The Optimal Body Size Index with Which to Determine Iodine Dose for Hepatic Dynamic CT: A Prospective Multicenter Study 1

Transcription

1 This copy is for personal use only. To order printed copies, contact Kazuo Awai, MD, PhD Masayuki Kanematsu, MD, PhD Tonsok Kim, MD, PhD Tomoaki Ichikawa, MD, PhD Yuko Nakamura, MD, PhD Atsushi Nakamoto, MD, PhD Kunihiro Yoshioka, MD, PhD Teruhito Mochizuki, MD, PhD Naofumi Matsunaga, MD, PhD Yasuyuki Yamashita, MD, PhD The Optimal Body Size Index with Which to Determine Iodine Dose for Hepatic Dynamic CT: A Prospective Multicenter Study 1 Purpose: Materials and Methods: To identify the body size parameter that exhibits the best correlation with aortic and hepatic enhancement at hepatic dynamic computed tomography (CT) in a large patient population enrolled in a multicenter study. This prospective study was approved by the ethics committee of each of the 31 participating institutions where 1342 patients were enrolled between April 2012 and September All patients provided either written or oral informed consent. All patients underwent hepatic dynamic CT, which included preenhanced, hepatic arterial phase (HAP), and portal venous phase (PVP) scanning, performed with the routine scanning protocol of each institution. Changes in CT number (in Hounsfield units) per gram of iodine in the aorta (ea/i) and liver (el/i) during HAP and PVP scanning were recorded for each patient. Hierarchical multivariate linear regression analysis was performed in which the outcome variable was either ea/i or el/i; the independent variables were age, sex, one body size parameter (height, body weight, body mass index, lean body weight [LBW], or body surface area), and liver function (aspartate aminotransferase, albumin, and total bilirubin levels). A two-level hierarchical model in which patients were level 1 and the institution was level 2 was used. Original Research n Gastrointestinal Imaging 1 From the Department of Diagnostic Radiology, Institute of Biomedical and Health Sciences, Hiroshima University, Kasumi, Minami-ku, Hiroshima , Japan (K.A., Y.N.); Department of Radiology, Gifu University Hospital, Gifu, Japan (M.K.); Department of Radiology, Osaka University Graduate School of Medicine, Suita, Japan (T.K., A.N.); Department of Radiology, University of Yamanashi, Chuo, Japan (T.I.); Memorial Heart Center, Iwate Medical University, Morioka, Japan (K.Y.); Department of Radiology, Ehime University School of Medicine, Toon, Japan (T.M.); Department of Radiology, Yamaguchi University School of Medicine, Ube, Japan (N.M.); and Department of Diagnostic Radiology, Kumamoto University, Kumamoto, Japan (Y.Y.). Received December 22, 2014; revision requested March 2, 2015; revision received May 8; accepted June 12; final version accepted July 6. Supported by a research grant from Bayer Yakuhin. Address correspondence to K.A. ( awai@hiroshima-u.ac.jp) Results: Conclusion: Hierarchical multivariate linear regression analysis revealed that in the population not stratified by sex, body size was significantly correlated with ea/i and el/i (P,.001) and that LBW exhibited the strongest correlation with ea/i and el/i (r = and r = , respectively). Sex-stratified analysis showed that LBW was more strongly correlated with ea/i and el/i in women (r = and r = , respectively) than in men (r = and r = , respectively) or in the nonstratified total population. Among body size parameters, LBW exhibited the strongest correlation with aortic and hepatic enhancement, especially in women. q RSNA, 2015 Online supplemental material is available for this article. q RSNA, 2015 Radiology: Volume 278: Number 3 March 2016 n radiology.rsna.org 773

2 Since Yamashita et al (1) reported that basing contrast medium (CM) dose on patient body weight (BW) yields high-quality hepatic dynamic computed tomography (CT) images, it has been the most widely used body size index to determine CM dose. However, in obese patients, the CM dose may be excessive because a large proportion of these patients BW consists of poorly perfused adipose tissue in which CM distributes poorly (2 5). Also, at a given BW, the percentage of body fat is greater in women than in men (6) and basing the CM dose solely on a patient s BW may result in an excessive dose. Consequently, various body size indexes, such as body mass index (BMI), lean body weight (LBW), and body surface area (BSA), have been proposed to determine the CM dose for hepatic dynamic CT (5,7 15); however, the best body size index remains to be identified. Liver function may affect hepatic enhancement on dynamic CT images. Although hepatic enhancement during the portal venous phase (PVP) was decreased in patients with cirrhosis (16,17), to our knowledge, no correlation between liver function and hepatic enhancement has been documented. If hepatic enhancement is decreased in patients with liver dysfunction, the CM dose may have to be modified in these patients. Advances in Knowledge nn Among body size parameters, such as height, body weight, body mass index, lean body weight (LBW), and body surface area, LBW exhibited the strongest correlation with aortic and hepatic enhancement (r = and r = , respectively) at hepatic dynamic CT. nn Sex-stratified analysis showed that LBW was more strongly correlated with aortic and hepatic enhancement in women (r = and r = , respectively) than in men (r = and r = , respectively). The purpose of our study was to identify the body size parameter that showed the best correlation with aortic and hepatic enhancement on hepatic dynamic CT images in a large patient population enrolled in a multicenter study. Materials and Methods This prospective study was funded by Bayer Yakuhin (Osaka, Japan). The sponsors had no role in the study concept, study design, data analysis and interpretation, or reporting of results. The study was approved by the ethics committee of each participating institution. Oral or written informed consent was obtained from all patients according to the policies of the local ethics committees; the study period ranged from April 2012 to September Patient Population Thirty-one institutions that routinely perform hepatic dynamic CT, including preenhanced, hepatic arterial phase (HAP), and PVP scanning, on imagers with at least 16 detector rows participated in this prospective study. We thought that inclusion of only those patients who were examined with identical scanning and CM injection protocols would not yield clinically robust data to determine the optimal body size index for hepatic dynamic CT; thus, we included patients who were examined with the protocol that was routinely used at each participating institution, with the understanding that these protocols would differ between institutions. We then subjected the results to hierarchical multivariate linear regression analysis to adjust for interinstitutional variability. Our inclusion criteria were as follows: (a) hepatic dynamic CT was performed in the evaluation or survey Implication for Patient Care nn LBW may be the best body size index with which to determine the contrast material dose for hepatic dynamic CT. of known hepatic tumors; (b) patient age of years; (c) no prior hepatic surgery or focal liver therapy, such as transarterial chemoembolization or radiofrequency ablation of liver tumors during the preceding 3 months because the liver regenerates to almost the original volume after focal liver therapy, such as hepatectomy or radiofrequency ablation of liver tumors (18); (d) no contraindication for iodinated CM and no renal failure (estimated glomerular filtration rate,45 ml/min/1.73 m 2 ); and (e) no severe thyroid disease. Consequently, we enrolled 1342 patients who underwent hepatic dynamic CT between April 2012 and September Of these 1342 patients, 35 were excluded: Fifteen patients were older than 85 years. In 14 patients, the estimated glomerular filtration rate was lower than 45 ml/ min/1.73 m 2. One patient was older than 85 years and had an estimated glomerular filtration rate lower than 45 ml/min/1.73 m 2. In five patients, image quality was poor due to inadequate breath holding. Thus, the final study population consisted of Published online before print /radiol Content code: Radiology 2016; 278: Abbreviations: BMI = body mass index BSA = body surface area BW = body weight CM = contrast media ea/i = CT number per gram of iodine in the aorta el/i = CT number per gram of iodine in the liver HAP = hepatic arterial phase LBW = lean body weight PVP = portal venous phase Author contributions: Guarantor of integrity of entire study, K.A.; study concepts/ study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; agrees to ensure any questions related to the work are appropriately resolved, all authors; literature research, K.A.; clinical studies, K.A., T.K., Y.N., A.N.; experimental studies, K.A.; statistical analysis, K.A.; and manuscript editing, K.A., M.K., T.I., Y.N., A.N., K.Y., T.M., N.M., Y.Y. Conflicts of interest are listed at the end of this article. 774 radiology.rsna.org n Radiology: Volume 278: Number 3 March 2016

3 1307 patients (818 men, 489 women) (Table 1). CT Scanning and CM Injection Protocols In the 31 participating institutions, 54 CT scanners from different manufacturers were used (Table E1 [online]). Table 1 Patient Demographic Data Characteristic No. of Available Patients All institutions were located in Japan. The section thickness and section interval were 5 mm, and the tube voltage was 120 kvp in all studies. Automatic tube current modulation was used in 1186 (90.7%) of the 1307 patients. Mean Standard Deviation Median Range Age All Men Women Height (cm) All Men Women BW (kg) All Men Women BMI (kg/m 2 ) All Men Women LBW (kg) All Men Women BSA (m 2 ) All Men Women Albumin level (g/dl) All Men Women Total bilirubin level (mg/dl) All Men Women Aspartate aminotransferase level (IU/L) All Men Women Alanine aminotransferase level (IU/L) All Men Women Note. SI conversion factor: To convert total bilirubin level to micromoles per liter, multiply by ; to convert aspartate aminotransferase level to microkatals per liter, multiply by ; to convert alanine aminotransferase level to to microkatals per liter, multiply by All helical studies were started at the top of the liver; unenhanced and three-phase CM-enhanced helical images of the entire liver were obtained. Patients were instructed to hold their breath with tidal inspiration during scanning. All patients underwent hepatic dynamic CT, including preenhanced, HAP, and PVP scanning, with the routine scanning protocol used at that particular institution. A power injector was used to inject iopamidol (300 or 370 mg of iodine per milliliter [mg I/ ml], Iopamiron 300 or Iopamiron 370; Bayer Yakuhin) via a forearm vein. In 229 (17.5%) of the 1307 patients, the CM dose was fixed (group 1). In the remaining 1078 patients (82.5%), it was based on the patient s BW (group 2). The dose delivery was based on the scanning protocol at each participating institution. The dose delivered in groups 1 and 2 is shown in Tables E2 and E3 [online]. The mean iodine injection rate was mg/sec (standard deviation), and the median injection rate was mg/sec (range, mg/sec). HAP scanning was started at a fixed time in 426 patients. In 881 patients, the scanning delay for HAP was determined with a bolus tracking system (Tables E4, E5 [online]). In all patients, the CT number was monitored around the L1 vertebral level. The region of interest cursor was placed in the abdominal aorta. In 632 patients, the timing for PVP scanning was determined by using a fixed scan delay time after the start of CM delivery; in 605, it was determined with a bolus tracking method; and in 70, it was unknown (Table E6 [online]). Quantitative Analysis The CT number of vessels and organs enhanced by iodine CM depends on the effective tube voltage, which differs among CT scanners even if the nominal tube voltage is the same. Thus, we scanned a phantom and corrected the CT number of the aorta and liver parenchyma on all scanners. Our phantom consisted of seven syringes Radiology: Volume 278: Number 3 March 2016 n radiology.rsna.org 775

4 (diameter, 21 mm; length, 55 mm) filled with different concentrations of iodine solution (ie, 0, 1, 3, 5, 10, 20, and 40 mg I/mL). The syringes were then inserted into the cylindrical container (diameter, 184 mm; length, 120 mm) filled with 2% agar (Fig E1 [online]). The wall of syringes and the container were made of polypropylene. We scanned the phantom with all CT scanners used in this study. To scan the phantom, we placed the syringes along the z-axis of the CT scanner. The scanning parameters were the same as those used for phantom and patient scanning on each of the scanners. At each session, a radiologist or radiologic technologist measured the CT number of the seven syringes by using circular region-of-interest cursors. We performed linear regression analysis for the iodine concentration and CT numbers and calculated the gradient of the regression line for each CT scanner. As we adopted data of the 64-detector scanner (LightSpeed 64 VCT; GE Healthcare, Milwaukee, Wis) as a standard in this study (Fig E2 [online]), we calculated the ratio of the gradients of the regression lines for this scanner and for the other scanners as a calibration coefficient. Finally, we corrected the CT number of the aorta and liver parenchyma obtained with the different CT scanners by using the calculated calibration coefficient. The attenuation of the abdominal aorta and hepatic parenchyma was determined by a radiologist or a radiologic technologist at each institution who used a circular region of interest cursor during unenhanced and HAP and PVP scanning. For each phase, the attenuation was measured on images at the level of the porta hepatis; the region of interest was approximately 1.0 cm 2 in the aorta and 2.0 cm 2 in the liver parenchyma. Contrast enhancement in the abdominal aorta during HAP and PVP was calculated as the absolute difference in the attenuation of the aorta on unenhanced images and during each phase of contrast-enhanced scanning. We calculated the CT number per gram of iodine in the aorta (ea/i) by dividing enhancement of the aorta by the administered iodine dose. Likewise, we calculated the CT number per gram of iodine in the liver (el/i) by dividing enhancement of the liver by the administered iodine dose. Both ea/i and el/i were measured in Hounsfield units per gram. Body Size Parameters We adopted height, BW, LBW, BSA, and BMI as the body size indexes. Patient height and BW were acquired from medical charts. Estimated LBW was obtained by using the following formula in men (7): LBW = ( BW)-128 [BW 2 /(100 H) 2 ], where BW is the body weight (in kilograms) and H is the height (in centimeters). Likewise, a similar equation was used in women: LBW = ( BW)-148 [BW 2 /(100 H) 2 ]. BSA was estimated by using the formula BW (100 H) (18); it is widely used as a determinant for pharmacokinetics and drug dosing (19,20). The BMI was obtained with the formula BW/H 2. Another parameter to consider was the blood volume (BV) (in milliliters). In men, it was calculated as follows (21): BV = BW (100 H) In women, it was calculated in a similar manner: BV = BW (100 H) This formula includes the same variable (in this case, BW [100 H] ) as the formula for BSA; therefore, because BV and BSA are strongly correlated, we did not include blood volume in our study. Statistical Analysis Multivariate linear regression analysis was used to assess the association between age, sex, one body size parameter, and liver function parameters (aspartate aminotransferase, albumin, and total bilirubin levels) (independent variables) and ea/i or el/i (outcome variables). Because of the nonnormal distribution of aspartate aminotransferase and total bilirubin, we used natural logarithmic transformation. In the analysis for el/i, we also incorporated the scan timing of the PVP as a dichotomous variable (,80 seconds and 80 seconds from the start of CM injection). We selected 80 seconds as the threshold value because it was the most frequently used scan timing for PVP in our study. In analyses that included the scan timing of PVP, we used data acquired in the 632 patients who were examined with a fixed scan delay. We applied a two-level hierarchical model in which patients were level 1 and the participating institution was level 2 because the observations made within the same institution would be expected to be correlated. We confirmed that scanning and CM injection protocols were consistent at each institution during the patient enrollment period and that the intrainstitutional correlation coefficients for ea/i and el/i were and 0.152, respectively, and thus were relatively large. In general, the two-level hierarchical model (linear mixed model) assumes normally distributed outcome variables at level 1 and normally distributed random effects at level 2. To justify these assumptions, we have provided residuals and histograms of estimated random effects of ea/i or el/i (outcome variables) (Figs E3, E4 [online]). We first analyzed the entire population and then performed sex-stratified analyses, as various body size parameters tend to be smaller in Japanese women than in Japanese men. For example, scatter diagrams stratified by sex show the relationship between LBW and ea/i and show that there is a distribution difference between men and women (Fig E5 [online]). We used receiver operating characteristic curve analysis to calculate the most appropriate cutoff value for the best body size parameter to obtain sufficient aortic and hepatic enhancement for hepatic dynamic CT studies. Yanaga et al (22) stated that at dynamic CT 776 radiology.rsna.org n Radiology: Volume 278: Number 3 March 2016

5 during HAP, aortic enhancement should be at least 280 HU for excellent depiction of hypervascular hepatocellular carcinoma. Heiken et al (23) reported that hepatic parenchymal enhancement should be at least 50 HU for high-quality hepatic CT. Thus, we adopted 280 HU for the aorta and 50 HU for the liver parenchyma as adequate enhancement values. Statistical analysis was performed with statistical software (SPSS Statistics 21; SPSS, Chicago, Ill). We considered P,.05 indicative of a significant difference. Results In hierarchical multivariate linear regression analysis, a standardized partial regression coefficient of each independent variable can be considered a correlation coefficient. This analysis showed that in the total study population, all body size parameters were significantly correlated with ea/i and el/i (P,.001) (Tables 2, 3). Among the body size parameters, LBW exhibited the strongest correlation with ea/i (r = ) and el/i (r = ). At sex-stratified analysis, all body size parameters were independently correlated with ea/i. The correlation between LBW and ea/i was stronger in women than in the total population (r = vs r = ); the same was true for the correlation between LBW and el/i (r = vs r = ). When sex, age, liver function parameters, and LBW were included as dependent variables in our analysis, sex, age, albumin level, total bilirubin level, and LBW were significantly correlated with ea/i; however, some of the correlation coefficients were small (age, r = 0.175; albumin level, r = 0.128; and total bilirubin level, r = ) (Table 4). When sex, age, liver function parameters, LBW, and scan timing of PVP were included as dependent variables, sex, LBW, and scan timing were significantly correlated with el/i (P,.001, P =.001, and P =.008, respectively; Table 5). Because the partial regression Table 2 Standardized Partial Regression Coefficient (Correlation Coefficient) of Each Body Size Index for ea/i Body size index No. of Patients Estimate coefficient (estimate) of the scan timing of PVP was negative, el/i tended to be higher in patients whose scan timing was less than 80 seconds than in Mean Error of Estimate P Value All patients Height , BW , LBW , BSA , BMI , Men Height , BW , LBW , BSA , BMI , Women Height , BW , LBW , BSA , BMI , Table 3 Standardized Partial Regression Coefficient Standardized Partial Regression Coefficient (Correlation Coefficient) of Each Body Size Index for el/i Body Size Index No. of Patients Estimate Mean Error of Estimate P Value All patients Height , BW , LBW , BSA , BMI , Men Height , BW , LBW , BSA , BMI , Women Height , BW , LBW , BSA , BMI , Standardized Partial Regression Coefficient patients whose scan timing was 80 seconds or more. Receiver operating characteristic curve analysis of the relationship Radiology: Volume 278: Number 3 March 2016 n radiology.rsna.org 777

6 Table 4 Results of Hierarchical Multivariate Linear Regression Analysis Where Sex, Age, Liver Function Parameters, and LBW Were Independent Variables and ea/i Was the Outcome Variable Characteristic Estimate between the administered iodine dose (in milligrams of iodine per kilogram of LBW [mg I/kg]) and aortic enhancement showed that the area under the receiver operating characteristic curve was (P,.001) for the total population, (P,.001) for men, and (P,.001) for women. Assessment of the relationship between the administered iodine dose and liver parenchyma enhancement revealed that the area under the receiver operating characteristic curve was (P,.001) for the total population, (P,.001) for men, and (P,.001) for women. Receiver operating characteristic curve analysis showed that the required iodine dose to achieve aortic enhancement of 280 HU (22) and hepatic enhancement of 50 HU (23) was mg I/kg and mg I/kg, respectively, based on total population Mean Error of Estimate P Value Intercept ,.001 Sex ,.001 Age , Aspartate aminotransferase level Serum albumin level , Total bilirubin level , LBW , Table 5 Standardized Partial Regression Coefficient Results of Hierarchical Multivariate Linear Regression Analysis Including Sex, Age, Liver Function Parameters, the Scan Timing of the PVP, and LBW as Independent Variables and el/i as a Dependent Variable Characteristic Estimate Mean Error of Estimate P Value Intercept ,.001 Sex ,.001 Age Aspartate aminotransferase level Serum albumin level Total bilirubin level LBW , Scan timing of PVP Standardized Partial Regression Coefficient data; mg I/kg and mg I/ kg, respectively, based on male population data; and mg I/kg and mg I/kg, respectively, based on female population data. Thus, for hepatic dynamic CT, we think that the optimal iodine dose is mg I/kg in men and mg I/kg in women. Discussion In our multicenter study, we attempted to identify the body size parameter or parameters most appropriate for use in determining the CM dose for hepatic dynamic CT. We found that the correlation coefficients between each body size index and aortic and hepatic enhancement decreased in the order of LBW, BSA, BW, BMI, and height, although all indexes were significantly correlated with aortic and hepatic enhancement. The LBW was most strongly correlated with both aortic and liver enhancement (ea/i, r = 0.561; el/i, r = 0.601). In women, the correlation between LBW and ea/i or el/i was particularly strong (r = and r = 0.948, respectively). This led us to conclude that LBW was the best body size index with which to determine the CM dose for hepatic dynamic CT. On the basis of our receiver operating characteristic curve analysis, the required iodine dose per LBW to achieve aortic enhancement of 280 HU (22) and hepatic enhancement of 50 HU (23) was 687 mg/kg and mg/ kg, respectively, based on total population data; mg/kg and mg/ kg, respectively, based on male population data; and mg/kg and mg/kg, respectively, based on female population data. Thus, for hepatic dynamic CT, we think that mg I/kg for men and mg I/kg for women are the optimal iodine doses. Controversy surrounds the best body size index with which to determine the CM dose for contrast-enhanced CT studies. According to Bae et al (8), for cardiovascular CT, the BSA allowed for better adjustment of the iodine dose across a wide range of body sizes than did the BW; however, the correlation coefficient between aortic attenuation and BSA or BW was almost the same (r = for BSA, r = for BW). Onishi et al (12) and Yanaga et al (24) found that determining CM dose on the basis of BSA reduced patient-to-patient variability in aortic enhancement. Kidoh et al (14) reported that among various body size indexes, the BSA most strongly correlated with aortic attenuation during the arterial phase (r = for BSA). On the other hand, Svensson et al (13) reported that the correlation coefficient between aortic enhancement and the BSA or LBW was the same (r = 20.54). With respect to hepatic enhancement during PVP, other researchers (7,10,11) documented that a contrast-enhancement protocol based on LBW reduced patient-to-patient variability. Kondo et al (15), who also performed direct comparison of CM injection protocols based on BW, LBW, 778 radiology.rsna.org n Radiology: Volume 278: Number 3 March 2016

7 and BSA, did not state whether LBW or BSA yielded the better protocol. In earlier studies (13,14), both BW and BSA most strongly correlated with hepatic enhancement, although the correlation coefficients between hepatic enhancement and BW or BSA were similar (r = for BW, r = for BSA). The pharmacokinetics of CM (3,4,25) explain why LBW was the body size index most strongly correlated with arterial and hepatic enhancement. The injected iodine CM in the vascular space is distributed rapidly to plasma and the extracellular fluid space of parenchymal organs because it does not permeate into the intracellular space. Consequently, the concentration of CM in parenchymal organs is closely related to the volume of the extracellular fluid space and plasma. Boer (26) studied 87 healthy subjects and reported that the LBW had the strongest relationship with the volume of plasma, blood, and extracellular volume among height, BW, LBW, and BSA indexes. Peters et al (27) also found that the estimated LBW was more strongly correlated with extracellular fluid volume than was the BSA. Thus, we suggest that LBW is most strongly correlated with hepatic enhancement. A key physiologic parameter that affects arterial enhancement is cardiac output (25,28), which is closely related to BSA (20,25,29). In fact, the formula used by Bae (21) to predict cardiac output features the same term (BW [100 H] ) as the formula used by Du Bois and Du Bois (19) to predict BSA. On the other hand, Taylor et al (30) reported that LBW has a stronger correlation with cardiac output than BSA. LBW is a major predictor of functional capacity in the field of pharmacology (31), and we suggest it as a determinant for the administration of iodinecontaining CM. We have shown that LBW is more strongly correlated with aortic and hepatic enhancement in women than in men. In men, unknown patientrelated factors besides volume of the extracellular fluid space and cardiac output contribute to aortic and hepatic enhancement. Sex differences in atherosclerotic changes may be involved, as the degree of carotid atherosclerosis is higher in men than in women (32,33), as is the severity of coronary artery atherosclerosis (34 36). Atherosclerotic changes may increase vascular resistance, and this may affect enhancement of the aorta and liver. Another explanation is differences in blood flow in fat tissues between men and women. In general, women harbor less subcutaneous fat (37) in which blood flow is lower than that in visceral fat (38). In our hierarchical multivariate linear regression analysis for el/i, sex, LBW, and scan timing of the PVP were significantly correlated with el/i. Liver function parameters exhibited no significant correlation with el/i. Vignaux et al (16,17) reported decreased hepatic enhancement during PVP in patients with cirrhosis; thus, their findings and our findings are contradictory. However, a decrease in liver function may not be linearly correlated with a decrease in hepatic parenchymal enhancement, and only severe liver dysfunction (ie, cirrhosis) rather than mild or moderate liver dysfunction may result in decreased hepatic enhancement. Furthermore, el/i tended to be higher in patients whose scan timing of PVP was less than 80 seconds, suggesting that for greater hepatic enhancement it should be less than 80 seconds from the start of CM injection. Our study had limitations. First, we did not directly compare aortic enhancement and liver enhancement among the contrast enhancement protocols in which the CM dose was based on LBW, BSA, BW, BMI, and height. Although randomized studies in which different contrast-enhanced protocols are used would be of value, such investigations would require many protocol groups, and a large number of patients would need to be available during a short time period. Second, the scanning and CM injection protocols differed among institutions. To compensate for this, we used a twolevel hierarchical multivariate linear regression model (39). Third, although we used phantoms to calibrate iodine enhancement for the different CT scanners before we initiated the clinical study, this may not have eliminated interinstitutional temporal differences. Fourth, we obtained the patient height and BW from medical charts; however, it may be necessary to measure height and BW at the time of the CT examination. Fifth, we estimated the LBW of each patient by using BW and height; however, an LBW calculated from the percentage of body fat determined with the aid of an analyzer scale may yield a more accurate analysis. We used different formulas for men and women, and they may not be applicable to the Japanese body habitus. Had we used more appropriate formulas for the Japanese LBW, our correlation coefficient between LBW and aortic or liver enhancement might have been different for men and women. However, at present, there are no specific formulas with which to calculate LBW in the Japanese population. Sixth, we used CT number per gram of iodine in the aorta and liver as enhancement parameters. However, there may be nonlinear relationships between the parameters of CM injection and enhancement. For example, there are nonlinear effects between the concentration of CM and enhancement even at equal administered iodine doses (40). Seventh, there are several prediction formulas for LBW (26,27,41) and BSA (19,42,43), and the application of prediction formulas other than the one used in this study may yield different results. Eighth, all participating institutions were located in Japan. Our findings might not be applicable to patients who are not Japanese, as their body habitus may be different. Ninth, there was no statistical method with which to compare a standardized partial regression coefficient (correlation coefficient) among different model equations. Thus, although the correlation coefficient was highest for LBW, we do not know whether the difference was significant. In conclusion, with our methods, LBW exhibited the strongest correlation with aortic and hepatic enhancement. We suggest it as the best body Radiology: Volume 278: Number 3 March 2016 n radiology.rsna.org 779

8 size index with which to determine the appropriate CM dose at hepatic dynamic CT. Acknowledgment: Participants in this multicenter trial were drawn from radiology departments of the following institutions: Hiroshima University Hospital, Asahikawa Medical University, Gifu University Hospital, Hamamatsu University Hospital, Kanazawa Medical University Hospital, Mie University Hospital, Kumamoto University Hospital, University of Fukui Hospital, Nagoya City University Hospital, Hospital of the University of Occupational and Environmental Health, Osaka Medical College Hospital, St Marianna School of Medicine University Yokohama City Seibu Hospital, Gunma University Hospital, Iwate Medical University Hospital, University of Tsukuba Hospital, Toho University Ohashi Medical Center, Chiba University Hospital, Tokyo Women s Medical University Medical Center East, Saitama Medical Center Jichi Medical University, Oita University Hospital, Shinshu University Hospital, Osaka University Hospital, Osaka City University Hospital, Juntendo University Urayasu Hospital, Aichi Medical University Hospital, University of Yamanashi Hospital, Kyoto University Hospital, Wakayama Medical University Hospital, Kurume University Medical Center, and Teikyo University School of Medicine Hospital, Mizonokuchi. All institutions are located in Japan. Disclosures of Conflicts of Interest: K.A. Activities related to the present article: gave a lecture for Bayer Yakuhin. Activities not related to the present article: received grants from Toshiba Medical Systems, Eizai, and Daiichi Sankyo; gave lectures for Eizai and Daiichi Sankyo. Other relationships: disclosed no relevant relationships. M.K. disclosed no relevant relationships. T.K. disclosed no relevant relationships. T.I. disclosed no relevant relationships. Y.N. disclosed no relevant relationships. A.N. disclosed no relevant relationships. K.Y. disclosed no relevant relationships. T.M. disclosed no relevant relationships. N.M. disclosed no relevant relationships. Y.Y. disclosed no relevant relationships. References 1. Yamashita Y, Komohara Y, Takahashi M, et al. Abdominal helical CT: evaluation of optimal doses of intravenous contrast material a prospective randomized study. Radiology 2000;216(3): Awai K, Hiraishi K, Hori S. Effect of contrast material injection duration and rate on aortic peak time and peak enhancement at dynamic CT involving injection protocol with dose tailored to patient weight. Radiology 2004;230(1): Bae KT, Heiken JP, Brink JA. Aortic and hepatic peak enhancement at CT: effect of contrast medium injection rate pharma- cokinetic analysis and experimental porcine model. Radiology 1998;206(2): Bae KT, Heiken JP, Brink JA. Aortic and hepatic contrast medium enhancement at CT. Part I. Prediction with a computer model. Radiology 1998;207(3): Kondo H, Kanematsu M, Goshima S, et al. Abdominal multidetector CT in patients with varying body fat percentages: estimation of optimal contrast material dose. Radiology 2008;249(3): Li C, Ford ES, Zhao G, Balluz LS, Giles WH. Estimates of body composition with dualenergy x-ray absorptiometry in adults. Am J Clin Nutr 2009;90(6): Ho LM, Nelson RC, Delong DM. Determining contrast medium dose and rate on basis of lean body weight: does this strategy improve patient-to-patient uniformity of hepatic enhancement during multi-detector row CT? Radiology 2007;243(2): Bae KT, Seeck BA, Hildebolt CF, et al. Contrast enhancement in cardiovascular MDCT: effect of body weight, height, body surface area, body mass index, and obesity. AJR Am J Roentgenol 2008;190(3): Yanaga Y, Awai K, Nakaura T, et al. Effect of contrast injection protocols with dose adjusted to the estimated lean patient body weight on aortic enhancement at CT angiography. AJR Am J Roentgenol 2009;192(4): Kondo H, Kanematsu M, Goshima S, et al. Body size indexes for optimizing iodine dose for aortic and hepatic enhancement at multidetector CT: comparison of total body weight, lean body weight, and blood volume. Radiology 2010;254(1): Kondo H, Kanematsu M, Goshima S, et al. Aortic and hepatic enhancement at multidetector CT: evaluation of optimal iodine dose determined by lean body weight. Eur J Radiol 2011;80(3):e273 e Onishi H, Murakami T, Kim T, et al. Abdominal multi-detector row CT: effectiveness of determining contrast medium dose on basis of body surface area. Eur J Radiol 2011;80(3): Svensson A, Nouhad J, Cederlund K, et al. Hepatic contrast medium enhancement at computed tomography and its correlation with various body size measures. Acta Radiol 2012;53(6): Kidoh M, Nakaura T, Oda S, et al. Contrast enhancement during hepatic computed tomography: effect of total body weight, height, body mass index, blood volume, lean body weight, and body surface area. J Comput Assist Tomogr 2013;37(2): Kondo H, Kanematsu M, Goshima S, et al. Body size indices to determine iodine mass with contrast-enhanced multi-detector computed tomography of the upper abdomen: does body surface area outperform total body weight or lean body weight? Eur Radiol 2013;23(7): Vignaux O, Gouya H, Augui J, et al. Hepatofugal portal flow in advanced liver cirrhosis with spontaneous portosystemic shunts: effects on parenchymal hepatic enhancement at dual-phase helical CT. Abdom Imaging 2002;27(5): Vignaux O, Legmann P, Coste J, Hoeffel C, Bonnin A. Cirrhotic liver enhancement on dual-phase helical CT: comparison with noncirrhotic livers in 146 patients. AJR Am J Roentgenol 1999;173(5): Yamanaka N, Okamoto E, Kawamura E, et al. Dynamics of normal and injured human liver regeneration after hepatectomy as assessed on the basis of computed tomography and liver function. Hepatology 1993;18(1): Du Bois D, Du Bois EF. A formula to estimate the approximate surface area if height and weight be known Nutrition 1989;5(5): ; discussion Sawyer M, Ratain MJ. Body surface area as a determinant of pharmacokinetics and drug dosing. Invest New Drugs 2001;19(2): Bae KT. Intravenous contrast medium administration and scan timing at CT: considerations and approaches. Radiology 2010; 256(1): Yanaga Y, Awai K, Nakayama Y, et al. Optimal dose and injection duration (injection rate) of contrast material for depiction of hypervascular hepatocellular carcinomas by multidetector CT. Radiat Med 2007;25(6): Heiken JP, Brink JA, McClennan BL, Sagel SS, Crowe TM, Gaines MV. Dynamic incremental CT: effect of volume and concentration of contrast material and patient weight on hepatic enhancement. Radiology 1995;195(2): Yanaga Y, Awai K, Nakaura T, et al. Contrast material injection protocol with the dose adjusted to the body surface area for MDCT aortography. AJR Am J Roentgenol 2010;194(4): Bae KT, Heiken JP, Brink JA. Aortic and hepatic contrast medium enhancement at CT. II. Effect of reduced cardiac output in a porcine model. Radiology 1998;207(3): Boer P. Estimated lean body mass as an index for normalization of body fluid volumes 780 radiology.rsna.org n Radiology: Volume 278: Number 3 March 2016

9 in humans. Am J Physiol 1984;247(4 Pt 2): F632 F Peters AM, Snelling HL, Glass DM, Love S, Bird NJ. Estimated lean body mass is more appropriate than body surface area for scaling glomerular filtration rate and extracellular fluid volume. Nephron Clin Pract 2010; 116(1):c75 c Makita O, Yamashita Y, Arakawa A, et al. Diffuse perfusion abnormality of the liver parenchyma on angiography-assisted helical CT in relation to cirrhosis and previous treatments: a potential diagnostic pitfall for detecting hepatocellular carcinoma. Clin Imaging 2000;24(5): de Simone G, Devereux RB, Daniels SR, et al. Stroke volume and cardiac output in normotensive children and adults. assessment of relations with body size and impact of overweight. Circulation 1997;95(7): Taylor HL, Brozek J, Keys A. Basal cardiac function and body composition with special reference to obesity. J Clin Invest 1952;31(11): Morgan DJ, Bray KM. Lean body mass as a predictor of drug dosage. implications for drug therapy. Clin Pharmacokinet 1994;26(4): Ota H, Reeves MJ, Zhu DC, et al. Sex differences of high-risk carotid atherosclerotic plaque with less than 50% stenosis in asymptomatic patients: an in vivo 3T MRI study. AJNR Am J Neuroradiol. 2013;34(5): , S Sinning C, Wild PS, Echevarria FM, et al. Sex differences in early carotid atherosclerosis (from the community-based Gutenberg- Heart Study). Am J Cardiol 2011;107(12): Khosa F, Khan AN, Nasir K, et al. Comparison of coronary plaque subtypes in male and female patients using 320-row MDCTA. Atherosclerosis 2013;226(2): Lansky AJ, Ng VG, Maehara A, et al. Gender and the extent of coronary atherosclerosis, plaque composition, and clinical outcomes in acute coronary syndromes. JACC Cardiovasc Imaging 2012;5(3,Suppl):S62 S Shaw LJ, Shaw RE, Merz CN, et al. Impact of ethnicity and gender differences on angiographic coronary artery disease prevalence and in-hospital mortality in the American College of Cardiology-National Cardiovascular Data Registry. Circulation 2008;117(14): Viljanen AP, Lautamäki R, Järvisalo M, et al. Effects of weight loss on visceral and abdominal subcutaneous adipose tissue bloodflow and insulin-mediated glucose uptake in healthy obese subjects. Ann Med 2009;41(2): Remigio-Baker RA, Allison MA, Schreiner PJ, et al. Difference by sex but not by race/ ethnicity in the visceral adipose tissue-depressive symptoms association: the Multi- Ethnic Study of Atherosclerosis. Psychoneuroendocrinology 2014;47: Kandel DB, Kiros GE, Schaffran C, Hu MC. Racial/ethnic differences in cigarette smoking initiation and progression to daily smoking: a multilevel analysis. Am J Public Health 2004;94(1): Awai K, Inoue M, Yagyu Y, et al. Moderate versus high concentration of contrast material for aortic and hepatic enhancement and tumor-to-liver contrast at multi-detector row CT. Radiology 2004;233(3): Hume R. Prediction of lean body mass from height and weight. J Clin Pathol 1966; 19(4): Livingston EH, Lee S. Body surface area prediction in normal-weight and obese patients. Am J Physiol Endocrinol Metab 2001; 281(3):E586 E Mosteller RD. Simplified calculation of bod y-surface area. N Engl J Med 1987;317(17): Radiology: Volume 278: Number 3 March 2016 n radiology.rsna.org 781