Patient-Tailored Scan Delay for Multiphase Liver CT: Improved Scan Quality and Lesion Conspicuity With a Novel Timing Bolus Method

Transcription

1 Gastrointestinal Imaging Original Research Schneider et al. Multiphase Liver CT Gastrointestinal Imaging Original Research J. Gabriel Schneider 1 Zhen J. Wang Wilbur Wang Judy Yee Yanjun Fu Benjamin M. Yeh Schneider JG, Wang ZJ, Wang W, Yee J, Fu Y, Yeh BM Keywords: abdomen, aorta, CT, flow dynamics, gastrointestinal, hemodynamics, liver, technical aspects DOI:1.2214/AJR Received July 28, 212; accepted after revision June 14, All authors: Department of Radiology, University of California San Francisco, Box 628, M372, 55 Parnassus Ave, San Francisco, CA Address correspondence to B. M. Yeh (ben.yeh@ucsf.edu). AJR 214; 22: X/14/ American Roentgen Ray Society Scan for Multiphase Liver CT: Improved Scan Quality and Lesion Conspicuity With a Novel Timing Bolus Method OBJECTIVE. The purpose of this study was to compare scan quality and lesion conspicuity for late arterial and portal venous phase liver CT scans using fixed versus patient-tailored scan delay derived with an evidence-based timing bolus method. MATERIALS AND METHODS. We retrospectively identified the cases of 73 patients who underwent both multiphase liver CT with fixed late arterial and portal venous phase scan delay times of 45 and 8 seconds and subsequent multiphase liver CT with patient-tailored scan delay determined with a timing bolus and a previously reported relation between the time to peak aortic and liver enhancement. Both late arterial and portal venous phase scans were graded in terms of scan quality. Hepatic lesion conspicuity (difference in attenuation between lesion and liver parenchyma) for hypervascular lesions (late arterial phase) and hypovascular lesions (portal venous phase) was recorded. RESULTS. Patient-tailored scan delay reflected a wide range of times to peak aortic enhancement (mean, 24 seconds; range, seconds) and yielded a greater proportion of optimal scans compared with fixed scan delay for both late arterial phase (92% versus 74%, p <.1) and portal venous phase (86% versus 7%, p <.5) scans. Mean hypervascular lesion conspicuity was greater for lesions imaged with patient-tailored scan delay rather than fixed scan delay (84. versus 57. HU, p <.1). CONCLUSION. Compared with examinations with fixed scan delay, multiphase liver CT that incorporates patient-tailored scan delay produces more optimally timed late arterial and portal venous phase CT scans with greater lesion conspicuity. I mproved evaluation of hepatic lesions is of central importance for accurate medical imaging and proper patient care. In particular, the increasing incidence of hepatocellular carcinoma and the propensity of the liver to seed both hypervascular and hypovascular metastases, compounded with the known limited sensitivity of CT for the detection of liver malignancy [1], give urgency to the development of optimized hepatic imaging techniques. For IV contrast-enhanced multiphase liver CT, the choice of scan delay time strongly influences the conspicuity of hepatic lesions [2 6]. Although it remains common practice to use fixed scan delay times for the late arterial and portal venous phases of hepatic attenuation, such scan delays are based on population averages and may be inappropriate for patients with atypical blood circulation times [7 1]. For example, the time and magnitude of peak aortic enhancement vary with left ventricular ejection fraction, coronary artery disease, age, and weight [11 14]. Thus multiphase liver CT scans that are individually patient tailored may improve image quality, and potentially lesion detection, by accounting for interpatient variation. Variable success has been reported in tailoring multiphase liver CT to the patient with automated bolustriggering methods, which have generally coupled scan initiation to a fixed delay after the achievement of a preset threshold CT attenuation value in the aorta or liver [15 22]. Alternatively, a timing bolus may be used to predict appropriate CT scan delay. This method is most widely used in CT angiography, in which precise individualized timing of scan delay is critical [23 26]. Unlike automated bolus-triggering protocols, which trigger scanning based on a fixed delay after a threshold level of aortic enhancement, timing bolus examinations proportionally adjust scan delay to the time to peak aortic enhancement [27 3]. The time to peak aortic enhancement has been found to be linearly related to peak visceral or- 318 AJR:22, February 214

2 Multiphase Liver CT gan attenuation (liver and pancreas) and thus can be used to calculate individualized late arterial and portal venous scan delay times for CT [27]. It remains to be seen, however, whether patient-tailored scan delay based on the linear relations between peak aortic attenuation and organ enhancement improve CT scan quality and liver lesion conspicuity compared with fixed scan delay for multiphase liver imaging. The purpose of this study was to retrospectively evaluate image quality and lesion conspicuity in the late arterial and portal venous phases of multiphase liver CT using a novel timing bolus technique that involves proportionally lengthening scan delay on the basis of time to peak enhancement of the aorta. Materials and Methods Patients Our retrospective single-institution study was approved by our institutional review board. This study was compliant with HIPAA, and the requirement for written informed consent was waived. An electronic database search was performed to identify all patients who had undergone both timing bolus multiphase liver CT with patient-tailored scan delay during a 4-month period at our institution and a previous routine multiphase liver CT examination with fixed scan delay. The cases of 73 patients (15 women [age range, years; mean, 59 years], 58 men [age range, 43 8 years; mean, 59 years]) who had undergone both studies were identified. The overall mean time between CT examinations was 1.1 months (range, 2 29 months). The mean time between CT examinations for the 35 patients with liver tumors was 4.2 months (range, months). The primary indications for imaging included follow-up or evaluation of known liver tumors (n = 35), routine screening for cirrhosis (n = 24), evaluation of pancreatic lesions (n = 6), other abdominal masses (n = 4), and other (n = 4). For all patients with liver lesions found at CT, one author reviewed all available medical records to determine the cause of the liver lesions. CT Technique All CT scans were obtained with MDCT scanners (LightSpeed, GE Healthcare). Preliminary unenhanced CT images were obtained through the liver with the following parameters: contiguous 5-mm slice thickness; gantry rotation time,.8 seconds; table speed, 27 mm/rotation; tube potential, 12 kvp; and automatic tube current modulation to achieve a noise index of 11.6 HU. The fixeddelay multiphase liver CT images were acquired at 45 seconds for the late arterial phase and 8 seconds for the portal venous phase scans after the initiation of IV injection of 15 ml iohexol (Omnipaque 35, GE Healthcare) followed by a 3-mL normal saline flush administered at 5 ml/s. These fixed-delay CT images were obtained with the following parameters: contiguous 2.5-mm slice thickness; gantry rotation time,.8 seconds; table speed, 27 mm/rotation; tube potential, 12 kvp; and automatic tube current modulation to achieve a noise index of 11.6 HU. The follow-up multiphase liver CT scans with patient-tailored scan delay were obtained after our hospital-wide multiphase CT protocols were changed to incorporate a timing bolus method for the determination of scan delay. The timing bolus was given as a 3-mL iohexol injection followed by a 3-mL normal saline flush administered at 5 ml/s while the aorta was repeatedly imaged in a single 1-mm section thickness every 2 seconds for up to 6 seconds at the level of L1 with a low radiation dose CT technique (axial slice acquisition; tube potential, kvp; tube current, 4 ma). The technologist was allowed to terminate the timing bolus image acquisition manually when peak aortic enhancement was seen. The CT technologist then determined the time to peak aortic enhancement. The CT scan delay times for the diagnostic bolus were calculated as follows. For the late arterial phase, the scan delay was (1.64 time to peak aortic enhancement) 3 seconds. For the portal venous phase, the scan delay was (3.32 time to peak aortic enhancement) 17 seconds. These scan delays were based on previously reported relations between aortic and end-organ times to peak enhancement [27] and adjusted for the longer duration of the diagnostic bolus injection than the timing bolus injection The late arterial scan delay formula was based on the relation between time to peak pancreatic and peak aortic enhancement because, much like pancreatic tissue, hypervascular liver lesions are believed to derive their blood supply from the arterial rather than the portal venous blood supply [31]. The portal venous phase scan delay was chosen to be that of peak hepatic parenchymal attenuation because maximal parenchymal attenuation would likely maximize the conspicuity of hypovascular lesions. The diagnostic CT scan with the specific patient-tailored late arterial and portal venous scan delays was obtained with 12 ml IV iohexol followed by a 3-mL normal saline flush administered at 5 ml/s. Because the duration of the contrast injection for the diagnostic scan (12 ml at 5 ml/s) was 18 seconds longer than that of the timing bolus (3 ml at 5 ml/s), each of the two diagnostic scan delays (late arterial and portal venous) were empirically lengthened 9 seconds (one half of the difference in injection duration) as a corrective factor. Otherwise, scan parameters identical to the fixed-delay CT examinations were used. The total quantity of IV contrast material administered for the timing bolus determined CT examinations (12-mL diagnostic scan bolus plus 3-mL timing bolus) was identical to that used for the fixed-delay examinations (15 ml). Data Analysis The acquired images were viewed at random at a PACS (Impax 4.2, AGFA) by an attending radiologist with 7 years of abdominal imaging subspecialty experience who was blinded to patient identification and clinical information and the method of scan delay. The reader graded the late arterial phase images as early, optimal, or late and the portal venous phase images as very early, early, or optimal. For late arterial phase images, early was defined as only partial or no enhancement of the portal vein without hepatic vein enhancement. Optimal late arterial images were defined as having both strong enhancement of the portal vein and no enhancement of the hepatic veins. Late was defined as enhancement of the hepatic vein [2]. Very early portal venous phase images were defined as having portal vein enhancement and nonuniform or no enhancement of the hepatic veins. Early was defined as portal vein enhancement more than 2 HU greater than hepatic vein enhancement. Optimal portal venous images were defined as having enhancement of the portal vein and hepatic veins with similar attenuation (< 2 HU attenuation difference). In a separate interpretation session 2 months after the first, the reader evaluated both late arterial phase images showing hypervascular liver lesions (defined as lesions larger than 1 cm in diameter enhancing to a higher CT attenuation value than liver parenchyma) and portal venous phase images showing hypovascular lesions (defined as lesions larger than 1 cm in diameter enhancing to a lower CT attenuation value than liver parenchyma). The reader was blinded to the method of scan delay and viewed the images of the two time points side by side. The reader recorded lesion conspicuity as the absolute value of the difference between the CT attenuation of the lesion and the liver parenchyma. Only lesions visible on both scans were evaluated. For patients with multiple lesions, only the most conspicuous lesion of each type (hypovascular and hypervascular) not treated locally in the interim was included in data analysis. The same lesions were measured for fixed and patient-tailored scan delay, and the section in which the lesion was most conspicuous in a given examination was chosen. The CT attenuation of hypervascular lesions was recorded in an oval region of interest (ROI) measuring at least.4 mm 2 placed in the highest noncalcified enhancing area of the lesion for each scan. The CT attenuation of hypovascular lesions were recorded with AJR:22, February

3 Schneider et al Attenuation (HU) 5 Scan Fig. 1 Box plot shows liver parenchymal attenuation on late arterial phase scans (n = 73) obtained with fixed versus patient-tailored scan delay. There is no significant difference in average late arterial phase liver parenchymal attenuation (p =.13). Lower and upper bounds of boxes delineate 25th and 75th percentiles of data; line in box, median; whiskers, fifth the largest possible oval ROI in the center of the lesion placed at least 1 mm from all margins of the lesion. The CT attenuation of the liver parenchyma was measured as the mean of two approximately 1-cm 2 oval ROIs, one obtained in the left and one in the right lobe of the liver, taken from a transverse image that included the left main portal vein. ROIs for measuring liver parenchymal attenuation were placed in similar locations between the two sets of CT scans. All ROIs excluded visible vascular structures and areas of obvious artifact. In addition, for both fixed delay and patienttailored delay examinations, the maximum transverse diameter of the abdomen was measured and recorded as a surrogate measure of any significant change in body mass between CT examinations. Statistical Analysis Statistical analysis was performed with Stata software (version 8., Stata). The McNemar test for matched pairs was performed to describe the relation between quality of hepatic lesions in the late arterial phase for fixed-delay versus timing bolus guided CT scans. A paired Student t test was used to compare fixed delay and patient-tailored delay in determining mean liver parenchymal attenuation Conspicuity (HU) 2 Scan in the portal venous phase and hypervascular and hypovascular hepatic lesion conspicuity. Lesion conspicuity was defined as the absolute value of the difference between CT attenuation of the lesion and that of the liver parenchyma. For all tests, p <.5 was considered statistically significant. Results Patient Differences Between Scans In the interval between the fixed-delay and the patient-tailored delay examinations, 61 patients had no treatment, four underwent systemic therapy, and eight underwent local hepatic ablation. The mean maximum transverse diameter of the abdomen was not significantly different between scanning with patient-tailored and scanning with fixed delay (33.8 versus 33.5 cm, p =.6). The median absolute difference in the maximum transverse diameter of the abdomen between CT examinations was.6 cm (range, 3.5 cm). A Conspicuity (HU) Scan Fig. 2 Conspicuity of liver lesions. A, Box plot shows hypervascular liver lesion (n = 21) conspicuity during late arterial phase of enhancement for scans obtained with fixed versus patient-tailored scan delay. Lesion conspicuity is defined as CT attenuation difference between hypervascular lesion and background liver parenchyma. Mean hypervascular lesion conspicuity is higher for scans obtained with patient-tailored (84 HU) than fixed (57 HU) scan delay (p <.1). Lower and upper bounds of box delineate 25th and 75th percentiles of data; line in box, median; whiskers, fifth B, Line graph shows conspicuity of hypervascular liver lesions for each patient (n = 21) for fixed versus patienttailored scan delay. Eighteen of 21 (86%) patients have increased hypervascular liver lesion conspicuity with patient-tailored late arterial phase scan delay compared with fixed scan delay of 45 seconds. Peak Aortic Time and Scan For the 73 patients, the measured mean time to peak aortic enhancement was 24 seconds (range, seconds). These measurements translated to a mean late arterial phase scan delay of 45 seconds (range, seconds) and a mean portal venous scan delay of 76 seconds (range, seconds). Late Arterial Phase Images A greater fraction of late arterial phase scans obtained with patient-tailored scan delay were judged to be optimal (67/73 [92%]) compared with those obtained with a fixed scan delay of 45 seconds (54/73 [74%]) (p <.1). Two of the patient-tailored scans (2.7%) were acquired before the ideal late arterial phase (too early), and four (5.5%) were late, compared with seven (9.6%) early and 12 (16%) late scans with fixed scan delay. For late arterial phase images, the mean liver parenchymal attenuation was not significantly different between examinations performed with patient-tailored versus fixed scan delay (94.3 ± 15.8 HU versus 91.8 ± 24.5 HU, p =.13) (Fig. 1). Twenty-one of 73 patients (29%) were identified as having hypervascular liver lesions on late arterial phase CT scans. The hypervascular liver lesions in the 21 patients were hepatocellular carcinoma B 32 AJR:22, February 214

4 Multiphase Liver CT (n = 12), neuroendocrine tumor metastasis (n = 4), hemangioma (n = 2), and benign nongrowing nodules (n = 3). The scans obtained with patient-tailored scan delays showed significantly higher lesion conspicuity in the late arterial phase compared with those obtained with fixed scan delay, independent of scan quality (84. ± 68 HU versus 57. ± 37 HU, p <.1) (Figs. 2, 3A, and 3B). Eighteen of 21 patients had greater hypervascular lesion conspicuity when the CT scans were obtained with patient-tailored rather than fixed scan delay. For the 21 hypervascular liver lesions, the mean difference in lesion diameter between sets of CT scans was.42 cm (range,.5 to 1.7 cm). Portal Venous Phase Images A greater fraction of portal venous phase scans obtained with patient-tailored scan delay were judged to be optimal (63 of 73 [86%]) compared with those obtained with a fixed scan delay of 8 seconds (51 of 73 [7%], p <.5). Nine (12.3%) patient-tailored scans were early, and one (1.4%) was very early, compared with 2 (27%) early and two (2.7%) very early fixed-delay scans. For portal venous phase images, the mean liver parenchymal attenuation was not significantly different between examinations obtained with patient-tailored versus fixed scan delay (127 ± 25 versus 124 ± 23 HU, p =.9) (Fig. 4). Twenty of the 73 patients (27%) were found to have hypovascular lesions. The hypovascular liver lesions in the 2 patients were cysts (n = 5), metastatic lesions (n = 7), previous ablation sites (n = 6), and benign unknown lesions (n = 2). The mean hypovascular lesion conspicuity did not differ significantly between scans obtained with patient-tailored and those obtained with fixed scan delay (78.3 ± 3 versus 79.7 ± 33 HU, p =.39) (Figs. 3C, 3D, and 5). For the 2 hypovascular liver lesions, the mean difference in lesion size between sets of CT scans was.9 cm (range,.8 to 2.2 cm). For the seven patients with hypovascular metastasis, lesion conspicuity was also not significantly different between patient-tailored and fixed scan delay (57 versus 54.4 HU, p =.83). Discussion Our results show that there is wide interpatient variation in time to peak aortic enhancement and that use of patient-tailored scan delay, which accounts for this highly variable individual parameter, results in higher quality of late arterial phase liver CT images than does use of fixed scan delay. We found that proportional lengthening of scan delay based on known linear relations between peak aortic and peak end-organ enhancement times significantly improves hypervascular lesion conspicuity in the late arterial phase. Importantly, use of a smaller diagnostic contrast bolus for the timing bolus method (12 ml) did not reduce portal venous phase liver attenuation or hypovascular lesion conspicuity compared with the fixed scan delay method (15 ml), possibly owing to more accurate timing of the latter phase. A C Fig year-old man with pancreatic neuroendocrine tumor liver metastasis. Hypervascular metastatic lesions are more conspicuous with patient-tailored versus fixed scan delay in late arterial phase but not in portal venous phase. A, Standard 45-second-delay late arterial phase CT image shows hypervascular metastatic (arrow) lesion measures 127 HU and background liver parenchyma measures 5 HU. B, Patient-tailored scan delay late arterial phase CT image shows metastatic lesion in A measures 15 HU and background liver parenchyma measures 68 HU. Thus conspicuity of lesion increased from 77 to 88 HU on patient-tailored timing bolus scan in late arterial phase. C, Standard 8-second scan delay portal venous phase CT image shows ablated metastatic lesion measures 3 HU and background liver parenchyma measures 95 HU. D, Patient-tailored scan delay portal venous phase CT image shows metastatic lesion in C measures 4 HU and background liver parenchyma measures 15 HU. Thus conspicuity of lesion remains same (65 HU) on both types of scans in portal venous phase. Previous reports on patient-tailored hepatic CT scan delay have focused on the use of automated bolus-triggered scan delay for optimizing liver CT [15 22]. Two studies showed that automated bolus-triggered scans had improved arterial phase image quality, but neither study specifically examined the late arterial phase [17, 2]. One group [2] determined that 54 67% of fixed-delay arterial phase scans were appropriately timed, compared with 83 93% of automated bolus-triggered scans. These values are similar to our late arterial phase findings on a timing bolus method. B D AJR:22, February

5 Schneider et al. Attenuation (HU) Scan Fig. 4 Box plot shows liver parenchymal attenuation during portal venous phase scans (n = 73) obtained with fixed versus patient-tailored scan delay. There was no significant difference in average portal venous phase liver parenchymal attenuation for fixed versus patient-tailored scans (p =.9). Lower and upper bounds of box delineate 25th and 75th percentiles of data; line in box, median; whiskers, fifth Other investigators [3, 15, 19] have evaluated whether use of automated bolus triggering improves the conspicuity of hypervascular hepatic lesions. A previous study [3] of automated bolus-triggered hepatic arterial phase CT showed a mean contrast-attenuation difference between hepatocellular carcinomas and liver parenchyma of 43 HU during a first arterial phase acquisition (mean time, 23 seconds) and 31 HU during a second arterial phase acquisition (mean time, 4 seconds). In a more recent study [15], the investigators evaluated automated bolus-triggered liver CT with 18-second delay after threshold aortic enhancement and found a mean contrast-attenuation difference of 47.2 HU between hepatocellular carcinomas and background liver parenchyma. Compared with these previous investigators, we observed a higher conspicuity of hypervascular lesions (mean, HU). Unfortunately, the results cannot be directly compared because we used a higher dose of denser contrast material. Nevertheless, our study built on previous work by showing Conspicuity (HU) 15 5 Scan Fig. 5 Box plot shows hypovascular liver lesion (n = 2) conspicuity during portal venous phase of enhancement for scans obtained with fixed versus patient-tailored delay. Lesion conspicuity is defined as absolute value of CT attenuation difference between hypovascular lesion and background liver parenchyma. Mean hypovascular lesion conspicuity is similar for scans obtained with patient-tailored (78 HU) and fixed (8 HU) scan delay (p =.39). Lower and upper bounds of box delineate 25th and 75th percentiles of data; line in box, median; whiskers, fifth that patient-specific scan delay as determined with a novel timing bolus calculation produces better scan quality and higher hypervascular lesion conspicuity than those achieved with fixed scan delay. To our knowledge no study has directly compared timing bolus peak aortic attenuation methods and automated threshold bolus-triggered methods, and further work is needed to determine whether one method substantially outperforms the other. A potential limitation of automated bolus-triggered scan delay is that triggering occurs on a threshold of enhancement [15 22] rather than at maximal enhancement. In addition, automated bolus-triggered scan delay entails a fixed delay after a threshold of enhancement is reached, regardless of the length of time it took to achieve that threshold [27], and may be inappropriate for patients with very rapid or very slow circulation times. Some limitations to the use of a timing bolus should also be recognized. A minimal amount of additional radiation is required to monitor the timing bolus for the time to peak aortic enhancement. However, the timing bolus can be obtained with a very low radiation dose by the use of thick sections and low tube current. The additional radiation exposure is approximately equivalent to that for obtaining fewer than six diagnostic CT images (up to 3 timing bolus slices at kvp and 4 ma for a total of 12 ma compared with images obtained at 12 kvp with approximate average tube current of 2 ma for a typical CT slice). The risk from a slight increase in radiation dose must be weighed against a higher probability of obtaining a suboptimal CT scan if a timing bolus or bolus detection method is not used. Another potential limitation of using a timing bolus method is that the infusion of iodinated contrast material for the timing bolus may increase the background liver attenuation and theoretically diminish lesion conspicuity. However, we found that images acquired in the patient-tailored examinations showed excellent hypervascular lesion conspicuity, suggesting that the presence of iodinated contrast material in the small timing bolus does not interfere with subsequent diagnostic image quality. Another benefit of using a small timing bolus is the potential improvement in the safety of hepatic CT, which typically requires a high contrast injection rate. A timing bolus examination serves as a test injection for the IV catheter and may alert the technologist to the presence of faulty catheter insertion, thereby reducing the likelihood of large-volume contrast extravasation. Our study had several limitations. First, there was a variable amount of time between the initial multiphase scan and the subsequent patient-tailored scan, ranging from several weeks to slightly more than 2 years. During this time, it is possible that changes in both the hepatic disease (cirrhosis, hepatocellular carcinoma) and the cardiovascular system (secondary to liver disease) may have influenced results. However, the inclusion of patients with hepatic disease (cirrhosis, hepatocellular carcinoma) in our study supports the use of a timing bolus technique in this type of patient cohort. In addition, patient body mass index and cardiac output were not available for analysis to determine significant changes in cardiovascular status between scans; however, the maximum transverse diameter of the abdomen did not vary significantly between scans. The second limitation was that the time to peak aortic enhancement was determined by the individual technologist who reviewed the timing bolus images, not with automated com- 322 AJR:22, February 214

6 Multiphase Liver CT puter analysis, which may be more accurate. However, our technologists are accustomed to performing and interpreting timing boluses for many CT applications, so manual assessment likely does not introduce significant variability into the scan delay times used. Last, only one reader interpreted the images. However, our criteria for image evaluation were largely objective because they were based heavily on CT attenuation values. Conclusion Multiphase liver CT examinations that account for patient-tailored intravascular variation produce higher-quality late arterial phase CT scans with greater hypervascular lesion conspicuity than on fixed-delay CT scans. These findings highlight the importance of tailoring scan delay to the individual patient s physiologic status. References 1. Murakami T, Kim T, Takamura M, et al. Hypervascular hepatocellular carcinoma: detection with double arterial phase multi-detector row helical CT. Radiology 21; 218: Frederick MG, McElaney BL, Singer A, et al. Timing of parenchymal enhancement on dual-phase dynamic helical CT of the liver: how long does the hepatic arterial phase predominate? AJR 1996; 166: Goshima S, Kanematsu M, Kondo H, et al. MDCT of the liver and hypervascular hepatocellular carcinomas: optimizing scan delays for bolus-tracking techniques of hepatic arterial and portal venous phases. AJR 26; 187:[web]W25 W32 4. Kim T, Murakami T, Hori M, et al. Small hypervascular hepatocellular carcinoma revealed by double arterial phase CT performed with single breath-hold scanning and automatic bolus tracking. AJR 22; 178: Murakami T, Kim T, Kawata S, et al. Evaluation of optimal timing of arterial phase imaging for the detection of hypervascular hepatocellular carcinoma by using triple arterial phase imaging with multidetector-row helical computed tomography. Invest Radiol 23; 38: Yu JS, Kim KW, Jo BJ, et al. Test-bolus injection for optimization of arterial phase imaging during contrast-enhanced hepatic MR imaging. Yonsei Med J 2; 41: Bae KT. Peak contrast enhancement in CT and MR angiography: when does it occur and why? ology 23; 227: Kanematsu M, Goshima S, Kondo H, et al. Optimizing scan delays of fixed duration contrast injection in contrast-enhanced biphasic multidetector-row CT for the liver and the detection of hypervascular hepatocellular carcinoma. J Comput Assist Tomogr 25; 29: Cox IH, Foley WD, Hoffmann RG. Right window for dynamic hepatic CT. Radiology 1991; 181: 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: Bae KT, Heiken JP, Brink JA. Aortic and hepatic contrast medium enhancement at CT. Part II. Effect of reduced cardiac output in a porcine model. Radiology 1998; 27: Bae KT, Heiken JP, Brink JA. Aortic and hepatic contrast medium enhancement at CT. Part I. Prediction with a computer model. Radiology 1998; 27: Francois CJ, Shors SM, Bonow RO, Finn JP. Analysis of cardiopulmonary transit times at contrast material-enhanced MR imaging in patients with heart disease. Radiology 23; 227: Paulson EK, Fisher AJ, DeLong DM, Parker DD, Nelson RC. Helical liver CT with computer-assisted bolus-tracking technology: is it possible to predict which patients will not achieve a threshold of enhancement? Radiology 1998; 29: Sultana S, Awai K, Nakayama Y, et al. Hypervascular hypercellular carcinomas: bolus tracking with a 4-detector CT scanner to time arterial phase imaging. Radiology 27; 243: Itoh S, Ikeda M, Achiwa M, Satake H, Iwano S, Ishigaki T. Late-arterial and portal-venous phase imaging of the liver with a multislice CT scanner in patients without circulatory disturbances: automatic bolus tracking or empirical scan delay? Eur Radiol 24; 14: Mehnert F, Pereira PL, Trubenbach J, Kopp AF, Claussen CD. Biphasic spiral CT of the liver: automatic bolus tracking or time delay? Eur Radiol 21; 11: Dinkel HP, Fieger M, Knupffer J, Moll R, Schindler G. Optimizing liver contrast in helical liver CT: value of a real-time bolus-triggering technique. Eur Radiol 1998; 8: Chan R, Kumar G, Abdullah B, et al. Optimising the scan delay for arterial phase imaging of the liver using the bolus tracking technique. Biomed Imaging Interv J 211; 7:e12 Oestmann JW, Grabbe E. Dual-phase helical CT of the liver: effects of bolus tracking and different volumes of contrast material. Radiology 1996; 21: Sandstede JJ, Tschammler A, Beer M, Vogelsang C, Wittenberg G, Hahn D. Optimization of automatic bolus tracking for timing of the arterial phase of helical liver CT. Eur Radiol 21; 11: Shimizu T, Misaki T, Yamamoto K, Sueyoshi K, Narabayashi I. Helical CT of the liver with computer-assisted bolus-tracking technology: scan delay of arterial phase scanning and effect of flow rates. J Comput Assist Tomogr 2; 24: Cademartiri F, Nieman K, van der Lugt A, et al. Intravenous contrast material administration at 16-detector row helical CT coronary angiography: test bolus versus bolus-tracking technique. Radiology 24; 233: van Hoe L, Marchal G, Baert AL, Gryspeerdt S, Mertens L. Determination of scan delay time in spiral CT-angiography: utility of a test bolus injection. J Comput Assist Tomogr 1995; 19: Vogl TJ, Abolmaali ND, Diebold T, et al. Techniques for the detection of coronary atherosclerosis: multi-detector row CT coronary angiography. Radiology 22; 223: Achenbach S, Giesler T, Ropers D, et al. Detection of coronary artery stenoses by contrast-enhanced, retrospectively electrocardiographically-gated, multislice spiral computed tomography. Circulation 21; 13: Chu LL, Joe BN, Westphalen AC, Webb EM, Coakley FV, Yeh BM. Patient-specific time to peak abdominal organ enhancement varies with time to peak aortic enhancement at MRI. Radiology 27; 245: Earls JP, Rofsky NM, DeCorato DR, Krinsky GA, Weinreb JC. Hepatic arterial-phase dynamic gadolinium-enhanced MR imaging: optimization with a test examination and a power injector. Radiology 1997; 22: Kanematsu M, Semelka RC, Matsuo M, et al. Gadolinium-enhanced MR imaging of the liver: optimizing imaging delay for hepatic arterial and portal venous phases a prospective randomized study in patients with chronic liver damage. Radiology 22; 225: Goshima S, Kanematsu M, Kondo H, et al. Optimal acquisition delay for dynamic contrast-enhanced MRI of hypervascular hepatocellular carcinoma. AJR 29; 192: Baron RL. Understanding and optimizing use of contrast material for CT of the liver. AJR 1994; Pharmacokinetic study in a porcine model. Radi- 2. Kopka L, Rodenwaldt J, Fischer U, Mueller DW, 163: AJR:22, February