High Opacification of Hilar Pulmonary Vessels with a Small Amount of Nonionic Contrast Medium for General Thoracic CT: A Prospective Study

Transcription

1 Pierre Loubeyre 1 Isabelle Debard 2 Chantal Nemoz 3 Van André Tran Minh 2 Received April 5, 2001; accepted after revision December 28, Service de Radiologie, l Hôpital Cantonal Universitaire de Genève, rue Micheli-du-Crest, CH-1211 Genève 14, Switzerland. Address correspondence to P. Loubeyre. 2 Service de Radiologie, Centre Hospitalier Lyon Sud, EA 643, Chemin du Grand Revoyet, Pierre Benite Cedex, France. 3 Unité de Pharmacologie Clinique, Université Claude Bernard Lyon 1, 8 Ave., Rockefeller, Lyon, Cedex 08, France. AJR 2002;178: X/02/ American Roentgen Ray Society High Opacification of Hilar Pulmonary Vessels with a Small Amount of Nonionic Contrast Medium for General Thoracic CT: A Prospective Study OBJECTIVE. The main goal of this study was to determine whether in thoracic helical CT for general indications a high opacification of hilar pulmonary vessels could be obtained using a small amount of nonionic contrast medium. SUBJECTS AND METHODS. Two hundred consecutive patients referred for contrastenhanced thoracic CT for malignancies or infections prospectively entered the study. They were randomly assigned to one of two simple bolus injection protocols (100 patients in each protocol): 60 ml of a nonionic contrast agent (250 mg I/mL) injected at a 3 ml/sec flow rate, or 80 ml of the same contrast agent injected at a 4 ml/sec flow rate. No saline flush or bolus triggering system was used. Hilar and mediastinal vessel enhancement was qualitatively (using a 4-point scale) and quantitatively (arterial attenuation values) assessed. Perivenous artifacts were also assessed. RESULTS. No extensive perivenous artifacts were noted. No significant difference was noted regarding pulmonary venous enhancement. Excellent opacification of the pulmonary veins was observed in 66% of patients injected at 3 ml/sec and in 56% of patients injected at 4 ml/sec ( p > 0.192). A highly significant difference was noted for pulmonary artery enhancement. Excellent opacification of the pulmonary arteries was noted in 83% of patients injected with a 3mL/sec flow rate and in 61% of patients injected with a 4mL/sec flow rate ( p = 0.001). CONCLUSION. A high level of opacification of hilar pulmonary vessels, with no major perivenous artifacts, can be obtained with a small amount of nonionic contrast medium using a simple bolus injection. A dministration of IV contrast material is mandatory in many general thoracic CT indications to obtain images of diagnostic quality [1 4]. The anatomic regions that are difficult to analyze when discriminating between normal and abnormal processes such as soft-tissue masses or lymph nodes are the pulmonary hila. A high level of opacification of hilar pulmonary vessels is required during the entire scan for accurate analysis of pulmonary hila. A high level of opacification of the thoracic aorta is less necessary for general thoracic indications because the thoracic aorta is more readily recognizable. The high cost and potential harmfulness of contrast media force radiologists to search for ways to further decrease contrast volume. The main purpose of this study was to determine whether a high level of opacification of hilar pulmonary vessels can be obtained with a small amount of contrast material using a simple bolus injection. Two bolus injection protocols using small amounts of contrast medium were evaluated. Pulmonary vessel opacification and perivenous artifacts were assessed. Thoracic aorta opacification was also assessed. Qualitative and quantitative vascular enhancement assessments were performed. Subjects and Methods Patients Two hundred consecutive patients referred for contrast-enhanced thoracic helical CT were prospectively enrolled in the study. Patients who required helical CT for the detection of pulmonary embolism or aortic dissection or patients unable to undergo short periods of breath-holding did not enter the study. The patients included 65 women and 135 men who were years old (mean AJR:178, June

2 Loubeyre et al. TABLE 1 Grading of Perivenous Artifacts and Arterial Opacification Score Perivenous Artifacts Arterial Opacification age, 58 years). Patients were randomly assigned to one of two protocols of injection (100 patients in each injection protocol). Thirty women and 70 men (mean age, 57.3 years) were enrolled in one arm of the study (protocol A), and 35 women and 65 men (mean age, 58.7 years) were enrolled in the other arm of the study (protocol B). The two arms of the study were not significantly different when considering age and sex. Parameters such as weight and heart rate were not assessed. Indications for helical CT were malignancies (n = 124), uncharacterized pulmonary or mediastinal masses and nodules (n = 40), and pulmonary infections (n = 36). No informed consent was obtained because the contrast agent used in this study is routinely used in radiologic practice and because of the small amount of contrast agent injected. Nor was institutional review board approval required. Injection of Contrast Material The type of contrast agent, duration of injection, and delay time between injection and initiation of CT scanning were the same for both protocols. Injection flow rates, and thus the amount of contrast medium, varied. For protocol A, 80 ml of iomeprol 250 (Iomeron; Bracco Diagnostics, Paris, France) was administered. The iodine concentration was 250 mg I/mL, osmolality was 435 mosm/kg of water, and viscosity measured at 37 C was 2.9 mpa/sec. The injection flow rate was 4 ml/sec. For protocol B, 60 ml of iomeprol 250 was administered using a 3 ml/sec injection flow rate. A left antecubital venous access was achieved for all patients using a 20-gauge venous catheter with the patient s arms positioned beside the head. The contrast medium was injected using a power injector (Multilevel CT Injector; Medrad, Pittsburgh, PA), and injection duration was 20 sec. Helical CT 1 Extensive; completely obscuring adjacent anatomy Poor; no visible enhancement; insufficient for diagnosis 2 Moderate; partially obscuring adjacent structures Fair; minimal enhancement; may be insufficient for diagnosis 3 Minimal; without notable obscuring of adjacent structures Good; diagnostic quality 4 None; clear anatomic detail Excellent; high degree of vascular opacification Ten seconds after initiation of the contrast medium injection, helical CT was performed using a Somatom Plus 4 CT scanner (Siemens Medical Systems, Erlangen, Germany) with 8-mm collimation, a pitch of 1.2, a 1.0-sec gantry rotation period, 206 ma, and 140 kvp. The entire thorax was imaged in a craniocaudal direction, in a single 25- to 30-sec breath-hold, from the lung apices to the lateral costophrenic sulci. Images were reconstructed using a 180 linear interpolation algorithm at 8-mm intervals on a matrix. All images were printed on hard-copy films at a window level appropriate for vascular visualization (window center, 50 H; window width, 350 H). Qualitative Assessment Two experienced thoracic radiologists independently evaluated the helical CT examinations, which were randomized and presented to reviewers in a different order. The radiologists were unaware of the patients dose group. Reviewers graded perivenous artifacts and arterial enhancement on a 4-point scale (Table 1). The baseline for scoring the images had been established by the two reviewers during a previous study [5]. Perivenous artifacts were scored 1 for extensive, 2 for moderate, 3 for minimal, or 4 for absent. Vascular opacification was considered 1, poor (Fig. 1); 2, fair (Fig. 2); 3, good (Fig. 3); or 4, excellent (Fig. 4). The anatomic region chosen to assess perivenous artifacts (i.e., blooms and streaks) was the region adjacent to the superior vena cava. Anatomic regions chosen for qualitative assessment of vascular enhancement were the right pulmonary artery (main and interlobar arteries), the left pulmonary artery (main and interlobar arteries), the right superior pulmonary vein, the left superior pulmonary vein, the ascending aorta, and the descending aorta. In case of discordance between reviewers scores, the lower score was considered the final score. Quantitative Assessment Images were reviewed on a viewing console. For each CT examination, single circular vascular regions of interest (diameter, 1.4 cm) were positioned in the ascending aorta, the descending aorta, and the pulmonary trunk at the origin of the right pulmonary artery. The three measurements were obtained on the same scan. For each CT examination and each measurement location, an arterial attenuation value and its standard deviation were recorded. For each measurement location, the mean and standard deviation of arterial attenuation values were calculated for each injection protocol. Arterial attenuation values were not measured in superior pulmonary veins because of volume averaging caused by the small diameter of these vessels. Statistical Analysis For both injection protocols, the mean arterial attenuation values in the pulmonary trunk, in the ascending aorta, and in the descending aorta were compared using the Student s t test. For each vessel, because scores and injection amount were ordinal data, scores were compared using the Wilcoxon s rank sum test. To compare the proportion of cases of excellent opacification (score = 4), scores were grouped into two categories (scores < 4 and scores = 4) and compared using Pearson s chi-square test with 1 degree of freedom. The relationship between scores and attenuation values was evaluated with a nonparametric correlation coefficient; Spearman s rho and Kendall s tau gave the same results. An arterial attenuation threshold above which vascular opacification was scored as excellent (score = 4) was calculated for the pulmonary trunk, the ascending aorta, and the descending aorta. For each vessel, we calculated, for each observed attenuation value, the proportion of cases with a score of 4 among those having an attenuation greater than that value. Then we adjusted these proportions versus the corresponding attenuation values using a quadratic model, which gave the best results. The arterial attenuation threshold is the adjusted attenuation value corresponding to 95% of cases above that value with a score of 4. We calculated the 95% confidence limits for individual predicted values. For all statistical tests, p values were calculated. Results No episodes of adverse allergic reactions or extravasation occurred during the study. No extensive perivenous artifacts were noted with either injection protocol. Artifacts were less present with a 4 ml/sec flow rate than with 3 ml/sec, but the difference was not statistically significant (Table 2). No significant difference was noted for pulmonary venous opacification or ascending aorta opacification. Excellent opacification of the pulmonary veins was noted in 66% of patients injected with a 3 ml/sec flow rate and in 56% of patients injected with a 4 ml/sec flow rate. Excellent opacification of the ascending aorta was noted in 51% of patients injected with a 3 ml/sec flow rate and in 59% of patients injected with a 4 ml/ sec flow rate (Table 2). A significant difference was noted for the descending aorta opacification. Excellent opacif AJR:178, June 2002

3 CT of Hilar Pulmonary Vessels Fig. 1. Contrast-enhanced helical CT scan in 61-year-old man with lymphoma. Pulmonary enhancement was scored 1 (on a scale of 4), poor. ication was noted in 40% of patients injected at 3 ml/sec and in 56% of patients injected at 4 ml/sec (Table 2). A highly significant difference was noted for pulmonary arterial opacification. Excellent opacification was noted in 83% of patients injected at 3 ml/sec and in 61% of patients injected at 4 ml/sec (Table 2). Arterial attenuation values were significantly higher in the ascending and descending aortas with a 4 ml/sec flow rate, and arterial attenuation values were significantly higher in the pulmonary trunk with a 3 ml/ sec flow rate (Table 3). The relationship between scores of arterial opacification and arterial attenuation values measured in the vessels was highly significant (Table 4). The arterial attenuation threshold above which excellent opacification (score = 4) was seen is given in Table 5 for each vessel. Ninety-five percent of cases of excellent opacification had an observed arterial attenuation greater than this threshold (Table 5). Fig. 2. Contrast-enhanced helical CT scan in 46- year-old man with lymphoma. Pulmonary enhancement was scored 2 (on a scale of 4), fair. Discussion Various bolus injection timings, injection flow rates, contrast volumes, and iodine concentrations are used in contrast-enhanced thoracic CT. Small to large volumes of contrast agents are used in general thoracic CT examinations [3, 4, 6]. During contrast-enhanced helical CT examinations for general thoracic evaluations, good opacification of thoracic vascular structures can be obtained with a small volume of contrast medium [4, 5]. Because pulmonary hila are difficult to analyze when discriminating between vessels and abnormal processes, a high level of opacification of hilar pulmonary vessels is mandatory. The main goal of this study was to determine whether a high level of opacification of hilar pulmonary vessels can be obtained using a small amount of contrast medium during contrast-enhanced helical CT for general thoracic examinations. Aorta opacification was also assessed. Two injection protocols were assessed. With a 60-mL contrast medium bolus injected at 3 ml/sec, excellent pulmonary artery opacification was noted in 83% of patients and excellent pulmonary venous opacification was noted in 66% in patients. Insufficient pulmonary artery opacification was noted in only 4% of patients and insufficient pulmonary venous opacification was noted in only 8% of patients. Excellent opacification of the ascending aorta was noted in 70% of patients. These results were obtained with an iodine concentration of 250 mg I/mL. Perivenous artifacts were scored as moderate or minimal and were never considered a problem, which can be explained by the low viscosity of the contrast medium and the low pooling of the contrast agent in the left brachiocephalic vein and the superior vena cava. Our results confirm that reduction of contrast agent volume for general thoracic CT does not affect the diagnostic quality of vessel opacification [4, 5, 7]. Moreover, our results indicate that 60 ml injected at a rate of 3mL/sec allows a statistically significant greater opacification of pulmonary vessels than 80 ml injected at 4 ml/sec. These results appear paradoxic. For both protocols. the injection duration and the delay between Fig. 3. Contrast-enhanced helical CT scan in 53-year-old man with pulmonary malignancy. Pulmonary enhancement was scored 3 (on a scale of 4), good. Fig. 4. Contrast-enhanced helical CT scan in 43-yearold woman with pulmonary infection. Pulmonary enhancement was scored 4 (on a scale of 4), excellent. the beginning of the contrast injection and the start of CT were the same, but the amount of iodine entering the blood pool per second was greater with 80 ml (4 ml/sec). Some explanations can be proposed. The first explanation is that the time between the beginning of the contrast injection and the peak of pulmonary artery enhancement is shorter with the 4 ml/sec injection rate. Thus, in some patients, CT scans at the level of the pulmonary arteries are probably obtained after the pulmonary artery enhancement peak. With the 4 ml/sec flow rate, a delay of 10 sec between the beginning of the bolus injection and the start of scanning could be too long to optimize the time window for data acquisition of enhanced pulmonary vessels. This hypothesis is supported by the fact that arterial attenuation measured in the aorta was greater with AJR:178, June

4 Loubeyre et al. TABLE 2 Scores for Perivenous Artifacts and for Vessel Enhancement for Two Injection Protocols Score Perivenous Artifacts Right Pulmonary Artery Left Pulmonary Artery Right Superior Pulmonary Vein Left Superior Pulmonary Vein Ascending Aorta Descending Aorta 3 ml/sec 4 ml/sec 3 ml/sec 4 ml/sec 3 ml/sec 4 ml/sec 3 ml/sec 4 ml/sec 3 ml/sec 4 ml/sec 3 ml/sec 4 ml/sec 3 ml/sec 4 ml/sec p a < < p b a Using Wilcoxon s rank sum test to compare scores and protocols of injection. b Using Pearson s chi-square test to compare proportions of cases of excellent opacification (score = 4) in two protocols. Mean Arterial Attenuation (in Hounsfield Units) in Pulmonary Trunk, TABLE 3 Ascending Aorta, and Descending Aorta for Each Rate of Injection Vessel 3 ml/sec 4 ml/sec p a Pulmonary trunk 259 (6) 236 (74) 0.02 Ascending aorta 197 (54) 215 (63) 0.03 Descending aorta 181 (52) 201 (60) 0.01 Note. Numbers in parentheses are SDs. a Using Student s t test. TABLE 4 Spearman s Correlation Coefficients Between Scores and Arterial Attenuation for Each Vessel and for Each Rate of Injection Vessel 3 ml/sec 4 ml/sec Pulmonary trunk Ascending aorta Descending aorta p for each coefficient < < TABLE 5 Vessel Arterial Attenuation Threshold (in Hounsfield Units) Above Which 95% of Cases of Vascular Enhancement Were Scored Excellent, for Each Vessel Arterial Attenuation Threshold (H) 95% Confidence Limits (H) No. of Cases Having Attenuation >Threshold Pulmonary trunk , Ascending aorta , Descending aorta , the 4 ml/sec flow rate. A second hypothesis is that, in some cases, a high flow rate (4 ml/sec) creates hemodynamic and rheologic effects that could modify the transit time of the contrast medium or modify the cardiac output [8]. This study was performed using a nonionic contrast medium to reduce the prevalence of nausea [9, 10]. To simplify the injection procedure, we did not attempt to push the contrast material with a saline solution [4, 11]. This technique has been proven to provide satisfactory thoracic vascular enhancement [4, 11]. In our study, the qualitative assessment of vascular enhancement proved to be accurate. The calculated threshold arterial attenuation above which vascular enhancement was noted to be excellent (score = 4) proved to be the same for the pulmonary trunk, the ascending aorta, and the descending aorta. Our study had some limitations. Our results cannot be extrapolated to specific thoracic CT indications such as the detection of pulmonary embolism. For that specific indication, the mandatory amount of contrast material is greater. The results of our study are valid for the scan collimation, the pitch, and the contrast agent used in this study. The influence of factors such as weight and heart rate was not assessed. We did not measure agreement between observers. In a previous study based on the same criteria of qualitative assessment, agreement between observers was good when considering the vascular enhancement [5]. To avoid overestimating the final scores, only the lower score of the two reviewers was taken into account. We did not use a bolus triggering system to optimize the timing of image acquisition [7] because this system is not yet available on all scanners. Our goal was to assess a simple injection procedure. With a short delay time of 10 sec between the beginning of the bolus injection and the start of scanning, we obtained high opacification of pulmonary vessels in most cases. Our injection protocol could probably be optimized for each patient using a bolus triggering system, and the amount of contrast medium could be further reduced with multidetector scanners. In conclusion, during contrast-enhanced helical CT for general thoracic indications, a high level of opacification of hilar pulmonary vessels can be obtained with a small volume of nonionic contrast medium and a simple injection procedure, with no major 1380 AJR:178, June 2002

5 CT of Hilar Pulmonary Vessels perivenous artifacts. Optimization of the delay between the beginning of injection and the start of scanning can be achieved using a bolus triggering system. References 1. Rubin GD, Lane MJ, Bloch DA, Leung AN, Stark P. Optimization of thoracic spiral CT: effects of iodinated contrast medium concentration. Radiology 1996;201: Costello P, Dupuy DE, Ecker CP, Tello R. Spiral CT of the thorax with reduced volume of contrast material: a comparative study. Radiology 1992;183: Leung AN. Spiral CT of the thorax in daily practice: optimization of technique. J Thorac Imaging 1997;12: Haage P, Schmitz-Rode T, Hübner D, Piroth W, Günther RW. Reduction of contrast material dose and artifacts by saline flush using a double power injector in helical CT of the thorax. AJR 2000;174: Loubeyre P, Debard I, Nemoz C, Tran Minh VA. Using thoracic helical CT to assess iodine concentration in a small volume of nonionic contrast medium during vascular opacification: a prospective study. AJR 2000;174: Nakayama M, Yamashita Y, Oyama Y, Ando M, Kadota M, Takahashi M. Hand exercise during contrast medium delivery at thoracic helical CT: a simple method to minimize perivenous artifact. J Comput Assist Tomogr 2000;24: Kirchner J, Kickuth R, Laufer U, Noack M, Liermann D. Optimized enhancement in helical CT: experiences with a real-time bolus tracking system in 628 patients. Clin Radiol 2000;55: Pugh ND. Haemodynamic and rheological effects of contrast media: the role of viscosity and osmolality. Eur Radiol 1996;6[suppl]:S13 S15 9. Katayama H, Yamaguchi K, Takeshima T, Seez P, Matsura K. Adverse reactions to ionic and nonionic contrast media: a report from the Japanese Committee on the Safety of Contrast Media. Radiology 1990;175: Federle MP, Willis LL, Swanson DP. Ionic versus nonionic contrast media: a prospective study of the effect of rapid bolus injection on nausea and anaphylactoid reactions. J Comput Assist Tomogr 1998;22: Hopper KD, Mosher TJ, Kasales CJ, Tenhave TR, Tully DA, Weaver JS. Thoracic spiral CT: delivery of contrast material pushed with injectable saline solution in a power injector. Radiology 1997;205: Applications for the 2003 ARRS Scholar Award are now being accepted; the deadline is November Log on to for more information. AJR:178, June