- Email: [email protected]
Micro-CT Imaging with a Hepatocyte-selective Contrast Agent for Detecting Liver Metastasis in Living Mice
Micro-CT Imaging with a Hepatocyte-selective Contrast Agent for Detecting Liver Metastasis in Living Mice1 Hye Won Kim, MD, Quan-Yu Cai, MD, Hong Yeoung Jun, Kwon Su Chon, PhD, Seung Hoon Park, MD Seung Jae Byun, MD, Myung Soo Lee, MD, Jae Min Oh, MD, Hun Soo Kim, MD, Kwon-Ha Yoon, MD
Rationale and Objectives. Micro-computed tomography (CT) is a important tool for longitudinal imaging of tumor development. The detection and monitoring of tumors in the liver in live animals using micro-CT is challenging. We evaluated the feasibility of high-resolution micro-CT enhanced with a hepatocyte-selective contrast agent for detecting liver metastases in a live murine model. Materials and Methods. Hepatic metastases were induced in 10 BALB/C mice. Two mice each were randomly selected on days 3, 5, 7, 10, and 13 after CT26 colon adenocarcinoma cells were injected into the portal vein; micro-CT imaging was performed at 10 minutes and 4 hours after intravenous administration of a hepatocyte-selective contrast agent at a dose of 0.4 mL/mouse. The attenuation values of the normal liver and the tumors were obtained. The number of metastases was counted and their sizes were measured on the micro-CT images. Gross or histopathologic evaluation was performed for correlating the liver tumors with the micro-CT images. Results. A total of 74 separate tumor sites larger than 300 m in diameter were detected on pathologic examination of the mice that were sacrificed 7 days after cell injection. On micro-CT, 66 of 74 tumors were detected (83.8%). The smallest tumor detected on micro-CT was 300 m. There were eight false-negative readings on micro-CT. The sizes of the individual liver metastases measured by micro-CT and on the excised specimen were highly correlated (P ⬍ .001). The correlation between the CT scan measurement and the actual measurement was r ⫽ 0.8354 (P ⬍ .0001). Conclusions. High-resolution micro-CT enhanced with a hepatocyte-selective contrast agent can be a promising tool for detecting liver metastases in a live murine model. Key Words. Contrast media; in vivo scanning; liver neoplasm; metastasis; X-ray microtomography. ©
AUR, 2008
The liver is a common site of metastases from various primary tumors, and so it is an important area of metastasis research (1). Rodent models that mimic the Acad Radiol 2008; 15:1282–1290 1
From the Departments of Radiology and the Institute for Radiological Imaging Science (H.W.K., Q.-Y.C., H.Y.J., K.S.C., S.H.P., K.-H.Y.), Surgery (S.J.B.), Internal Medicine (M.S.L.), Anatomy (J.M.O.), and Pathology (H.S.K.), Wonkwang University School of Medicine, 344-2 Sinyong-dong, Iksan, Jeonbuk 570-749, Korea. Supported by Wonkwang University in 2008. Received February 17, 2008; accepted March 19, 2008. Address correspondence to: K.-H.Y. e-mail: [email protected]
© AUR, 2008 doi:10.1016/j.acra.2008.03.021
1282
development of metastatic disease in the human liver are increasingly being recognized as powerful tools for the development of anticancer drugs and for evaluating the efficacy of novel therapeutics in the preclinical setting (2,3). Noninvasive longitudinal imaging of rodent models would decrease experimental variability and provide a more accurate assessment of metastatic progression and the efficacy of therapeutic interventions (4,5). Several longitudinal imaging modalities are now used, including magnetic resonance imaging, x-ray computed tomography (CT), positron emission tomography, and fluorescent and bioluminescent imaging.
Academic Radiology, Vol 15, No 10, October 2008
Micro-CT has been applied in biomedical research mainly for studying bones and lung disease, for which the natural contrast between bone and air and the surrounding soft tissues are provided (6,7). Micro-CT increased the spatial resolution noninvasively and it allowed very high precision for localization of bony changes. It has been proposed by some researchers that small-animal x-ray CT is an important tool for longitudinal imaging of tumor development, for visualizing blood vessels and angiogenesis, and for following the response of tumors to preclinical therapeutic intervention (8,9). However, the detection and monitoring of tumor in the liver in live animals using micro-CT is still challenging because of the poor natural contrast between the tumor and the liver parenchyma. When performing micro-CT imaging for the liver of small animals, it is not possible to use the conventional water-soluble contrast agents that are used in humans because they clear rapidly from the blood. For the detection of experimental liver tumors, a hepatocyte-specific contrast agent (1,3-bis [7-(3-amino2,4,6-triiodophenyl) heptanoyl]-2-oleoyl-glycerol) can be used to overcome this limitation (10). Previous reports have demonstrated the efficacy of clinical CT using the hepatocyte-specific contrast agent for the detection of liver metastasis in rabbits and rats (11,12). However, the capability of micro-CT to detect liver metastasis tumors in live mice is largely unknown. Ohta et al (13) and Almajdub et al (14) have recently suggested that micro-CT can be used to detect small size liver tumors in mice. Weber et al (15) have evaluated the ability of micro-CT using a hepatocyte-specific contrast agent for detecting and monitoring liver tumors in murine hepatic tumor models. In these studies, the ability of micro-CT using a hepatocyte-specific contrast agent for detecting small liver metastasis has not been directly assessed, even if these studies to date support the possibility of using micro-CT for detecting small liver tumors in mice. Therefore, the feasibility of using micro-CT with a hepatocyte-selective contrast agent has not been currently established for detecting early liver metastasis in a small animal model. We hypothesized that micro-CT with adequate contrast enhancement will allow detection of early liver metastasis in a mouse model. Therefore, we studied the feasibility of high-resolution micro-CT enhanced with a hepatocyteselective contrast agent for detecting experimental liver metastases at an early stage in live mice.
DETECTING LIVER METASTASIS IN LIVING MICE
METHODS AND MATERIALS Tumor Model Thirteen male Balb/c mice (Central Lab. Animal Inc, Seoul, Korea), age 6 weeks, were used for this study. The mice ranged in weight from 22 to 25 g. All the mice lived in a system equipped with day–night light cycling and the mice were provided with standard mouse chow. The CT-26 murine colon adenocarcinoma cell line was acquired from ATCC (Manassas, VA). The cells were grown as monolayer cultures in DMEM-10% fetal bovine serum (Invitrogen, Grand Island, NY) supplemented with 1% glutamine, and 1% antibiotics. The cells were maintained in a 37°C incubator with 5% CO2-humidified air. To simulate liver metastases, 10 mice were anesthetized with intraperitoneally injected ketamine (100 g/g body weight) and xylazine (10 g/g body weight). CT-26 cells (5 ⫻ 103 in 0.1 mL phosphate-buffered saline) were injected into the mesenteric vein using a 30-gauge needle after laparotomy. All the animal studies were carried out in accordance with the regulations set forth by the institutional review board of our university.
Micro-CT System The imaging was performed using a volumetric CT scanner (NFR-Polaris-G90; NanoFocusRay, Iksan, Korea), which is a micro-CT system designed to acquire high resolution imaging of live mice. The micro-CT system is composed primarily of an anode x-ray source with a 5-m focal spot, a complementary metal oxide semiconductor flat panel detector coupled with a thallium-doped cesium iodide scintillator, a linearly moving couch, a rotational gantry coupled with a positioning encoder, and a parallel processing system for the image data. The images were acquired at 50 kVp, 65 A, and 470 ms/frame, and there were 360 views. The estimated radiation dose was approximately 84 mGy using this image acquisition protocol. The spatial resolution of the micro-CT system was calculated as 44.25 m at 10% point on the modulation transfer factor. The images were reconstructed using the Feldkamp cone-beam reconstruction algorithm. The reconstruction image size was 1024 ⫻ 1024 pixels, and 512 slices were acquired. The final reconstructed data were converted to the Digital Imaging and Communication in Medicine format to create three-dimensional–rendered imaging using three-dimensional–rendering software (Lucion, Infinite, Seoul, Korea).
1283
KIM ET AL
Micro-CT Imaging The mice were placed in an induction chamber with 4% isoflurane in oxygen to induce anesthesia. During imaging, the mice were kept under anesthesia with 1.5% isoflurane in oxygen and they were allowed to recover between image acquisitions. Three BALB/C mice were imaged at the baseline and then serial micro-CT images were obtained immediately and at 1, 2, 4, 8, and 24 hours after bolus injection of the 150-nm particle diameter-iodinated triglyceride contrast agent (150 mg of iodine per kilogram of body weight; Fenestra LC ART, Toronto, ON, Canada) at a dose of 0.4 mL/mouse via a distal tail vein. These experimental data were used as the reference for the micro-CT imaging protocol for the detection of liver metastasis. Two mice were randomly selected on days 3, 5, 7, 10, and 13 after tumor cell injection, and micro-CT imaging was performed after intravenous administration of the contrast agent. Each mouse was injected with the contrast agent at a dose of 0.4 mL/mouse into the distal tail vein using a 30-gauge needle. Based on the CT images of the control mice, we acquired the micro-CT imaging of the tumor models at 10 minutes and 4 hours after contrast agent administration. All CT image data were acquired in live, free-breathing, anesthetized mice without respiratory or cardiac gating. Image Analysis To assess the enhancement achieved with the contrast agent, the CT signal intensity was expressed in Hounsfield units (HU). The mean CT number was determined in regions of interest that measured 1.5 mm in diameter (⬃700 pixels) on various slices in the liver, inferior vena cava, and tumors after injection of the contrast agent. The mean CT number and standard deviation were collected for each region of interest for all the images and these values were then compared for each time point in all mice. Micro-CT images of the tumor model were co-evaluated by two observers, who were blinded to the histopathologic results. The 10-minute and 4-hour images were evaluated simultaneously to avoid false-positive tumor detection. The observers were given a diagram that depicted the representative axial slices of mouse hepatic anatomy. The observers identified a suspected lesion by drawing it on the appropriate image. The number of metastases was counted and their size was measured on the micro-CT images. The diameter was defined as the longest distance across the tumor.
1284
Academic Radiology, Vol 15, No 10, October 2008
Histopathologic Correlation The mice were euthanized after micro-CT imaging. The abdomen was opened, the portal vein was cannulated for perfusion with 10% buffered formalin in situ, and the livers were fixed for 24 hours. The fixed livers were removed and then embedded in paraffin. The whole livers were sectioned axially at 4-m thickness with a 100-m interval and then we performed standard hematoxylin and eosin staining, and the tumors were measured microscopically. Correlation of the histopathologic characteristics of the tumors with the CT images was performed. Statistical Analysis Statistical analysis was performed using SPSS software (version 11.5, SPSS, Inc., Chicago, IL). Comparisons of the tumor size were done using paired t-tests. Evaluation of the correlation of the sizes of the tumors on micro-CT images with the sizes of the tumors, as determined on histopathologic examination, was done via regression analysis. For all the statistical analyses, a P value less than .05 was considered to indicate a statistically significant difference.
RESULTS Histopathologic Findings of the Metastatic Liver Tumors All the animals survived long enough to complete the full protocol. Mesenteric vein injection of 5 ⫻ 103 CT-26 cells in the BALB/C mice reproducibly generated multiple liver tumors in all the animals. At day 3 after cell injection, no tumor was found on 40⫻ magnification, and some necrosis areas were found on the peripheral areas of the mice’s livers (Fig 1a). At day 5, after cell injection, the necrotic areas had shrunk a good deal and some small tumors (smaller than 300 m) were observed, and two tumors larger than 300 m were found in one mouse liver (Fig 1b). At day 7 after cell injection, the necrotic areas were completely dissolved and were no longer observed on the histopathologic images. A total of 74 tumors larger than 300 m were observed (Fig 1c). At days 10 and 13 after cell injection, numerous sized tumors smaller than 300 m were observed (Fig 1d). The small size tumors were not included in the data analysis because measuring the number and size of these tumors was actually impossible. There were more tumors localized in left lobe than in the right lobe, and most of the tumors were localized in the peripheral portion of the liver.
Academic Radiology, Vol 15, No 10, October 2008
DETECTING LIVER METASTASIS IN LIVING MICE
Figure 1. Photographs of the hematoxylin and eosin staining of the hepatic metastasis murine model (40⫻ magnification). (a) Some necrotic areas were found on the peripheral areas of the livers at day 3 after cell injection. (b) A small tumor (⬍300 m) was observed and the necrotic areas had shrunk at day 5 after cell injection. (c) Discrete tumors larger than 300 m were observed and the necrotic areas were completely dissolved at day 7 after cell injection. (d) Multiple innumerable tumors were observed at day 10 after cell injection.
Liver Parenchyma and Tumor Attenuation on the Micro-CT Images The CT imaging appearance of three normal BALB/C mice was evaluated as a function of time after injection of the contrast agent (Fig 2). The baseline image shows the lack of contrast between soft tissues, which is inherent to CT. Immediately after injecting the contrast agent, the intrapulmonary vessels were well visualized as bright lines with an observed contrast enhancement of 104.16 HU over the baseline values. Contrast enhancement in the liver reached a maximum value of 156.14 HU over baseline at the 4-hour time point. At 24 hours after injection, the vessels returned to their original CT number values, and the contrast enhancement was located in the liver. Inside the liver, the vessels were still visible, but they were dark relative to the liver parenchyma (data not show). On the tumor micro-CT imaging, a similar enhancement pattern was observed in each mouse. The 10-minute
scans showed high attenuation of blood vessels in the liver, and 4-hour scans showed high attenuation of the liver parenchyma. Contrast enhancement of the liver and tumor at 10 minutes and 4 hours after contrast agent injection was measured as an attenuation difference (Table 1, Fig 3). The mean liver attenuation values were 126.3 ⫾ 25.15 and 191.14 ⫾ 34.24 HU for the 10 minute and 4 hour scans, respectively. The liver to lesion attenuation differences were 45.48 ⫾ 25.28 and 113.32 ⫾ 18.11 HU for the 10-minute and 4-hour scans, respectively. The mean inferior vena cava attenuation values were 139.16 ⫾ 20.79 and 79.88 ⫾ 10.01 HU for the 10-minute and 4-hour scans, respectively. Micro-CT Images Correlated with the Histopathology Micro-CT with the contrast agent enhancement showed a high capacity for detecting liver tumors in live mice. On micro-CT imaging, the tumor was detected from the
1285
KIM ET AL
Academic Radiology, Vol 15, No 10, October 2008
Figure 2. Serial micro– computed tomography (CT) images acquired after a hepatocyte-specific contrast agent injection. The baseline image before contrast agent injection shows the lack of contrast between soft tissues (a). On the micro-CT image with 10-minute scan, the pulmonary and hepatic vessels were well visualized as bright lines (b). Four-hour scan shows high attenuation of the liver parenchyma, which was reached a maximum value of contrast enhancement (c).
Table 1 Mean CT Number and Standard Deviation Measured in the Livers and Tumors of the Mice
Time (h) Precontrast 10 min 4h
Liver Parenchyma Tumor Mean ⫾ Statistical Mean ⫾ SD (HU) SD (HU) Analysis (P value) 77.8 ⫾ 5.76 126.3 ⫾ 25.15 191.14 ⫾ 34.24
71.7 ⫾ 7.73 80.82 ⫾ 21.72 84.05 ⫾ 14.28
.132062 .011595 .00021
CT: computed tomography; SD: standard deviation; HU: Hounsfield units.
seventh day after cell injection, and innumerable tumors were detected on days 10 –13 on micro-CT images (Fig 4). The measured size of the smallest tumor detected on micro-CT was 300 m. Small metastatic tumors less than 300 m were observed on the histologic examination, but these tumors were not detected on micro-CT imaging. A total of 74 separate tumor sites in the two mice sacrificed on the seventh day after cell injection was detected on the pathologic examination. Thirty-two lesions were larger than 500 m in diameter, and 42 lesions were less than 500 m and larger than 300 m. On micro-CT, 66 of the 74 tumors larger than 300 m were detected. There were eight false-negative readings on micro-CT. All of the eight undetected tumors were less than 500 m in size. Three tumors localized on the dome area and two tumors on the liver surface were undetected. Three tumors localized in the hilum of the portal vein or the perivascular areas were undetected. The overall detectability on micro-CT was 83.8% (Table 2). All the tumors larger than 500 m in size were detected on micro-CT
1286
Figure 3. Contrast enhancement of the liver parenchyma, the inferior vena cava (IVC), and the tumor at 10 minutes and 4 hours after contrast agent injection. The graph represents the significant difference of contrast enhancement between the liver parenchyma and tumor at 4 hours after contrast agent injection. HU, Housefield Units.
imaging, so the detectability of these tumors was 100% (Fig 5). The smallest tumor detected on micro-CT was 300 m in diameter (Fig 6). On the 1:1 correlation of the tumor between micro-CT and the histopathology, the mean tumor size as measured on micro-CT (621.77 ⫾ 216.37 m) was larger than that measured on histopathologic evaluation (494.12 ⫾ 197.09 m). Regression testing showed the intercept of the regression line was 0.12 (95% confidence interval ⫺0.8896 to 1.12) with a slope of 0.8373 (95% confidence interval 0.7413– 0.9332). The correlation between the CT scan
DETECTING LIVER METASTASIS IN LIVING MICE
Academic Radiology, Vol 15, No 10, October 2008
Figure 4. Micro– computed tomographic image at day 13 after cell injection. After injection of a hepatocyte-specific contrast agent, numerous low-density nodules in the liver (arrows) were detected (a). On gross morphology of the liver, innumerable metastatic nodules were found at the surface of the liver (b).
Table 2 Detectability of the Experimental Liver Tumors Tumor Size
No. on Micro–CT No. on Pathology Detectability
⬎500 m 300–500 m Total
32 34 66
32 42 74
100% 80.9% 83.8%
CT: computed tomography.
measurement and the actual measurement was r ⫽ 0.8354 (P ⬍ .0001) (Fig 7).
DISCUSSION This study focused on using hepatobiliary contrast agent enhanced micro-CT for the early detection and monitoring of liver metastases. Our results show that micro-CT using a hepatocyte-selective contrast agent was suitable for detecting and monitoring multiple liver tumors in live mice as early as 7 days after injecting the cancer cells. On the micro-CT images, the liver to tumor attenuation differences were higher than 100 HU when contrast enhancement in the liver reached a maximum value at the 4-hour time point. The hepatic blood vessels were able to be demarcated from the enhanced liver parenchyma. Micro-CT imaging with the hepatocyte-selective contrast agents is an effective means to detect liver metastases as small as 300 m, and all the tumors larger than 500 m in size tumors were detected on the
micro-CT imaging. The tumor size measured on the micro-CT images were well correlated with that measured on the histopathologic evaluation. Although micro-CT has a high resolution, the actual resolution might be lower in the in vivo studies. Unavoidable motion artifacts from animal breathing and internal organ movement are always present, and this does not allow achievement of the highest possible resolution. The smallest tumors might not be detected. For our results, eight tumors smaller than 500 m and all the tumors smaller than 300 m in size were not detected on the micro-CT scan. Tumor detectability could be increased using respiratory-gated micro-CT (9,10), but the radiation dose and acquisition time are also increased, which may not be suitable for serial monitoring. We used deep anesthesia with a high concentration of oxygen in the anesthetic gas mixture to minimize the cardiac and respiratory motion. Our data indicated that it was possible to obtain high-quality liver images and to accurately estimate the tumor size with the acquired non-respiratory gated micro-CT images. In general, comparatively long acquisition times are required for the highest resolution scans; thus, the rapid renal clearance of conventional water-soluble imaging agents precludes their use in this setting. The recently developed hepatocyte-selective contrast agent is a synthetic chylomicron remnant microemulsion that is transported selectively to hepatocytes via an apolipoprotein E receptor–mediated pathway. The contrast agent was shown to provide superior long-lasting visualization on
1287
KIM ET AL
Academic Radiology, Vol 15, No 10, October 2008
Figure 5. Micro– computed tomographic image (a) at day 7 after cell injection. After injection of a hepatocyte-specific contrast agent, a low-density nodule of 500 m in size is seen (arrow) in the right lobe of the liver, which was correlated with histopathology (b).
Figure 6. Micro– computed tomographic image (a) at day 7 after cell injection after injection of a hepatocyte-specific contrast agent shows a low-density nodule of 300 m in size (arrow) in the left lobe of the liver, which was correlated with the histopathology (b).
micro-CT imaging studies, and especially those of the liver. Because primary and secondary liver tumors are deficient in hepatic lipase, they should not take up the contrast agent like a normal liver, and so they will be seen as hypoattenuating on micro-CT. The 4-hour scans showed high attenuation of the liver parenchyma (191.14 ⫾ 34.24 HU) and high liver to lesion attenuation differences (113.32 ⫾ 18.11 HU). Using micro-CT with hepatocyteselective contrast material has advantages for lesion detection in a murine model, and particularly for small lesions, compared to conventional CT techniques, which
1288
include non-enhanced CT and helical CT with conventional water-soluble contrast materials. One advantage of our micro-CT system is the low radiation dose. In general, a total body radiation dose of 6 Gy is considered lethal for a small rodent (16). Cavanaugh et al (17) concluded that radiation damage is unlikely to occur with their measured 150 mGy per scan. Because it is unclear the way threshold x-ray irradiation affects organ and tissue metabolism, the x-ray exposure to the animal must be minimized to perform multiple sequential scanning. In our micro-CT system, the scan time
Academic Radiology, Vol 15, No 10, October 2008
Figure 7. Regression analysis shows that the correlation between the computed tomographic (CT) scan measurement and the actual measurement was r ⫽ 0.8354 (P ⬍ .001). Units: micrometers.
was 6 minutes and the estimated radiation dose was approximately 81.5 mGy using our imaging protocol. Other researchers reported that for the radiation doses used in their studies (77–93 mGy) when performing serial scans, there was no effect on tumor growth and no difference in mouse survival (18,19). The low radiation dose when performing micro-CT imaging is suitable for serial monitoring of tumor growth and long-term follow-up of therapeutic effect on tumors, without affecting tissue metabolism and tumor growth. One potential difficulty of micro-CT coupled with hepatocyte-selective contrast agent enhancement involves the ability to distinguish small blood vessels from hepatic tumor on the delayed time images. On the 4-hour and 24-hour point images after injection, the contrast agent has been cleared from the vasculature, as happened in our study. Because this contrast agent does not enhance tumor tissue, the intensity of a small lesion may be similar to that of blood vessels. On CT images, small circular hypodense areas may represent either a blood vessel in cross section or a focal lesion. In our study, we acquired dual phase micro-CT imaging at 10 minutes and 4 hours after contrast agent administration for detecting tumors. The 10-minute images showed good enhancement of the hepatic vessels, but poor enhancement of the liver parenchyma, which was not enough for detecting small tumors. When we reviewed the both time point images (ie, 10 minute and 4 hour) simultaneously, most of the tumors that were located in the peripheral perivascular area could be detected. Some researchers have introduced a method of administration of a second dose of contrast agent to
DETECTING LIVER METASTASIS IN LIVING MICE
opacify the hepatic vessels to distinguish them from the tumor (11). However, experiments with additional contrast agents were restricted to feasibility experiments, because injection of contrast agents represented a serious additional stress for the diseased mice. In our study, the mean tumor size measured on micro-CT was larger than that on the histopathologic evaluation (621.77 ⫾ 216.37 vs. 494.12 ⫾ 197.09 m, respectively). Because our results have referred to a visual correlation, it is obvious that measurement of the size of liver tumors in a living, breathing animal is not straightforward. An accurate estimate on micro-CT is only possible for motionless objects, which was not the case in our in vivo experiments. With micro-CT, the actual movement in every part of the cross section is not known. Tissue motion is much more pronounced in areas of the mouse that are involved with breathing. As a result, some of the tumors may become blurred and so look larger or smaller than their actual size. A major drawback of the classic histologic techniques is the interference of shrinkage as the extent of this cannot be quantified. Further, contours and shapes may become distorted by movement artifacts. Nevertheless, a good correlation between tumor size on the micro-CT and the histopathologic slices was evident in our study. Consequently, our results show that micro-CT is a useful tool to initially screen for liver tumors and to avoid technical processing of tissues for histopathology, and this is mainly because of micro-CT’s high detectability and that the modality is also noninvasive. An interesting future development would be to increase the speed of in vivo scanning of livers and to reduce the motion artifact in the scans. Another interesting improvement for scanning would be the possibility to acquire images at a higher resolution. Another important challenge is to increase the soft tissue contrast; this is mainly a problem of the reconstruction algorithm. For this latter issue, either the dose of radiation should be increased, or a more sensitive detector should be used. In conclusion, our study shows that high resolution micro-CT enhanced with hepatocyte-selective contrast agent can be a promising tool for detecting the early stages of tumor development in the liver of living mice. Micro-CT can be successfully applied for longterm studies on the growth, development, and drug treatment of liver tumors in mice, and for reducing the number of mice that are required for carcinogenicity studies.
1289
KIM ET AL
REFERENCES 1. Liu Y, Matsui O. Changes of intratumoral microvessels and blood perfusion during establishment of hepatic metastases in mice. Radiology 2007; 243:386 –395. 2. Gordon EM, Liu PX, Chen ZH, et al. Inhibition of metastatic tumor growth in nude mice by portal vein infusions of matrix-targeted retroviral vectors bearing a cytocidal cyclin G1 construct. Cancer Res 2000; 60:3343–3347. 3. Suh KW, Piantadosi S, Yazdi HA, et al. Treatment of liver metastases from colon carcinoma with autologous tumor vaccine expressing granulocyte-macrophage colony-stimulating factor. J Surg Oncol 1999; 72:218 –224. 4. Graham KC, Wirtzfeld LA, MacKenzie LT, et al. Three-dimensional high-frequency ultrasound imaging for longitudinal evaluation of liver metastases in preclinical models. Cancer Res 2005; 65:5231–5237. 5. Chang CH, Jan ML, Fan KH, et al. Longitudinal evaluation of tumor metastasis by an FDG-microPet/microCT dual-imaging modality in a lung carcinoma-bearing mouse model. Anticancer Res 2006; 26:159 – 166. 6. Paulus MJ, Gleason SS, Kennel SJ, et al. High resolution X-ray computed tomography: an emerging tool for small animal cancer research. Neoplasia 2000; 2:62–70. 7. Li XF, Zanzonico P, Ling CC, et al. Visualization of experimental lung and bone metastases in live nude mice by X-ray micro-computed tomography. Technol Cancer Res Treat 2006; 5:147–155. 8. Marx J. Imaging. Animal models: live and in color. Science 2003; 302:1880 –1882. 9. Maehara N. Experimental microcomputed tomography study of the 3D microangioarchitecture of tumors. Eur Radiol 2003; 13:1559 –1565.
1290
Academic Radiology, Vol 15, No 10, October 2008
10. Weichert JP, Lee FT, Jr, Longino MA, et al. Computed tomography scanning of hepatic tumors with polyiodinated triglycerides. Acad Radiol 1996; 3(suppl 2):S229 –S231. 11. Weichert JP, Lee FT, Jr, Chosy SG, et al. Combined hepatocyte-selective and blood-pool contrast agents for the CT detection of experimental liver tumors in rabbits. Radiology 2000; 216:865– 871. 12. Weichert JP, Longino MA, Spigarelli MG, et al. Computed tomography scanning of Morris hepatomas with liver-specific polyiodinated triglycerides. Acad Radiol 1996; 3:412– 417. 13. Ohta S, Lai EW, Morris JC, et al. MicroCT for high-resolution imaging of ectopic pheochromocytoma tumors in the liver of nude mice. Int J Cancer 2006; 119:2236 –2241. 14. Almajdub M, Nejjari M, Poncet G, et al. In-vivo high-resolution X-ray microtomography for liver and spleen tumor assessment in mice. Contrast Media Mol Imaging 2007; 2:88 –93. 15. Weber SM, Peterson KA, Durkee B, et al. Imaging of murine liver tumor using microCT with a hepatocyte-selective contrast agent: accuracy is dependent on adequate contrast enhancement. J Surg Res 2004; 119:41– 45. 16. Ritman EL. Micro-computed tomography-current status and developments. Annu Rev Biomed Eng 2004; 6:185–208. 17. Cavanaugh D, Johnson E, Price RE, et al. In vivo respiratory-gated micro-CT imaging in small-animal oncology models. Mol Imaging 2004; 3:55– 62. 18. Taschereau R, Chow PL, Chatziioannou AF. Monte carlo simulations of dose from microCT imaging procedures in a realistic mouse phantom. Med Phys 2006; 33:216 –224. 19. Winkelmann CT, Figueroa SD, Rold TL, et al. Microimaging characterization of a B16-F10 melanoma metastasis mouse model. Mol Imaging 2006; 5:105–114.
Rationale and Objectives. Micro-computed tomography (CT) is a important tool for longitudinal imaging of tumor development. The detection and monitoring of tumors in the liver in live animals using micro-CT is challenging. We evaluated the feasibility of high-resolution micro-CT enhanced with a hepatocyte-selective contrast agent for detecting liver metastases in a live murine model. Materials and Methods. Hepatic metastases were induced in 10 BALB/C mice. Two mice each were randomly selected on days 3, 5, 7, 10, and 13 after CT26 colon adenocarcinoma cells were injected into the portal vein; micro-CT imaging was performed at 10 minutes and 4 hours after intravenous administration of a hepatocyte-selective contrast agent at a dose of 0.4 mL/mouse. The attenuation values of the normal liver and the tumors were obtained. The number of metastases was counted and their sizes were measured on the micro-CT images. Gross or histopathologic evaluation was performed for correlating the liver tumors with the micro-CT images. Results. A total of 74 separate tumor sites larger than 300 m in diameter were detected on pathologic examination of the mice that were sacrificed 7 days after cell injection. On micro-CT, 66 of 74 tumors were detected (83.8%). The smallest tumor detected on micro-CT was 300 m. There were eight false-negative readings on micro-CT. The sizes of the individual liver metastases measured by micro-CT and on the excised specimen were highly correlated (P ⬍ .001). The correlation between the CT scan measurement and the actual measurement was r ⫽ 0.8354 (P ⬍ .0001). Conclusions. High-resolution micro-CT enhanced with a hepatocyte-selective contrast agent can be a promising tool for detecting liver metastases in a live murine model. Key Words. Contrast media; in vivo scanning; liver neoplasm; metastasis; X-ray microtomography. ©
AUR, 2008
The liver is a common site of metastases from various primary tumors, and so it is an important area of metastasis research (1). Rodent models that mimic the Acad Radiol 2008; 15:1282–1290 1
From the Departments of Radiology and the Institute for Radiological Imaging Science (H.W.K., Q.-Y.C., H.Y.J., K.S.C., S.H.P., K.-H.Y.), Surgery (S.J.B.), Internal Medicine (M.S.L.), Anatomy (J.M.O.), and Pathology (H.S.K.), Wonkwang University School of Medicine, 344-2 Sinyong-dong, Iksan, Jeonbuk 570-749, Korea. Supported by Wonkwang University in 2008. Received February 17, 2008; accepted March 19, 2008. Address correspondence to: K.-H.Y. e-mail: [email protected]
© AUR, 2008 doi:10.1016/j.acra.2008.03.021
1282
development of metastatic disease in the human liver are increasingly being recognized as powerful tools for the development of anticancer drugs and for evaluating the efficacy of novel therapeutics in the preclinical setting (2,3). Noninvasive longitudinal imaging of rodent models would decrease experimental variability and provide a more accurate assessment of metastatic progression and the efficacy of therapeutic interventions (4,5). Several longitudinal imaging modalities are now used, including magnetic resonance imaging, x-ray computed tomography (CT), positron emission tomography, and fluorescent and bioluminescent imaging.
Academic Radiology, Vol 15, No 10, October 2008
Micro-CT has been applied in biomedical research mainly for studying bones and lung disease, for which the natural contrast between bone and air and the surrounding soft tissues are provided (6,7). Micro-CT increased the spatial resolution noninvasively and it allowed very high precision for localization of bony changes. It has been proposed by some researchers that small-animal x-ray CT is an important tool for longitudinal imaging of tumor development, for visualizing blood vessels and angiogenesis, and for following the response of tumors to preclinical therapeutic intervention (8,9). However, the detection and monitoring of tumor in the liver in live animals using micro-CT is still challenging because of the poor natural contrast between the tumor and the liver parenchyma. When performing micro-CT imaging for the liver of small animals, it is not possible to use the conventional water-soluble contrast agents that are used in humans because they clear rapidly from the blood. For the detection of experimental liver tumors, a hepatocyte-specific contrast agent (1,3-bis [7-(3-amino2,4,6-triiodophenyl) heptanoyl]-2-oleoyl-glycerol) can be used to overcome this limitation (10). Previous reports have demonstrated the efficacy of clinical CT using the hepatocyte-specific contrast agent for the detection of liver metastasis in rabbits and rats (11,12). However, the capability of micro-CT to detect liver metastasis tumors in live mice is largely unknown. Ohta et al (13) and Almajdub et al (14) have recently suggested that micro-CT can be used to detect small size liver tumors in mice. Weber et al (15) have evaluated the ability of micro-CT using a hepatocyte-specific contrast agent for detecting and monitoring liver tumors in murine hepatic tumor models. In these studies, the ability of micro-CT using a hepatocyte-specific contrast agent for detecting small liver metastasis has not been directly assessed, even if these studies to date support the possibility of using micro-CT for detecting small liver tumors in mice. Therefore, the feasibility of using micro-CT with a hepatocyte-selective contrast agent has not been currently established for detecting early liver metastasis in a small animal model. We hypothesized that micro-CT with adequate contrast enhancement will allow detection of early liver metastasis in a mouse model. Therefore, we studied the feasibility of high-resolution micro-CT enhanced with a hepatocyteselective contrast agent for detecting experimental liver metastases at an early stage in live mice.
DETECTING LIVER METASTASIS IN LIVING MICE
METHODS AND MATERIALS Tumor Model Thirteen male Balb/c mice (Central Lab. Animal Inc, Seoul, Korea), age 6 weeks, were used for this study. The mice ranged in weight from 22 to 25 g. All the mice lived in a system equipped with day–night light cycling and the mice were provided with standard mouse chow. The CT-26 murine colon adenocarcinoma cell line was acquired from ATCC (Manassas, VA). The cells were grown as monolayer cultures in DMEM-10% fetal bovine serum (Invitrogen, Grand Island, NY) supplemented with 1% glutamine, and 1% antibiotics. The cells were maintained in a 37°C incubator with 5% CO2-humidified air. To simulate liver metastases, 10 mice were anesthetized with intraperitoneally injected ketamine (100 g/g body weight) and xylazine (10 g/g body weight). CT-26 cells (5 ⫻ 103 in 0.1 mL phosphate-buffered saline) were injected into the mesenteric vein using a 30-gauge needle after laparotomy. All the animal studies were carried out in accordance with the regulations set forth by the institutional review board of our university.
Micro-CT System The imaging was performed using a volumetric CT scanner (NFR-Polaris-G90; NanoFocusRay, Iksan, Korea), which is a micro-CT system designed to acquire high resolution imaging of live mice. The micro-CT system is composed primarily of an anode x-ray source with a 5-m focal spot, a complementary metal oxide semiconductor flat panel detector coupled with a thallium-doped cesium iodide scintillator, a linearly moving couch, a rotational gantry coupled with a positioning encoder, and a parallel processing system for the image data. The images were acquired at 50 kVp, 65 A, and 470 ms/frame, and there were 360 views. The estimated radiation dose was approximately 84 mGy using this image acquisition protocol. The spatial resolution of the micro-CT system was calculated as 44.25 m at 10% point on the modulation transfer factor. The images were reconstructed using the Feldkamp cone-beam reconstruction algorithm. The reconstruction image size was 1024 ⫻ 1024 pixels, and 512 slices were acquired. The final reconstructed data were converted to the Digital Imaging and Communication in Medicine format to create three-dimensional–rendered imaging using three-dimensional–rendering software (Lucion, Infinite, Seoul, Korea).
1283
KIM ET AL
Micro-CT Imaging The mice were placed in an induction chamber with 4% isoflurane in oxygen to induce anesthesia. During imaging, the mice were kept under anesthesia with 1.5% isoflurane in oxygen and they were allowed to recover between image acquisitions. Three BALB/C mice were imaged at the baseline and then serial micro-CT images were obtained immediately and at 1, 2, 4, 8, and 24 hours after bolus injection of the 150-nm particle diameter-iodinated triglyceride contrast agent (150 mg of iodine per kilogram of body weight; Fenestra LC ART, Toronto, ON, Canada) at a dose of 0.4 mL/mouse via a distal tail vein. These experimental data were used as the reference for the micro-CT imaging protocol for the detection of liver metastasis. Two mice were randomly selected on days 3, 5, 7, 10, and 13 after tumor cell injection, and micro-CT imaging was performed after intravenous administration of the contrast agent. Each mouse was injected with the contrast agent at a dose of 0.4 mL/mouse into the distal tail vein using a 30-gauge needle. Based on the CT images of the control mice, we acquired the micro-CT imaging of the tumor models at 10 minutes and 4 hours after contrast agent administration. All CT image data were acquired in live, free-breathing, anesthetized mice without respiratory or cardiac gating. Image Analysis To assess the enhancement achieved with the contrast agent, the CT signal intensity was expressed in Hounsfield units (HU). The mean CT number was determined in regions of interest that measured 1.5 mm in diameter (⬃700 pixels) on various slices in the liver, inferior vena cava, and tumors after injection of the contrast agent. The mean CT number and standard deviation were collected for each region of interest for all the images and these values were then compared for each time point in all mice. Micro-CT images of the tumor model were co-evaluated by two observers, who were blinded to the histopathologic results. The 10-minute and 4-hour images were evaluated simultaneously to avoid false-positive tumor detection. The observers were given a diagram that depicted the representative axial slices of mouse hepatic anatomy. The observers identified a suspected lesion by drawing it on the appropriate image. The number of metastases was counted and their size was measured on the micro-CT images. The diameter was defined as the longest distance across the tumor.
1284
Academic Radiology, Vol 15, No 10, October 2008
Histopathologic Correlation The mice were euthanized after micro-CT imaging. The abdomen was opened, the portal vein was cannulated for perfusion with 10% buffered formalin in situ, and the livers were fixed for 24 hours. The fixed livers were removed and then embedded in paraffin. The whole livers were sectioned axially at 4-m thickness with a 100-m interval and then we performed standard hematoxylin and eosin staining, and the tumors were measured microscopically. Correlation of the histopathologic characteristics of the tumors with the CT images was performed. Statistical Analysis Statistical analysis was performed using SPSS software (version 11.5, SPSS, Inc., Chicago, IL). Comparisons of the tumor size were done using paired t-tests. Evaluation of the correlation of the sizes of the tumors on micro-CT images with the sizes of the tumors, as determined on histopathologic examination, was done via regression analysis. For all the statistical analyses, a P value less than .05 was considered to indicate a statistically significant difference.
RESULTS Histopathologic Findings of the Metastatic Liver Tumors All the animals survived long enough to complete the full protocol. Mesenteric vein injection of 5 ⫻ 103 CT-26 cells in the BALB/C mice reproducibly generated multiple liver tumors in all the animals. At day 3 after cell injection, no tumor was found on 40⫻ magnification, and some necrosis areas were found on the peripheral areas of the mice’s livers (Fig 1a). At day 5, after cell injection, the necrotic areas had shrunk a good deal and some small tumors (smaller than 300 m) were observed, and two tumors larger than 300 m were found in one mouse liver (Fig 1b). At day 7 after cell injection, the necrotic areas were completely dissolved and were no longer observed on the histopathologic images. A total of 74 tumors larger than 300 m were observed (Fig 1c). At days 10 and 13 after cell injection, numerous sized tumors smaller than 300 m were observed (Fig 1d). The small size tumors were not included in the data analysis because measuring the number and size of these tumors was actually impossible. There were more tumors localized in left lobe than in the right lobe, and most of the tumors were localized in the peripheral portion of the liver.
Academic Radiology, Vol 15, No 10, October 2008
DETECTING LIVER METASTASIS IN LIVING MICE
Figure 1. Photographs of the hematoxylin and eosin staining of the hepatic metastasis murine model (40⫻ magnification). (a) Some necrotic areas were found on the peripheral areas of the livers at day 3 after cell injection. (b) A small tumor (⬍300 m) was observed and the necrotic areas had shrunk at day 5 after cell injection. (c) Discrete tumors larger than 300 m were observed and the necrotic areas were completely dissolved at day 7 after cell injection. (d) Multiple innumerable tumors were observed at day 10 after cell injection.
Liver Parenchyma and Tumor Attenuation on the Micro-CT Images The CT imaging appearance of three normal BALB/C mice was evaluated as a function of time after injection of the contrast agent (Fig 2). The baseline image shows the lack of contrast between soft tissues, which is inherent to CT. Immediately after injecting the contrast agent, the intrapulmonary vessels were well visualized as bright lines with an observed contrast enhancement of 104.16 HU over the baseline values. Contrast enhancement in the liver reached a maximum value of 156.14 HU over baseline at the 4-hour time point. At 24 hours after injection, the vessels returned to their original CT number values, and the contrast enhancement was located in the liver. Inside the liver, the vessels were still visible, but they were dark relative to the liver parenchyma (data not show). On the tumor micro-CT imaging, a similar enhancement pattern was observed in each mouse. The 10-minute
scans showed high attenuation of blood vessels in the liver, and 4-hour scans showed high attenuation of the liver parenchyma. Contrast enhancement of the liver and tumor at 10 minutes and 4 hours after contrast agent injection was measured as an attenuation difference (Table 1, Fig 3). The mean liver attenuation values were 126.3 ⫾ 25.15 and 191.14 ⫾ 34.24 HU for the 10 minute and 4 hour scans, respectively. The liver to lesion attenuation differences were 45.48 ⫾ 25.28 and 113.32 ⫾ 18.11 HU for the 10-minute and 4-hour scans, respectively. The mean inferior vena cava attenuation values were 139.16 ⫾ 20.79 and 79.88 ⫾ 10.01 HU for the 10-minute and 4-hour scans, respectively. Micro-CT Images Correlated with the Histopathology Micro-CT with the contrast agent enhancement showed a high capacity for detecting liver tumors in live mice. On micro-CT imaging, the tumor was detected from the
1285
KIM ET AL
Academic Radiology, Vol 15, No 10, October 2008
Figure 2. Serial micro– computed tomography (CT) images acquired after a hepatocyte-specific contrast agent injection. The baseline image before contrast agent injection shows the lack of contrast between soft tissues (a). On the micro-CT image with 10-minute scan, the pulmonary and hepatic vessels were well visualized as bright lines (b). Four-hour scan shows high attenuation of the liver parenchyma, which was reached a maximum value of contrast enhancement (c).
Table 1 Mean CT Number and Standard Deviation Measured in the Livers and Tumors of the Mice
Time (h) Precontrast 10 min 4h
Liver Parenchyma Tumor Mean ⫾ Statistical Mean ⫾ SD (HU) SD (HU) Analysis (P value) 77.8 ⫾ 5.76 126.3 ⫾ 25.15 191.14 ⫾ 34.24
71.7 ⫾ 7.73 80.82 ⫾ 21.72 84.05 ⫾ 14.28
.132062 .011595 .00021
CT: computed tomography; SD: standard deviation; HU: Hounsfield units.
seventh day after cell injection, and innumerable tumors were detected on days 10 –13 on micro-CT images (Fig 4). The measured size of the smallest tumor detected on micro-CT was 300 m. Small metastatic tumors less than 300 m were observed on the histologic examination, but these tumors were not detected on micro-CT imaging. A total of 74 separate tumor sites in the two mice sacrificed on the seventh day after cell injection was detected on the pathologic examination. Thirty-two lesions were larger than 500 m in diameter, and 42 lesions were less than 500 m and larger than 300 m. On micro-CT, 66 of the 74 tumors larger than 300 m were detected. There were eight false-negative readings on micro-CT. All of the eight undetected tumors were less than 500 m in size. Three tumors localized on the dome area and two tumors on the liver surface were undetected. Three tumors localized in the hilum of the portal vein or the perivascular areas were undetected. The overall detectability on micro-CT was 83.8% (Table 2). All the tumors larger than 500 m in size were detected on micro-CT
1286
Figure 3. Contrast enhancement of the liver parenchyma, the inferior vena cava (IVC), and the tumor at 10 minutes and 4 hours after contrast agent injection. The graph represents the significant difference of contrast enhancement between the liver parenchyma and tumor at 4 hours after contrast agent injection. HU, Housefield Units.
imaging, so the detectability of these tumors was 100% (Fig 5). The smallest tumor detected on micro-CT was 300 m in diameter (Fig 6). On the 1:1 correlation of the tumor between micro-CT and the histopathology, the mean tumor size as measured on micro-CT (621.77 ⫾ 216.37 m) was larger than that measured on histopathologic evaluation (494.12 ⫾ 197.09 m). Regression testing showed the intercept of the regression line was 0.12 (95% confidence interval ⫺0.8896 to 1.12) with a slope of 0.8373 (95% confidence interval 0.7413– 0.9332). The correlation between the CT scan
DETECTING LIVER METASTASIS IN LIVING MICE
Academic Radiology, Vol 15, No 10, October 2008
Figure 4. Micro– computed tomographic image at day 13 after cell injection. After injection of a hepatocyte-specific contrast agent, numerous low-density nodules in the liver (arrows) were detected (a). On gross morphology of the liver, innumerable metastatic nodules were found at the surface of the liver (b).
Table 2 Detectability of the Experimental Liver Tumors Tumor Size
No. on Micro–CT No. on Pathology Detectability
⬎500 m 300–500 m Total
32 34 66
32 42 74
100% 80.9% 83.8%
CT: computed tomography.
measurement and the actual measurement was r ⫽ 0.8354 (P ⬍ .0001) (Fig 7).
DISCUSSION This study focused on using hepatobiliary contrast agent enhanced micro-CT for the early detection and monitoring of liver metastases. Our results show that micro-CT using a hepatocyte-selective contrast agent was suitable for detecting and monitoring multiple liver tumors in live mice as early as 7 days after injecting the cancer cells. On the micro-CT images, the liver to tumor attenuation differences were higher than 100 HU when contrast enhancement in the liver reached a maximum value at the 4-hour time point. The hepatic blood vessels were able to be demarcated from the enhanced liver parenchyma. Micro-CT imaging with the hepatocyte-selective contrast agents is an effective means to detect liver metastases as small as 300 m, and all the tumors larger than 500 m in size tumors were detected on the
micro-CT imaging. The tumor size measured on the micro-CT images were well correlated with that measured on the histopathologic evaluation. Although micro-CT has a high resolution, the actual resolution might be lower in the in vivo studies. Unavoidable motion artifacts from animal breathing and internal organ movement are always present, and this does not allow achievement of the highest possible resolution. The smallest tumors might not be detected. For our results, eight tumors smaller than 500 m and all the tumors smaller than 300 m in size were not detected on the micro-CT scan. Tumor detectability could be increased using respiratory-gated micro-CT (9,10), but the radiation dose and acquisition time are also increased, which may not be suitable for serial monitoring. We used deep anesthesia with a high concentration of oxygen in the anesthetic gas mixture to minimize the cardiac and respiratory motion. Our data indicated that it was possible to obtain high-quality liver images and to accurately estimate the tumor size with the acquired non-respiratory gated micro-CT images. In general, comparatively long acquisition times are required for the highest resolution scans; thus, the rapid renal clearance of conventional water-soluble imaging agents precludes their use in this setting. The recently developed hepatocyte-selective contrast agent is a synthetic chylomicron remnant microemulsion that is transported selectively to hepatocytes via an apolipoprotein E receptor–mediated pathway. The contrast agent was shown to provide superior long-lasting visualization on
1287
KIM ET AL
Academic Radiology, Vol 15, No 10, October 2008
Figure 5. Micro– computed tomographic image (a) at day 7 after cell injection. After injection of a hepatocyte-specific contrast agent, a low-density nodule of 500 m in size is seen (arrow) in the right lobe of the liver, which was correlated with histopathology (b).
Figure 6. Micro– computed tomographic image (a) at day 7 after cell injection after injection of a hepatocyte-specific contrast agent shows a low-density nodule of 300 m in size (arrow) in the left lobe of the liver, which was correlated with the histopathology (b).
micro-CT imaging studies, and especially those of the liver. Because primary and secondary liver tumors are deficient in hepatic lipase, they should not take up the contrast agent like a normal liver, and so they will be seen as hypoattenuating on micro-CT. The 4-hour scans showed high attenuation of the liver parenchyma (191.14 ⫾ 34.24 HU) and high liver to lesion attenuation differences (113.32 ⫾ 18.11 HU). Using micro-CT with hepatocyteselective contrast material has advantages for lesion detection in a murine model, and particularly for small lesions, compared to conventional CT techniques, which
1288
include non-enhanced CT and helical CT with conventional water-soluble contrast materials. One advantage of our micro-CT system is the low radiation dose. In general, a total body radiation dose of 6 Gy is considered lethal for a small rodent (16). Cavanaugh et al (17) concluded that radiation damage is unlikely to occur with their measured 150 mGy per scan. Because it is unclear the way threshold x-ray irradiation affects organ and tissue metabolism, the x-ray exposure to the animal must be minimized to perform multiple sequential scanning. In our micro-CT system, the scan time
Academic Radiology, Vol 15, No 10, October 2008
Figure 7. Regression analysis shows that the correlation between the computed tomographic (CT) scan measurement and the actual measurement was r ⫽ 0.8354 (P ⬍ .001). Units: micrometers.
was 6 minutes and the estimated radiation dose was approximately 81.5 mGy using our imaging protocol. Other researchers reported that for the radiation doses used in their studies (77–93 mGy) when performing serial scans, there was no effect on tumor growth and no difference in mouse survival (18,19). The low radiation dose when performing micro-CT imaging is suitable for serial monitoring of tumor growth and long-term follow-up of therapeutic effect on tumors, without affecting tissue metabolism and tumor growth. One potential difficulty of micro-CT coupled with hepatocyte-selective contrast agent enhancement involves the ability to distinguish small blood vessels from hepatic tumor on the delayed time images. On the 4-hour and 24-hour point images after injection, the contrast agent has been cleared from the vasculature, as happened in our study. Because this contrast agent does not enhance tumor tissue, the intensity of a small lesion may be similar to that of blood vessels. On CT images, small circular hypodense areas may represent either a blood vessel in cross section or a focal lesion. In our study, we acquired dual phase micro-CT imaging at 10 minutes and 4 hours after contrast agent administration for detecting tumors. The 10-minute images showed good enhancement of the hepatic vessels, but poor enhancement of the liver parenchyma, which was not enough for detecting small tumors. When we reviewed the both time point images (ie, 10 minute and 4 hour) simultaneously, most of the tumors that were located in the peripheral perivascular area could be detected. Some researchers have introduced a method of administration of a second dose of contrast agent to
DETECTING LIVER METASTASIS IN LIVING MICE
opacify the hepatic vessels to distinguish them from the tumor (11). However, experiments with additional contrast agents were restricted to feasibility experiments, because injection of contrast agents represented a serious additional stress for the diseased mice. In our study, the mean tumor size measured on micro-CT was larger than that on the histopathologic evaluation (621.77 ⫾ 216.37 vs. 494.12 ⫾ 197.09 m, respectively). Because our results have referred to a visual correlation, it is obvious that measurement of the size of liver tumors in a living, breathing animal is not straightforward. An accurate estimate on micro-CT is only possible for motionless objects, which was not the case in our in vivo experiments. With micro-CT, the actual movement in every part of the cross section is not known. Tissue motion is much more pronounced in areas of the mouse that are involved with breathing. As a result, some of the tumors may become blurred and so look larger or smaller than their actual size. A major drawback of the classic histologic techniques is the interference of shrinkage as the extent of this cannot be quantified. Further, contours and shapes may become distorted by movement artifacts. Nevertheless, a good correlation between tumor size on the micro-CT and the histopathologic slices was evident in our study. Consequently, our results show that micro-CT is a useful tool to initially screen for liver tumors and to avoid technical processing of tissues for histopathology, and this is mainly because of micro-CT’s high detectability and that the modality is also noninvasive. An interesting future development would be to increase the speed of in vivo scanning of livers and to reduce the motion artifact in the scans. Another interesting improvement for scanning would be the possibility to acquire images at a higher resolution. Another important challenge is to increase the soft tissue contrast; this is mainly a problem of the reconstruction algorithm. For this latter issue, either the dose of radiation should be increased, or a more sensitive detector should be used. In conclusion, our study shows that high resolution micro-CT enhanced with hepatocyte-selective contrast agent can be a promising tool for detecting the early stages of tumor development in the liver of living mice. Micro-CT can be successfully applied for longterm studies on the growth, development, and drug treatment of liver tumors in mice, and for reducing the number of mice that are required for carcinogenicity studies.
1289
KIM ET AL
REFERENCES 1. Liu Y, Matsui O. Changes of intratumoral microvessels and blood perfusion during establishment of hepatic metastases in mice. Radiology 2007; 243:386 –395. 2. Gordon EM, Liu PX, Chen ZH, et al. Inhibition of metastatic tumor growth in nude mice by portal vein infusions of matrix-targeted retroviral vectors bearing a cytocidal cyclin G1 construct. Cancer Res 2000; 60:3343–3347. 3. Suh KW, Piantadosi S, Yazdi HA, et al. Treatment of liver metastases from colon carcinoma with autologous tumor vaccine expressing granulocyte-macrophage colony-stimulating factor. J Surg Oncol 1999; 72:218 –224. 4. Graham KC, Wirtzfeld LA, MacKenzie LT, et al. Three-dimensional high-frequency ultrasound imaging for longitudinal evaluation of liver metastases in preclinical models. Cancer Res 2005; 65:5231–5237. 5. Chang CH, Jan ML, Fan KH, et al. Longitudinal evaluation of tumor metastasis by an FDG-microPet/microCT dual-imaging modality in a lung carcinoma-bearing mouse model. Anticancer Res 2006; 26:159 – 166. 6. Paulus MJ, Gleason SS, Kennel SJ, et al. High resolution X-ray computed tomography: an emerging tool for small animal cancer research. Neoplasia 2000; 2:62–70. 7. Li XF, Zanzonico P, Ling CC, et al. Visualization of experimental lung and bone metastases in live nude mice by X-ray micro-computed tomography. Technol Cancer Res Treat 2006; 5:147–155. 8. Marx J. Imaging. Animal models: live and in color. Science 2003; 302:1880 –1882. 9. Maehara N. Experimental microcomputed tomography study of the 3D microangioarchitecture of tumors. Eur Radiol 2003; 13:1559 –1565.
1290
Academic Radiology, Vol 15, No 10, October 2008
10. Weichert JP, Lee FT, Jr, Longino MA, et al. Computed tomography scanning of hepatic tumors with polyiodinated triglycerides. Acad Radiol 1996; 3(suppl 2):S229 –S231. 11. Weichert JP, Lee FT, Jr, Chosy SG, et al. Combined hepatocyte-selective and blood-pool contrast agents for the CT detection of experimental liver tumors in rabbits. Radiology 2000; 216:865– 871. 12. Weichert JP, Longino MA, Spigarelli MG, et al. Computed tomography scanning of Morris hepatomas with liver-specific polyiodinated triglycerides. Acad Radiol 1996; 3:412– 417. 13. Ohta S, Lai EW, Morris JC, et al. MicroCT for high-resolution imaging of ectopic pheochromocytoma tumors in the liver of nude mice. Int J Cancer 2006; 119:2236 –2241. 14. Almajdub M, Nejjari M, Poncet G, et al. In-vivo high-resolution X-ray microtomography for liver and spleen tumor assessment in mice. Contrast Media Mol Imaging 2007; 2:88 –93. 15. Weber SM, Peterson KA, Durkee B, et al. Imaging of murine liver tumor using microCT with a hepatocyte-selective contrast agent: accuracy is dependent on adequate contrast enhancement. J Surg Res 2004; 119:41– 45. 16. Ritman EL. Micro-computed tomography-current status and developments. Annu Rev Biomed Eng 2004; 6:185–208. 17. Cavanaugh D, Johnson E, Price RE, et al. In vivo respiratory-gated micro-CT imaging in small-animal oncology models. Mol Imaging 2004; 3:55– 62. 18. Taschereau R, Chow PL, Chatziioannou AF. Monte carlo simulations of dose from microCT imaging procedures in a realistic mouse phantom. Med Phys 2006; 33:216 –224. 19. Winkelmann CT, Figueroa SD, Rold TL, et al. Microimaging characterization of a B16-F10 melanoma metastasis mouse model. Mol Imaging 2006; 5:105–114.