Effect of Gadoxetic Acid Injection Duration on Tumor Enhancement in Arterial Phase Liver MRI

Effect of Gadoxetic Acid Injection Duration on Tumor Enhancement in Arterial Phase Liver MRI

ARTICLE IN PRESS Original Investigation Effect of Gadoxetic Acid Injection Duration on Tumor Enhancement in Arterial Phase Liver MRI Takahiro Tsuboy...

1MB Sizes 0 Downloads 26 Views

ARTICLE IN PRESS

Original Investigation

Effect of Gadoxetic Acid Injection Duration on Tumor Enhancement in Arterial Phase Liver MRI Takahiro Tsuboyama, MD, PhD, Gregor Jost, PhD, Hubertus Pietsch, PhD, Noriyuki Tomiyama, MD, PhD

Rationale and Objectives: Rapid injection of gadoxetic acid has been shown not to increase tumor enhancement in arterial phase liver MRI for unknown reasons. This study aimed to investigate the effect of injection durations on peak contrast concentration in tumors and to correlate it with signal enhancement in gadoxetic acid-enhanced arterial phase MRI. Materials and Methods: Gadoxetic acid-enhanced arterial phase MRI was obtained using a bolus-tracking technique with injection durations of 1, 3, and 6s in six rabbits with VX2 liver tumors. The peak concentration of gadoxetic acid in the aorta and tumor was estimated by iopromide-enhanced time-resolved CT using the same injection volume and durations with those for MRI. Signal enhancement on MRI and peak enhancement on CT were compared and correlated. Results: There was no significant difference in MR signal enhancement of tumors among the 3 injection durations (p = 0.87). In CT, shorter injection durations significantly increased peak contrast concentration in the aorta (p < 0.01) but produced equivalent peak contrast concentration in tumors (p = 0.24). The longer injections resulted in the stronger correlation between peak contrast concentration in CT and MR signal enhancement in tumors (r = 0.31, 0.43, and 0.86 with 1s-, 3s-, and 6s-injection, respectively) with a statistical significance only found with 6s-injection (p = 0.03). Conclusion: Estimation of contrast concentration by CT demonstrated that shorter injections did not increase peak contrast concentration in tumors despite increased peak concentration in the aorta. Furthermore, tumor signal enhancement in gadoxetic acid-enhanced arterial phase MRI was less correlated with the peak contrast concentration with shorter injections. Key words: Gadoxetic acid; Injection duration/Injection rate; Tumor enhancement; Magnetic resonance imaging; Liver. © 2019 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved. Abbreviations: MRI magnetic resonance imaging, CT computed tomography, SI signal intensity, ROI region of interest, ER enhancement ratio, CR contrast ratio, HU Hounsfield units, AUC area under the curve

INTRODUCTION

G

adoxetic acid-enhanced magnetic resonance imaging (MRI) can provide both dynamic and hepatobiliary phase imaging, and during the dynamic phase imaging, the arterial phase plays an essential role in characterizing focal liver lesions (13). Due to the pronounced uptake into hepatocytes and the high r1-relaxivity, the recommended diagnostic dose and volume of gadoxetic acid is one-fourth (0.025 mmol/ kg) and half (0.1 ml/kg), respectively, compared to multiAcad Radiol 2019; &:1–8 From the Department of Radiology, National Hospital Organization Osaka National Hospital, 2-1-14 Hoenzaka, Chuo-ku, Osaka 540-0006, Japan (T.T.); MR and CT Contrast Media Research, Bayer Pharma AG, Berlin, Germany (G.J., H.P.); Department of Radiology, Osaka University Graduate School of Medicine, Osaka, Japan (N.T.). Received August 25, 2019; revised October 14, 2019; accepted October 17, 2019. Address correspondence to: TT. e-mail: [email protected] © 2019 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.acra.2019.10.018

purpose conventional gadolinium-based contrast agents (0.1 mmol/kg and 0.2 ml/kg). Therefore, given the compact bolus, bolus timing techniques such as bolus-tracking techniques or multiple arterial phase acquisitions are demanded to obtain well-timed arterial phase imaging (48). Previous studies suggest that a decreased injection rate of 1 ml/s instead of conventional 23 ml/s is also useful to avoid acquisition timing errors because the prolonged injection duration stretches the contrast bolus (79). In fact, slow injection has been shown to increase aortic enhancement (7,1012). Theoretically, rapid contrast injection should increase peak contrast concentration in the aorta and subsequently yield higher peak aortic and tumor enhancement as in clinical dynamic liver computed tomography (CT) (1315). Therefore, if the peak enhancement could be precisely captured in MRI, rapid injection of gadoxetic acid would provide higher signal enhancement. Despite the tremendous advances in scan techniques to capture the peak enhancement especially for tumors which occurs about 10s

1

ARTICLE IN PRESS TSUBOYAMA ET AL

later than the aortic bolus peak (48,16), rapid injection has been shown not to increase tumor or organ enhancement (1012,17). Since signal enhancement in MRI is influenced not only by scan timing but also by many other factors such as nonlinear relationship between a signal intensity and gadolinium concentration, r1-relaxivity, and pulse sequence parameters (1821), we speculate that timing errors are not the only cause of the lowered signal enhancement with rapid injection for tumors. Most importantly, effect of injection durations on gadoxetic acid concentration in tumors remains unknown. Time-resolved dynamic CT allows precise and quantitative evaluation of contrast agent kinetics in vessels and tumors because CT attenuation values have a linear correlation with iodine concentration (2123). Therefore, we think that dynamic changes in gadoxetic acid concentration in the aorta and tumors can be estimated with this technique by administrating iodinated contrast agents with the same injection volume and durations with those for MRI. Moreover, CT enables assessment of small arteries such as peritumoral arteries thanks to its high temporal and spatial resolution, which may help to understand the contrast agent flow from the aorta to tumors. Therefore, the purpose of this animal study was to investigate the effect of injection durations on peak contrast agent concentration in tumors and to correlate it with signal enhancement in gadoxetic acid-enhanced arterial phase MRI. MATERIALS AND METHODS Animals

This animal study was approved by the local authorities and performed in accordance with German animal regulations. Six male New Zealand white rabbits (Charles River, Sulzfeld, Germany) with a mean body weight of 4.2 kg (range: 3.84.4 kg) were examined. They were inoculated with 1 £ 1-mm pieces of VX2 carcinoma tissue into the left medial liver lobe by laparotomy two weeks before the image examinations. All tumors were successfully developed and ranged in diameter from 1.6 to 2.2 cm (mean: 1.9 cm). For MRI and CT scans, the rabbits were investigated under general anesthesia induced with 10 mg/kg of xylazine (Rompun, Bayer Vital GmbH, Leverkusen, Germany) and 30 mg/kg of ketamine (Pharmacia, Karlsruhe, Germany). Anesthesia was maintained by intravenous infusion of 0.9 ml/kg/h Propofol (Ratiopharm, Ulm, Germany). They were orotracheally intubated, ventilated with 55% oxygen, and placed in the scanner in a supine position. Ventilation was stopped during image acquisitions to enable breath-hold imaging. Contrast Agent Injection Protocols

A standard dose (0.025 mmol Gd/kg) of gadoxetic acid (Primovist, Bayer Vital GmbH) and 185 mg I/kg of iopromide (Ultravist 370, Bayer Vital GmbH) were used for dynamic liver MRI and CT, respectively. Gadoxetic acid was diluted 5-fold with saline to enable the usage of a commercially available power injector 2

Academic Radiology, Vol &, No &&, && 2019

(Spectris Solaris, Bayer Pharmaceuticals, Pittsburgh, PA) for the small animals. The injection volume of gadoxetic acid and iopromide were the same (0.5 ml/kg, range: 1.92.2 ml). Contrast media were injected via an ear vein followed by 8 ml of a saline chaser using the automated injection system. Each rabbit underwent dynamic MRI followed by dynamic CT on 3 consecutive days, allowing a time interval of at least 4 hours between the examinations. The following 3 injection protocols were assigned in a randomized order to the 3 examination days: Short injection (injection duration: 1s, injection rate: 1.92.2 ml/s), standard injection (injection duration: 3s, injection rate: 0.630.73 ml/s), and long injection (injection duration: 6s, injection rate: 0.320.37 ml/s). Considering the fast circulation of the rabbits, injection duration of 3s was set as the standard injection according to the previous study of gadoxetic acid-enhanced MRI for rabbits (12). MRI Examination

A commercially available 1.5T MR system (Magnetom Avanto, Siemens Healthcare, Erlangen, Germany) with 4-channel knee coil was used. After axial pre-contrast image acquisition (3 times), dynamic liver MRI was performed in 5 postcontrast phases according to a previously described method, in which the 1st and 2nd phases corresponded to arterial phase (12). Injection of contrast agent and coronal bolus tracking scans (thoracic region) were started simultaneously, and immediately after detection of contrast enhancement on the pulmonary artery, 5 continuous axial dynamic scans were obtained. Fat-saturated T1-weighted 3D gradient echo sequence (Volumetric Interpolated Breath-bold Examination, VIBE) was used for all scans with the parameters shown in Table 1. Time-resolved CT Examination

Time-resolved CT was performed with a shuttle mode using a dual-source CT scanner (SOMATOM Definition, Siemens Healthcare). The tumor was positioned at the center of the scan range. The scan was started simultaneously with the start of contrast media injection and continued for 50 seconds with 50 repetitive bidirectional table movements. CT parameters were shown in Table 1. Gadoxetic Acid-enhanced MRI Evaluation

Signal intensities (SIs) were measured by 1 radiologist with 17 years of experience in abdominal MRI using a standard workstation (MMWP, Siemens Healthcare). SI of the aorta was measured by placing circular regions of interest (ROIs) of 0.15 cm2 at three different levels (9th, 18th, and 27th of 36 slices), and the three measured SIs were then averaged. The SI of tumors was measured by tracing the portions of ring enhancement. There was a little variation in the ROI size of the tumor for the 3 injection protocols due to the free hand placement (mean size = 0.73, 0.69, and 0.72 cm2 for short, standard, and long injection protocols, respectively). The SI

ARTICLE IN PRESS Academic Radiology, Vol &, No &&, && 2019

EFFECT OF GADOXETIC ACID INJECTION DURATION

TABLE 1. Summary of MR and CT Examination

Time-resolved CT Evaluation

MR examination

The 50 dynamic-phase image sets were sent to the workstation and the radiologist measured attenuation values in Hounsfield units (HU) by placing ROIs at the level of the largest area of the tumor which was located approximately at the center of the scan range. Circular ROIs of 0.15 cm2, 0.02 cm2, and 0.35 cm2 were placed in the aorta, peritumoral arteries (small arteries located at the periphery of the tumor, Fig. 1), and liver parenchyma, respectively. The density of liver parenchyma was measured in the same way as the MRI evaluation. For tumors, free hand ROIs with mean size of 0.41, 0.44, and 0.42 cm2 for short, standard and long injection protocols, respectively, were placed to include the portions of ring enhancement devoid of tumor vessels. Enhancement of each object was calculated as the difference in the attenuation value on each phase from that on the 1st phase and time-enhancement curves were generated. On the time-enhancement curves, (1) peak aortic enhancement, (2) the area under the curve (AUC) of the aorta, (3) aortic bolus length, (4) peak peritumoral artery enhancement, (5) AUC for peritumoral artery, and (6) peak tumor enhancement were evaluated (Fig. 1) and compared between the three injection protocols. Aortic bolus length was defined by the period when aortic enhancement was 200 HU or greater. To investigate the cumulative effect of vessel enhancement to yield the peak tumor enhancement, the AUC was calculated cumulatively for each

Pulse sequence Acquisition planes Repetition time (ms) Echo time (ms) Flip angle (degree) Matrix Field of view (mm) Slice thickness (mm) Slices (n) Parallel imaging factor Scan time (s) The center of the k-space

3D T1-weighted gradient echo with fat saturation Axial 4.73 2.38 10 256*85 310*102 2 36 2 6.2 1.9s after the start

CT examination Scan method Acquisition planes kVp/mAs Scan range (mm) Field of view (mm) Slice thickness (mm) Temporal resolution (s)

Helical scan with a shuttle mode Axial 80/100 48 140 2 1

of the surrounding liver parenchyma was measured by placing two circular ROIs of 0.35 cm2 into the liver adjacent to the tumor. The measured 2 SIs were then averaged. ROIs were copied and placed on the same position of precontrast, 1st, and 2nd phase images. Then, the enhancement ratio (ER) and the tumor-liver contrast ratio (CR) on each phase were calculated with the following formulas: ER ¼ ðSIpost SIpre Þ=SIpre

ð1Þ

where SIpost is the postcontrast SI and SIpre is the average of the three precontrast SIs; tumorliver CR ¼ SItumor =SIliver

ð2Þ

where SItumor is the SI of the tumors and SIliver is the SI of the surrounding liver parenchyma (12). The 1st and 2nd phase images of all dynamic MRI were evaluated by the radiologist if they were scanned in the appropriate timing as early or late arterial phase imaging. Appropriate timing was judged by the following findings: strong aortic enhancement, absent enhancement of portal vein, and absent or minimal enhancement of liver and tumor for early arterial phase imaging; maximal tumor-liver contrast due to strong tumor enhancement and minimal liver enhancement, and moderate to strong enhancement of portal vein for late arterial phase imaging (7,10). Aortic enhancement (i.e., ER of the aorta) was assessed in the early arterial phase, and tumor enhancement (i.e., ER of the tumors) and tumor-liver CR were assessed on the late arterial phase. ER and CR were compared between the 3 injection protocols.

Figure 1. An example of time-enhancement curves of the aorta, peritumoral artery, tumor and liver on time-resolved CT. Inset is a CT image showing aortic enhancement (long arrow), peritumoral artery enhancement (short arrow), and tumor enhancement (arrowheads).

3

ARTICLE IN PRESS TSUBOYAMA ET AL

time point from the start of the scan to the timepoint of peak tumor enhancement using the trapezoidal rule (24). Correlation between peak aortic and peak tumor enhancement was evaluated for each injection duration. Blood iodine concentrations were calculated using the previously published phantom data that 15 mgI/ml of iopromide increased 568.5 HU (i.e., 1 mgI/ml = 37.9 HU) at tube voltage of 80 kV (25).

Statistical Analysis

The Friedman test was used for multigroup comparisons, and a test for pairwise comparison of variables was performed according to Conover when there was a significant difference among groups (26). Linear regression analysis was used to evaluate the correlation between peak aortic and peak tumor enhancement on CT and coefficient of determination (R2) was calculated. Correlation between ER of the tumors on late arterial phase MRI and peak tumor enhancement on CT was determined by Pearson’s correlation coefficient for each injection duration. A p value less than 0.05 was considered statistically significant. MedCalc 18.11 (MedCalc Software bvba, Ostend, Belgium) was used for all statistical analyses.

Academic Radiology, Vol &, No &&, && 2019

RESULTS Gadoxetic Acid-enhanced Arterial Phase MRI

Appropriate early arterial phase imaging could be successfully obtained at the 1st phase in all rabbits with standard and long injection protocols, but only in 3 of the 6 rabbits with short injection protocol (Fig. 2). As a result, aortic enhancement on the early arterial phase was compared in the 3 rabbits, and there was no significant difference (median aortic ER = 4.61, 4.45, and 4.77 with short, standard, and long injection, respectively, p = 0.790) among the three injection durations (Fig. 3a). Appropriate late arterial phase imaging could be obtained in all rabbits with all injection protocols. The 2nd phase corresponded to the late arterial phase in all rabbits with standard and long injection protocols. With short injection protocol, the late arterial phase imaging was obtained at the 1st phase in three rabbits and at the 2nd phase in the other three rabbits. Tumor enhancement and tumor-liver contrast on the late arterial phase were compared in the six rabbits, and there were no significant differences in tumor enhancement (median tumor ER = 0.63, 0.59, and 0.51 with short, standard, and long injection, respectively, p = 0.869) and tumor-liver contrast (median CR = 1.18, 1.21, and 1.20 with short, standard, and long injection, respectively, p = 0.555) among the three injection durations (Fig 3b, c). Time-enhancement Curves Obtained From CT

Figure 2. A representative case (a rabbit with a VX2 liver tumor). On gadoxetic acid-enhanced MRI, 1st phase missed early arterial phase and corresponded to late arterial phase with short injection, while 1st and 2nd phase corresponded to early and late arterial phase, respectively, with standard and long injections. Short injection produced slightly lower tumor enhancement on the late arterial phase than standard and long injections (short arrows, enhancement ratio=0.61, 0.74, and 0.73 for short, standard, and long injection, respectively). Time-resolved CT showed higher peak aortic enhancement with shorter injections (long arrows, 902.6, 703.6, and 566.7 HU with short, standard, and long injection, respectively), but slightly higher peak tumor enhancement with longer injections (short arrows, 77.3, 83.4, and 86.4 HU with short, standard, and long injection, respectively). Cine imaging of the CT data were shown in Supplementary video 1.

4

Shorter injection produced significantly higher peak aortic enhancement (median, 929.8, 772.5, and 476.2 HU with short, standard, and long injection, respectively, p < 0.001) (Figs 2 and 4a). The short injection exhibited a significantly shorter aortic bolus length compared to standard and long injection (median, 4.6, 6.3, 6.5s with short, standard, and long injection, respectively, p < 0.001), while no difference was found between standard and long injection. There was no significant difference in the AUC of the aorta among the three injection protocols (median, 2831.4, 3278.4, and 3049.0 HU*s with short, standard, and long injection, respectively, p = 0.053), but the AUC tended to be lower with short injection (Fig 4b). For the peritumor artery, there was no significant difference in the peak enhancement between short and standard injection, but long injection showed significantly lower enhancement (median, 148.6, 167.3, and 106.4 HU with short, standard, and long injection, respectively, p < 0.001) (Fig 4c). Although there were no significant differences in the AUC among the three injection protocols (median, 536.6, 794.7, and 808.7 HU*s with short, standard, and long injection, respectively, p = 0.237), the AUC tended to be lower with short injection (Fig 4d) Peak tumor enhancement was not significantly different among the three injection protocols (median, 59.0, 69.6, and 56.2 HU with short, standard, and long injection, respectively, p = 0.237) (Figs 2, 4e, Supplementary video 1). Peak tumor enhancement decreased as the injection duration was shortened in 3 rabbits (Figs 2, 4e, Supplementary video 1, 2), whereas it was the highest with the standard injection and the lowest with the long injection in the other three rabbits (Fig 4e).

ARTICLE IN PRESS Academic Radiology, Vol &, No &&, && 2019

EFFECT OF GADOXETIC ACID INJECTION DURATION

Figure 3. Box plots of the enhancement ratio (ER) of the aorta in early arterial phase (N = 3) (a), ER of the tumor in late arterial phase (N = 6) (b), and tumor-liver contrast ratio (CR) in late arterial phase (N = 6) (c) on gadoxetic acid-enhanced MRI.

Figure 4. Box plots of peak enhancement (a) and the area under the curve (AUC) (b) of the aorta, peak enhancement (c) and AUC (d) of the peritumoral artery, and peak enhancement of the tumors (e) on CT (*=p < 0.05). Peak tumor enhancement of each rabbit was shown with lines.

Calculated peak blood concentration of iopromide in the aorta, peritumoral artery, and tumors were summarized in Table 2. Correlation Between Peak Aortic and Tumor Enhancement on CT

Results on linear regression analyses are summarized in Table 3. Overall, there was no correlation between peak aortic and peak tumor enhancement (R2 = 0.14). However, when the correlation was analyzed for each injection duration individually, the correlation became stronger with longer

injection (R2 = 0.22, 0.62, and 0.79 for short, standard, and long injection, respectively). The regression lines showed steeper slopes with longer injections (Fig 5). Correlation Between Tumor ER on Late Arterial Phase MRI and Peak Tumor Enhancement in CT

Overall, there was no significant correlation between peak tumor enhancement in CT and tumor signal enhancement on late arterial phase MRI (r = 0.45, p = 0.062). However, when the correlation was analyzed for each injection duration 5

ARTICLE IN PRESS Academic Radiology, Vol &, No &&, && 2019

TSUBOYAMA ET AL

TABLE 2. Peak Blood Concentration of Contrast Media Calculated from Time-resolved CT Injection duration

Aorta Peritumoral artery Tumor

1s

3s

6s

24.53 mgI/ml (15.1 times) 3.92 mgI/ml (94.4 times) 1.56 mgI/ml (237.7 times)

20.38 mgI/ml (18.2 times) 4.41 mgI/ml (83.8 times) 1.84 mgI/ml (201.5 times)

12.56 mgI/ml (29.4 times) 2.81 mgI/ml (131.8 times) 1.48 mgI/ml (249.5 times)

Data are median. Dilution ratios of the injected contrast media (370mgI/ml) at each site were shown in parentheses.

TABLE 3. Summary of Results on Correlations Between Peak Aortic and Tumor Enhancement on CT Protocol

Slope

Intercept

R2

All injections (n = 18) Short injection (n =6) Standard injection (n =6) Long injection (n =6)

0.0339 0.0652 0.1525 0.2354

39.6639 2.8557 -44.8475 -53.3632

0.1434 0.2200 0.6174 0.7888

R2 indicates a coefficient of determination.

individually, the correlation became stronger with longer injection (r = 0.31, 0.43, and 0.86 for short, standard, and long injection, respectively). Significant correlation was found only with long injection (p = 0.549, 0.392, and 0.029 with short, standard, and long injection, respectively) (Fig 6).

DISCUSSION To investigate the contrast agent kinetics in gadoxetic acidenhanced arterial phase MRI, we performed time-resolved CT using the same injection volume and durations with those for MRI. As a result, CT revealed that shorter contrast injections significantly increased peak enhancement (i.e., peak contrast concentration) in the aorta but not in the tumors. Therefore, it

is suggested that rapid injection of gadoxetic acid could increase the contrast concentration in the aorta but not in the tumors. Our study showed that aortic enhancement in early arterial phase MRI was not changed by the injection durations, whereas the peak contrast concentration in the aorta determined on CT was significantly higher with shorter injections. This discrepancy might be caused by an acquisition timing error due to a difficulty in sampling the k-space center even with a bolus tracking technique during the narrow aortic bolus produced by the short injection (12). In fact, the 1st phase of MRI missed the appropriate timing for early arterial phase in three of six rabbits with the short injection in our study. Saturation effects at higher blood concentrations of gadoxetic acid at the aorta with short injection might also explain the lowered MR signal enhancement (11). Our study confirmed that the injection durations did not significantly affect arterial tumor enhancement in MRI as previously reported (1012,17). Surprisingly also in CT, shorter injections did not increase peak contrast concentration in the tumors despite significantly higher peak contrast concentration in the aorta. Furthermore, the short 1s-injection resulted in a lower peak contrast concentration in tumors for all rabbits compared with the standard 3s-injection. These findings suggested that rapid injection of gadoxetic acid did not increase MR signal enhancement of the tumors because

Figure 5. Scatterplots and regression lines showing correlation of peak aortic with peak tumor enhancement on CT.

6

ARTICLE IN PRESS Academic Radiology, Vol &, No &&, && 2019

Figure 6. Scatterplots and trend lines showing correlation of peak enhancement on time-resolved CT with enhancement ratio on gadoxetic acid-enhanced late arterial phase MRI for the tumors.

peak contrast concentration in tumors was not increased or even decreased with rapid injection. One reason for this decreased peak contrast concentration in tumors might be that too short aortic bolus might not be able to carry enough contrast agent to the peripheral arteries like the peritumoral arteries, and thus to the tumors. This hypothesis can be supported by the following observations in our study. First, the significantly higher aortic peak contrast concentration with the short injection was not transferred to the periphery because there was no significant difference in peak contrast concentration in peritumoral arteries between short and standard injection durations. Second, the AUC for the time-enhancement curves of the aorta and peritumoral artery tended to be lower with short injection in both the aorta and peritumoral artery. Third, we found that the correlation between peak aortic and tumor contrast concentration changed considerably depending on injection durations. The weaker correlation found for shorter injection durations suggests that the higher aortic peak contrast concentration is not associated with higher peak contrast concentrations in the tumor. This finding can also explain the discrepancy between gadoxetic acid-enhanced MRI and clinical dynamic liver CT in the effect of injection durations. Since injection duration is much longer for CT than that for MRI, higher aortic enhancement with rapid injection can be associated with higher tumor enhancement in CT. However, even in clinical CT, one previous study found that organ enhancement in the arterial phase was superior with 35s- than with 25s-injection while aortic enhancement was superior with 25s- than with 35s-injection (27). A sufficient length of the aortic bolus seems to be one important factor to obtain high enhancement for solid tissues like organs and tumors. In our CT analysis, long 6s-injection led to variable findings. The peak contrast concentration in tumors was the highest in three of six rabbits and the lowest in the other 3 rabbits with long injection. The long injection provided a prolonged aortic bolus which might be an important factor to carry much contrast agent to the tumors as mentioned above but on the other hand suffered decreased peak aortic

EFFECT OF GADOXETIC ACID INJECTION DURATION

contrast concentration which could be disadvantageous to tumor enhancement. The inconsistent effect of the long injection might be occurred by a balance between the advantage and the disadvantage of the long injection under different hemodynamic status of the rabbits or a different grade of tumor vascularity. Our study suggests that the standard injection can provide the most stable tumor enhancement. Shortening injection durations, i.e., increasing injection rates, has no merit because it does not increase peak tumor enhancement. It also might increase the risk of timing errors when conventional single arterial phase is applied because the appropriate late arterial phase was variably achieved during our double arterial phase MRI scan. The reduced correlation between peak contrast concentration in CT and MR signal enhancement in tumors with rapid injection might also be caused by the unstable timing for tumor enhancement. Prolonging injection durations, i.e., decreasing injection rates, carries a chance to increase tumor enhancement as well as a risk to decrease tumor enhancement, which results in a higher variability. There are several limitations in our study. First, the sample size is small (N = 6). Second, our animal study was different from the clinical situation in humans in many ways such as smaller injected contrast volume, faster circulation in rabbits, and different MRI and CT scan parameters. However, we applied the same scan method for MRI as in a previous study (12), and we used only appropriately timed images for our analysis. Third, since we diluted gadoxetic acid to expand the volume 5-fold to adjust for small animals, the effect of the dilution must to be considered. Although some previous studies showed improved aortic enhancement or lesion-liver contrast by dilution of gadoxetic acid, the suboptimal high injection rate (3 ml/s) applied for the undiluted control group might explain their results (10,28). The dilution method compared with appropriate injection (1 ml/s) of undiluted gadoxetic acid reduced artifacts (29) but did not show any advantages in enhancement effects (10,30). Lastly, since we used iopromideenhanced time-resolved CT to estimate the gadolinium concentration as previously reported (31), differences in properties of gadoxetic acid and iopromide need to be considered. Molecular weight of gadoxetic acid (726 Da) and iopromide (791 Da) are equivalently low. Gadoxetic acid has weak protein binding, which may impact the distribution of gadoxetic acid from vascular to the interstitial spaces. Viscosity is much higher in iopromide (9.5 mPa.s at 37°C) than gadoxetic acid (1.2 mPa.s at 37°C) (32). Given the lower viscosity of pure water (0.7 mPa¢s at 37°C) (33), the dilution might further decrease the viscosity of gadoxetic acid in our study. Those differences in viscosity may have biased the local concentrations and enhancement levels between the two agents. Behrendt et al. compared vascular enhancement in thoracic area using an identical iodine delivery rate with different contrast concentrations (viscosity ranged from 1.5 to 9.5 mPa.s). Although the results are not straightforward, that is, contrast media with medium viscosity showed higher enhancement than those with low or high viscosity, viscosity was suggested to have some 7

ARTICLE IN PRESS TSUBOYAMA ET AL

influence on enhancement effects (32). In contrast, Sandstede et al. found no difference in abdominal organ enhancement in arterial phase by using different contrast concentrations, suggesting that viscosity might affect little in the abdominal area because contrast media are considerably diluted when they reach abdomen as shown in our study (34). Nonetheless, direct comparison of enhancement effects between CT and MRI is difficult because relationship between gadolinium concentration and MR signal is nonlinear (19) whereas that between iodine concentration and CT density is linear. In conclusion, estimation of contrast concentration by CT demonstrated that shorter injection durations did not increase peak contrast concentration in tumors despite increased peak concentration in the aorta. Furthermore, tumor signal enhancement in gadoxetic acid-enhanced arterial phase MRI was less correlated with the peak contrast concentration with shorter injections. REFERENCES 1. Chen L, Zhang L, Liang M, et al. Magnetic resonance imaging with gadoxetic acid disodium for the detection of hepatocellular carcinoma: a meta-analysis of 18 studies. Acad Radiol 2014; 21:1603–1613. 2. van Lunenburg JTJ, Tripathi V, Chan VSH, et al. Sequence and observer variability in gadoxetic acid-enhanced MRI lesion measurement in hepatocellular carcinoma. Acad Radiol, doi: 0.1016/j.acra.2019.05.021[Epub ahead of print]. 3. Elsayes KM, Hooker JC, Agrons MM, et al. 2017 version of LI-RADS for CT and MR imaging: an update. Radiographics 2017; 37:1994–2017. 4. Sofue K, Marin D, Jaffe TA, et al. Can combining triple-arterial phase acquisition with fluoroscopic triggering provide both optimal early and late hepatic arterial phase images during gadoxetic acid-enhanced MRI? J Magn Reson Imaging 2016; 43:1073–1081. 5. Park YS, Lee CH, Kim JW, et al. Application of high-speed T1 sequences for high-quality hepatic arterial phase magnetic resonance imaging: intraindividual comparison of single and multiple arterial phases. Invest Radiol 2017; 52:605–611. 6. Goshima S, Kanematsu M, Kondo H, et al. Evaluation of optimal scan delay for gadoxetate disodium-enhanced hepatic arterial phase MRI using MR fluoroscopic triggering and slow injection technique. AJR Am J Roentgenol 2013; 201:578–582. 7. Haradome H, Grazioli L, Tsunoo M, et al. Can MR fluoroscopic triggering technique and slow rate injection provide appropriate arterial phase images with reducing artifacts on gadoxetic acid-DTPA (Gd-EOB-DTPA)-enhanced hepatic MR imaging? J Magn Reson Imaging 2010; 32:334–340. 8. Huh J, Kim SY, Yeh BM, et al. Troubleshooting arterial-phase MR images of gadoxetate disodium-enhanced liver. Korean J Radiol 2015; 16:1207–1215. 9. Zech CJ, Bartolozzi C, Bioulac-Sage P, et al. Consensus report of the fifth international forum for liver MRI. AJR Am J Roentgenol 2013; 201:97–107. 10. Tamada T, Ito K, Yoshida K, et al. Comparison of three different injection methods for arterial phase of Gd-EOB-DTPA enhanced MR imaging of the liver. Eur J Radiol 2011; 80:e284–e288. 11. Chung SH, Kim MJ, Choi JY, et al. Comparison of two different injection rates of gadoxetic acid for arterial phase MRI of the liver. J Magn Reson Imaging 2010; 31:365–372. 12. Tsuboyama T, Jost G, Kim T, et al. Experimental studies on artifacts and tumor enhancement on gadoxetic acid-enhanced arterial phase liver MRI. Acta Radiol 2018; 59:1029–1037. 13. Bae KT. Intravenous contrast medium administration and scan timing at CT: considerations and approaches. Radiology 2010; 256:32–61. 14. Kim T, Murakami T, Takahashi S, et al. Effects of injection rates of contrast material on arterial phase hepatic CT. AJR Am J Roentgenol 1998; 171:429–432.

8

Academic Radiology, Vol &, No &&, && 2019

15. 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; 25:278–288. 16. Kagawa Y, Okada M, Kumano S, et al. Optimal scanning protocol of arterial dominant phase for hypervascular hepatocellular carcinoma with gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid-enhanced MR. J Magn Reson Imaging 2011; 33:864–872. 17. Kim SM, Heo SH, Kim JW, et al. Hepatic arterial phase on gadoxetic acid-enhanced liver MR imaging: a randomized comparison of 0.5 mL/s and 1 mL/s injection rates. Korean J Radiol 2014; 15:605–612. 18. Zech CJ, Vos B, Nordell A, et al. Vascular enhancement in early dynamic liver MR imaging in an animal model: comparison of two injection regimen and two different doses Gd-EOB-DTPA (gadoxetic acid) with standard Gd-DTPA. Invest Radiol 2009; 44:305–310. 19. Bokacheva L, Rusinek H, Chen Q, et al. Quantitative determination of Gd-DTPA concentration in T1-weighted MR renography studies. Magn Reson Med 2007; 57:1012–1018. 20. Rohrer M, Bauer H, Mintorovitch J, et al. Comparison of magnetic properties of MRI contrast media solutions at different magnetic field strengths. Invest Radiol 2005; 40:715–724. 21. Hadizadeh DR, Jost G, Pietsch H, et al. Intraindividual quantitative and qualitative comparison of gadopentetate dimeglumine and gadobutrol in time-resolved contrast-enhanced 4-dimensional magnetic resonance angiography in minipigs. Invest Radiol 2014; 49:457–464. 22. Fischer MA, Leidner B, Kartalis N, et al. Time-resolved computed tomography of the liver: retrospective, multi-phase image reconstruction derived from volumetric perfusion imaging. Eur Radiol 2014; 24:151–161. 23. Sommer WH, Becker CR, Haack M, et al. Time-resolved CT angiography for the detection and classification of endoleaks. Radiology 2012; 263:917–926. 24. Jost G, Pietsch H, Grenacher L. Dynamic contrast-enhanced computed tomography to assess antitumor treatment effects. Invest Radiol 2013; 48:715–721. 25. Lell MM, Jost G, Korporaal JG, et al. Optimizing contrast media injection protocols in state-of-the art computed tomographic angiography. Invest Radiol 2015; 50:161–167. 26. Conover WJ. Practical nonparametric statistics. 3rd edition New York, NY: John Wiley & Sons, 1999. 27. Tsuge Y, Kanematsu M, Goshima S, et al. Abdominal vascular and visceral parenchymal contrast enhancement in MDCT: effects of injection duration. Eur J Radiol 2011; 80:259–264. 28. Motosugi U, Ichikawa T, Sou H, et al. Dilution method of gadolinium ethoxybenzyl diethylenetriaminepentaacetic acid (Gd-EOB-DTPA)enhanced magnetic resonance imaging (MRI). J Magn Reson Imaging 2009; 30:849–854. 29. Kim YK, Lin WC, Sung K, et al. Reducing artifacts during arterial phase of gadoxetate disodium-enhanced MR imaging: dilution method versus reduced injection rate. Radiology 2017; 283:429–437. 30. Iyama A, Nakaura T, Iyama Y, et al. Spiral flow-generating tube for saline chaser improves aortic enhancement in Gd-EOB-DTPA-enhanced hepatic MRI. Eur Radiol 2019; 29:2009–2016. 31. Ishida M, Sakuma H, Murashima S, et al. Absolute blood contrast concentration and blood signal saturation on myocardial perfusion MRI: estimation from CT data. J Magn Reson Imaging 2009; 29:205–210. 32. Behrendt FF, Pietsch H, Jost G, et al. Identification of the iodine concentration that yields the highest intravascular enhancement in MDCT angiography. AJR Am J Roentgenol 2013; 200:1151–1156. 33. Smedby O. Viscosity of some contemporary contrast media before and after mixing with whole blood. Acta Radiol 1992; 33:600–605. 34. Sandstede JJ, Werner A, Kaupert C, et al. A prospective study comparing different iodine concentrations for triphasic multidetector row CT of the upper abdomen. Eur J Radiol 2006; 60:95–99.

SUPPLEMENTARY MATERIALS Supplementary material associated with this article can be found in the online version at doi:10.1016/j.acra.2019.10.018.