HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) All Versions of this Article: 2423051191v1 242/3/832    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Waaijer, A. Articles by van Leeuwen, M. S. Search for Related Content PubMed PubMed Citation Articles by Waaijer, A. Articles by van Leeuwen, M. S.

Circle of Willis at CT Angiography: Dose Reduction and Image Quality—Reducing Tube Voltage and Increasing Tube Current Settings¹

¹ From the Department of Radiology, University Medical Center Utrecht, Heidelberglaan 100, 3584 CX Utrecht, the Netherlands. Received July 22, 2005; revision requested September 23; revision received March 26, 2006; accepted May 2; final version accepted August 7. Address correspondence to A.W. (e-mail: annetwaaijer{at}gmail.com).

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 
Purpose: To prospectively assess the effects of lower tube voltage and various effective tube currents on image quality for computed tomographic (CT) angiography of the circle of Willis.

Materials and Methods: Institutional review board approval was obtained. Patients or family provided written informed consent. Signal-to-noise ratios (SNRs) were determined in a head phantom for various effective tube currents with tube voltages of 90, 120, and 140 kVp. Patients were referred for CT angiography because of acute subarachnoid hemorrhage (n = 20) or family history of cerebral aneurysms (n = 20). In each group, 10 patients were scanned with 120 kVp and 200 mAs_eff and 10 were scanned with 90 kVp and 330 mAs_eff (CT dose index volumes, 27.2 mGy and 20.6 mGy, respectively). CT numbers were measured in the internal carotid artery at the T junction and compared with a t test. Two radiologists used a five-point scale to subjectively score arterial enhancement, depiction of small arterial detail, image noise, venous contamination, and interference of subarachnoid blood. Mann-Whitney U test was used for statistical analysis.

Results: In the phantom, SNR² was proportional to effective tube current and CT dose index volume. With an identical effective tube current, SNR² was lower at 90 kVp than at 120 or 140 kVp. With identical CT dose index volume, tube voltage of 90 kVp resulted in a 45%–52% increase of SNR² compared with SNR² at 120 kVp. In patients, mean attenuation in the internal carotid artery T junction was higher with 90 kVp (340 HU) than with 120 kVp (252 HU, P < .001). Although dose at 90 kVp was 30% lower than dose at 120 kVp, scores for arterial enhancement and depiction of small arterial detail were higher at 90 kVp than at 120 kVp (4.0 vs 3.2 and 3.6 vs 3.1, respectively; P < .005).

Conclusion: In head phantoms, lower tube voltage improved SNR at equal radiation doses. For CT angiography of the circle of Willis, this translated into superior image quality at 90 kVp.

© RSNA, 2007

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 
Computed tomographic (CT) angiography of the circle of Willis is a well-established minimally invasive diagnostic procedure used to detect cerebral aneurysms (1–4). In patients with subarachnoid hemorrhage (SAH), CT angiography can be performed immediately after unenhanced CT to facilitate prompt treatment planning (5–7). Aneurysms and the vessels that supply them are small; therefore, arterial contrast material enhancement needs to be optimized with use of appropriate contrast material injection and scanning parameters.

There is a trend toward an increased radiation dose with thin-section multi–detector row CT compared with the dose used for single–detector row CT (8). Evaluation of potential options for dose reduction is therefore important and has become the focus of research (9–12).

X-ray absorption of iodine increases substantially with low effective beam energy as long as the effective energy remains above the k edge of iodine (33 keV) (13). When the tube voltage is decreased below the 120–140-kVp level that is typically used for diagnostic CT, the arterial enhancement increases substantially with a constant intravascular iodine concentration. This increased intraarterial enhancement should provide a means for dose reduction without influencing signal-to-noise ratios (SNRs). However, decreasing the tube voltage without changing the tube current may not be sufficient for this goal: At lower tube voltage settings, the dose decreases rapidly if the tube current settings are not adjusted. This results in a substantial increase in image noise that may or may not be compensated for by an increase in arterial enhancement (14). Despite various phantom studies on CT scanning with low tube voltage settings, the relationship between SNR and dose for different tube voltage settings in patients is not well established (15,16).

Two studies on the use of lower tube voltage for CT angiography of the circle of Willis in a clinical setting had contradictory results. Ertl-Wagner et al (17) found that higher tube voltage resulted in better image quality; however, Bahner et al (18) found that image quality benefited from lower tube voltage. Thus, the purpose of our study was to prospectively assess the effect of lower tube voltage and various effective tube currents on image quality for CT angiography of the circle of Willis.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 
First, we performed a phantom study to establish the iodine attenuation curve for the CT scanner with tube voltages of 90, 120, and 140 kVp; this enabled us to quantify the increase in enhancement at lower tube voltage settings. Next, the phantom was scanned with the 90-, 120-, and 140-kVp settings while effective tube current and tube voltage were varied independently. Subsequently, two groups of patients who were scheduled to undergo CT angiography for aneurysm detection were included in a clinical study to compare image quality at 90 and 120 kVp.

CT Scanner
All experiments and clinical studies were performed by using a 16–detector row CT scanner (MX 8000 IDT; Philips, Cleveland, Ohio) with 16 x 0.75-mm collimation (eg, 16 detectors with 0.75-mm section thickness) and a rotation time of 0.75 second. A slightly smoothing filter (B filter; Philips Medical Systems, Best, the Netherlands) was used in combination with a 512 x 512 matrix. Images were reconstructed with a section thickness of 1.0 mm and a reconstruction increment of 0.5 mm. To ensure consistent image quality, the scanner was calibrated every week.

Phantom Study
We used a proprietary cylindrical polymethylmethacrylate (PMMA) head phantom with a diameter of 185 mm. A 25-mm-diameter tube filled with contrast material could be inserted into the center of the phantom.

Iodine attenuation curve and statistical analysis.—We prepared 13 iodine concentrations (0.0, 0.5, 0.75, 1.0, 1.5, 2.0, 4.0, 6.0, 7.5, 12.0, 25.0, 30.0, and 50.0 mg of iodine per milliliter [mg I/mL]) by diluting contrast material (300 mg I/mL; Ultravist 300, Schering, Berlin, Germany) with distilled water. All contrast material solutions were prepared just prior to scanning to prevent settling of the solution. We used tube voltage settings of 90, 120, and 140 kVp and chose the effective tube current settings so that the head phantom CT dose index volume was as close as possible to 27.2 mGy for all three protocols. This was achieved by using a pitch of 0.3, a field of view of 200 mm, and 450 mAs_eff with 90 kVp, 200 mAs_eff with 120 kVp, and 135 mAs_eff with 140 kVp.

Attenuation was measured by one observer (A.W.), who drew a region of interest (approximately 100 mm²) in the middle of the tube. The mean attenuation (measured in Hounsfield units) for this region of interest was plotted against the iodine concentration. Data were analyzed with linear regression analysis and by calculating Pearson correlation coefficients.

Variation of tube voltage and tube current settings.—To measure SNR, we chose the iodine concentration that best yielded the contrast enhancement seen in patients. (A concentration of 12 mg I/mL yielded enhancement of approximately 250–420 HU, depending on tube voltage.) The phantom was scanned with three tube voltages—90, 120, and 140 kVp—in combination with nine tube current settings that ranged from 25 to 370 mAs_eff.

Dose measurements.—To estimate patient dose at the various tube voltage and tube current settings, we recorded the CT numbers for CT dose index volume indicated on the scanner interface and compared them with our own measurements obtained in a standard 160-mm CT dose index head phantom. To relate SNR to local dose, we also obtained dose measurements in the central hole of the 185-mm CT dose index head phantom that we used to calculate SNR.

All dose measurements were obtained by two observers (A.W. and M.P.) with a 10-cm pencil ionization chamber (Solidose 400; RTI Electronics, Goteborg, Sweden). Doses were measured 10 times, and an average was calculated.

For the standard 160-mm CT dose index phantom, a weighted average of the measurements obtained in the four peripheral positions and in the central position was used to calculate the weighted CT dose index, which in a static setup is identical to the CT dose index volume (19). Linear regression analysis was used to determine a calibration factor for the scanner that was used to transform the effective tube current into a CT dose index volume for the three tube voltage settings.

Local CT dose index in the 185-mm PMMA phantom was measured in the central hole. Again, we performed linear regression analysis to determine a calibration factor that was used to convert effective tube current to local CT dose index values for the three tube voltage settings.

SNR measurements and statistical analysis.—For each scan of the 185-mm PMMA phantom that contained contrast material, we determined the mean CT number (attenuation) and standard deviation (image noise) by placing a region of interest (approximately 100 mm²) in the center of the tube. Measurements were obtained by one observer (A.W.) at five equidistant levels along the z-axis to correct for variations in noise over the phantom. SNR was calculated by dividing the mean CT number in the tube by the standard deviation.

Because CT numbers are normalized to water (0 HU), the CT number of a contrast material–enhanced structure actually corresponds to the contrast between this structure and water. Therefore, the SNR used in the aforementioned definition can also be interpreted as a contrast-to-noise ratio in which contrast is defined as the difference in CT numbers between the contrast-enhanced structure and water.

In theory, image noise () and radiation dose (d) in CT are inversely related (² 1/d) (20). Dose as determined with the CT dose index is proportional to the effective tube current; therefore, SNR² can be expected to show a linear relationship with effective tube current and CT dose index (13). Because of this linear relationship, we used SNR² to quantify the differences in SNR between the three tube voltage settings. Both the relationship between SNR² and effective tube current and the relationship between SNR² and CT dose index (measured CT dose index volume and local CT dose index) were established for the three tube voltage settings by using linear regression analysis and by determining the Pearson correlation coefficient.

Clinical Study
Patients.—We included 40 patients who were referred for (a) CT angiography of the circle of Willis because of SAH or (b) screening because of a family history that was positive for intracranial aneurysms. In each group, we recruited 20 consecutive patients between October 2003 and March 2004. In both groups, the first 10 patients were consecutively scanned with 120 kVp and the next 10 were scanned with 90 kVp.

Poor-quality CT angiograms due to patient movement or failure of contrast material injection were excluded from the study. Mean age of the patients was 61 years ± 12.9 (standard deviation) (age range, 22–85 years); there were 17 men and 23 women. This study was performed with institutional review board approval. Informed consent was obtained from all patients or their family members.

Scanning protocol.—We chose our standard scanning parameters for the circle of Willis according to the recommendations of the CT unit manufacturer. Consequently, we used a collimation of 16 x 0.75 mm, a pitch of 1.0, a rotation time of 0.75 second, and exposure settings of 120 kVp and 200 mAs_eff, which resulted in a CT dose index volume of 27.2 mGy. The scanning volume extended from the posterior arch of the C1 vertebra to the vertex of the brain. Images were reconstructed with a section thickness of 1.0 mm and a reconstruction increment of 0.5 mm; the field of view was 160 mm.

At 90 kVp, we used the same scanning and reconstruction parameters; however, we used a different effective tube current. The tube current settings were intended to result in a CT dose index that was as close as possible to that at 120 kVp. However, at a pitch of 1.0, the maximum tube current available on the scanner was only 330 mAs_eff (not 450 mAs_eff as would have been theoretically required); therefore, the resultant CT dose index volume was 20.6 mGy and was thus 33% lower than that at 120 kVp.

Contrast material (Ultravist 300) was injected for 17 seconds (50 mL was injected at a rate of 5 mL/sec and 20 mL was injected at a rate of 3 mL/sec). This injection was followed by injection of a 30-mL saline chaser administered at a rate of 3 mL/sec. A test bolus was applied to determine the individual scan delay.

Image assessment and statistical analysis.—One observer (A.W., 1 year of full-time experience in head CT angiography) under the supervision of an author (M.S.v.L.) measured attenuation in the internal carotid artery at the T junction by drawing a 2.5–3.5-mm² region of interest just within the vessel lumen. Mean CT number and standard deviation were noted. Attenuation of the brain parenchyma was noted by placing a 250-mm² region of interest within the center of one occipital lobe and avoiding vascular structures. CT numbers were rounded to the nearest whole number. SNR was calculated by dividing the mean enhancement at the internal carotid artery T junction (attenuation) by the standard deviation in the brain parenchyma (image noise).

Two radiologists working independently (B.K.V., G.A.P.d.K.) who had more than 5 years of experience in evaluation of the circle of Willis (>300 CT angiographic examinations per year) subjectively rated image quality on a five-point scale in which the center of the scale (rating of 3) corresponded to the image quality expected with standard CT angiography of the circle of Willis. A score of 1 indicated nondiagnostic image quality; a score of 2, substandard image quality; a score of 3, standard image quality; a score of 4, better than standard image quality; and a score of 5, excellent image quality. The two observers also independently scored the following image quality parameters: arterial enhancement, visibility of small arterial detail (based on depiction of small arteries, such as the ophthalmic, anterior choroid, anterior and posterior communicating, and superior cerebellar arteries), interference of venous structures (venous contamination), and image noise. A lower tube voltage may also increase venous enhancement. Thus, we also evaluated the influence of interference of venous structures on image quality. In patients with SAH, we also evaluated interference of subarachnoid blood, which was defined as the visibility of vascular detail in regions where perivascular hemorrhage was present. Observers were blinded as to whether scanning was performed with 120 kVp or 90 kVp. They were free to choose window and level settings as they deemed appropriate.

For evaluation of subjective scoring, the Wilcoxon signed rank test was used to analyze the differences between 90 and 120 kVp for both observers. For further analysis—except for the evaluation of subjective scores for interference of subarachnoid blood, in which only patients with SAH were included—we pooled data from both observers and both clinical indications. The Mann-Whitney U test for two independent samples was used to analyze differences in arterial attenuation and SNR, as well as subjective scoring differences between scanning performed with 90 kVp and scanning performed with 120 kVp. P values less than .05 were considered to indicate a significant difference.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 
Iodine Attenuation Curve in the Phantom
We found a linear relationship between iodine concentration and attenuation for all tube voltages (Fig 1). Pearson correlation coefficients were as follows: r 0.999 for 90 kVp and r = 0.999 for both 120 kVp and 140 kVp (P < .001) (Fig 1). Attenuation at 90 kVp (33 HU/mg I) was 43% higher than attenuation at 120 kVp (23 HU/mg I) and 74% higher than attenuation at 140 kVp (19 HU/mg I).

View larger version (10K): [in this window] [in a new window] [Download PPT slide]   Figure 1: Graph shows CT number as a function of iodine concentration. The linear relationships for three tube voltage settings were as follows: at 90 kVp, y = 32.5x + 13.3 (r = 0.999); at 120 kVp, y = 22.8x + 13.4 (r = 0.999); and at 140 kVp, y = 19.1x + 12.0 (r = 0.999).

View larger version (9K): [in this window] [in a new window] [Download PPT slide]   Figure 2a: Graphs show the relationship between (a) effective tube current (mAseff) and noise, (b) effective tube current and SNR2, (c) CT dose index volume (CTDIvol) and SNR2, and (d) local CT dose index (CTDIlocal) and SNR2. SNR was calculated by dividing the mean enhancement in a region of interest, as measured at the internal carotid artery T junction (attenuation), by the standard deviation in the brain parenchyma (noise), as measured in the occipital lobe. In b, the relationship between effective tube current and SNR2 was as follows: for 140 kVp, SNR2 = 1.94 · mAseff (r = 1.000); for 120 kVp, SNR2 = 1.87 · mAseff (r = 0.998); and for 90 kVp, SNR2 = 1.27 · mAseff (r = 0.998). In c, the relationship between dose, expressed as CT dose index volume, and SNR2 was as follows: for 90 kVp, SNR2 = 20.8 · CTDIvol (r = 0.999); for 120 kVp, SNR2 = 14.3 · CTDIvol (r = 0.993); and for 140 kVp, SNR2 = 10.3 · CTDIvol (r = 0.999). In d, the relationship between local CT dose index as measured within the 185-mm PMMA phantom and SNR2 was as follows: for 90 kVp, SNR2 = 27.5 · CTDIlocal (r = 0.996); for 120 kVp, SNR2 = 18.2 · CTDIlocal (r = 0.997); and for 140 kVp, SNR2 = 12.8 · CTDIlocal (r = 0.999).

View larger version (9K): [in this window] [in a new window] [Download PPT slide]   Figure 2b: Graphs show the relationship between (a) effective tube current (mAseff) and noise, (b) effective tube current and SNR2, (c) CT dose index volume (CTDIvol) and SNR2, and (d) local CT dose index (CTDIlocal) and SNR2. SNR was calculated by dividing the mean enhancement in a region of interest, as measured at the internal carotid artery T junction (attenuation), by the standard deviation in the brain parenchyma (noise), as measured in the occipital lobe. In b, the relationship between effective tube current and SNR2 was as follows: for 140 kVp, SNR2 = 1.94 · mAseff (r = 1.000); for 120 kVp, SNR2 = 1.87 · mAseff (r = 0.998); and for 90 kVp, SNR2 = 1.27 · mAseff (r = 0.998). In c, the relationship between dose, expressed as CT dose index volume, and SNR2 was as follows: for 90 kVp, SNR2 = 20.8 · CTDIvol (r = 0.999); for 120 kVp, SNR2 = 14.3 · CTDIvol (r = 0.993); and for 140 kVp, SNR2 = 10.3 · CTDIvol (r = 0.999). In d, the relationship between local CT dose index as measured within the 185-mm PMMA phantom and SNR2 was as follows: for 90 kVp, SNR2 = 27.5 · CTDIlocal (r = 0.996); for 120 kVp, SNR2 = 18.2 · CTDIlocal (r = 0.997); and for 140 kVp, SNR2 = 12.8 · CTDIlocal (r = 0.999).

View larger version (10K): [in this window] [in a new window] [Download PPT slide]   Figure 2c: Graphs show the relationship between (a) effective tube current (mAseff) and noise, (b) effective tube current and SNR2, (c) CT dose index volume (CTDIvol) and SNR2, and (d) local CT dose index (CTDIlocal) and SNR2. SNR was calculated by dividing the mean enhancement in a region of interest, as measured at the internal carotid artery T junction (attenuation), by the standard deviation in the brain parenchyma (noise), as measured in the occipital lobe. In b, the relationship between effective tube current and SNR2 was as follows: for 140 kVp, SNR2 = 1.94 · mAseff (r = 1.000); for 120 kVp, SNR2 = 1.87 · mAseff (r = 0.998); and for 90 kVp, SNR2 = 1.27 · mAseff (r = 0.998). In c, the relationship between dose, expressed as CT dose index volume, and SNR2 was as follows: for 90 kVp, SNR2 = 20.8 · CTDIvol (r = 0.999); for 120 kVp, SNR2 = 14.3 · CTDIvol (r = 0.993); and for 140 kVp, SNR2 = 10.3 · CTDIvol (r = 0.999). In d, the relationship between local CT dose index as measured within the 185-mm PMMA phantom and SNR2 was as follows: for 90 kVp, SNR2 = 27.5 · CTDIlocal (r = 0.996); for 120 kVp, SNR2 = 18.2 · CTDIlocal (r = 0.997); and for 140 kVp, SNR2 = 12.8 · CTDIlocal (r = 0.999).

View larger version (11K): [in this window] [in a new window] [Download PPT slide]   Figure 2d: Graphs show the relationship between (a) effective tube current (mAseff) and noise, (b) effective tube current and SNR2, (c) CT dose index volume (CTDIvol) and SNR2, and (d) local CT dose index (CTDIlocal) and SNR2. SNR was calculated by dividing the mean enhancement in a region of interest, as measured at the internal carotid artery T junction (attenuation), by the standard deviation in the brain parenchyma (noise), as measured in the occipital lobe. In b, the relationship between effective tube current and SNR2 was as follows: for 140 kVp, SNR2 = 1.94 · mAseff (r = 1.000); for 120 kVp, SNR2 = 1.87 · mAseff (r = 0.998); and for 90 kVp, SNR2 = 1.27 · mAseff (r = 0.998). In c, the relationship between dose, expressed as CT dose index volume, and SNR2 was as follows: for 90 kVp, SNR2 = 20.8 · CTDIvol (r = 0.999); for 120 kVp, SNR2 = 14.3 · CTDIvol (r = 0.993); and for 140 kVp, SNR2 = 10.3 · CTDIvol (r = 0.999). In d, the relationship between local CT dose index as measured within the 185-mm PMMA phantom and SNR2 was as follows: for 90 kVp, SNR2 = 27.5 · CTDIlocal (r = 0.996); for 120 kVp, SNR2 = 18.2 · CTDIlocal (r = 0.997); and for 140 kVp, SNR2 = 12.8 · CTDIlocal (r = 0.999).

Clinical Study
For the clinical scans, mean attenuations at the internal carotid artery T junction were 340 HU (range, 189–446 HU) at 90 kVp and 253 HU (range, 185–349 HU) at 120 kVp, with a mean difference of 87 HU (95% confidence interval: 53 HU, 121 HU; P < .001). These data indicated a 35% increase in arterial attenuation at 90 kVp as compared with arterial attenuation at 120 kVp (Fig 3). Brain parenchyma attenuation at 90 kVp (37 HU) was not significantly different from that at 120 kVp (35 HU) (Table 2). The SNR at 120 kVp (19.8) was not significantly different from that at 90 kVp (21.6).

View larger version (8K): [in this window] [in a new window] [Download PPT slide]   Figure 3: Box plots show arterial attenuation in 20 patients scanned with 90 kVp and 20 patients scanned with 120 kVp. Error bars indicate the range of values of arterial attenuation, boxes contain all values within the 25th to 75th percentiles (interquartile range), and thick black lines represent the median.

View larger version (115K): [in this window] [in a new window] [Download PPT slide]   Figure 4a: (a, c) Unenhanced CT images and (b, d) corresponding CT angiograms obtained in two patients with SAH. Unenhanced CT was performed with the same protocol in both patients. CT angiography was performed with 120 kVp and 200 mAseff in b and with 90 kVp and 330 mAseff in d. Both angiograms are 20-mm-thick slab maximum intensity projections obtained with equal window width and level settings (470 and 200 HU, respectively); these images show that 90-kVp scanning resulted in increased visibility of vascular structures surrounded by subarachnoid blood. In a and b, the patient had a right middle cerebral artery aneurysm (arrow), with a total score of 2 assigned by both observers. Observers A and B assigned scores of 2 and 3, respectively, for interference of subarachnoid blood. In c and d, the patient had a left middle cerebral artery aneurysm (arrow) with total scores of 5 (assigned by observer A) and 4 (assigned by observer B). Observers A and B assigned scores of 5 and 4, respectively, for interference of subarachnoid blood.

View larger version (148K): [in this window] [in a new window] [Download PPT slide]   Figure 4b: (a, c) Unenhanced CT images and (b, d) corresponding CT angiograms obtained in two patients with SAH. Unenhanced CT was performed with the same protocol in both patients. CT angiography was performed with 120 kVp and 200 mAseff in b and with 90 kVp and 330 mAseff in d. Both angiograms are 20-mm-thick slab maximum intensity projections obtained with equal window width and level settings (470 and 200 HU, respectively); these images show that 90-kVp scanning resulted in increased visibility of vascular structures surrounded by subarachnoid blood. In a and b, the patient had a right middle cerebral artery aneurysm (arrow), with a total score of 2 assigned by both observers. Observers A and B assigned scores of 2 and 3, respectively, for interference of subarachnoid blood. In c and d, the patient had a left middle cerebral artery aneurysm (arrow) with total scores of 5 (assigned by observer A) and 4 (assigned by observer B). Observers A and B assigned scores of 5 and 4, respectively, for interference of subarachnoid blood.

View larger version (108K): [in this window] [in a new window] [Download PPT slide]   Figure 4c: (a, c) Unenhanced CT images and (b, d) corresponding CT angiograms obtained in two patients with SAH. Unenhanced CT was performed with the same protocol in both patients. CT angiography was performed with 120 kVp and 200 mAseff in b and with 90 kVp and 330 mAseff in d. Both angiograms are 20-mm-thick slab maximum intensity projections obtained with equal window width and level settings (470 and 200 HU, respectively); these images show that 90-kVp scanning resulted in increased visibility of vascular structures surrounded by subarachnoid blood. In a and b, the patient had a right middle cerebral artery aneurysm (arrow), with a total score of 2 assigned by both observers. Observers A and B assigned scores of 2 and 3, respectively, for interference of subarachnoid blood. In c and d, the patient had a left middle cerebral artery aneurysm (arrow) with total scores of 5 (assigned by observer A) and 4 (assigned by observer B). Observers A and B assigned scores of 5 and 4, respectively, for interference of subarachnoid blood.

View larger version (145K): [in this window] [in a new window] [Download PPT slide]   Figure 4d: (a, c) Unenhanced CT images and (b, d) corresponding CT angiograms obtained in two patients with SAH. Unenhanced CT was performed with the same protocol in both patients. CT angiography was performed with 120 kVp and 200 mAseff in b and with 90 kVp and 330 mAseff in d. Both angiograms are 20-mm-thick slab maximum intensity projections obtained with equal window width and level settings (470 and 200 HU, respectively); these images show that 90-kVp scanning resulted in increased visibility of vascular structures surrounded by subarachnoid blood. In a and b, the patient had a right middle cerebral artery aneurysm (arrow), with a total score of 2 assigned by both observers. Observers A and B assigned scores of 2 and 3, respectively, for interference of subarachnoid blood. In c and d, the patient had a left middle cerebral artery aneurysm (arrow) with total scores of 5 (assigned by observer A) and 4 (assigned by observer B). Observers A and B assigned scores of 5 and 4, respectively, for interference of subarachnoid blood.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

Some reports regarding optimization of image quality in CT angiography of the brain by means of adapting tube voltage have been published (17,18). While most articles advocate 120 kVp for CT angiography of the brain (1–5), we found that enhancement was higher on CT angiograms acquired with 90 kVp and that experienced radiologists rated image quality of CT images acquired with 90 kVp as being better than that of CT angiograms acquired with 120 kVp, even though we had to use a reduced CT dose index volume of 20.6 mGy for 90-kVp scans instead of the conventional 27.2-mGy CT dose index volume used for 120-kVp scans. The improved subjective scores were found for vascular enhancement, visibility of small arterial detail, and—in patients with SAH—interference of subarachnoid blood. This is concordant with the findings of Bahner et al (18), which showed improved vascular opacification with a low tube voltage in patients with intracranial vascular malformations who were examined with tube voltages of 80 and 120 kVp during follow-up after radiation therapy. In this study, the effective tube current was also adjusted to the lowest tube voltage setting, which resulted in an even lower CT dose index volume than that in our study (13.5 mGy at 80 kVp versus 21.9 mGy at 120 kVp).

Our clinical results seem to be in conflict with the findings of Ertl-Wagner et al (17), who reported better image quality for cerebral CT angiograms obtained with a higher tube voltage. The principal difference between our study and that of Ertl-Wagner et al (17) is that Ertl-Wagner and co-workers chose not to adapt effective tube current settings when they reduced tube voltage, so there was an almost four-fold increase in radiation dose at the highest tube voltage setting. This led to substantial noise reduction, which can explain the improved image quality at a higher tube voltage.

Our findings in the clinical study are in concordance with our findings in the phantom study in that SNR was substantially lower at 90 kVp than at 120 kVp if the effective tube current settings were kept constant. However, when we compared the SNR at identical doses (CT dose index volume or local CT dose index), we found SNR was substantially higher at 90 kVp than at 120 or 140 kVp. In fact, SNR² was 45%–52% higher at 90 kVp than at 120 kVp. Thus, we believe that the difference in study design (ie, varying tube voltage at constant effective tube current, as done by Ertl-Wagner et al [17], versus varying tube voltage at constant radiation dose, as in the current study) can explain the seemingly contradictory results.

In our clinical study, we found no difference between 90 and 120 kVp regarding subjective scoring of venous contamination or image noise. The lack of difference in the image noise score probably had to do with the fact that we allowed readers to adapt window settings to the arterial enhancement: On average, 90-kVp scans had better enhancement and therefore required wider window settings. Such wider window settings, on the other hand, reduce the subjective appreciation of image noise (21).

We chose to calculate the SNR instead of the contrast-to-noise ratio because the attenuation of PMMA is dependent on tube voltage, thus, incorporation of background attenuation would have strongly influenced contrast-to-noise ratios in the phantom study. Also, background attenuation in patients can vary between liquor density, brain tissue, and hemorrhage and thus influence contrast-to-noise ratio. Therefore, we also used the SNR to compare the phantom study findings with the clinical study findings.

Our study had limitations. First, scanning protocols could not be compared intraindividually, as it was deemed unethical to scan each patient with both 120 and 90 kVp. Second, for optimum comparison, CT angiography at 90 kVp should have been performed at the same dose (CT dose index volume) used with 120 kVp, but tube limitations prevented this. Our results do however imply that equal or better results can be obtained even with a dose reduction of approximately 30%. However, it should be realized that these results, and especially the dose-noise relationship, are specific for CT images and do not necessarily apply to conventional radiographic images.

Third, we were unable to determine how much sensitivity and specificity for detection of aneurysms would increase with 90-kVp scanning because angiography was not used as an independent reference standard in all patients. Finally, we did not evaluate the effect of low tube voltage in the presence of metallic material (vascular clips or coils) that could cause stronger streak artifacts.

In conclusion, the use of a lower tube voltage setting substantially improves SNR if radiation dose is kept constant. For evaluation of the circle of Willis in patients, this translates to superior image quality for 90-kVp scanning compared with the image quality achieved with 120-kVp scanning, even if patient dose is reduced by approximately 30%. For this reason, we recommend low tube voltage scanning with properly adjusted tube current for CT angiography of the circle of Willis in patients with acute SAH or for screening of patients with a family history positive for intracranial aneurysms.

	ADVANCES IN KNOWLEDGE

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 

• With use of an identical effective tube current setting, signal-to-noise ratios (SNRs) at 120 and 140 kVp are almost identical; however, SNR is substantially lower at 90 kVp.
  
• With an identical dose (CT dose index volume), the use of 90 kVp results in an SNR² increase of approximately 50% compared with the SNR² at 120 kVp and an increase of approximately 120% compared with the SNR² at 140 kVp.
  
• In patients, the use of 90 kVp improves image quality at lower patient dose compared with image quality at 120 kVp.
  

	ACKNOWLEDGMENTS

 
We thank Professor F. W. Zonneveld for his useful comments regarding this article.

	FOOTNOTES

 

Abbreviations: PMMA = polymethylmethacrylate • SAH = subarachnoid hemorrhage • SNR = signal-to-noise ratio

Authors stated no financial relationship to disclose.

Author contributions: Guarantors of integrity of entire study, A.W., M.P.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, A.W., M.P., C.J.G.B., M.S.v.L.; clinical studies, A.W., M.P., B.K.V., G.A.P.d.K., M.S.v.L.; experimental studies, A.W., M.P., C.J.G.B.; statistical analysis, A.W.; and manuscript editing, all authors

	References

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 

1. Kato Y, Katada K, Hayakawa M, et al. Can 3D-CTA surpass DSA in diagnosis of cerebral aneurysm? Acta Neurochir (Wien) 2001;143:245–250.
2. Velthuis BK, Rinkel GJ, Ramos LM, et al. Subarachnoid hemorrhage: aneurysm detection and preoperative evaluation with CT angiography. Radiology 1998;208:423–430.[Abstract/Free Full Text]
3. Villablanca JP, Jahan R, Hooshi P, et al. Detection and characterization of very small cerebral aneurysms by using 2D and 3D helical CT angiography. AJNR Am J Neuroradiol 2002;23:1187–1198.[Abstract/Free Full Text]
4. Wintermark M, Uske A, Chalaron M, et al. Multislice computerized tomography angiography in the evaluation of intracranial aneurysms: a comparison with intraarterial digital subtraction angiography. J Neurosurg 2003;98:828–836.[Medline]
5. Velthuis BK, van Leeuwen MS, Witkamp TD, Ramos LM, Berkelbach van der Sprenkel JW, Rinkel GJ. Computerized tomography angiography in patients with subarachnoid hemorrhage: from aneurysm detection to treatment without conventional angiography. J Neurosurg 1999;91:761–767.[Medline]
6. Verro P, Tanenbaum LN, Borden NM, Sen S, Eshkar N. CT angiography in acute ischemic stroke: preliminary results. Stroke 2002;33:276–278.[Abstract/Free Full Text]
7. Ruigrok YM, Rinkel GJ, Buskens E, Velthuis BK, van Gijn J. Perimesencephalic hemorrhage and CT angiography: a decision analysis. Stroke 2000;31:2976–2983.[Abstract/Free Full Text]
8. Brix G, Nagel HD, Stamm G, et al. Radiation exposure in multi-slice versus single-slice spiral CT: results of a nationwide survey. Eur Radiol 2003;13:1979–1991.[CrossRef][Medline]
9. Wintermark M, Maeder P, Verdun FR, et al. Using 80 kVp versus 120 kVp in perfusion CT measurement of regional cerebral blood flow. AJNR Am J Neuroradiol 2000;21:1881–1884.[Abstract/Free Full Text]
10. Sigal-Cinqualbre AB, Hennequin R, Abada HT, Chen X, Paul JF. Low-kilovoltage multi–detector row chest CT in adults: feasibility and effect on image quality and iodine dose. Radiology 2004;231:169–174.[Abstract/Free Full Text]
11. Cohnen M, Fischer H, Hamacher J, Lins E, Kotter R, Modder U. CT of the head by use of reduced current and kilovoltage: relationship between image quality and dose reduction. AJNR Am J Neuroradiol 2000;21:1654–1660.[Abstract/Free Full Text]
12. Mullins ME, Lev MH, Bove P, et al. Comparison of image quality between conventional and low-dose nonenhanced head CT. AJNR Am J Neuroradiol 2004;25:533–538.[Abstract/Free Full Text]
13. Farr RF, Allisy-Roberts PJ. Radiation physics. In: Physics for medical imaging. London, England: Elsevier Science, 2002; 19–22.
14. Zonneveld FW. Principles of computed tomography. In: Computed tomography of the temporal bone and orbit. Munich, Germany: Urban & Schwarzenberg, 1987; 12–13.
15. Zatz LM. The effect of the kVp level on EMI values: selective imaging of various materials with different kVp settings. Radiology 1976;119:683–688.[Abstract]
16. Keller MR, Kessler RM, Brooks RA, Kirkland LR. Optimum energy for performing CT iodinated constrast studies. Br J Radiol 1980;53:576–579.[Abstract/Free Full Text]
17. Ertl-Wagner BB, Hoffmann RT, Bruning R, et al. Multi–detector row CT angiography of the brain at various kilovoltage settings. Radiology 2004;231:528–535.[Abstract/Free Full Text]
18. Bahner ML, Bengel A, Brix G, Zuna I, Kauczor HU, Delorme S. Improved vascular opacification in cerebral computed tomography angiography with 80 kVp. Invest Radiol 2005;40:229–234.[CrossRef][Medline]
19. DeMarco JJ, Cagnon CH, Cody DD, et al. A Monte Carlo based method to estimate radiation dose from multidetector CT (MDCT): cylindrical and anthropomorphic phantoms. Phys Med Biol 2005;50:3989–4004.[CrossRef][Medline]
20. Prokop M. Image analysis. In: Prokop M, Galanski M. Computed tomography of the body. Stuttgart, Germany: Georg Thieme Verlag, 2003; 143.
21. Husstedt H, Prokop M, Becker H. Window width as a dosage-relevant factor in high-contrast structures in CT [in German]. Rofo 1998;168:139–143.[Medline]

This article has been cited by other articles:

				D. Marin, R. C. Nelson, E. Samei, E. K. Paulson, L. M. Ho, D. T. Boll, D. M. DeLong, T. T. Yoshizumi, and S. T. Schindera Hypervascular Liver Tumors: Low Tube Voltage, High Tube Current Multidetector CT during Late Hepatic Arterial Phase for Detection--Initial Clinical Experience Radiology, June 1, 2009; 251(3): 771 - 779. [Abstract] [Full Text] [PDF]

				C.-Y. Yang, Y.-F. Chen, C.-W. Lee, A. Huang, Y. Shen, C. Wei, and H.-M. Liu Multiphase CT Angiography versus Single-Phase CT Angiography: Comparison of Image Quality and Radiation Dose AJNR Am. J. Neuroradiol., August 1, 2008; 29(7): 1288 - 1295. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) All Versions of this Article: 2423051191v1 242/3/832    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Waaijer, A. Articles by van Leeuwen, M. S. Search for Related Content PubMed PubMed Citation Articles by Waaijer, A. Articles by van Leeuwen, M. S.