Vascular Enhancement and Image Quality of CT Venography: Comparison of Standard and Low Kilovoltage Settings

Cardiopulmonary Imaging Original Research Fujikawa et al. Vascular Enhancement and Image Quality of CTV Cardiopulmonary Imaging Original Research Atsuko Fujikawa 1 Shin Matsuoka Kenji Kuramochi Tatsuo Yoshikawa Kunihiro Yagihashi Yasuyuki Kurihara Yasuo Nakajima Fujikawa A, Matsuoka S, Kuramochi K, et al. Keywords: CT venography, deep venous thrombosis, low kilovoltage setting, peak kilovoltage, radiation dose DOI:10.2214/AJR.10.5424 Received August 2, 2010; accepted after revision February 23, 2011. 1 All authors: Department of Radiology, St. Marianna University School of Medicine, 2-16-1 Sugao, Miyamae-Ku, Kawasaki City, Kanagawa 216-8511, Japan. Address correspondence to A. Fujikawa (a2fujikawa@marianna-u.ac.jp). AJR 2011; 197:838 843 0361 803X/11/1974 838 American Roentgen Ray Society Vascular Enhancement and Image Quality of CT Venography: Comparison of Standard and Low Kilovoltage Settings OBJECTIVE. The objective of our study was to investigate the vascular enhancement and image quality of CT venography (CTV) with a lower peak kilovoltage (kvp) setting than the standard setting. MATERIALS AND METHODS. In this retrospective study, the clinical records of 100 consecutive patients with suspected pulmonary embolism were analyzed. All patients underwent pulmonary CT angiography and CTV of the abdomen, pelvis, and lower extremities using 64-MDCT with automatic tube current modulation: 50 patients underwent CT at 120 kvp, the standard kvp setting, and 50 patients were scanned at 100 kvp; we refer to these groups as the standard-kvp group and the low-kvp group, respectively. Vessel enhancement and image noise were assessed in the inferior vena cava (IVC), femoral vein, and popliteal vein. Two radiologists who were blinded to the kvp setting placed the regions of interest on vessels by consensus and assessed image quality using a 5-point visual scale. Effective dose was estimated using the dose-length product. The Wilcoxon rank test was used to evaluate differences between the two groups using statistics software (JMP, version 5.1). A p value of less than 0.05 was considered to indicate statistical significance. RESULTS. Mean vascular enhancement was significantly higher in the low-kvp group than in the standard-kvp group: IVC, 138.4 ± 12.2 (SD) HU versus 164.5 ± 17.4 HU, respectively; femoral vein, 130.2 ± 18.0 HU versus 152.0 ± 24.5 HU; and popliteal vein, 136.7 ± 17.5 HU versus 158.3 ± 26.0 HU. Although the images of the low-kvp group had significantly higher image noise, there were no significant differences in image quality in the IVC and popliteal vein. The mean effective dose for the low-kvp protocol was significantly lower than that for the standard-kvp protocol. CONCLUSION. Lowering the kvp setting for CTV examinations improved vascular enhancement while providing sufficient image quality. D eep venous thrombosis (DVT) and pulmonary embolism (PE) are considered to be parts of the same pathologic process. Since the combination of CT venography (CTV) and pulmonary CT angiography (CTA) was initially described in 1998 [1], several researchers have reported highly accurate diagnoses of DVT in patients with PE [1, 2]. This method enables the simultaneous diagnosis of PE and DVT without the administration of additional contrast material. The fact that the Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED) II study investigators reported a higher sensitivity (90%) for PE using combined pulmonary CTA and CTV than using pulmonary CTA alone (83%) seems to validate this method [2]. The disadvantage of combined pulmonary CTA and CTV is the increased radiation dose to which the patient is exposed. Several methods have been proposed to reduce the radiation dose from pulmonary CTA and cardiac CTA [3 5], of which lowering the peak kilovoltage (kvp) setting is very effective. Goodman et al. [6] reported that obtaining discontiguous images for CTV reduced radiation dose. Other investigators have reported that a reduction in the kvp setting from 120 to 100 kvp during pulmonary CTA results in an average reduction of the estimated effective dose of approximately 44% [7]. Furthermore, lowering the kvp setting has been suggested as a technique to improve contrast enhancement [7, 8]. However, to our knowledge, the efficacy of a low-kilovoltage protocol for CTV has not been previously analyzed. On 838 AJR:197, October 2011

Vascular Enhancement and Image Quality of CTV the basis of previous reports of the efficacy of lowering the kvp setting in pulmonary CTA, we hypothesized that lowering the kvp setting in CTV would improve vascular enhancement and reduce radiation exposure without a significant degradation of image quality. Thus, the purpose of our study was to compare contrast enhancement and image quality of veins on CTV images obtained using the standard kvp setting and CTV images obtained using a lower kvp setting and to determine whether CTV with a low kvp setting can improve vascular enhancement without degrading the image quality. Materials and Methods Subjects Our study comprised a retrospective evaluation of MDCT studies performed in routine clinical practice before and after changing the standard scanning protocol for the evaluation of PE and DVT at our institute. The change in protocol, which involved a reduction in kvp, was adopted because this technique was reported to improve contrast enhancement while also reducing radiation dose [7 11]. In our institution, combined pulmonary CTA and CTA have been performed using 100 kvp since September 2009. We could make the prospective aspect of this research, but this protocol change was not made for an academic purpose; instead, the change was initiated because of its utility in clinical practice. This retrospective study was performed with institutional review board approval, and the requirement for informed consent was waived. The clinical records (age, sex, body mass index [BMI]) of 100 consecutive patients (74 women and 26 men; mean age ± SD, 64 ± 15 years; age range, 23 86 years) suspected of having PE were analyzed in this study. Fifty patients (36 women and 14 men; mean age, 64 ± 15 years; age range, 23 86 years) were scanned using the 120- kvp protocol; we refer to this group as the standard-kvp group. Another group of 50 patients (38 women and 12 men; mean age, 64 ± 15 years; age range, 34 86 years) underwent CT after the scanning protocol had been changed to 100 kvp; we refer to this group as the low-kvp group. Patients who had undergone amputation of a lower extremity and those younger than 20 years were excluded from this evaluation. In addition, we excluded cases with severe degradation of image quality caused by artifacts (e.g., artifacts from metallic implants, motion). MDCT All patients were scanned with a 64-MDCT scanner (Aquilion 64, Toshiba Medical Solutions). After pulmonary CTA was performed, CTV was performed in the craniocaudal direction from the diaphragm to the ankles during a single inspiratory breath-hold. CTV was performed approximately 3 minutes 30 seconds after the IV injection of contrast material began. The scanning parameters were 120 kvp (standard-kvp group) or 100 kvp (low-kvp group), a 0.5-mm collimation, a 0.5-second rotation time, a pitch of 53 or 64, 1-mm-thick reconstructions, and 0.8-mm reconstruction increments with automatic tube current modulation. Automatic tube current modulation (z-axis modulation with Real E.C. technique, Toshiba Medical Solutions) was used with the noise level set at 8 SD for the 120-kVp group and at 10 SD for the 100-kVp group. The noise level settings were based on a phantom study performed at our institution. A different tube current was needed to achieve the noise level at 8 SD as follows: 120 kvp, 50 ma; 100 kvp, 80 ma; and 80 kvp, 175 ma. When the noise level was set at 10 SD, the 100-kVp protocol needed 50 ma. Considering CT tube unloading, 100 kvp with the noise level set at 10 SD was valid. Although the signal-to-noise ratio (SNR) of the 100-kVp protocol with the noise level at 10 SD was higher than the SNR of the 120-kVp protocol with the noise level at 8 SD, the contrast-to-noise ratio of the 100-kVp protocol with the noise level at 10 SD was almost equal to that of the 120-kVp protocol with the noise level at 8 SD. The SNR of the 100-kVp protocol with the noise level at 8 SD was lower than at 10 SD, but the low noise level setting increases radiation dose. For all studies, nonionic contrast material (iohexol, Omnipaque 300, GE Healthcare) was administered via an antecubital vein with an automatic power injector. We used 150 ml of iohexol for patients who weighed more than 50 kg and 3 ml/kg of body weight of iohexol for patients who weighed 50 kg or less. The first 80 ml of contrast material was injected at a rate of 3 4 ml/s, and the remaining contrast material was injected at a rate of 1.5 ml/s. Individual contrast optimization was achieved using bolus tracking in the main pulmonary artery. Assessment of Vascular Enhancement and Image Quality All images were reviewed and interpreted on PACS workstations (Ziostation, Ziosoft). Image analysis was performed using transverse images with a window level of 100 HU and a window width of 450 HU. Vascular enhancement was evaluated quantitatively using attenuation values in the inferior vena cava (IVC) at the level of the left renal vein, right femoral vein at the level of the femoral head, and right popliteal vein at the level of the popliteal fossa. The mean attenuation values for each vein were measured using a region of interest (ROI). For the evaluation of vascular enhancement, two radiologists (one with 4 years of experience in chest CT and the other with 17 years of experience) who were blinded to the kvp setting placed the ROIs on vessels by consensus and manually drew an elliptic ROI within the inner edge of each vessel. If a clot in the vessel or severe streak artifacts from a metallic prosthesis were present, they measured the vein on the left side. Image noise was objectively quantified by measuring the SD of the attenuation value in an ROI of each vein. The same two radiologists scored overall image quality by consensus using a 5-point scale (Figs. 1 and 2) as follows: 1, nondiagnostic (no diagnosis possible); 2, poor (inadequate to allow diagnosis of the presence or absence of a clot); 3, fair (optimal enhancement sufficient for diagnosis); 4, good (optimal enhancement to allow confident diagnosis of the presence or absence of a clot); or 5, excellent (optimal enhancement superior to a score of 4 that allows confident diagnosis of the presence or absence of a clot). In addition, a 5-point scale for the evaluation of subjective image quality based on previous reports was adopted [7, 8, 12]. The radiologists who scored image quality were blinded to the kvp setting. Images were selected randomly from both patient groups. Regarding image noise in the abdomen, reviewers also analyzed image quality of the liver as follows: 1, nondiagnostic (no diagnosis possible); 2, poor (confidence in making diagnosis degraded); 3, fair (moderate but sufficient for diagnosis); 4, good; or 5, excellent (enabling excellent differentiation of even small structures). The reviewers were blinded to the kvp setting used to obtain the images. Fig. 1 56-year-old woman with liver cirrhosis who presented for imaging to evaluate for suspected pulmonary embolism. Axial 100-kVp CT venography image obtained at level of inferior vena cava was assigned image quality score of 5 (i.e., excellent, optimal enhancement superior to a score of 4 that allows confident diagnosis of presence or absence of clot superior to a score of 4). Image shows very tortuous splenic venous system. AJR:197, October 2011 839

Fujikawa et al. Image noise in a low-kv technique is influenced by BMI [13]. We analyzed BMI versus noise in each patient group using the Spearman rank correction. BMI was correlated with noise in each group and in each ROI position. Assessment of Attenuation Values for Deep Venous Thrombosis The detection of DVT was based on CTV images. The mean attenuation value for each DVT was measured using an ROI. For the evaluation of vascular enhancement, we manually drew an elliptic ROI within the clots in each vessel. If multiple foci of DVT were present in the same patient, we chose the largest clot. Evaluation of Radiation Dose To compare radiation exposure from CTV performed using the standard 120-kVp protocol with that from CTV performed using the low 100-kVp protocol, we recorded the dose-length product (DLP) as a CT radiation dose descriptor. The DLP was provided by the scanner system. The DLPs of 35 patients in the low-kvp group were measured using both the 100 and 120 kvp settings. Estimated DLP measurements did not include pulmonary CTA. A Fig. 2 Axial 100-kVp CT venography images obtained at level of femoral vein. A, 47-year-old woman with suspected pulmonary embolism. Image quality score of 4 (good, optimal enhancement to allow confident diagnosis of presence or absence of clot) was assigned. B, 73-year-old woman with suspected pulmonary embolism. Image quality score of 3 (fair, optimal enhancement, sufficient for diagnosis) was assigned. C, 73-year-old woman with suspected pulmonary embolism. Image quality score of 2 (poor, inadequate to allow diagnosis of presence or absence of clot) as assigned. Statistical Analysis Statistical analysis was performed using commercially available software (JMP, version 5.1, SAS Institute). A p < 0.05 was considered to indicate a statistically significant difference. The Wilcoxon rank test was applied to determine whether venous enhancement in each vein showed significant differences between the standard-kvp group and the low-kvp group and was used to compare the scoring of image quality between groups. The Wilcoxon rank test was also used to compare patient characteristics including age, sex, BMI, the number of measurable veins, and the prevalence of venous thrombi between the two groups. In addition, differences in vascular enhancement, image noise, and image quality between patients with and those without DVT were also tested using the Wilcoxon rank test. The Spearman rank correction was applied to determine whether BMI was correlated with the noise in each group and in each ROI position. Results There were no statistically significant differences in age, sex, or BMI between the standard-kvp and low-kvp groups (Table 1). The IVC, femoral vein, and popliteal vein were measurable in all patients (standard-kvp group, n = 50; low-kvp group, n = 50). Four femoral veins (standard-kvp group, n = 2; low-kvp group, n = 2) and two popliteal veins (standard-kvp group, n = 1; lowkvp group, n = 1) were measured on the left side because of severe artifacts or a diffuse thrombus on the right side. Eleven patients were excluded from this evaluation. One exclusion was based on patient age and nine, on the presence of metallic artifact. The final case was excluded because the patient had diffuse clots in both legs. There were no exclusions because of amputation. TABLE 1: Patient Characteristics B There were 20 DVTs (40% of patients) and 14 PEs (28%) in the standard-kvp group and 19 DVTs (38% of patients) and 10 PEs (20%) in the low-kvp group. There was no significant difference in the prevalence of DVTs or PEs between the two groups (DVT, p = 0.83; PE, p = 0.34). All 33 DVTs (standard-kvp group, n = 18; low-kvp group, n = 15) were measurable. The IVC, femoral vein, and popliteal vein showed significantly higher attenuation values in the low-kvp group than in the standardkvp group (p < 0.001) (Table 2). The mean attenuation value in the IVC was 138.4 ± 12.2 HU in the standard-kvp group and 164.5 ± 17.4 HU in the low-kvp group; in the femoral vein, 130.2 ± 18.0 HU in the standard-kvp group and 152.0 ± 24.5 HU in the low-kvp group; and in the popliteal vein, 136.7 ± 17.5 HU in the standard-kvp group and 158.3 ± 26.0 HU in the low-kvp group. The mean attenuation value of the DVTs was 63.8 ± 3.8 HU in the standard-kvp group and 64.7 ± 6.1 HU in the low-kvp group. There was no significant difference between the two groups (p = 0.928). Image noise in all measured veins in the low-kvp group was significantly higher than Characteristic 120 kvp (n = 50) 100 kvp (n = 50) p Age (y) 1 Mean ± SD 64 ± 15 64 ± 15 Sex (no. of patients) 0.64 Male-female ratio 14:36 12:38 Body mass index 0.92 Mean ± SD 22.9 ± 3.4 22.8 ± 3.5 With pulmonary embolism (no. of patients) 14 10 0.34 With deep venous thrombosis (no. of patients) 20 19 0.83 C 840 AJR:197, October 2011

Vascular Enhancement and Image Quality of CTV TABLE 2: Vascular Enhancement in Standard Peak Kilovoltage (kvp) and Low-kVp Groups Vascular Enhancement (HU) Region of Interest 120 kvp (n = 50) 100 kvp (n = 50) p Inferior vena cava 138.42 ± 12.15 164.47 ± 17.43 < 0.001 Femoral vein 130.17 ± 18.00 152.04 ± 24.51 < 0.001 Popliteal vein 136.72 ± 17.52 158.25 ± 26.04 < 0.001 Note Vascular enhancement in all measured veins showed significantly higher attenuation values in the low-kvp group than in the standard-kvp group (p < 0.001). TABLE 3: Image Noise in Standard Peak Kilovoltage (kvp) and Low-kVp Groups Vascular Enhancement (HU) Region of Interest 120 kvp (n = 50) 120 kvp (n = 50) p Inferior vena cava 26.25 ± 5.55 40.08 ± 8.37 < 0.001 Femoral vein 23.95 ± 4.77 36.77 ± 8.35 < 0.001 Popliteal vein 15.35 ± 4.86 18.31 ± 5.60 0.0052 Note Image noise in the low-kvp group was significantly higher than that in the standard-kvp group. that in the standard-kvp group (Table 3). However, there were no significant differences in the subjective score of image quality at the level of the IVC (p = 0.245) or the popliteal vein (p = 0.487). At the level of the femoral vein, the visual score of the low-kvp group (mean ± SD, 3.92 ± 0.63) was significantly lower than that of the standard-kvp group (4.18 ± 0.63); however, this difference between the groups was relatively weak (p = 0.0476) (Table 4). The mean image quality score of the liver was 4.3 ± 0.7 in the standard-kvp group and 4.12 ± 0.6 in the low-kvp group. This difference between the two groups was significant (p = 0.02). BMI was correlated with noise in each ROI position in the 120-kVp group (IVC vs BMI, p = 0.0267; femoral vein vs BMI, p = 0.0004; popliteal vein vs BMI, p = 0.0205) and in the 100-kVp group (IVC vs BMI, p = 0.0005; femoral vein vs BMI, p = 0.0031; and popliteal vein vs BMI, p = 0.0106). The correlations between noise and BMI in each group were moderate to strong. The mean DLP in the low-kvp group (1202 ± 273.5 mgy cm; range, 748.3 2330 mgy cm) was significantly lower than that in the standard-kvp group (1774.9 ± 426.0 mgy cm; range, 1130 3570 mgy cm) (p < 0.001). Discussion In our study, vascular enhancement on CTV was greater with the low kvp setting than with the standard kvp setting. Although the image noise was higher in the low-kvp group than in the standard-kvp group, the mean visual scores of image quality at the level of the IVC and the popliteal vein were not statistically different between the two groups. Moreover, the CT radiation dose of CTV was significantly lower in the low-kvp group than in the standard-kvp group. To our knowledge, there have been no previous investigations of the vascular enhancement, image quality, and radiation dose of CTV with a low kvp setting. TABLE 4: Image Quality Score in Standard Peak Kilovoltage (kvp) and Low-kVp Groups Image Quality Score Region of Interest 120 kvp (n = 50) 100 kvp (n = 50) p Inferior vena cava 4.54 ± 0.50 4.40 ± 0.57 0.2449 Femoral vein 4.18 ± 0.62 3.92 ± 0.63 0.0476 Popliteal vein 4.38 ± 0.75 4.5 ± 0.64 0.4973 Note Regarding image quality scores for the inferior vena cava and popliteal vein, there was no significant difference in subjective scores between the two patient groups, whereas the image quality scores for the femoral vein deteriorated in the low-kvp group compared with the standard-kvp group (p = 0.0476). There are several advantages to reducing the kvp setting in CTV. Lowering the kvp setting is an effective method of reducing the radiation dose [14]. Moreover, lowering the kvp setting leads to increased vascular enhancement because the attenuation of iodine-based contrast agents increases with reduced x-ray energy distribution. This increase in attenuation occurs because of the high relative atomic number of iodine [11] as well as the k-edge in such energy levels. Indeed, in this study, the mean attenuation values in all measured veins with the low kvp setting were significantly higher than those in the standard-kvp group. These results are consistent with the findings of previous pulmonary CTA studies [7, 8, 12]. The contrast between clot and venous enhancement also could be improved by lowering the kvp setting because the mean attenuation value of DVTs did not differ significantly between the two groups. The disadvantage of a lower kvp setting is the increase in image noise. Boone et al. [15] evaluated the relationship between image noise and radiation dose reduction in CT and reported that image noise increases at lower tube current (ma) and lower kvp settings. In our study, image noise was significantly higher in the low-kvp group than in the standard-kvp group. However, no significant difference in image quality at the level of the IVC and the popliteal vein between the groups was observed. Previous studies also found no significant degradation in image quality after lowering the kvp setting for pulmonary CTA [7, 8]. We used 1-mm image reconstructions for CTV because that is routine practice at our institution. However, if we had used thicker images, the image quality would have been higher and the image noise would have been lower. Regarding image noise in the abdomen, we also analyzed the image quality score of the liver. Although the mean image quality score of the liver was statistically lower in the low-kvp group than in the standard-kvp group (p = 0.02), the mean image quality score of the low-kvp group was 4.12. This high mean image quality score shows that the deterioration of image quality did not much influence diagnostic accuracy. Nakayama et al. [13] reported that image noise is mainly related to the characteristics of the CT scanner used and to the characteristics of individual patients, such as body habitus. They evaluated the relationship between SNR of the aorta and the patient s body weight at 120 and 90 kvp settings and AJR:197, October 2011 841

Fujikawa et al. reported that a technique with a low kvp should be used in patients weighing less than 80 kg. In our study, only one patient in each group weighed more than 80 kg, and there was no statistically difference in BMI between two groups (p = 0.917). In our study, the correlations between noise and BMI in each group were moderate to strong. Noise is known to increase at low kvp settings, so we should be more concerned about BMI when using a low kvp setting. In contrast, in our study, there was a weak but significant difference in image quality at the femoral vein between the two groups. However, the mean image quality score for the 100-kVp group was 3.92, which was sufficient for diagnosis. Dual-energy CT enables improved detection of small PEs, particularly by emphasizing the presence of subsegmental pulmonary iodine mapping defects. Probably the same result would be seen on CTV. In addition, dual-energy CT can remove bones and intraluminal plaques from angiography datasets on the basis of spectral differentiation separating iodine from calcium and can reduce metal artifact. In our study, artifacts from bones and metallic prostheses were suspected to cause deterioration in image quality of the femoral vein; dual-energy CT would solve these problems. The use of CTV for the diagnosis of PE is controversial, in part because the additional pelvic radiation dose from CTV is of concern. Although some researchers have argued that imaging the lower extremities of patients with PE is not necessary [16 18], a moderate potential benefit is probable for high-risk patients with signs or symptoms of DVT or a history of DVT [6, 19 21]. Some reports have also suggested that CTV should not be routinely performed in all patients evaluated for PE but that CTV may be useful only in patients with a high probability of PE. Sonography is the primary imaging modality for imaging the lower extremities for the diagnosis of DVT [22]. Although sonography of the lower extremities has approximately the same diagnostic accuracy as CTV for the detection of DVT [23], the improvement in contrast enhancement obtained with the low kvp setting in CTV may enhance the diagnostic accuracy of CTV for DVT. Moreover, although the absence of radiation exposure in sonography may render this modality preferable to detect DVT, our data indicate that a radiation dose reduction can be achieved by lowering the kvp setting in CTV. In addition, recent studies have shown that radiation from CTV can be reduced by 75% or more using a modified CTV technique [6, 20, 24]. Goodman et al. [6] reported that radiation from CTV can be reduced by 75% or more using discontinuous imaging, reduced anatomic coverage, and automated milliampere (ma) adjustments. Begemann et al. [25] reported that a wider collimation can reduce radiation dose. Thus, CTV with a lower kvp setting in conjunction with such modified techniques could reduce radiation exposure. The disadvantages of sonography include dependence on the operator s skill and cost. Our study has several limitations. First, the sample size of our study was relatively small and our results were obtained from a retrospective analysis. To confirm these findings a prospective analysis with a large patient sample is required. Second, cardiac function may affect vascular enhancement at CTV. Garg et al. [26] reported that complex hemodynamic factors due to a combination of abnormal cardiac contractility and peripheral arterial disease result in poor delivery of contrast bolus to the deep venous system. In particular, massive PE leads to severe right cardiac failure. We did not consider the effect of cardiac function in each patient; however, there was no statically significant difference in the observed PEs and DVTs between the standard-kvp and low-kvp groups. Thus, we speculate that the influence of the difference in cardiac function between the two groups is minimal. Third, sample size was relatively small for the evaluation of the attenuation value of clots because the attenuation of clots has been reported to be variable [1, 27]. If we had tried to make a statistically significant observation about the density of clots between the two groups, a larger sample would have been needed. Baldt et al. [28] reported that the attenuation of clots showed a trend of negative correlation with days after the onset of disease. Clots imaged within 8 days of disease onset showed an average attenuation of 66 HU, whereas those that had been present more than 8 days showed an average attenuation of 55 HU. In consideration of these reports, a larger sample size would be needed to evaluate the attenuation of clots between the two groups. In conclusion, we confirmed that CTV with the low kvp setting improves vascular enhancement with acceptable image quality for the detection of DVT. We recommend the CTV protocol with a low kvp for the evaluation of DVT in clinical practice. References 1. Loud PA, Katz DS, Klippenstein DL, Shah RD, Grossman ZD. Combined CT venography and pulmonary angiography in suspected thromboembolic disease: diagnostic accuracy for deep venous evaluation. AJR 2000; 174:61 65 2. Stein PD, Fowler SE, Goodman LR, et al. Multidetector computed tomography for acute pulmonary embolism. N Engl J Med 2006; 354:2317 2327 3. Bischoff B, Hein F, Meyer T, et al. Impact of a reduced tube voltage on CT angiography and radiation dose: results of the PROTECTION I study. JACC Cardiovasc Imaging 2009; 2:940 946 4. Hurwitz LM, Yoshizumi TT, Goodman PC, et al. Radiation dose savings for adult pulmonary embolus 64-MDCT using bismuth breast shields, lower peak kilovoltage, and automatic tube current modulation. AJR 2009; 192:244 253 5. Kalva SP, Sahani DV, Hahn PF, Saini S. Using the K-edge to improve contrast conspicuity and to lower radiation dose with a 16-MDCT: a phantom and human study. J Comput Assist Tomogr 2006; 30:391 397 6. Goodman LR, Stein PD, Beemath A, et al. CT venography for deep venous thrombosis: continuous images versus reformatted discontinuous images using PIOPED II data. AJR 2007; 189:409 412 7. Heyer CM, Mohr PS, Lemburg SP, Peters SA, Nicolas V. Image quality and radiation exposure at pulmonary CT angiography with 100- or 120- kvp protocol: prospective randomized study. Radiology 2007; 245:577 583 8. Schueller-Weidekamm C, Schaefer-Prokop CM, Weber M, Herold CJ, Prokop M. CT angiography of pulmonary arteries to detect pulmonary embolism: improvement of vascular enhancement with low kilovoltage settings. Radiology 2006; 241: 899 907 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 2000; 21:1881 1884 10. Huda W, Scalzetti EM, Levin G. Technique factors and image quality as functions of patient weight at abdominal CT. Radiology 2000; 217: 430 435 11. 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 12. Matsuoka S, Hunsaker AR, Gill RR, et al. Vascular enhancement and image quality of MDCT pulmonary angiography in 400 cases: comparison of standard and low kilovoltage settings. AJR 2009; 192:1651 1656 13. Nakayama Y, Awai K, Funama Y, et al. Abdominal 842 AJR:197, October 2011

Vascular Enhancement and Image Quality of CTV CT with low tube voltage: preliminary observa- C. Compression sonography in patients with inde- pelvis be imaged? Radiology 2008; 246:605 611 tions about radiation dose, contrast enhancement, terminate or low-probability lung scans: lack of 25. Begemann PG, Bonacker M, Kemper J, et al. image quality, and noise. Radiology 2005; 237: usefulness in the absence of both symptoms of Evaluation of the deep venous system in patients 945 951 deep-vein thrombosis and thromboembolic risk with suspected pulmonary embolism with multi- 14. Kubo T, Lin PJ, Stiller W, et al. Radiation dose factors. AJR 1996; 166:285 289 detector CT: a prospective study in comparison to reduction in chest CT: a review. AJR 2008; 190: 20. Hunsaker AR, Zou KH, Poh AC, et al. Routine Doppler sonography. J Comput Assist Tomogr 335 343 15. Boone JM, Geraghty EM, Seibert JA, Wootton- Gorges SL. Dose reduction in pediatric CT: a rational approach. Radiology 2003; 228:352 360 16. Perrier A, Bounameaux H. Accuracy or outcome in suspected pulmonary embolism. N Engl J Med 2006; 354:2383 2385 17. Perrier A, Roy PM, Sanchez O, et al. Multidetectorrow computed tomography in suspected pulmonary embolism. N Engl J Med 2005; 352:1760 1768 18. Righini M, Le Gal G, Aujesky D, et al. Diagnosis of pulmonary embolism by multidetector CT alone or combined with venous ultrasonography of the leg: a randomised non-inferiority trial. Lancet 2008; 371:1343 1352 19. Rosen MP, Sheiman RG, Weintraub J, McArdle pelvic and lower extremity CT venography in patients undergoing pulmonary CT angiography. AJR 2008; 190:322 326 21. Dodd JD. Evidence-based practice in radiology: steps 3 and 4 appraise and apply diagnostic radiology literature. Radiology 2007; 242:342 354 22. Zierler BK. Ultrasonography and diagnosis of venous thromboembolism. Circulation 2004; 109: I9 I14 23. Goodman LR, Stein PD, Matta F, et al. CT venography and compression sonography are diagnostically equivalent: data from PIOPED II. AJR 2007; 189:1071 1076 24. Kalva SP, Jagannathan JP, Hahn PF, Wicky ST. Venous thromboembolism: indirect CT venography during CT pulmonary angiography should the 2003; 27:399 409 26. Garg K, Kemp JL, Russ PD, Baron AE. Thromboembolic disease: variability of interobserver agreement in the interpretation of CT venography with CT pulmonary angiography. AJR 2001; 176: 1043 1047 27. Cham MD, Yankelevitz DF, Shaham D, et al. Deep venous thrombosis: detection by using indirect CT venography The Pulmonary Angiography-Indirect CT Venography Cooperative Group. Radiology 2000; 216:744 751 28. Baldt MM, Zontsich T, Stumpflen A, et al. Deep venous thrombosis of the lower extremity: efficacy of spiral CT venography compared with conventional venography in diagnosis. Radiology 1996; 200:423 428 AJR:197, October 2011 843