF) map, mean transit time (MTT) map and/or permeability surface (PS) map. For each patient of groups 1 and 3, similar region of interest (ROIs) of approximately 100 mm2 were placed on the same anatomic features: (1) normal cortical gray matter (four ROIs for each section), (2) normal white matter (four ROIs for each section), (3) head of the caudate nucleus (an ROI for each side), and (4) lentiform nuclei (a ROI for each side). Then the value of all perfusion parameters was given automatically by the software.
X-ray dosimetry
The volume CT dose index (CTDIvol) and dose length product (DLP) were provided automatically by the CT scanner.
Statistical analysis
Our data was analyzed by MLWin2.02. For statistical analysis, the Wald Chi-square test was applied for comparison of the CTA image quality between different protocols and for analysis of the variability between observers. T test was applied for analyzing CTP parameters of the two methods. A value of P<0.05 was considered statistically significant.
Results
In group 1, CTA and CTP images of the whole brain could be obtained simultaneously with a single scan on 16 MDCT (Fig.β1). All lesions were described clearly on the CTP images of the whole brain, including the locations, sizes, borders and haemodynamic changes. Although the temporal resolution of CTP was compromised in multiphase dynamic helical scan, the quality of the whole brain CTP images was acceptable (Fig.β2). Compared with the routine method (group 3), the new protocol tended to overestimate BV and BF and underestimate MTT. However, there was no statistically significant difference between the two methods. PS value was overestimated by using the new protocol and it was statistical different between the two methods (Table 2).
The CTA images of group 1 were good enough for diagnosis (Fig. 3). The image quality of the main cerebral arteries and their first-order to third-order branches was scored. The details are listed in Tableβ3. No statistical difference in the CTA image quality was found between groups 1 and 2 (Wald χ2=0.3617, P=0.54756). In addition, the variability between two observers has no statistical difference (Wald χ2=0.5383, P=0.46314).
The CTDIvol and DLP of the new protocol were 161.98 mGy and 2394.0 mGy·cm. If the CTP and CTA images were obtained by the two separate standard scans, the total CTDIvol and DLP of two scans were 707.68 mGy and 2419.56 mGy·cm, respectively.
Discussion
Recent years, the rapid development of the CT technology not only allows us to reduce the examination time but also provides more functional information to contribute to a more accurate diagnosis. We can assess a patient in detail and quickly by combination of the non-contrast CT, CTA and CTP. However, the combination of multiple CT techniques increased both radiation dose and contrast medium dose in patient on the 16 MDCT. Optimizing the protocol to achieve good image quality at lower radiation exposure, reduced acquisition time and lower contrast medium dose and to gain more comprehensive information is a challenge for all of us.
Currently, two protocols which combine CTA with CTP exam are common in clinical practice on the 16 MDCT. One is perfused-blood-volume CT, by which the CTA map and whole brain BV map could be obtained simultaneously in a single bolus injection of contrast medium of 90-120 mL. But it could not yield other maps of CT perfusion parameters (BF, MTT and PS) [
15-
17]. The other is routine CTA plus CTP protocol which can obtain CTA and CTP images in two separate scans. This protocol could yield CTA maps and multiple CT perfusion parameter (BV, BF, MTT and PS) maps. The contrast medium is required to inject twice with total dosage of 130-150 mL in this protocol [
18], and a longer scan time is also needed. In addition, the spatial coverage of CTP is limited to 20 mm due to the limitation of the 16 MDCT scanner. Even by using the “toggling-table” technique [
13,
19], the spatial coverage could only expand to 40 mm on 16 MDCT or double coverage on the 64 MDCT. According to our results, we could get CTA maps as well as the whole brain CTP maps of multiple perfusion parameters in a single scan and in single bolus injection of contrast medium of 50 mL by using multiphase dynamic helical scan without significantly compromising the CTA and CTP image quality.
In routine CTP protocol, the coverage was only 20 mm and the temporal resolution was 0.5 s. In order to increase the spatial coverage of CT perfusion in multiphase dynamic helical scan, we compromised the temporal resolution to 3.1 s (2.1 s for each helical scan, 1 s for table feed). From the results, we could find that lower temporal resolution led to an overestimation of BV, BF, and PS and underestimation of MTT. However, no statistical difference was found between the groups 1 and 3 except PS. Wintermark and his colleagues got the same results [
20]. In their study, they found that the lowest allowable temporal resolution to avoid overestimation or underestimation of the CTP parameters was 3 s for 50 mL bolus injection of contrast medium and 4 s for 60 mL. It is possible that the value of PS may be more dependable if we increase the dose of the contrast medium. From the CTP images of the patients in group 1, we concluded that although the temporal resolution was somewhat compromised, useful and predictive perfusion parameters of the whole brain can, nevertheless, be derived.
In multiphase dynamic helical scan, we shortened the spiral scan acquisition time to 2.1 s to ensure the temporal resolution of CTP. It would also result in the compromise of the spatial resolution of the CTA because the source images of CTA were reconstructed from the data of spiral scan. To compensate the spatial resolution, we reconstructed the source images using the thinnest section thickness and largest overlapping thickness to get the best image quality. According to our results, the majority of the main cerebral arteries and nearly all of their first-order and second-order branches were depicted good enough for diagnosis although the spatial resolution was compromised somewhat. Although the exhibition of the third-order branches in group 1 was not as good as that in group 2, we could obtain the dynamic CTA of the process from arterial input to venous output if we reconstructed CTA images from several to even all of the phases. The information is helpful to observe hemodynamic changes of some lesions such as vascular abnormalities. In addition, since the new protocol did not require the bolus testing technique or smart prep technique, it simplified the scan procedure. Similar to the traditional CTA protocol, the image quality of the main cerebral arteries from the new scan protocol also depended on post-processing techniques. For example, the edge of the ICA was not smooth on 3D CT angiography usually because it is difficult to differentiate vessels from bone structure in the portion close to the skull base.
According to the Food and Drug Administration (FDA) in the USA, two reference dose quantification, volume CTDI (CTDIvol) and DLP, are proposed for CT to measure the radiation exposure dose. In our protocol, CTDIvol and DLP were a little lower than that of routine CTA plus CTP protocol, although the differences were not significant. It indicated the higher efficiency of the new protocol because of less contrast medium and radiation doses.
The newly developed method — the multiphase dynamic helical scan can obtain the CTP and CTA images simultaneously by a single scan on the 16 MDCT, which can increase the coverage of CTP from 2 cm to the whole brain with lower radiation exposure and less contrast medium dose without significantly compromising CTP and CTA image quality. It can be implemented in routine clinical examination of central nerve system abnormalities, especially be used to guide emergency management of acute stroke. By multiphase dynamic helical scan, we can get multimodal CT images as fast as we can on the 16 MDCT which is largely available in the emergency. Thus, we can quickly distinguish the infarct core and penumbra on CTP images and find the presence of vessel occlusion on CTA which is important for thrombolysis.
Higher Education Press and Springer-Verlag Berlin Heidelberg
PDF
(233KB)
