Skip to main content
NCBI home page
The British Journal of Radiology logoLink to The British Journal of Radiology
. 2017 Jan 20;90(1070):20160634. doi: 10.1259/bjr.20160634

Optimal scan timing for artery–vein separation at whole-brain CT angiography using a 320-row MDCT volume scanner

Takashi Shirasaka 1,✉, Akio Hiwatashi 2, Koji Yamashita 2, Masatoshi Kondo 1, Hiroshi Hamasaki 1, Yamato Shimomiya 1, Yasuhiko Nakamura 1, Yoshinori Funama 3, Hiroshi Honda 2
PMCID: PMC5685121  PMID: 27995807

Abstract

Objective:

A 320-row multidetector CT (MDCT) is expected for a good artery–vein separation in terms of temporal resolution. However, a shortened scan duration may lead to insufficient vascular enhancement. We assessed the optimal scan timing for the artery–vein separation at whole-brain CT angiography (CTA) when bolus tracking was used at 320-row MDCT.

Methods:

We analyzed 60 patients, who underwent whole-brain four-dimensional CTA. Difference in CT attenuation between the internal carotid artery (ICA) and the superior sagittal sinus (Datt) was calculated in each phase. Using a visual evaluation score for the depiction of arteries and veins, we calculated the difference between the mean score for the intracranial arteries and the mean score for the veins (Dscore). We assessed the time at which the maximum Datt and Dscore were simultaneously observed.

Results:

The maximum Datt was observed at 6.0 s and 8.0 s in the arterial-dominant phase and at 16.0 s and 18.0 s in the venous-dominant phase after the contrast media arrival time at the ICA (Taa). The maximum Dscore was observed at 6.0 s and 8.0 s in the arterial-dominant phase and at 16.0 s in the venous-dominant phase after the Taa. There were no statistically significant differences in Datt (p = 0.375) or Dscore (p = 0.139) between these scan timings.

Conclusion:

The optimal scan timing for artery–vein separation at whole-brain CTA was 6.0 s or 8.0 s for the arteries and 16.0 s for the veins after the Taa.

Advances in knowledge:

Optimal scan timing allowed us to visualize intracranial arteries or veins with minimal superimposition.

INTRODUCTION

CT angiography (CTA) is widely used to evaluate the intracranial arteries and veins.1–7 Simultaneous enhancement of the arteries and veins on brain CTA allows us to evaluate these structures precisely. However, enhancing the venous structures may interfere with detailed evaluation of the arterial structures and vice versa.1,2,8,9 Therefore, a demonstration of artery–vein separation at brain CTA10 is desirable for increasing diagnostic accuracy and improving surgical planning.

The scan timing after the arrival of contrast media (CM) is one factor that affects the separation of intracranial arteries and veins. Matsumoto et al10 reported a technique for obtaining a separate demonstration of the arterial and venous phases on brain three-dimensional (3D)-CTA at 16-row multidetector CT (MDCT). However, 16-row MDCT is inferior to the novel wide coverage detector CT in terms of temporal resolution. A 320-row MDCT is thus expected for the artery–vein separation on brain 3D-CTA. This is because the scanner equipped with a detector coverage of 160 mm per rotation leads to a scan of <1.0 s for the whole brain.11 As the degree of cardiac output or the presence of vascular obstruction could affect differences in contrast media arrival in various areas of the cerebral vasculature, we could miss the optimal scan timing with sufficient enhancement for entire intracranial arteries or veins, given such a short scan duration. Therefore, it is important to optimize the scan delay12 to achieve acceptable CTA using bolus tracking.6,13 However, to the best of our knowledge, the optimal scan timing for artery–vein separation at whole-brain CTA using 320-row MDCT has not yet been established. The artery–vein separation is available to analyze temporally repeated whole-brain images like four-dimensional (4D)-CTA. Therefore, the purpose of this study was to determine the optimal scan timing for artery–vein separation at whole-brain CTA based on whole-brain 4D-CTA images.

METHODS AND MATERIALS

Patients

This prospective study was approved by our institutional review board, and written informed consent was obtained from each patient. The initial candidates for the study were 111 consecutive patients who underwent whole-brain 4D-CTA between April 2012 and March 2013. The inclusion criterion was as follows: patients who underwent whole-brain 4D-CTA for the pre-operative evaluation of intracranial masses. In our clinical practice, we obtained 4D-CTA routinely for the pre-operative evaluation of intracranial masses.14 The exclusion criteria were as follows: patients with tumours and cerebrovascular disease (occlusion, stenosis, aneurysm, venous sinus thrombosis, moyamoya disease or arteriovenous malformation), multiple metastatic tumours, post-operative status and severe motion artefacts during scanning. In a total of 51 subjects, 27 patients with cerebrovascular disease, 8 patients with multiple metastatic tumours, 10 patients with post-operative status and 6 patients with severe motion artefacts during scanning were excluded owing to the exclusion criteria. Finally, 60 subjects (32 females and 28 males; age range, 17–86 years; median age, 56 years) were enrolled in this study. There were 20 meningiomas, 11 gliomas, 9 schwannomas, 7 lymphomas, 5 pituitary adenomas, 4 metastatic tumours, 1 haemangioblastoma, 1 epidermoid tumour, 1 cavernous malformation and 1 demyelinating disease.

Data acquisition

Whole-brain 4D-CTA examinations were performed using a 320-row MDCT volume scanner (Aquilion™ ONE Vision Edition; Toshiba Medical Systems Corporation, Otawara, Japan). A total of 20 acquisitions were obtained during a total examination time of 60.0 s. Acquisition was started at 5.0 s after the start of CM injection. The first 16 acquisitions were obtained every 2.0 s, and the next 4 acquisitions were obtained every 5.0 s (Figure 1). Wintermark et al15 studied the optimal temporal sampling interval and CM volume without altering the quantitative accuracy of brain perfusion CT. They reported that with an injection rate of 4 ml s−1 and 40 ml of the CM volume, a 3.0-s interval was the maximal allowable temporal sampling, and the time to entrance of the CM into the artery and time to exit of the CM in the vein was about 5 s and 27 s, respectively. On the basis of their results and a prior experimental study,16 scan start time and duration were determined so as not to miss the attenuation upslope of arteries and downslope of the veins. Images were acquired in volume scan mode with 320 × 0.5-mm collimation. The first 16 acquisitions were obtained at 80 kVp, 100 mA and a rotation time of 1.0 s. The next four acquisitions were obtained at 80 kVp, 50 mA and a rotation time of 1.0 s. For whole-brain 4D-CTA, the volume of the CM was adjusted according to the patient body weight; all patients received 245 mg kg−1 of non-ionic CM (iopamidol) (Iopamiron 370; Bayer HealthCare, Osaka, Japan) injected over a fixed duration of 10.0 s, followed by a 40-ml saline flush injected at the same rate as the CM.13,17,18 A dual-shot injector (Nemoto Kyorindo, Tokyo, Japan) was used to inject both the CM and saline flush via a 20-gauge i.v. injection catheter placed in an antecubital vein.

Figure 1.

Figure 1.

Open in a new tab

A graph showing the scanning protocol for whole-brain four-dimensional CT angiography: the scanning speed was 1.0 s per rotation. First, 16 acquisitions were obtained every 2.0 s. Next, four acquisitions were obtained every 5.0 s. Solid lines are indicating the time–density curve (TDC) for the internal carotid artery (ICA) and dashed lines are indicating the TDC for the superior sagittal sinus (SSS). Datt, difference in CT attenuation between ICA and SSS; HU, Hounsfield unit; Taa, contrast media arrival time at the ICA; Tav, contrast media arrival time at the SSS; Tia, interval time between Taa and Tpa; Tiv, time interval time between Tav and Tpv; Tpa, peak enhancement time at the ICA; Tpv, peak enhancement time at the SSS; Ttra, transit time of contrast media from ICA to SSS.

Measurement of CT attenuation

We considered the optimal scan timing for artery–vein separation as the time when the maximum attenuation difference between the intracranial artery and vein was observed. For monitoring the CM arrival at the circle of Willis for 3D-CTA using bolus tracking, we investigated the time from CM arrival at the internal carotid artery (ICA). For quantitative analyses, we investigated the time when the maximum difference in CT attenuation between the intracranial arteries and veins was observed. CT attenuation in the supraclinoid segment of the ICA, and in the superior sagittal sinus (SSS) at the level of the obelion, was considered representative of the vessels and was measured in each acquisition for all patients. Measurement was performed in the contralateral side without a mass lesion. Regions of interest with a diameter of 0.5 mm in the ICA and 2.8 mm in the SSS were placed at the centres of the target vessels by one of the authors (TS) with 13 years' experience as a radiological technologist at the CT console. Time–density curves (TDCs) for the ICA and for the SSS were obtained for each patient (Figure 1). In the TDCs, we basically estimated the contrast media arrival time as the observed time ascent of the ICA attenuation from the baseline.19 The contrast media arrival time at the ICA (Taa) and contrast media arrival time at the SSS (Tav), peak enhancement time at the ICA (Tpa) and peak enhancement time at the SSS (Tpv) and interval time at the ICA (Tia) and interval time at the SSS (Tiv) were defined as follows:

Ta = 2.0 s earlier than the acquisition time when CT attenuation exceeded 150 HU and Ti = time difference between Ta and Tp (Tp − Ta) (Figure 1). In addition, the transit time (Ttra) from the ICA to the SSS was calculated as the difference between Tpa and Tpv.

The difference in CT attenuation between the ICA and the SSS (Datt) was calculated for each acquisition: Datt = CT attenuation at the ICA − CT attenuation at the SSS. The arterial-dominant phase was defined as Datt > 0 and the venous-dominant phase was defined as Datt < 0 (Figure 1). In the arterial- and venous-dominant phases, the time after Taa was measured when the maximum Datt was observed.

Visual evaluation

The degree of simultaneous description of the intracranial arteries or veins in the craniocaudal direction was evaluated at each dominant phase. For qualitative analyses, two board-certified neuroradiologists (AH and KY with 17 and 13 years' experience) independently evaluated the depiction of the major intracranial arteries and veins on the contralateral side without a mass lesion. We extensively selected the vessels in the craniocaudal direction for the evaluation of simultaneous depiction of the vessels. The vessels evaluated were as follows: the intracranial arteries (ICA, M1–M2 segments in the middle cerebral artery, anterior communicating artery, pericallosal artery, vertebral artery and basilar artery tip) and veins (SSS, sphenoparietal sinus and jugular bulb). Maximum intensity projection images were reconstructed from the 4D-CTA data for each patient. Of the 20 acquired data sets, the first 2 data sets were used as the mask images for bone subtraction. As a result, the observers reviewed 18 maximum intensity projection images for each patient (Figure 1). The observers were allowed to adjust the window level and window width on the monitor—preset settings of the window level and window width were 400 HU and 800 HU—and no limit was placed on the reading time. For each image, the observers judged the presence or absence of vessels using stereoscopic anteroposterior and lateral views and rated their confidence level using a 4-grade scale where 0 = definitely absent or uncertain (non-diagnostic), 1 = probably present but discontinuous (non-diagnostic), 2 = definitely present and continuous (diagnostic) and 3 = definitely present and continuous with sufficient enhancement (diagnostic). In each phase, the mean scores were calculated for the intracranial arteries and veins, and the mean score of the two readers was also calculated for a comprehensive evaluation. The number of patients with mean scores of >2 for the arteries and veins was also evaluated. The difference between the mean score for the intracranial arteries and the mean score for the veins (Dscore) was calculated for each phase: Dscore = mean score for the intracranial arteries − mean score for the intracranial veins.

Statistical analyses

Statistical analyses were conducted using commercially available software (SPSS® v. 11.0.1; IBM Corp., New York, NY; formerly SPSS Inc., Chicago, IL). Dunnett's test was used to evaluate the Datt in each arterial- and venous-dominant phase. We defined the time when the maximum attenuation difference between the ICA and the SSS was observed as the control. The Kruskal–Wallis test was used for comparison of the Dscore in each arterial- and venous-dominant phase. When the overall differences were statistically significant, post hoc analyses were performed using the Steel test. Differences of p < 0.05 were considered statistically significant. The Cohen's kappa test was used to assess the degree of agreement between the observers, with a kappa value of 0.01–0.20 for slight agreement, 0.21–0.40 for fair, 0.41–0.60 for moderate, 0.61–0.80 for substantial and 0.81–1.00 for almost perfect agreement.

RESULTS

Measurement of CT attenuation

Figure 2 shows the mean TDCs of the ICA and SSS after Taa. The ICA and SSS were visualized for 20.0 s after Taa in all patients. Therefore, evaluation was performed within 20.0 s after Taa. Tia was 9.2 ± 1.1 s (mean ± standard deviation) (range: 6.0–12.0 s). Tiv was 9.9 ± 1.2 s (range: 8.0–12.0 s). Ttra was 5.7 ± 1.4 s. The arterial-dominant phase continued until 12.0 s, and the venous-dominant phase extended from 12.0 to 20.0 s after Taa. The maximum Datt in the arterial-dominant phase was observed at 6.0 s (Datt = 387.1 HU) followed by the value at 8.0 s (Datt = 364.7 HU) after Taa. These values were significantly higher than those in the other arterial-dominant phases (p < 0.001). However, there was no statistically significant difference in Datt between 6.0 s and 8.0 s (p = 0.38). The maximum Datt in the venous-dominant phase was observed at 16.0 s (Datt = −334.9 HU) followed by the value at 18.0 s (Datt = −305.8 HU) after Taa. These values were significantly higher than those in other venous-dominant phases (p < 0.001). However, there was no statistically significant difference in Datt between 16.0 s and 18.0 s (p = 0.15) (Table 1).

Figure 2.

Figure 2.

Open in a new tab

Coronal maximum intensity projection images reconstructed from four-dimensional CT angiography data in 78-year-old female: each figure is showing the image at (a) 0 s, (b) 4.0 s, (c) 6.0 s and (d) 10.0 s after contrast media arrival time at the internal carotid artery.

Table 1.

Difference in CT attenuation between the internal carotid artery and the superior sagittal sinus (Datt), number of patients with mean score ≥2 and difference between the mean score for the intracranial arteries and the mean score for the veins (Dscore) in arterial- and venous-dominant phases. Dunnett's test was used to evaluate the Datt and the Kruskal–Wallis test followed by Steel test was used to evaluate the Dscore in each arterial- and venous-dominant phase

Dominant phase Time after Taa (s) Datt (HU) (mean ± SD) p-value Number of patients with mean score ≥2 (%)
 Dscore (mean ± SD) p-value
Intracranial arteries Veins
Arterial-dominant phase 0.0 19.0 ± 37.4 a 0/60 (0) 0/60 (0) 0.15 ± 0.27 a
2.0 154.8 ± 71.7 a 0/60 (0) 0/60 (0) 1.03 ± 0.42 a
4.0 312.1 ± 87.8 a 33/60 (55) 0/60 (0) 1.80 ± 0.39 a
6.0 387.1 ± 85.4 b 59/60 (98) 0/60 (0) 1.97 ± 0.51 b
8.0 364.7 ± 114.2 0.375 60/60 (100) 2/60 (3) 1.88 ± 0.47 0.139
10.0 245.2 ± 144.5 a 60/60 (100) 25/60 (42) 0.97 ± 0.80 a
12.0 9.3 ± 150.3 a 40/60 (67) 47/60 (78) −0.32 ± 0.81 a
Venous-dominant phase 14.0 −233.2 ± 134.7 a 7/60 (12) 58/60 (97) −1.50 ± 0.76 a
16.0 −334.9 ± 96.2 b 0/60 (0) 58/60 (97) −1.92 ± 0.58 b
18.0 −305.8 ± 87.6 0.15 0/60 (0) 38/60 (63) −1.72 ± 0.58 c
20.0 −223.7 ± 96.5 a 0/60 (0) 15/60 (25) −1.35 ± 0.59 a
Open in a new tab

SD, standard deviation; Taa, contrast media arrival time at the internal carotid artery.

a

Indicates p < 0.001.

b

Indicates control.

c

Indicates p < 0.05.

Visual evaluation

With regard to the arteries, 60 (100%) of the 60 patients had a mean score of >2 at 8.0 s and 10.0 s after Taa. The percentage was 98% at 6 s after Taa. With regard to the veins, 58 (97%) of the 60 patients had a mean score of >2 at 14.0 s and 16.0 s after Taa (Table 1). Figure 3 shows the images at the time after Taa in the arterial-dominant phase. The maximum Dscore in the arterial-dominant phase was observed at 6.0 s (Dscore = 1.97) followed by the value at 8.0 s (Dscore = 1.88) after Taa. These values were significantly higher than those in other arterial-dominant phases (p < 0.001). However, there was no statistically significant difference in Dscore between 6.0 s and 8.0 s (p = 0.14). The maximum Dscore in the venous-dominant phase was observed at 16.0 s (Dscore = −1.92) after Taa. This value was significantly higher than those in other venous-dominant phases (p < 0.05) (Table 1). A substantial interrater agreement was observed (κ = 0.613).

Figure 3.

Figure 3.

Open in a new tab

A graph showing mean time–density curves (TDCs) up to 20.0 s from contrast media arrival time at the internal carotid artery (ICA): the solid line is the TDC for the ICA and dashed line is the TDC for the superior sagittal sinus (SSS). HU, Hounsfield unit.

DISCUSSION

The present study was conducted to determine the optimal scan timing for artery–vein separation at whole-brain CTA. For our injection protocol, we recommend a scan timing of 6.0 s or 8.0 s for the arterial phase and 16.0 s for the venous phase after Taa. These scan timings allow the intracranial arteries or veins to be demonstrated in their entirety with minimal superimposition in each phase. This makes it easy to reconstruct 3D volume-rendered images with fusion of the arteries and veins.

In the present study, Ttra was 5.7 s, which is in accordance with the findings of a previous study regarding the cerebral circulation time calculated from patients without any vascular pathology.20 Therefore, we considered that our results obtained on the contralateral side without a mass lesion could be adapted to other patients requiring 3D-CTA. For artery–vein separation, a period of at least Ttra, the cerebral circulation time, is required between the arterial phase and the venous phase. One could argue that scanning at peak enhancement of the ICA and SSS can achieve stronger enhancement; however, Datt is reduced (Figure 2 and Table 1). As a result, simultaneous opacification of the arteries and veins occurs, making it difficult to separate these vessels on brain CTA. In addition, it is difficult to visualize small vessels with 3D volume rendering using the threshold technique if Datt is small. On the other hand, the arterial phase can be obtained before CM arrival in the veins and the venous phase can be obtained after the first pass through the arteries; however, the optimal timing for enhancing the arteries and veins is lost (Figure 2 and Table 1). Smit et al21 reported the usefulness of a simultaneous evaluation technique for the arteries and veins with their CT perfusion data. However, their technique is associated with modest venous superimposition in the evaluation of the circle of Willis. We therefore conducted the present study to determine the optimal timing for the arterial and venous phases at whole-brain CTA.

In our quantitative and qualitative assessment, the optimal arterial scan timing should be present between 6.0 s and 8.0 s after Taa. However, owing to the limited temporal resolution of 4D-CTA as performed in this study (1.0-s acquisition, 1.0-s gap), a more precise value could not be determined. Further studies with higher temporal resolutions are recommended, although this may be difficult owing to concerns regarding radiation exposure. In the arterial-dominant phase, attenuation of the vein begins to rise after around 6.0 s from Taa. The Datt continues in equilibrium until after 8.0 s, a period of 2.0 s, from Taa (Figure 2). Therefore, we could scan the arterial phase in the time period between 6.0 s and 8.0 s from Taa. If the artery–vein separation was performed for the whole-brain CTA using bolus tracking at different scanners with a narrow detector (e.g. 64-row MDCT), we could adapt the scan timing in consideration of a period of 2.0 s. Das et al22 reported that cerebral CTA images with significant reduction of venous contamination could be achieved with a low CM dose using a 320-detector CT. In their study, they determined the peak time of the CTA scan using the test bolus technique. However, there was no quantitative evaluation of the degree of venous contamination relating to the Datt. We evaluated the difference in attenuation between the intracranial arteries and veins. If our results were adapted to the test bolus technique, whole-brain CTA could achieve an additional reduction of venous contamination.

In this study, we investigated the optimal scan timing for the case of bolus tracking. A previous study preferred a fixed delay time for bolus tracking owing to the lack of consistency in optimal triggering.23 In another study of brain CTA at 320-row MDCT, a test bolus was also used for a similar reason.22 However, the combination of our proposed scan timing and detection of the CM arrival at the ICA, confirming the ascent of the ICA attenuation from the baseline, could resolve this defect in 3D-CTA with bolus tracking. We injected the CM for over 10.0 s, which is in line with the Tia and Tiv. Awai et al17 reported that the time from CM arrival to peak enhancement at the aorta depends on the injection duration. In the brain, inflow to the venous system is simply due to outflow from the arteries. Therefore, our results showing that the TDC of the SSS corresponded to that of the ICA with Ttra are in agreement with the findings of a previous study. Based on our results, we believe that we can achieve optimal CTA with good artery–vein separation, if we can accurately detect CM arrival at the ICA using bolus tracking.

The present study suffers from a number of limitations. First, we analyzed 4D-CTA data scanned with a tube voltage of 80 kVp. Our results were analyzed under sufficient vascular attenuation, because CT attenuation can be influenced by tube voltage.24,25 If these results were adapted to the 3D-CTA commonly employed for higher tube voltage in clinical cases, we would need to perform additional studies of the CM dose to maintain sufficient enhancement. Second, we assessed the scan timing for artery–vein separation using a fixed CM injection duration of 10.0 s. Our results showed that appropriate setting of the scan timing allowed observation of the intracranial arteries or veins in their entirety, with minimal superimposition in each phase, when this short CM injection duration was adapted. If different injection durations were employed, we would need to perform additional studies. Third, we did not directly assess the differentiation between the artery and vein. However, we considered this acceptable evaluation because our assessment was performed on the basis of each sufficient depiction of the major intracranial arteries and veins. Additional studies are needed to evaluate more distal arteries limited by the opacification of veins. Fourth, we tried to separate intracranial arteries and veins for patients without vascular lesions. We will try to investigate the optimal timing for patients with arteriovenous malformations, aneurysms or stenoses.

CONCLUSION

The optimal scan timing to achieve good artery–vein separation at whole-brain CTA using 320-row MDCT with a CM injection duration of 10.0 s was found to be 6.0 s or 8.0 s for the arteries and 16.0 s for the veins after CM arrival at the ICA. These recommended scan timings allow the intracranial arteries or veins to be demonstrated in their entirety with minimal superimposition in each phase.

Contributor Information

Takashi Shirasaka, Email: shirasa@med.kyushu-u.ac.jp.

Akio Hiwatashi, Email: hiwatasi@radiol.med.kyushu-u.ac.jp.

Koji Yamashita, Email: yamakou@med.kyushu-u.ac.jp.

Masatoshi Kondo, Email: m-kondo@med.kyushu-u.ac.jp.

Hiroshi Hamasaki, Email: hama1230@med.kyushu-u.ac.jp.

Yamato Shimomiya, Email: shimomiya.rad.technologist@gmail.com.

Yasuhiko Nakamura, Email: yas-nkmr@med.kyushu-u.ac.jp.

Yoshinori Funama, Email: funama@kumamoto-u.ac.jp.

Hiroshi Honda, Email: honda@radiol.med.kyushu-u.ac.jp.

REFERENCES

Articles from The British Journal of Radiology are provided here courtesy of Oxford University Press

RESOURCES