CLINICAL STUDY
Hepatic Blood Volume Imaging with the Use of Flat-Detector CT Perfusion in the Angiography Suite: Comparison with Results of Conventional Multislice CT Perfusion Zhi-guo Zhuang, PhD, Xue-bin Zhang, PhD, Jing-feng Han, PhD, Janina Beilner, MD, Yu Deuerling-Zheng, MSc, Jia-chang Chi, MD, Ji Wang, MD, Li-jun Qian, MD, Yan Zhou, PhD, and Jian-rong Xu, MD
ABSTRACT Purpose: To prospectively determine the feasibility of flat-detector (FD) computed tomography (CT) perfusion to measure hepatic blood volume (BV) in the angiography suite in patients with hepatocellular carcinoma (HCC). Materials and Methods: Twenty patients with HCC were investigated with conventional multislice and FD CT perfusion. CT perfusion was carried out on a multislice CT scanner, and FD CT perfusion was performed on a C-arm angiographic system, before transarterial chemoembolization procedures. BV values of conventional and FD CT perfusion were measured within tumors and liver parenchyma. The arterial perfusion portion of CT perfusion BV was extracted from CT perfusion BV by multiplying it by a hepatic perfusion index. Relative values (RVs) for CT perfusion arterial BV and FD CT perfusion BV (FD BV) were defined by dividing BV of tumor by BV of parenchyma. Relationships between BV and RV values of these two techniques were analyzed. Results: In all patients, both perfusion procedures were technically successful, and all 33 HCCs larger than 10 mm were identified with both imaging methods. There were strong correlations between the absolute values of FD BV and CT perfusion arterial BV (tumor, r ¼ 0.903; parenchyma, r ¼ 0.920; both P o .001). Bland–Altman analysis showed a mean difference of 0.15 ⫾ 0.24 between RVs for CT perfusion arterial BV and FD BV. Conclusions: The feasibility of FD CT perfusion to assess BV values of liver tumor and surrounding parenchyma in the angiographic suite was demonstrated.
ABBREVIATIONS ALP = arterial liver perfusion, BV = blood volume, DSA = digital subtraction angiography, FD = flat detector, FD BV = blood volume on flat-panel CT perfusion, HCC = hepatocellular carcinoma, ICC = intraclass correlation coefficient, ROI = region of interest, RV = relative value
From the Department of Radiology (Z.Z., X.Z., J.C., J.W., L.Q., Y.Z., J.X.), Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University; Healthcare Sector (J.H., J.B.), Siemens Limited China, Shanghai, China; and Healthcare Sector (Y.D.-Z.), Siemens AG, Forchheim, Germany. Received October 12, 2013; final revision received January 17, 2014; accepted January 18, 2014. Address correspondence to: J.X., No. 160, Pujian Road, Pudong New District, Shanghai 200127, China; E-mail: [email protected] Z.Z. and X.Z. contributed equally to this work. This work was supported by National Natural Science Foundation of China Grant 81201172 (to Z.Z.), Medical and Engineering Program of Shanghai Jiaotong University Grant YG2012MS16 (to L.Q.), and National Natural Science Foundation of China Grant 81171325 (to Y.Z.). J.H. and J.B. are employees of Siemens Limited China (Shanghai, China). Y.D.-Z. is an employee of Siemens AG (Forchheim, Germany). None of the other authors have identified a conflict of interest. & SIR, 2014 J Vasc Interv Radiol 2014; 25:739–746 http://dx.doi.org/10.1016/j.jvir.2014.01.021
Because most liver diseases lead to significant changes in hepatic microcirculation regionally, globally, or both, quantification of hepatic perfusion can improve the assessment and management of liver diseases (1,2). In recent years, various imaging techniques, such as computed tomography (CT), xenon-enhanced CT, isotope scintigraphy, magnetic resonance (MR) imaging, and ultrasonography, as well as positron emission tomography with the use of oxygen-15–labeled water, have been applied in the evaluation of hepatic perfusion (1–3). Many studies have demonstrated the capacity of CT perfusion imaging to evaluate hepatic tissue blood flow, blood volume (BV), and hepatic perfusion index (HPI) (2,4–6). BV in CT perfusion imaging is reportedly useful for assessment of liver tumor vascularity and
740
’
Hepatic Blood Volume Imaging: Flat Detector vs Conventional CT
angiogenesis (4,7), prediction of tumor response to therapies (8), quantification of hemodynamic alterations in liver transplantation (9), early response assessment, and early detection of recurred tumors after radiologic interventions (10–12). Liver functional imaging during interventional treatment has been rarely available. With the development of flat-detector (FD) technology, FD-equipped angiographic systems are increasingly used in daily routine. They can offer high-quality three-dimensional digital subtraction angiography (DSA) images and soft-tissue cross-sectional images in the angiography suite with FD CT (13). Nowadays, there is evidence that FD CT could also provide a means to generate functional images. Studies have proved the feasibility of cerebral FD CT perfusion and verified the accuracy of cerebral BV acquired by FD CT perfusion compared with conventional CT perfusion (14,15). The ability of immediate measurement of cerebral BV in an angiography suite has a great potential to save time from initial assessment to intervention and enhance the management of patients with ischemic strokes (14,16,17). Based on the successful application of FD CT perfusion in neurologic interventions, it is reasonable to propose that hepatic BV might be measured by FD CT perfusion. In the present study, our purpose was to prospectively determine the feasibility of measuring the BV values of liver parenchyma or tumor tissues in patients with hepatocellular carcinoma (HCC) by using FD CT perfusion in the angiography suite.
Zhuang et al
’
JVIR
Twenty patients satisfying our inclusion/exclusion criteria were consecutively enrolled in this prospective study. The patients’ demographic data are listed in Table 1. The diagnosis of HCC was verified by performing a needle biopsy and histopathologic examination of resected specimen, or on the basis of the presence of a tumor larger than 1 cm in diameter with typical imaging findings in the setting of cirrhosis (18). The study was approved by the ethics committee at our institution. Written informed consent was obtained from all patients before their participation in the study.
CT Perfusion Technique With a 128-row CT scanner (SOMATOM Definition ASþ; Siemens, Forchheim, Germany), CT perfusion examinations were carried out at least 2 days after diagnostic CT. All patients fasted for 8 hours and rested calmly for 30 minutes before the CT perfusion examination to minimize physiologic variations in hepatic perfusion. After a preliminary unenhanced localizer scan covering the epigastric region, a scanning range of 175 mm covering almost the entire liver as well as the spleen was chosen. Dynamic CT was performed with the following parameters: 150 mAs, 80 kV, rotation time of 0.30 seconds, pitch of 0.6, slice acquisition of 128 0.6 mm and 5-mm reconstructed section thickness, scan time of 1.8 seconds, cycle time of 1.75 seconds, and examination time of 48.67 seconds. Dynamic scanning was initiated after a 6-second delay from the beginning Table 1 . Patients’ Demographic Data (N ¼ 20)
MATERIALS AND METHODS Patients From June 2012 to December 2012, 113 patients with hepatic cirrhosis with HCCs referred to our department for transarterial chemoembolization. Patients included in this study satisfied all of the following criteria: the patients (i) were older than 18 years; (ii) had Child– Pugh class A or B disease; (iii) had no contraindications to iodinated contrast media; and (iv) were able to hold their breath for 22 seconds for FD CT perfusion and maintain shallow breathing for approximately 50 seconds in CT perfusion examination after training. Excluded from the study were patients with (i) Child– Pugh class C disease; (ii) uncorrectable coagulopathy (International Normalized Ratio 4 1.5); (iii) portal vein thrombosis; (iv) a total bilirubin level higher than 4.0 mg/dL; (v) a serum creatinine level higher than 1.7 mg/ dL; (vi) thrombocytopenia (platelet count o 50,000/μL); (vii) previous therapeutic procedures; and (viii) hepatic arterial variants. For each patient, a routine diagnostic four-phase CT was performed. The hepatic arterial anatomy was determined by CT angiography reconstructed from arterial-phase images of CT.
Characteristics
No. of Patients
Sex Male
13
Female Age*
7
o 60 y
12
Z 60 y No. of tumors 1
8 13
2–3 43
6 1
Tumor size o 3 cm 3–5 cm
5 7
4 5 cm
8
Child–Pugh class A
15
B Cause of cirrhosis Hepatitis B
5 13
Hepatitis C
2
Hepatitis B and C Alcohol-related
1 4
*Mean age was 59.6 y ⫾ 12.2 y (standard deviation).
Volume 25
’
Number 5
’
May
’
2014
of the injection of 40 mL nonionic contrast medium (Ultravist 370 mgI/mL; Schering, Berlin, Germany) at the rate of 5.0 mL/s, followed by 40 mL saline solution flush at the same rate, through an 18-gauge intravenous cannula inserted into the antecubital vein with an automatic dual-headed injector (OptiVantage DH; Mallinckrodt, St. Louis, Missouri). Before imaging, all patients were instructed to keep breathing shallowly during CT perfusion. In addition, a band compressing the abdomen that limited breath-related liver excursions was used. The estimated effective radiation dose of CT perfusion, calculated by multiplying the dose–length product, obtained from the patient protocol, with a conversion coefficient for the abdomen (k ¼ 0.015 mSv mGy1 cm1), was 28.7 mSv. The data were processed on a commercially available workstation (syngo MultiModality Workplace VE40; Siemens) with CT perfusion software (syngo VPCT Body; Siemens). If respiratory motion–associated misregistrations were detected, the motion was compensated manually by using a rigid registration technique. Regions of interest (ROIs) were placed on the abdominal aorta at the level of the celiac axis (range, 52–105 mm2), the main portal vein (range, 24–124 mm2), and the spleen (range, 2,650–5,769 mm2) to generate time– density curves. Then, functional color map images of each perfusion parameter of the liver were generated. The mathematical technique has previously been fully described elsewhere (19). To calculate BV values of liver tumors and parenchyma, ROIs for tumors (range, 86– 7,863 mm2) were manually drawn within the tumor section with the maximum cross-sectional area on the CT perfusion map of BV. In the presence of multiple tumors, ROIs were drawn for all tumors larger than 10 mm, whereas tumors smaller than 10 mm in diameter were excluded from evaluation. Similarly, two ROIs were drawn in the background liver parenchyma (range,
741
130–1,255 mm2) in each patient. ROIs of the background liver were drawn as far away as possible from the tumor and were kept distant from large intrahepatic vessels and the borders of the organ (Fig 1a).
FD CT Perfusion Protocol All FD CT perfusion procedures were performed by two experienced interventional radiologists 3 days after CT perfusion on a robotic angiographic system (Artis zeego VC13; Siemens) during transarterial chemoembolization. According to previous CT angiography, which was used to assess the morphology of the celiac artery and superior mesenteric artery, a 5-F RH catheter (Terumo, Tokyo, Japan) or 3-F coaxial microcatheter (Terumo) was inserted into the proper hepatic artery. The FD CT perfusion acquisition consisted of two rotations: an initial rotation (mask run) followed by a second rotation after contrast medium injection (fill run). The data acquisition was carried out by using the following parameters: acquisition time of 5 seconds, examination time of 22 seconds, 90 kV, 512 512 matrix, projection on 30 40 cm flat panel size, 2001 total angle, 0.81 per frame, 250 frames total, dose of 0.36 μGy per frame. A total of 36 mL of contrast medium (Ultravist 370 mgI/ mL; Schering) diluted to 25% was injected by power injector (Angiomat 6000; Liebel-Flarsheim, Cincinnati, Ohio) at 3 mL/s after the mask run. With a 7-second xray delay, the fill run was then performed. The time delay was adopted to perfuse the liver in a steady state for the BV acquisition. After the FD CT perfusion acquisition, the chemoembolization procedure was performed as routinely done.
Postprocessing of FD BV Imaging FD BV postprocessing was performed by using prototype software (Siemens) installed on a research workstation (syngo X-Workplace VB15; Siemens). The
Figure 1. CT perfusion BV (a) and FD BV (b) maps from the same patient show conformity of the images. Positioning of the ROIs is displayed: 1, tumor; 2, liver tissue.
742
’
Hepatic Blood Volume Imaging: Flat Detector vs Conventional CT
software adopted the same postprocessing workflow as previously described (14). The mask and the fill run were reconstructed and subtracted. The motion between the two runs was corrected by the nonrigid registration algorithm. The steady-state arterial input function value was calculated from an automated histogram analysis of the vessel tree. A final scaling was then applied to the dataset to account for the arterial input value. In the end, a smoothing filter was applied to image to reduce pixel noise. The matching positions, both tumor and liver tissue ROIs with similar sizes to CT perfusion maps (range, 89–7,871 mm2 in tumors and 135–1,273 mm2 in liver tissue), were selected, and the BV values were measured from FD BV maps (Fig 1b).
Data Analysis and Statistics Detection of tumors in CT perfusion and FD CT perfusion maps were performed by two independent radiologists with 5 and 6 years of experience in reading perfusion images. If matching lesions were found in both images, the two radiologists were required to reach consensus to select the corresponding section and ROIs to measure CT perfusion BV and FD BV values. Each radiologist was blinded to the values of the other, and the outcomes of both radiologists were recorded by another assistant. It is known that there exists dual blood supply for the liver: hepatic arterial and portal vein supply. Pure hepatic arterial blood perfusion could be distinguished from portal vein supply by measuring the dynamic change in contrast medium perfusion of the liver after injection via the hepatic artery (20). In the present study, CT perfusion BV represented the total BV supplied by the hepatic artery and portal vein because the contrast medium was injected intravenously, whereas FD BV was solely arterial BV because it was performed by administering contrast medium intraarterially. To make the comparison of CT perfusion BV and FD BV more precise, we extracted the arterial perfusion portion from the total liver BV in CT perfusion examinations as follows. From the postprocessing results of CT perfusion, the following parameters can be obtained: BV, blood flow, arterial liver perfusion (ALP), portal vein perfusion, and HPI. The blood perfusion from the hepatic artery and portal vein can be distinguished by HPI, which represents the arterial portion to the total liver perfusion. For this reason, we assumed that the arterial BV values measured from CT perfusion could be obtained by multiplying CT perfusion BV by HPI. Values are presented as mean ⫾ standard deviation. Differences of BV values between the two examinations were analyzed by paired t test. Differences of BV values between tumor and liver parenchyma in the same examination were analyzed by the independent-sample t test. Pearson correlation coefficients were calculated between CT perfusion arterial BV and FD BV for all
Zhuang et al
’
JVIR
liver parenchyma and tumor tissues, respectively. Concordance between the measurements of CT perfusion arterial BV or FD BV in parenchyma from the two ROIs was assessed by using intraclass correlation coefficients (ICC). In the present study, the relative value (RV) between liver tumor and parenchyma for CT perfusion arterial BV was defined as CT perfusion arterial BV of tumor divided by CT perfusion arterial BV of parenchyma. The RV between liver tumor and parenchyma for FD BV was defined as FD BV of tumor divided by FD BV of parenchyma. Agreement between RVs for CT perfusion arterial BV and FD BV was assessed by using the approach of Bland and Altman (21).
RESULTS FD BV and CT perfusion BV acquisitions were successfully performed in all patients. CT perfusion BV and FD BV maps and data quality were suitable for evaluation in all patients. On both BV maps, liver parenchyma and tumor lesions can be recognized in the corresponding positions (Fig 1). In the CT perfusion BV maps, 33 proven HCC lesions larger than 10 mm were recognized in all 20 patients. FD BV findings of tumor showed oneto-one correlation with CT perfusion in all patients. It is worth mentioning that FD BV maps showed more hypervascularized regions smaller than 10 mm than CT perfusion BV maps in four patients (Fig 2). Because lesions smaller than 10 mm were difficult to diagnose qualitatively and quantitatively in the present study, these small hypervascularized regions were not further analyzed. Finally, BV values in 33 tumor spots and 40 liver parenchyma areas were measured. A total of 12 patients had FD CT perfusion with a 5-F catheter, compared with eight with a 3-F microcatheter. There were no significant differences between the two groups in FD BV values (P ¼ .415 and P ¼ .850 for tumor and liver parenchyma, respectively; Table 2). There were significant differences between CT perfusion arterial BV and FD BV in tumor and liver parenchyma (all P o .001). In addition, HPI, CT perfusion BV, CT perfusion arterial BV, and FD BV in tumor were significantly higher than in parenchyma (all P o .001; Table 3). There were strong correlations between the absolute values of FD BV and CT perfusion arterial BV (r ¼ 0.903, P o .001 for tumor [Fig 3a]; and r ¼ 0.920, P o .001 for liver parenchyma [Fig 3b]). The ICC of the measurements obtained from the two ROIs in liver parenchyma were 0.96 for CT perfusion arterial BV and 0.95 for FD BV. Because there was a close correlation between the two ROIs, the two measurements were averaged as the CT perfusion arterial BV or FD BV of parenchyma in each patient for further analysis of RV. The Bland–Altman plots of RV for CT perfusion arterial BV and FD BV are displayed in
Volume 25
’
Number 5
’
May
’
2014
743
Figure 2. CT perfusion BV (a) and FD BV (b) maps from a patient with multifocal lesions. The two BV maps match closely except that an additional small arterial hypervascularized region (arrow) was found on the FD BV map.
Table 2 . Differences in BV Values of Flat-Detector CT Perfusion between Patients with a 5-F Catheter and a 3-F Microcatheter BV Measurement Tumor (mL/100 g) Liver parenchyma (mL/100 g)
Catheter (n ¼ 12)
Microcatheter (N ¼ 8)
P Value
25.2 ⫾ 5.1
26.8 ⫾ 6.0
.415
4.9 ⫾ 1.2
4.7 ⫾ 1.1
.850
BV ¼ blood volume.
Table 3 . Parameters of Perfusion Generated by CT Perfusion and FD CT Perfusion Liver Parenchyma
P Value
HPI (%)
90.5 ⫾ 10.7
25.1 ⫾ 6.2
o .001
CT perfusion BV (mL/100 g) CT perfusion arterial BV (mL/100 g)
16.5 ⫾ 2.5 15.0 ⫾ 3.1
11.1 ⫾ 1.2 2.8 ⫾ 0.8
o .001 o .001
25.9 ⫾ 5.6
4.8 ⫾ 1.2
o .001
10.9 ⫾ 3.1
2.0 ⫾ 0.6
Parameter
FD BV (mL/100 g) Difference: CT perfusion arterial BV vs FD BV (mL/100 g)
Tumor
NA
Values presented as means ⫾ standard deviation. BV ¼ blood volume, FD ¼ flat detector, HPI ¼ hepatic perfusion index, NA ¼ not applicable.
Figure 4. This analysis showed that the mean difference in RV between CT perfusion arterial BV and FD BV was 0.15 ⫾ 0.24, meaning that RV for CT perfusion arterial BV was slightly lower than that for FD BV.
DISCUSSION Previous studies have validated that FD CT perfusion could provide a means to measure the cerebral BV at an accuracy comparable to that of CT perfusion (14–16). However, this technique has not been applied to measure liver perfusion. In the present study, compared with the data from CT perfusion imaging, we showed that FD CT perfusion has potential to be used to monitor BV within the targeted tumors and adjacent liver tissue during transarterial chemoembolization in the angiography suite. According to previous CT perfusion studies
(4,7,22–25), BV values ranged from 3.2 to 11.7 mL/100 g in cirrhotic liver parenchyma and from 4.9 to 24.0 mL/ 100 g in HCC. HPI values ranged from 10.6% to 65% in cirrhotic liver parenchyma and from 75.3% to 92.5% in HCC. Our results are therefore within these reported ranges. On the contrary, our methods of CT perfusion and FD CT perfusion were also proven to be reliable by the excellent ICCs of CT perfusion arterial BV and FD BV values obtained from the two separate ROIs in liver parenchyma. In the present study, FD BV had a very good correlation with CT perfusion arterial BV. In addition, the Bland–Altman plot showed good agreement in RV between CT perfusion arterial BV and FD BV, indicating that the BV values obtained from FD CT perfusion were believable, and there were similar capacities to display HCC between the two techniques. However, the absolute values of FD BV were higher than CT
744
’
Hepatic Blood Volume Imaging: Flat Detector vs Conventional CT
Zhuang et al
’
JVIR
Figure 3. Correlation between BV values from FD CT perfusion and CT perfusion imaging. (a) Scatter plots of FD BV and CT perfusion arterial BV in tumor (r ¼ 0.903, P o .001). (b) Scatter plots of FD BV and CT perfusion arterial BV in liver parenchyma (r ¼ 0.920, P o .001).
Figure 4. Bland–Altman plots representing the degree of agreement in RVs between CT perfusion arterial BV and FD BV. The Bland–Altman interval was [0.63 (0.15) 0.33].
perfusion arterial BV values. Possible reasons for this phenomenon might be different injection protocols and different algorithms adopted for the two methods. Direct hepatic arterial injection might cause higher BV than intravenous injection because the vascular bed might have been further opened when the arterial injection protocol with additional pressure was applied, whereas CT perfusion BV involves only the use of conventional intravenous injections. In addition, tumor spots were often highly vascularized and supplied mainly by the hepatic artery, and therefore their FD BV values were even higher than on CT perfusion BV. Certainly, in liver parenchyma, because of dominant portal venous source in dual blood supply, FD BV was lower than CT perfusion BV.
There is strong evidence that CT perfusion imaging may be valuable in monitoring the outcome of transarterial chemoembolization in cirrhotic patients with HCC (11,12). Ippolito et al (11) have shown that, in patients with HCC treated with transarterial chemoembolization, CT perfusion BV showed a tendency to be greater in the relapse site than in the primary lesion area of iodized oil deposition. In an animal study, Choi et al (10) made similar observations with CT perfusion in VX2 tumor rabbit models. Similarly, Chen et al (12) also demonstrated that changes in CT perfusion BV of viable tumors are associated with different responses of HCC to transarterial chemoembolization. As mentioned earlier, much work has been done regarding the value of conventional CT perfusion for chemoembolization in HCC management. However, little was known about intrachemoembolization physiologic monitoring. Although some recent literature (26–28) has demonstrated the use of transcatheter intraarterial perfusion MR imaging to detect intraprocedural perfusion changes of HCC and adjacent surrounding liver tissue during transarterial chemoembolization, a special and dedicated Miyabi system (a conventional DSA unit integrated with an MR imaging system) was needed for this technique (26–28). Nowadays, the FD CT perfusion technology can also provide perfusion examination within the catheterization laboratory intraprocedurally. It might offer information to determine the optimal endpoint of chemoembolization for a better prognosis in the future. FD CT perfusion has several other potential advantages. First, FD CT perfusion enables direct monitoring of local arterial BV of liver tumors during transarterial chemoembolization. One potential application is the replanning of catheter position in the angiographic suite. Prechemoembolization FD CT perfusion could be used
Volume 25
’
Number 5
’
May
’
2014
to determine if the targeted hepatic artery is supplying the tumor, and hence decide whether the initial catheter position needs to be replanned. Similar transcatheter intraarterial perfusion MR imaging has already been shown to alter the initial catheter position anticipated based on conventional DSA results in 40%–50% of patients during transcatheter HCC therapy (27,28). Second, in the present study, compared with CT perfusion BV, more hypervascularized regions smaller than 10 mm were detected on the FD BV maps. FD BV imaging may enhance tumor conspicuity in HCC because of the increased perfusion at the tumor rims, which makes liver tumors appear larger, thereby facilitating the detection of smaller lesions when contrast material is injected into the hepatic artery directly (29). Further studies are needed to prove whether these small hypervascularized regions are HCC lesions that were not detected by CT perfusion. Finally, the localized delivery of smaller doses of contrast material during FD CT perfusion scanning than in CT perfusion imaging could mitigate potential kidney injury. The radiation dose of our CT perfusion protocol was similar to that reported by Goetti et al (19). However, the radiation exposure was not available for FD CT perfusion in this study. Although Fiorella et al (30) reported that the FD CT perfusion dose was similar to that of standard CT perfusion when estimating cerebral BV, the exact effective radiation dose was not obtained in their study. We speculate that, in the present study, the FD CT perfusion radiation exposure may be lower than with CT perfusion as a result of the short scan time. The present study has several limitations. First, compared with CT perfusion, which can provide HPI, ALP, portal vein perfusion, and BV, only BV can be assessed with FD CT perfusion because of mechanical limitations. Ippolito et al (24) found that ALP and HPI were likely to be the most relevant parameters to assess neovascularization induced by tumor growth. However, BV could be affected by intrahepatic arterioportal shunts. Second, this technique cannot be used to assess perfusion values that result from the collateral arterial supply or the portal vein. Third, we assessed perfusion at only one time point during transarterial chemoembolization, namely before embolization. Future studies need to be performed to determine perfusion changes before and after chemoembolization. Consequently, the relationship between perfusion change and tumor response should be investigated. Fourth, the total radiation dose received by the patients may be high at present. In our future studies, low-dose perfusion imaging should be performed. Finally, we extracted the arterial perfusion portion from the total liver BV on CT perfusion by using HPI, which represented the percentage of arterial perfusion in the total liver blood flow. This assumption has not yet been validated fully. Further studies are needed to evaluate the relationship between HPI and the arterial contribution to total liver BV.
745
In summary, it was demonstrated in a series of patients with HCC that BV mapping with the use of FD CT perfusion is feasible and shows good correlation with that obtained with standard CT perfusion techniques. The good agreement between RVs for CT perfusion arterial BV and FD BV indicated that FD BV was comparable with CT perfusion arterial BV. This technique offers the ability to measure objective, quantitative changes in perfusion during transarterial chemoembolization therapy.
REFERENCES 1. Kanda T, Yoshikawa T, Ohno Y, et al. CT hepatic perfusion measurement: comparison of three analytic methods. Eur J Radiol 2012; 81: 2075–2079. 2. Pandharipande PV, Krinsky GA, Rusinek H, Lee VS. Perfusion imaging of the liver: current challenges and future goals. Radiology 2005; 234: 661–673. 3. Zhuang ZG, Qian LJ, Wang BX, et al. Computed tomography perfusion in living donor liver transplantation: an initial study of normal hemodynamic changes in liver grafts. Clin Transplant 2009; 23:692–699. 4. Sahani DV, Holalkere NS, Mueller PR, Zhu AX. Advanced hepatocellular carcinoma: CT perfusion of liver and tumor tissue–initial experience. Radiology 2007; 243:736–743. 5. Tsushima Y, Endo K. Portal perfusion measurement on dynamic CT in patients with liver cirrhosis. AJR Am J Roentgenol 2005; 185:813. 6. Van Beers BE, Leconte I, Materne R, Smith AM, Jamart J, Horsmans Y. Hepatic perfusion parameters in chronic liver disease: dynamic CT measurements correlated with disease severity. AJR Am J Roentgenol 2001; 176:667–673. 7. Ippolito D, Sironi S, Pozzi M, et al. Perfusion CT in cirrhotic patients with early stage hepatocellular carcinoma: assessment of tumor-related vascularization. Eur J Radiol 2010; 73:148–152. 8. Petralia G, Fazio N, Bonello L, D’Andrea G, Radice D, Bellomi M. Perfusion computed tomography in patients with hepatocellular carcinoma treated with thalidomide: initial experience. J Comput Assist Tomogr 2011; 35:195–201. 9. Qian LJ, Zhuang ZG, Cheng YF, Xia Q, Zhang JJ, Xu JR. Hemodynamic alterations in anterior segment of liver graft after right-lobe living-donor liver transplantation: computed tomography perfusion imaging findings. Abdom Imaging 2010; 35:522–527. 10. Choi SH, Chung JW, Kim HC, et al. The role of perfusion CT as a followup modality after transcatheter arterial chemoembolization: an experimental study in a rabbit model. Invest Radiol 2010; 45:427–436. 11. Ippolito D, Bonaffini PA, Ratti L, et al. Hepatocellular carcinoma treated with transarterial chemoembolization: dynamic perfusion-CT in the assessment of residual tumor. World J Gastroenterol 2010; 16: 5993–6000. 12. Chen G, Ma DQ, He W, Zhang BF, Zhao LQ. Computed tomography perfusion in evaluating the therapeutic effect of transarterial chemoembolization for hepatocellular carcinoma. World J Gastroenterol 2008; 14:5738–5743. 13. Iwazawa J, Ohue S, Hashimoto N, Abe H, Hamuro M, Mitani T. Detection of hepatocellular carcinoma: comparison of angiographic C-arm CT and MDCT. AJR Am J Roentgenol 2010; 195:882–887. 14. Ahmed AS, Zellerhoff M, Strother CM, et al. C-arm CT measurement of cerebral blood volume: an experimental study in canines. AJNR Am J Neuroradiol 2009; 30:917–922. 15. Struffert T, Deuerling-Zheng Y, Kloska S, et al. Cerebral blood volume imaging by flat detector computed tomography in comparison to conventional multislice perfusion CT. Eur Radiol 2011; 21:882–889. 16. Struffert T, Deuerling-Zheng Y, Kloska S, et al. Flat detector CT in the evaluation of brain parenchyma, intracranial vasculature, and cerebral blood volume: a pilot study in patients with acute symptoms of cerebral ischemia. AJNR Am J Neuroradiol 2010; 31:1462–1469. 17. van der Bom IM, Mehra M, Walvick RP, Chueh JY, Gounis MJ. Quantitative evaluation of C-arm CT cerebral blood volume in a canine model of ischemic stroke. AJNR Am J Neuroradiol 2012; 33:353–358.
746
’
Hepatic Blood Volume Imaging: Flat Detector vs Conventional CT
18. Bruix J, Sherman M. American Association for the Study of Liver Diseases. Management of hepatocellular carcinoma: an update. Hepatology 2011; 53:1020–1022. 19. Goetti R, Reiner CS, Knuth A, et al. Quantitative perfusion analysis of malignant liver tumors: dynamic computed tomography and contrastenhanced ultrasound. Invest Radiol 2012; 47:18–24. 20. Komemushi A, Tanigawa N, Kojima H, Kariya S, Sawada SCT. perfusion of the liver during selective hepatic arteriography: pure arterial blood perfusion of liver tumor and parenchyma. Radiat Med 2003; 21: 246–251. 21. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999; 8:135–160. 22. Ippolito D, Sironi S, Pozzi M, et al. Hepatocellular carcinoma in cirrhotic liver disease: functional computed tomography with perfusion imaging in the assessment of tumor vascularization. Acad Radiol 2008; 15:919–927. 23. Miles KA, Hayball MP, Dixon AK. Functional images of hepatic perfusion obtained with dynamic CT. Radiology 1993; 188:405–411. 24. Ippolito D, Capraro C, Casiraghi A, Cestari C, Sironi S. Quantitative assessment of tumour associated neovascularisation in patients with liver cirrhosis and hepatocellular carcinoma: role of dynamic-CT perfusion imaging. Eur Radiol 2012; 22:803–811.
Zhuang et al
’
JVIR
25. Hashimoto K, Murakami T, Dono K, et al. Assessment of the severity of liver disease and fibrotic change: the usefulness of hepatic CT perfusion imaging. Oncol Rep 2006; 16:677–683. 26. Larson AC, Wang D, Atassi B, et al. Transcatheter intraarterial perfusion: MR monitoring of chemoembolization for hepatocellular carcinoma– feasibility of initial clinical translation. Radiology 2008; 246:964–971. 27. Vogl TJ, Balzer JO, Mack MG, Bett G, Oppelt A. Hybrid MR interventional imaging system: combined MR and angiography suites with single interactive table. Feasibility study in vascular liver tumor procedures. Eur Radiol 2002; 12:1394–1400. 28. Wilson MW, Kerlan RK Jr, Fidelman NA, et al. Hepatocellular carcinoma: regional therapy with a magnetic targeted carrier bound to doxorubicin in a dual MR imaging/ conventional angiography suite–initial experience with four patients. Radiology 2004; 230:287–293. 29. Meijerink MR, van Waesberghe JH, van der Weide L, van den Tol P, Meijer S, van Kuijk C. Total-liver-volume perfusion CT using 3-D image fusion to improve detection and characterization of liver metastases. Eur Radiol 2008; 18:2345–2354. 30. Fiorella D, Turk A, Chaudry I, et al. A prospective, multicenter pilot study investigating the utility of flat detector derived parenchymal blood volume maps to estimate cerebral blood volume in stroke patients. J Neurointerv Surg 2013. http://dx.doi.org/10.1136/neurintsurg-2013-010840.
Hepatic Blood Volume Imaging with the Use of Flat-Detector CT Perfusion in the Angiography Suite: Comparison with Results of Conventional Multislice CT Perfusion Zhi-guo Zhuang, PhD, Xue-bin Zhang, PhD, Jing-feng Han, PhD, Janina Beilner, MD, Yu Deuerling-Zheng, MSc, Jia-chang Chi, MD, Ji Wang, MD, Li-jun Qian, MD, Yan Zhou, PhD, and Jian-rong Xu, MD
ABSTRACT Purpose: To prospectively determine the feasibility of flat-detector (FD) computed tomography (CT) perfusion to measure hepatic blood volume (BV) in the angiography suite in patients with hepatocellular carcinoma (HCC). Materials and Methods: Twenty patients with HCC were investigated with conventional multislice and FD CT perfusion. CT perfusion was carried out on a multislice CT scanner, and FD CT perfusion was performed on a C-arm angiographic system, before transarterial chemoembolization procedures. BV values of conventional and FD CT perfusion were measured within tumors and liver parenchyma. The arterial perfusion portion of CT perfusion BV was extracted from CT perfusion BV by multiplying it by a hepatic perfusion index. Relative values (RVs) for CT perfusion arterial BV and FD CT perfusion BV (FD BV) were defined by dividing BV of tumor by BV of parenchyma. Relationships between BV and RV values of these two techniques were analyzed. Results: In all patients, both perfusion procedures were technically successful, and all 33 HCCs larger than 10 mm were identified with both imaging methods. There were strong correlations between the absolute values of FD BV and CT perfusion arterial BV (tumor, r ¼ 0.903; parenchyma, r ¼ 0.920; both P o .001). Bland–Altman analysis showed a mean difference of 0.15 ⫾ 0.24 between RVs for CT perfusion arterial BV and FD BV. Conclusions: The feasibility of FD CT perfusion to assess BV values of liver tumor and surrounding parenchyma in the angiographic suite was demonstrated.
ABBREVIATIONS ALP = arterial liver perfusion, BV = blood volume, DSA = digital subtraction angiography, FD = flat detector, FD BV = blood volume on flat-panel CT perfusion, HCC = hepatocellular carcinoma, ICC = intraclass correlation coefficient, ROI = region of interest, RV = relative value
From the Department of Radiology (Z.Z., X.Z., J.C., J.W., L.Q., Y.Z., J.X.), Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University; Healthcare Sector (J.H., J.B.), Siemens Limited China, Shanghai, China; and Healthcare Sector (Y.D.-Z.), Siemens AG, Forchheim, Germany. Received October 12, 2013; final revision received January 17, 2014; accepted January 18, 2014. Address correspondence to: J.X., No. 160, Pujian Road, Pudong New District, Shanghai 200127, China; E-mail: [email protected] Z.Z. and X.Z. contributed equally to this work. This work was supported by National Natural Science Foundation of China Grant 81201172 (to Z.Z.), Medical and Engineering Program of Shanghai Jiaotong University Grant YG2012MS16 (to L.Q.), and National Natural Science Foundation of China Grant 81171325 (to Y.Z.). J.H. and J.B. are employees of Siemens Limited China (Shanghai, China). Y.D.-Z. is an employee of Siemens AG (Forchheim, Germany). None of the other authors have identified a conflict of interest. & SIR, 2014 J Vasc Interv Radiol 2014; 25:739–746 http://dx.doi.org/10.1016/j.jvir.2014.01.021
Because most liver diseases lead to significant changes in hepatic microcirculation regionally, globally, or both, quantification of hepatic perfusion can improve the assessment and management of liver diseases (1,2). In recent years, various imaging techniques, such as computed tomography (CT), xenon-enhanced CT, isotope scintigraphy, magnetic resonance (MR) imaging, and ultrasonography, as well as positron emission tomography with the use of oxygen-15–labeled water, have been applied in the evaluation of hepatic perfusion (1–3). Many studies have demonstrated the capacity of CT perfusion imaging to evaluate hepatic tissue blood flow, blood volume (BV), and hepatic perfusion index (HPI) (2,4–6). BV in CT perfusion imaging is reportedly useful for assessment of liver tumor vascularity and
740
’
Hepatic Blood Volume Imaging: Flat Detector vs Conventional CT
angiogenesis (4,7), prediction of tumor response to therapies (8), quantification of hemodynamic alterations in liver transplantation (9), early response assessment, and early detection of recurred tumors after radiologic interventions (10–12). Liver functional imaging during interventional treatment has been rarely available. With the development of flat-detector (FD) technology, FD-equipped angiographic systems are increasingly used in daily routine. They can offer high-quality three-dimensional digital subtraction angiography (DSA) images and soft-tissue cross-sectional images in the angiography suite with FD CT (13). Nowadays, there is evidence that FD CT could also provide a means to generate functional images. Studies have proved the feasibility of cerebral FD CT perfusion and verified the accuracy of cerebral BV acquired by FD CT perfusion compared with conventional CT perfusion (14,15). The ability of immediate measurement of cerebral BV in an angiography suite has a great potential to save time from initial assessment to intervention and enhance the management of patients with ischemic strokes (14,16,17). Based on the successful application of FD CT perfusion in neurologic interventions, it is reasonable to propose that hepatic BV might be measured by FD CT perfusion. In the present study, our purpose was to prospectively determine the feasibility of measuring the BV values of liver parenchyma or tumor tissues in patients with hepatocellular carcinoma (HCC) by using FD CT perfusion in the angiography suite.
Zhuang et al
’
JVIR
Twenty patients satisfying our inclusion/exclusion criteria were consecutively enrolled in this prospective study. The patients’ demographic data are listed in Table 1. The diagnosis of HCC was verified by performing a needle biopsy and histopathologic examination of resected specimen, or on the basis of the presence of a tumor larger than 1 cm in diameter with typical imaging findings in the setting of cirrhosis (18). The study was approved by the ethics committee at our institution. Written informed consent was obtained from all patients before their participation in the study.
CT Perfusion Technique With a 128-row CT scanner (SOMATOM Definition ASþ; Siemens, Forchheim, Germany), CT perfusion examinations were carried out at least 2 days after diagnostic CT. All patients fasted for 8 hours and rested calmly for 30 minutes before the CT perfusion examination to minimize physiologic variations in hepatic perfusion. After a preliminary unenhanced localizer scan covering the epigastric region, a scanning range of 175 mm covering almost the entire liver as well as the spleen was chosen. Dynamic CT was performed with the following parameters: 150 mAs, 80 kV, rotation time of 0.30 seconds, pitch of 0.6, slice acquisition of 128 0.6 mm and 5-mm reconstructed section thickness, scan time of 1.8 seconds, cycle time of 1.75 seconds, and examination time of 48.67 seconds. Dynamic scanning was initiated after a 6-second delay from the beginning Table 1 . Patients’ Demographic Data (N ¼ 20)
MATERIALS AND METHODS Patients From June 2012 to December 2012, 113 patients with hepatic cirrhosis with HCCs referred to our department for transarterial chemoembolization. Patients included in this study satisfied all of the following criteria: the patients (i) were older than 18 years; (ii) had Child– Pugh class A or B disease; (iii) had no contraindications to iodinated contrast media; and (iv) were able to hold their breath for 22 seconds for FD CT perfusion and maintain shallow breathing for approximately 50 seconds in CT perfusion examination after training. Excluded from the study were patients with (i) Child– Pugh class C disease; (ii) uncorrectable coagulopathy (International Normalized Ratio 4 1.5); (iii) portal vein thrombosis; (iv) a total bilirubin level higher than 4.0 mg/dL; (v) a serum creatinine level higher than 1.7 mg/ dL; (vi) thrombocytopenia (platelet count o 50,000/μL); (vii) previous therapeutic procedures; and (viii) hepatic arterial variants. For each patient, a routine diagnostic four-phase CT was performed. The hepatic arterial anatomy was determined by CT angiography reconstructed from arterial-phase images of CT.
Characteristics
No. of Patients
Sex Male
13
Female Age*
7
o 60 y
12
Z 60 y No. of tumors 1
8 13
2–3 43
6 1
Tumor size o 3 cm 3–5 cm
5 7
4 5 cm
8
Child–Pugh class A
15
B Cause of cirrhosis Hepatitis B
5 13
Hepatitis C
2
Hepatitis B and C Alcohol-related
1 4
*Mean age was 59.6 y ⫾ 12.2 y (standard deviation).
Volume 25
’
Number 5
’
May
’
2014
of the injection of 40 mL nonionic contrast medium (Ultravist 370 mgI/mL; Schering, Berlin, Germany) at the rate of 5.0 mL/s, followed by 40 mL saline solution flush at the same rate, through an 18-gauge intravenous cannula inserted into the antecubital vein with an automatic dual-headed injector (OptiVantage DH; Mallinckrodt, St. Louis, Missouri). Before imaging, all patients were instructed to keep breathing shallowly during CT perfusion. In addition, a band compressing the abdomen that limited breath-related liver excursions was used. The estimated effective radiation dose of CT perfusion, calculated by multiplying the dose–length product, obtained from the patient protocol, with a conversion coefficient for the abdomen (k ¼ 0.015 mSv mGy1 cm1), was 28.7 mSv. The data were processed on a commercially available workstation (syngo MultiModality Workplace VE40; Siemens) with CT perfusion software (syngo VPCT Body; Siemens). If respiratory motion–associated misregistrations were detected, the motion was compensated manually by using a rigid registration technique. Regions of interest (ROIs) were placed on the abdominal aorta at the level of the celiac axis (range, 52–105 mm2), the main portal vein (range, 24–124 mm2), and the spleen (range, 2,650–5,769 mm2) to generate time– density curves. Then, functional color map images of each perfusion parameter of the liver were generated. The mathematical technique has previously been fully described elsewhere (19). To calculate BV values of liver tumors and parenchyma, ROIs for tumors (range, 86– 7,863 mm2) were manually drawn within the tumor section with the maximum cross-sectional area on the CT perfusion map of BV. In the presence of multiple tumors, ROIs were drawn for all tumors larger than 10 mm, whereas tumors smaller than 10 mm in diameter were excluded from evaluation. Similarly, two ROIs were drawn in the background liver parenchyma (range,
741
130–1,255 mm2) in each patient. ROIs of the background liver were drawn as far away as possible from the tumor and were kept distant from large intrahepatic vessels and the borders of the organ (Fig 1a).
FD CT Perfusion Protocol All FD CT perfusion procedures were performed by two experienced interventional radiologists 3 days after CT perfusion on a robotic angiographic system (Artis zeego VC13; Siemens) during transarterial chemoembolization. According to previous CT angiography, which was used to assess the morphology of the celiac artery and superior mesenteric artery, a 5-F RH catheter (Terumo, Tokyo, Japan) or 3-F coaxial microcatheter (Terumo) was inserted into the proper hepatic artery. The FD CT perfusion acquisition consisted of two rotations: an initial rotation (mask run) followed by a second rotation after contrast medium injection (fill run). The data acquisition was carried out by using the following parameters: acquisition time of 5 seconds, examination time of 22 seconds, 90 kV, 512 512 matrix, projection on 30 40 cm flat panel size, 2001 total angle, 0.81 per frame, 250 frames total, dose of 0.36 μGy per frame. A total of 36 mL of contrast medium (Ultravist 370 mgI/ mL; Schering) diluted to 25% was injected by power injector (Angiomat 6000; Liebel-Flarsheim, Cincinnati, Ohio) at 3 mL/s after the mask run. With a 7-second xray delay, the fill run was then performed. The time delay was adopted to perfuse the liver in a steady state for the BV acquisition. After the FD CT perfusion acquisition, the chemoembolization procedure was performed as routinely done.
Postprocessing of FD BV Imaging FD BV postprocessing was performed by using prototype software (Siemens) installed on a research workstation (syngo X-Workplace VB15; Siemens). The
Figure 1. CT perfusion BV (a) and FD BV (b) maps from the same patient show conformity of the images. Positioning of the ROIs is displayed: 1, tumor; 2, liver tissue.
742
’
Hepatic Blood Volume Imaging: Flat Detector vs Conventional CT
software adopted the same postprocessing workflow as previously described (14). The mask and the fill run were reconstructed and subtracted. The motion between the two runs was corrected by the nonrigid registration algorithm. The steady-state arterial input function value was calculated from an automated histogram analysis of the vessel tree. A final scaling was then applied to the dataset to account for the arterial input value. In the end, a smoothing filter was applied to image to reduce pixel noise. The matching positions, both tumor and liver tissue ROIs with similar sizes to CT perfusion maps (range, 89–7,871 mm2 in tumors and 135–1,273 mm2 in liver tissue), were selected, and the BV values were measured from FD BV maps (Fig 1b).
Data Analysis and Statistics Detection of tumors in CT perfusion and FD CT perfusion maps were performed by two independent radiologists with 5 and 6 years of experience in reading perfusion images. If matching lesions were found in both images, the two radiologists were required to reach consensus to select the corresponding section and ROIs to measure CT perfusion BV and FD BV values. Each radiologist was blinded to the values of the other, and the outcomes of both radiologists were recorded by another assistant. It is known that there exists dual blood supply for the liver: hepatic arterial and portal vein supply. Pure hepatic arterial blood perfusion could be distinguished from portal vein supply by measuring the dynamic change in contrast medium perfusion of the liver after injection via the hepatic artery (20). In the present study, CT perfusion BV represented the total BV supplied by the hepatic artery and portal vein because the contrast medium was injected intravenously, whereas FD BV was solely arterial BV because it was performed by administering contrast medium intraarterially. To make the comparison of CT perfusion BV and FD BV more precise, we extracted the arterial perfusion portion from the total liver BV in CT perfusion examinations as follows. From the postprocessing results of CT perfusion, the following parameters can be obtained: BV, blood flow, arterial liver perfusion (ALP), portal vein perfusion, and HPI. The blood perfusion from the hepatic artery and portal vein can be distinguished by HPI, which represents the arterial portion to the total liver perfusion. For this reason, we assumed that the arterial BV values measured from CT perfusion could be obtained by multiplying CT perfusion BV by HPI. Values are presented as mean ⫾ standard deviation. Differences of BV values between the two examinations were analyzed by paired t test. Differences of BV values between tumor and liver parenchyma in the same examination were analyzed by the independent-sample t test. Pearson correlation coefficients were calculated between CT perfusion arterial BV and FD BV for all
Zhuang et al
’
JVIR
liver parenchyma and tumor tissues, respectively. Concordance between the measurements of CT perfusion arterial BV or FD BV in parenchyma from the two ROIs was assessed by using intraclass correlation coefficients (ICC). In the present study, the relative value (RV) between liver tumor and parenchyma for CT perfusion arterial BV was defined as CT perfusion arterial BV of tumor divided by CT perfusion arterial BV of parenchyma. The RV between liver tumor and parenchyma for FD BV was defined as FD BV of tumor divided by FD BV of parenchyma. Agreement between RVs for CT perfusion arterial BV and FD BV was assessed by using the approach of Bland and Altman (21).
RESULTS FD BV and CT perfusion BV acquisitions were successfully performed in all patients. CT perfusion BV and FD BV maps and data quality were suitable for evaluation in all patients. On both BV maps, liver parenchyma and tumor lesions can be recognized in the corresponding positions (Fig 1). In the CT perfusion BV maps, 33 proven HCC lesions larger than 10 mm were recognized in all 20 patients. FD BV findings of tumor showed oneto-one correlation with CT perfusion in all patients. It is worth mentioning that FD BV maps showed more hypervascularized regions smaller than 10 mm than CT perfusion BV maps in four patients (Fig 2). Because lesions smaller than 10 mm were difficult to diagnose qualitatively and quantitatively in the present study, these small hypervascularized regions were not further analyzed. Finally, BV values in 33 tumor spots and 40 liver parenchyma areas were measured. A total of 12 patients had FD CT perfusion with a 5-F catheter, compared with eight with a 3-F microcatheter. There were no significant differences between the two groups in FD BV values (P ¼ .415 and P ¼ .850 for tumor and liver parenchyma, respectively; Table 2). There were significant differences between CT perfusion arterial BV and FD BV in tumor and liver parenchyma (all P o .001). In addition, HPI, CT perfusion BV, CT perfusion arterial BV, and FD BV in tumor were significantly higher than in parenchyma (all P o .001; Table 3). There were strong correlations between the absolute values of FD BV and CT perfusion arterial BV (r ¼ 0.903, P o .001 for tumor [Fig 3a]; and r ¼ 0.920, P o .001 for liver parenchyma [Fig 3b]). The ICC of the measurements obtained from the two ROIs in liver parenchyma were 0.96 for CT perfusion arterial BV and 0.95 for FD BV. Because there was a close correlation between the two ROIs, the two measurements were averaged as the CT perfusion arterial BV or FD BV of parenchyma in each patient for further analysis of RV. The Bland–Altman plots of RV for CT perfusion arterial BV and FD BV are displayed in
Volume 25
’
Number 5
’
May
’
2014
743
Figure 2. CT perfusion BV (a) and FD BV (b) maps from a patient with multifocal lesions. The two BV maps match closely except that an additional small arterial hypervascularized region (arrow) was found on the FD BV map.
Table 2 . Differences in BV Values of Flat-Detector CT Perfusion between Patients with a 5-F Catheter and a 3-F Microcatheter BV Measurement Tumor (mL/100 g) Liver parenchyma (mL/100 g)
Catheter (n ¼ 12)
Microcatheter (N ¼ 8)
P Value
25.2 ⫾ 5.1
26.8 ⫾ 6.0
.415
4.9 ⫾ 1.2
4.7 ⫾ 1.1
.850
BV ¼ blood volume.
Table 3 . Parameters of Perfusion Generated by CT Perfusion and FD CT Perfusion Liver Parenchyma
P Value
HPI (%)
90.5 ⫾ 10.7
25.1 ⫾ 6.2
o .001
CT perfusion BV (mL/100 g) CT perfusion arterial BV (mL/100 g)
16.5 ⫾ 2.5 15.0 ⫾ 3.1
11.1 ⫾ 1.2 2.8 ⫾ 0.8
o .001 o .001
25.9 ⫾ 5.6
4.8 ⫾ 1.2
o .001
10.9 ⫾ 3.1
2.0 ⫾ 0.6
Parameter
FD BV (mL/100 g) Difference: CT perfusion arterial BV vs FD BV (mL/100 g)
Tumor
NA
Values presented as means ⫾ standard deviation. BV ¼ blood volume, FD ¼ flat detector, HPI ¼ hepatic perfusion index, NA ¼ not applicable.
Figure 4. This analysis showed that the mean difference in RV between CT perfusion arterial BV and FD BV was 0.15 ⫾ 0.24, meaning that RV for CT perfusion arterial BV was slightly lower than that for FD BV.
DISCUSSION Previous studies have validated that FD CT perfusion could provide a means to measure the cerebral BV at an accuracy comparable to that of CT perfusion (14–16). However, this technique has not been applied to measure liver perfusion. In the present study, compared with the data from CT perfusion imaging, we showed that FD CT perfusion has potential to be used to monitor BV within the targeted tumors and adjacent liver tissue during transarterial chemoembolization in the angiography suite. According to previous CT perfusion studies
(4,7,22–25), BV values ranged from 3.2 to 11.7 mL/100 g in cirrhotic liver parenchyma and from 4.9 to 24.0 mL/ 100 g in HCC. HPI values ranged from 10.6% to 65% in cirrhotic liver parenchyma and from 75.3% to 92.5% in HCC. Our results are therefore within these reported ranges. On the contrary, our methods of CT perfusion and FD CT perfusion were also proven to be reliable by the excellent ICCs of CT perfusion arterial BV and FD BV values obtained from the two separate ROIs in liver parenchyma. In the present study, FD BV had a very good correlation with CT perfusion arterial BV. In addition, the Bland–Altman plot showed good agreement in RV between CT perfusion arterial BV and FD BV, indicating that the BV values obtained from FD CT perfusion were believable, and there were similar capacities to display HCC between the two techniques. However, the absolute values of FD BV were higher than CT
744
’
Hepatic Blood Volume Imaging: Flat Detector vs Conventional CT
Zhuang et al
’
JVIR
Figure 3. Correlation between BV values from FD CT perfusion and CT perfusion imaging. (a) Scatter plots of FD BV and CT perfusion arterial BV in tumor (r ¼ 0.903, P o .001). (b) Scatter plots of FD BV and CT perfusion arterial BV in liver parenchyma (r ¼ 0.920, P o .001).
Figure 4. Bland–Altman plots representing the degree of agreement in RVs between CT perfusion arterial BV and FD BV. The Bland–Altman interval was [0.63 (0.15) 0.33].
perfusion arterial BV values. Possible reasons for this phenomenon might be different injection protocols and different algorithms adopted for the two methods. Direct hepatic arterial injection might cause higher BV than intravenous injection because the vascular bed might have been further opened when the arterial injection protocol with additional pressure was applied, whereas CT perfusion BV involves only the use of conventional intravenous injections. In addition, tumor spots were often highly vascularized and supplied mainly by the hepatic artery, and therefore their FD BV values were even higher than on CT perfusion BV. Certainly, in liver parenchyma, because of dominant portal venous source in dual blood supply, FD BV was lower than CT perfusion BV.
There is strong evidence that CT perfusion imaging may be valuable in monitoring the outcome of transarterial chemoembolization in cirrhotic patients with HCC (11,12). Ippolito et al (11) have shown that, in patients with HCC treated with transarterial chemoembolization, CT perfusion BV showed a tendency to be greater in the relapse site than in the primary lesion area of iodized oil deposition. In an animal study, Choi et al (10) made similar observations with CT perfusion in VX2 tumor rabbit models. Similarly, Chen et al (12) also demonstrated that changes in CT perfusion BV of viable tumors are associated with different responses of HCC to transarterial chemoembolization. As mentioned earlier, much work has been done regarding the value of conventional CT perfusion for chemoembolization in HCC management. However, little was known about intrachemoembolization physiologic monitoring. Although some recent literature (26–28) has demonstrated the use of transcatheter intraarterial perfusion MR imaging to detect intraprocedural perfusion changes of HCC and adjacent surrounding liver tissue during transarterial chemoembolization, a special and dedicated Miyabi system (a conventional DSA unit integrated with an MR imaging system) was needed for this technique (26–28). Nowadays, the FD CT perfusion technology can also provide perfusion examination within the catheterization laboratory intraprocedurally. It might offer information to determine the optimal endpoint of chemoembolization for a better prognosis in the future. FD CT perfusion has several other potential advantages. First, FD CT perfusion enables direct monitoring of local arterial BV of liver tumors during transarterial chemoembolization. One potential application is the replanning of catheter position in the angiographic suite. Prechemoembolization FD CT perfusion could be used
Volume 25
’
Number 5
’
May
’
2014
to determine if the targeted hepatic artery is supplying the tumor, and hence decide whether the initial catheter position needs to be replanned. Similar transcatheter intraarterial perfusion MR imaging has already been shown to alter the initial catheter position anticipated based on conventional DSA results in 40%–50% of patients during transcatheter HCC therapy (27,28). Second, in the present study, compared with CT perfusion BV, more hypervascularized regions smaller than 10 mm were detected on the FD BV maps. FD BV imaging may enhance tumor conspicuity in HCC because of the increased perfusion at the tumor rims, which makes liver tumors appear larger, thereby facilitating the detection of smaller lesions when contrast material is injected into the hepatic artery directly (29). Further studies are needed to prove whether these small hypervascularized regions are HCC lesions that were not detected by CT perfusion. Finally, the localized delivery of smaller doses of contrast material during FD CT perfusion scanning than in CT perfusion imaging could mitigate potential kidney injury. The radiation dose of our CT perfusion protocol was similar to that reported by Goetti et al (19). However, the radiation exposure was not available for FD CT perfusion in this study. Although Fiorella et al (30) reported that the FD CT perfusion dose was similar to that of standard CT perfusion when estimating cerebral BV, the exact effective radiation dose was not obtained in their study. We speculate that, in the present study, the FD CT perfusion radiation exposure may be lower than with CT perfusion as a result of the short scan time. The present study has several limitations. First, compared with CT perfusion, which can provide HPI, ALP, portal vein perfusion, and BV, only BV can be assessed with FD CT perfusion because of mechanical limitations. Ippolito et al (24) found that ALP and HPI were likely to be the most relevant parameters to assess neovascularization induced by tumor growth. However, BV could be affected by intrahepatic arterioportal shunts. Second, this technique cannot be used to assess perfusion values that result from the collateral arterial supply or the portal vein. Third, we assessed perfusion at only one time point during transarterial chemoembolization, namely before embolization. Future studies need to be performed to determine perfusion changes before and after chemoembolization. Consequently, the relationship between perfusion change and tumor response should be investigated. Fourth, the total radiation dose received by the patients may be high at present. In our future studies, low-dose perfusion imaging should be performed. Finally, we extracted the arterial perfusion portion from the total liver BV on CT perfusion by using HPI, which represented the percentage of arterial perfusion in the total liver blood flow. This assumption has not yet been validated fully. Further studies are needed to evaluate the relationship between HPI and the arterial contribution to total liver BV.
745
In summary, it was demonstrated in a series of patients with HCC that BV mapping with the use of FD CT perfusion is feasible and shows good correlation with that obtained with standard CT perfusion techniques. The good agreement between RVs for CT perfusion arterial BV and FD BV indicated that FD BV was comparable with CT perfusion arterial BV. This technique offers the ability to measure objective, quantitative changes in perfusion during transarterial chemoembolization therapy.
REFERENCES 1. Kanda T, Yoshikawa T, Ohno Y, et al. CT hepatic perfusion measurement: comparison of three analytic methods. Eur J Radiol 2012; 81: 2075–2079. 2. Pandharipande PV, Krinsky GA, Rusinek H, Lee VS. Perfusion imaging of the liver: current challenges and future goals. Radiology 2005; 234: 661–673. 3. Zhuang ZG, Qian LJ, Wang BX, et al. Computed tomography perfusion in living donor liver transplantation: an initial study of normal hemodynamic changes in liver grafts. Clin Transplant 2009; 23:692–699. 4. Sahani DV, Holalkere NS, Mueller PR, Zhu AX. Advanced hepatocellular carcinoma: CT perfusion of liver and tumor tissue–initial experience. Radiology 2007; 243:736–743. 5. Tsushima Y, Endo K. Portal perfusion measurement on dynamic CT in patients with liver cirrhosis. AJR Am J Roentgenol 2005; 185:813. 6. Van Beers BE, Leconte I, Materne R, Smith AM, Jamart J, Horsmans Y. Hepatic perfusion parameters in chronic liver disease: dynamic CT measurements correlated with disease severity. AJR Am J Roentgenol 2001; 176:667–673. 7. Ippolito D, Sironi S, Pozzi M, et al. Perfusion CT in cirrhotic patients with early stage hepatocellular carcinoma: assessment of tumor-related vascularization. Eur J Radiol 2010; 73:148–152. 8. Petralia G, Fazio N, Bonello L, D’Andrea G, Radice D, Bellomi M. Perfusion computed tomography in patients with hepatocellular carcinoma treated with thalidomide: initial experience. J Comput Assist Tomogr 2011; 35:195–201. 9. Qian LJ, Zhuang ZG, Cheng YF, Xia Q, Zhang JJ, Xu JR. Hemodynamic alterations in anterior segment of liver graft after right-lobe living-donor liver transplantation: computed tomography perfusion imaging findings. Abdom Imaging 2010; 35:522–527. 10. Choi SH, Chung JW, Kim HC, et al. The role of perfusion CT as a followup modality after transcatheter arterial chemoembolization: an experimental study in a rabbit model. Invest Radiol 2010; 45:427–436. 11. Ippolito D, Bonaffini PA, Ratti L, et al. Hepatocellular carcinoma treated with transarterial chemoembolization: dynamic perfusion-CT in the assessment of residual tumor. World J Gastroenterol 2010; 16: 5993–6000. 12. Chen G, Ma DQ, He W, Zhang BF, Zhao LQ. Computed tomography perfusion in evaluating the therapeutic effect of transarterial chemoembolization for hepatocellular carcinoma. World J Gastroenterol 2008; 14:5738–5743. 13. Iwazawa J, Ohue S, Hashimoto N, Abe H, Hamuro M, Mitani T. Detection of hepatocellular carcinoma: comparison of angiographic C-arm CT and MDCT. AJR Am J Roentgenol 2010; 195:882–887. 14. Ahmed AS, Zellerhoff M, Strother CM, et al. C-arm CT measurement of cerebral blood volume: an experimental study in canines. AJNR Am J Neuroradiol 2009; 30:917–922. 15. Struffert T, Deuerling-Zheng Y, Kloska S, et al. Cerebral blood volume imaging by flat detector computed tomography in comparison to conventional multislice perfusion CT. Eur Radiol 2011; 21:882–889. 16. Struffert T, Deuerling-Zheng Y, Kloska S, et al. Flat detector CT in the evaluation of brain parenchyma, intracranial vasculature, and cerebral blood volume: a pilot study in patients with acute symptoms of cerebral ischemia. AJNR Am J Neuroradiol 2010; 31:1462–1469. 17. van der Bom IM, Mehra M, Walvick RP, Chueh JY, Gounis MJ. Quantitative evaluation of C-arm CT cerebral blood volume in a canine model of ischemic stroke. AJNR Am J Neuroradiol 2012; 33:353–358.
746
’
Hepatic Blood Volume Imaging: Flat Detector vs Conventional CT
18. Bruix J, Sherman M. American Association for the Study of Liver Diseases. Management of hepatocellular carcinoma: an update. Hepatology 2011; 53:1020–1022. 19. Goetti R, Reiner CS, Knuth A, et al. Quantitative perfusion analysis of malignant liver tumors: dynamic computed tomography and contrastenhanced ultrasound. Invest Radiol 2012; 47:18–24. 20. Komemushi A, Tanigawa N, Kojima H, Kariya S, Sawada SCT. perfusion of the liver during selective hepatic arteriography: pure arterial blood perfusion of liver tumor and parenchyma. Radiat Med 2003; 21: 246–251. 21. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999; 8:135–160. 22. Ippolito D, Sironi S, Pozzi M, et al. Hepatocellular carcinoma in cirrhotic liver disease: functional computed tomography with perfusion imaging in the assessment of tumor vascularization. Acad Radiol 2008; 15:919–927. 23. Miles KA, Hayball MP, Dixon AK. Functional images of hepatic perfusion obtained with dynamic CT. Radiology 1993; 188:405–411. 24. Ippolito D, Capraro C, Casiraghi A, Cestari C, Sironi S. Quantitative assessment of tumour associated neovascularisation in patients with liver cirrhosis and hepatocellular carcinoma: role of dynamic-CT perfusion imaging. Eur Radiol 2012; 22:803–811.
Zhuang et al
’
JVIR
25. Hashimoto K, Murakami T, Dono K, et al. Assessment of the severity of liver disease and fibrotic change: the usefulness of hepatic CT perfusion imaging. Oncol Rep 2006; 16:677–683. 26. Larson AC, Wang D, Atassi B, et al. Transcatheter intraarterial perfusion: MR monitoring of chemoembolization for hepatocellular carcinoma– feasibility of initial clinical translation. Radiology 2008; 246:964–971. 27. Vogl TJ, Balzer JO, Mack MG, Bett G, Oppelt A. Hybrid MR interventional imaging system: combined MR and angiography suites with single interactive table. Feasibility study in vascular liver tumor procedures. Eur Radiol 2002; 12:1394–1400. 28. Wilson MW, Kerlan RK Jr, Fidelman NA, et al. Hepatocellular carcinoma: regional therapy with a magnetic targeted carrier bound to doxorubicin in a dual MR imaging/ conventional angiography suite–initial experience with four patients. Radiology 2004; 230:287–293. 29. Meijerink MR, van Waesberghe JH, van der Weide L, van den Tol P, Meijer S, van Kuijk C. Total-liver-volume perfusion CT using 3-D image fusion to improve detection and characterization of liver metastases. Eur Radiol 2008; 18:2345–2354. 30. Fiorella D, Turk A, Chaudry I, et al. A prospective, multicenter pilot study investigating the utility of flat detector derived parenchymal blood volume maps to estimate cerebral blood volume in stroke patients. J Neurointerv Surg 2013. http://dx.doi.org/10.1136/neurintsurg-2013-010840.
