The Role of Tumor Vascularity in Predicting Survival after Yttrium-90 Radioembolization for Liver Metastases

Transcription

1 The Role of Tumor Vascularity in Predicting Survival after Yttrium-90 Radioembolization for Liver Metastases Kent T. Sato, MD, Reed A. Omary, MD, Christopher Takehana, MD, Saad Ibrahim, MD, Robert J. Lewandowski, MD, Robert K. Ryu, MD, and Riad Salem, MD, MBA PURPOSE: It is unclear what role pretreatment tumor vascularity plays in determining outcomes after yttrium-90 radioembolization. A hypothesis was tested that radiographic vascularity of a tumor does not affect patient survival. MATERIALS AND METHODS: In this two-institution retrospective study, 137 patients with metastatic liver disease underwent 90 Y radioembolization. Primary sites were categorized as colon, neuroendocrine, and other. All patients underwent triphasic contrast-enhanced computed tomography (CT) or magnetic resonance imaging, as well as detailed hepatic angiography. Two board-certified interventional radiologists interpreted all images and evaluated them for the presence of enhancement. Median survival times, as well as 1- and 2-year survival rates, were compared between patients with hypervascular and hypovascular tumors on (i) cross-sectional imaging and (ii) angiography with use of the log-rank statistic ( 0.05). RESULTS: On angiography, 108 patients had hypervascular tumors and 29 had hypovascular tumors. Median survival times for the two subgroups were 300 days and 261 days, respectively (P.95). On CT, 24 patients had hypervascular tumors and 113 had hypovascular tumors. Median survival times for these subgroups were 306 days and 284 days, respectively (P.67). Eighty-four patients tumors that were hypovascular on CT were hypervascular on angiography. There were no statistical differences in survival between patients with hypervascular and hypovascular tumors, regardless if vascularity was defined based on CT or angiography. CONCLUSIONS: Radiographic vascular appearance of liver tumors, regardless of imaging modality, does not affect survival after radioembolization. Therefore, hypovascular tumors should not be considered contraindicated for radioembolization. J Vasc Interv Radiol 2009; 20: From the Department of Radiology, Northwestern University, 251 East Huron, Room 4-710R Feinberg, Chicago, IL Received November 10, 2008; final revision received August 21, 2009; accepted August 24, Address correspondence to K.T.S.; k-sato@northwestern.edu From the 2008 SIR Annual Meeting. R.S. serves as a paid consultant for MDS Nordion. None of the other authors have identified a conflict of interest. SIR, 2009 DOI: /j.jvir METASTATIC liver disease is the most common form of hepatic malignancy, as the liver is a common site of metastasis for many tumors. Surgical resection is the only curative treatment for metastatic liver disease, but fewer than 20% of patients are candidates for surgery (1). For patients with unresectable disease, systemic chemotherapy is available; however, when these therapies fail, there is no consensus on how to treat these patients. Liver-directed therapies such as transarterial chemoembolization, bland embolization, and radioembolization with yttrium-90 microspheres have been used for patients in whom first- and second-line therapies fail. Radioembolization has been shown to be beneficial and well tolerated for metastatic disease (2,3), and its use for the treatment of liver metastasis is steadily increasing. Radioembolization relies on two fundamental principles: (i) liver tumors receive the majority of their blood supply from the hepatic artery (4) and (ii) tumor neovascularity is more dense than native hepatic vasculature and therefore allows the radioembolic particles to be concentrated within the tumor relative to normal liver tissue. Therefore, radioembolization theoretically depends on the hypervascularity of the tumors, with the goals to maximize concentration of microspheres within the tumor and minimize radiation exposure to surrounding normal hepatic parenchyma. With the use of digital subtraction angiography or cross-sectional imaging, contrast agent enhancement of liver metastases can vary. Depending on 1564

2 Volume 20 Number 12 Sato et al 1565 the primary tumor, liver metastases can be hypervascular or hypovascular. The rich blood supply of hypervascular tumors suggests that they would concentrate the 90 Y microspheres more and should therefore show a better response than hypovascular tumors. However, radioembolization has been shown to be effective in the treatment of a variety of metastases (5 7) regardless of vascularity. This suggests that there may be more to vascularity than enhancement alone. It is unclear what role pretreatment tumor vascularity plays in determining outcomes after 90 Y radioembolization. In this study, we tested the hypothesis that hypovascular liver metastases, as defined by angiography or cross-sectional imaging (ie, computed tomography [CT] or magnetic resonance [MR] imaging), are not associated with reduced survival after 90 Y radioembolization compared with hypervascular tumors. MATERIALS AND METHODS Clinical Setting and Patient Sample Between 2003 and 2006, 137 patients with chemotherapy-refractory disease presenting with liver-dominant hepatic metastases from various primary malignancies were treated at two institutions. Institution A was a large communitybased hospital with a large oncology practice. Institution B was a large academic practice, also with a large volume of oncology procedures. Pretreatment imaging consisted of a CT examination whenever possible, but when patients could not undergo contrast-enhanced CT or presented to us with an MR imaging study on initial consultation, CT was not performed. Survival and follow-up data from both institutions were obtained and included in this analysis. The Institutional Review Boards from both institutions approved the study protocol, and all patients signed an informed consent allowing use of their data. All data were stored in a fashion compliant with the Health Insurance Portability and Accountability Act. The sites of the primary malignancy were categorized as colon, neuroendocrine, and noncolorectal/nonneuroendocrine (ie, other ). This group included patients with the following primary malignancies: breast, pancreas, lung, cholangiocarcinoma, melanoma, renal, esophageal, ovarian, adenocarcinoma of unknown origin, lymphoma, gastric, duodenal, bladder, angiosarcoma, squamous cell, thyroid, adrenal, and parotid. CT Technique All CT scans were performed on a Lightspeed scanner (GE Healthcare, Princeton, New Jersey) or a Somatom 16-slice scanner (Siemens Medical Solutions, Erlangen, Germany). A triphasic contrast-enhanced liver protocol was employed that used a standard power injector (Medrad, Indianola, Pennsylvania) connected to an intravenous catheter to inject 125 ml of Omnipaque 350 (GE Healthcare) at a rate of 4 5 ml/sec. Arterial-phase scans were obtained 40 seconds after the start of the intravenous injection and venous-phase imaging was performed 90 seconds after the start of injection. Angiography Technique Angiography was performed with Angiostar, Multistar, or Axiom Artis equipment (Siemens Medical Solutions). Hepatic arterial angiography was performed selectively in the right or left hepatic artery. Selective arterial catheterization was performed with a Renegade Hi-Flo microcatheter (Boston Scientific, Natick, Massachusetts). Injection rates varied depending on the caliber of the hepatic artery as well as the level of flow within it but were 1 3 ml/sec as injected through a power injector (Medrad, Indianola, Pennsylvania). The contrast agent used was Omnipaque 350 (GE Healthcare, Milwaukee, Wisconsin), and a typical volume of 6 12 ml was used per injection, depending on the size of the liver being examined. Radioembolization Technique Radioembolization was performed with TheraSphere (MDS Nordion, Ottawa, Ontario, Canada). TheraSphere has received a Humanitarian Device Exemption from the United States Food and Drug Administration for use in hepatocellular carcinoma. For this study, all patients received permission for off-label use per the published Food and Drug Administration guidelines for Humanitarian Device Exemptions in disease conditions other than the approved indication. Depending on the extent of disease within the liver at the time of presentation, patients received bilobar (ie, right and left lobe) or unilobar treatment. For bilobar treatment, the first lobe treatment was followed by the second lobe treatment approximately 30 days later. No whole-liver infusions were performed. Lobar treatment was targeted to deliver an absorbed dose of 120 Gy according to previously published dosimetry techniques (8). Infusion of TheraSphere was performed by one of four interventional radiologists (K.T.S., R.J.L., R.K.R., R.S.) with 5, 2, 8, and 8 years of experience, respectively. For those patients who received bilobar treatment, imaging was repeated just before the second lobar treatment (30 45 days) and again at the third month. Patients undergoing unilobar treatment were scheduled for 30-day imaging, which was repeated every 3 months until the patient s death. Laboratory and clinical parameters (including lymphocyte counts) were assessed at each clinic visit and classified according the National Cancer Institute Common Toxicity Criteria, version 3.0. Procedural complications were classified according to the standards established by the Society of Interventional Radiology (9). All patients were prescribed a 2-week course of proton pump inhibitors. Nondiabetic patients received a 5-day methylprednisolone dose pack to counteract fatigue. Data Analysis Two board-certified attending interventional radiologists retrospectively reviewed and interpreted the pretreatment CT and angiographic images simultaneously and evaluated them for the presence of enhancement by consensus. Sixteen patients did not have a pretreatment contrast-enhanced CT study because of allergies or renal dysfunction and underwent MR imaging. For the sake of data analysis, these patients were grouped with the CT group and evaluated based on the same criteria. Tumors were classified as hypervascular if they enhanced greater than the surrounding hepatic parenchyma on imaging. Tumors were deemed hypervascular if they showed a distinct tumor blush relative to the normal hepatic parenchyma on angiography. The primary endpoint of the study was survival specific to the groups defined as having hypervascular and hypovascular tumors on both imaging modalities. Me-

3 1566 Tumor Vascularity to Predict Survival after Liver Radioembolization December 2009 JVIR Figure 1. Images from a 45-year-old woman with metastatic neuroendocrine tumors who is being evaluated for treatment with 90 Y. (a) CT scan of the liver obtained during the arterial phase shows a hypervascular mass (arrow) within the left lobe of the liver. (b) Selective left hepatic arteriography shows multiple hypervascular enhancing masses (arrowheads) corresponding with the masses seen on CT. Figure 2. Images from a 67-year-old man with metastatic rectal carcinoma. (a) CT scan shows a large hypodense mass (arrow) in the right lobe of the liver, which does not enhance relative to the liver parenchyma during the arterial phase. (b) Selective common hepatic arteriography does not demonstrate any appreciable enhancement in the large mass. dian days of survival, as well as 1- and 2-year survival percentages, were compared between the hypervascular and hypovascular imaging groups as well as the hypervascular and hypovascular angiography groups using the log-rank statistic, with an value of Additionally, a subgroup analysis was performed for median days of survival in patients with hypervascular or hypovascular tumors on angiography and imaging scans with use of 2 analyses, with an value of RESULTS Patient Demographics One hundred thirty-seven patients underwent a total of 225 administrations of 90 Y TheraSphere via transcatheter intraarterial infusion. The median age of the patients was 61 years, and there were 66 men and 71 women in the patient cohort. Primary malignancies included colon (n 51), breast (n 21), neuroendocrine (n 19), cholangiocarcinoma (n 7), pancreas (n 6), lung (n 5), melanoma (n 5), renal (n 4), esophageal (n 3),

4 Volume 20 Number 12 Sato et al 1567 Figure 3. Images from a 59-year-old man with metastatic sigmoid colon carcinoma. (a) Arterial-phase CT does not clearly demonstrate a lesion, which is in the region of the arrowheads. (b) Equilibrium-phase image, which was windowed for more contrast, shows an area of hypodensity relative to the normal liver (arrowheads). (c) Selective right hepatic angiography shows a hypervascular enhancing mass in the right lobe of the liver (arrowheads). adenocarcinoma with unknown primary tumor (n 3), ovary (n 2), and adrenal, angiosarcoma, bladder, cervical, duodenal, gallbladder, gastric, lymphoma, parotid, squamous cell carcinoma, and thyroid (n 1 each). Cross-sectional Imaging Classification A total of 137 patients were studied. Of those, 121 had contrast-enhanced CT and 16 had contrast-enhanced MR imaging. On cross-sectional imaging, 24 of 137 patients (18%) were found to have hypervascular tumors (Fig 1) and 113 (82%) had hypovascular tumors (Fig 2). Eighty-four of the 113 patients (74%) whose tumors were deemed hypovascular on cross-sectional imaging were found to have hypervascular tumors on angiography (Fig 3). Digital Subtraction Angiographic Classification Of the 137 patients, 108 (79%) had lesions that appeared hypervascular on angiography, whereas only 29 (21%) had tumors that appeared hypovascular. Further evaluation of the data revealed that 29 of 137 patients (21%) had tumors that were hypovascular on both angiography and crosssectional imaging, whereas 24 patients (18%) had tumors that were found to be hypervascular on both imaging modalities. Survival After radioembolization, the angiographically hypervascular tumor group exhibited a 1-year survival rate of 48.5% and a 2-year survival rate of 25.9%. The median survival time for the hypervascular tumor group was 300 days. For the angiographically hypovascular tumor group, the 1-year survival rate was similar at 47.9% and the 2-year survival rate was 32.8%. The median survival time for this group was 261 days. There were no statistical differences in survival between patients with hypervascular tumors and those with hypovascular tumors, regardless of the modality used (P.05).

5 1568 Tumor Vascularity to Predict Survival after Liver Radioembolization December 2009 JVIR Table 1 Survival Outcomes among Patients with Hypervascular and Hypovascular Tumors Vascularity On cross-sectional imaging, only 18% of the patients exhibited hypervascular tumors, whereas 82% of tumors were considered hypovascular. After radioembolization, results were very similar to the angiography group. Patients with hypervascular tumors on cross-sectional imaging had a 1-year survival rate of 47.6% and a 2-year survival rate of 30.5%. The median survival rate for the hypervascular tumor group was 306 days. Patients with hypovascular tumors exhibited a 48.3% 1-year survival rate and a 31.1% 2-year survival rate. The median survival time for this group was 284 days (Table 1). In addition, the subgroup analysis of combined tumor vascularity as defined by angiography and cross-sectional imaging also demonstrated no statistical difference in survival between hypervascular and hypovascular tumors. Of the 137 patients, 24 exhibited tumor hypervascularity on cross-sectional imaging and angiography, and had a median survival of 305 days. Twenty-nine patients had hypovascular tumors on cross-sectional imaging and angiography, and had a median survival of 224 days (P.57; Table 2). DISCUSSION No. of Pts. Mean Survival (d) Median Survival (d) Overall Survival (%) On angiography Hypervascular Hypovascular On CT/MR imaging Hypervascular Hypovascular y 2y P Value Table 2 Median Survival after 90 Y Radioembolization as a Function of CT/MR Imaging Tumor Vascularity Vascularity No. of Pts. Median Survival (d)* Hypervascular Hypovascular * P value between groups:.3227 ( 2 test) In an attempt to refine selection criteria for patients undergoing radioembolization, identification of pretreatment variables that may predict the response of a tumor to radioembolization becomes crucial. In the present study of 137 patients, we found that the enhancement patterns of tumors did not appear to affect survival of these patients after radioembolization. Specifically, patients with hypervascular tumors did not have a significant survival advantage versus those with hypovascular tumors. The mechanism of action for radioembolization is flow-directed deposition of the radioactive microspheres into the vascular bed of liver tumors. Because of this, one would expect that patients with hypervascular tumors should have the radioactive microspheres concentrate to a greater degree in the tumors compared with those with hypovascular tumors, and therefore have a better response. However, because it has been shown that both hypovascular and hypervascular tumors respond well to radioembolization (8), our goal was to study the role of vascularity in the success of radioembolization. As there is no accepted objective measure of enhancement, the determination of tumor vascularity on crosssectional imaging and catheter angiography was subjectively determined by two interventional radiologists. Our results show that the response of tumors to radioembolization is independent of enhancement patterns. In addition, enhancement on one imaging modality does not correlate with enhancement on another modality. These confounding data further demonstrate the difficulty in using enhancement as a measure of tumor perfusion. Our results may in fact be secondary to an inadequate means to measure true tumor perfusion, indicating a discrepancy between actual tumor perfusion and radiographic enhancement. These findings were echoed in a study by Taniai et al (10), who examined the response of liver metastases to chemoembolization. Similarly to our study, they characterized metastatic colorectal tumors as hypervascular and hypovascular by angiography and contrastenhanced CT. They found that, although patients with tumors with early angiographic hypervascularity showed a slight trend toward improved survival, there was no significant difference in survival between patients hypervascular and hypovascular tumors. These findings raise the question of how much tumor enhancement really correlates with tumor vascularity. Tumor enhancement depends not only on tumor blood flow, blood volume, and contrast agent concentration, but also on vessel permeability. In some tumors, newly developed tumor vessels demonstrated increased fenestrations in the basement membrane that results in increased permeability (11). This increase in permeability is one reason these tumors exhibit intense enhancement on cross-sectional imaging. However, not all neovascularity demonstrates this high permeability, so some tumors may possess a network of new vasculature, but the tumors may not enhance as much as others. Recently, methods of measuring tissue perfusion other than tissue enhancement have been evaluated. Tissue perfusion is a reflection of overall blood flow to a volume of tissue over time. Contrast enhancement is a phenomenon that relies on two main factors. First, there must be perfusion of the tissue in question to carry the contrast material to the target tissue. Second, there is diffusion of the contrast material across the cell membranes, which allows it to be detected. Even with dynamic contrast scans, images are obtained from one point in time; not in a

6 Volume 20 Number 12 Sato et al 1569 dynamic fashion over a period of time. Transcatheter intraarterial perfusion MR imaging is a new technique that has been used to measure such objective tumor perfusion changes during transarterial chemoembolization (12 15). High perfusion values may represent not only a higher microvessel density, but also intrinsically greater angiogenic activity of the tumor as a secondary response to tissue hypoxia. This is significant in the setting of radioembolization, as tissue hypoxia has also been shown to offer resistance to radiation therapy (16). Hypoxia is known to upregulate vascular endothelial growth factor and hypoxia-inducible factor 1 (17 19). If a hypoxic tumor were to upregulate vascular endothelial growth factor expression, this may result in higher membrane permeability and the tumor may appear hypervascular on imaging because of the increased enhancement seen. Because not all tumors produce vascular endothelial growth factor, this is likely not the only factor affecting enhancement, but it does suggest that neovascularity by itself does not necessarily correspond with contrast enhancement. Further study into innovative methods of evaluating tissue perfusion may help us understand our findings better. There were several limitations to this study. First, the method of defining hypervascular and hypovascular tumors based on different imaging techniques was subjective. A more rigorous objective method such as transcatheter intraarterial perfusion MR imaging (15) would be preferable, but this technique is still not widely available. Second, our cohort of patients included patients with many different primary tumors. Regardless of vascularity, the underlying malignancy may have more of an effect on survival than vascularity. In the future, further subgroup analysis could be performed within each of the major tumor groups comparing vascularity and survival. Finally, although the determination of vascularity was based on consensus opinion by two boardcertified attending interventional radiologists, no intra- or interobserver variability was tested. In conclusion, radiographic vascular appearance of liver tumors, regardless of imaging modality, does not appear to significantly affect survival after treatment with radioembolization. Therefore, tumor hypovascularity should not be considered a contraindication for 90 Y radioembolization. References 1. Sasson AR, Sigurdson ER. Surgical treatment of liver metastases. Semin Oncol 2002; 29: Mulcahy MF, Lewandowski RJ, Ibrahim SM, et al. Radioembolization of colorectal hepatic metastases using yttrium-90 microspheres. Cancer 2009; 115: Lewandowski RJ, Thurston KG, Goin JE, et al. 90Y Microsphere (Thera- Sphere) treatment for unresectable colorectal cancer metastases of the liver: response to treatment at targeted doses of Gy as measured by [18F]fluorodeoxyglucose positron emission tomography and computed tomographic imaging. J Vasc Interv Radiol 2005; 16: Breedis C, Young G. The blood supply of neoplasms in the liver. Am J Pathol 1954; 30: Herba MJ, Thirlwell MP. Radioembolization for hepatic metastases. Semin Oncol 2002; 29: Wong CY, Salem R, Raman S, et al. Evaluating 90Y-glass microsphere treatment response of unresectable colorectal liver metastases by [18F]FDG PET: a comparison with CT or MRI. Eur J Nucl Med Mol Imaging 2002; 29: Stubbs RS, Cannan RJ, Mitchell AW. Selective internal radiation therapy (SIRT) with 90yttrium microspheres for extensive colorectal liver metastases. Hepatogastroenterology 2001; 48: Salem R, Thurston KG. Radioembolization with yttrium-90 microspheres: a state-of-the-art brachytherapy treatment for primary and secondary liver malignancies. Part 3: comprehensive literature review and future direction. J Vasc Interv Radiol 2006; 17: Omary RA, Bettmann MA, Cardella JF, et al. Quality improvement guidelines for the reporting and archiving of interventional radiology procedures. J Vasc Interv Radiol 2003; 14(suppl):S293 S Taniai N, Onda M, Tajiri T, et al. Good embolization response for colorectal liver metastases with hypervascularity. Hepatogastroenterology 2002; 49: Dvorak HF, Brown LF, Detmar M, Dvorak AM. Vascular permeability factor/vascular endothelial growth factor, microvascular hyperpermeability, and angiogenesis. Am J Pathol 1995; 146: Wang D, Bangash AK, Rhee TK, et al. Liver tumors: monitoring embolization in rabbits with VX2 tumors transcatheter intraarterial first-pass perfusion MR imaging. Radiology 2007; 245: Virmani S, Wang D, Harris KR, et al. Comparison of transcatheter intraarterial perfusion MR imaging and fluorescent microsphere perfusion measurements during transcatheter arterial embolization of rabbit liver tumors. J Vasc Interv Radiol 2007; 18: Lewandowski RJ, Wang D, Gehl J, et al. A comparison of chemoembolization endpoints using angiographic versus transcatheter intraarterial perfusion/mr imaging monitoring. J Vasc Interv Radiol 2007; 18: 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: Harada H, Kizaka-Kondoh S, Li G, et al. Significance of HIF-1-active cells in angiogenesis and radioresistance. Oncogene 2007; 26: Byrne AM, Bouchier-Hayes DJ, Harmey JH. Angiogenic and cell survival functions of vascular endothelial growth factor (VEGF). J Cell Mol Med 2005; 9: Hicklin DJ, Ellis LM. Role of the vascular endothelial growth factor pathway in tumor growth and angiogenesis. J Clin Oncol 2005; 23: Rhee TK, Young JY, Larson AC, et al. Effect of transcatheter arterial embolization on levels of hypoxia-inducible factor-1alpha in rabbit VX2 liver tumors. J Vasc Interv Radiol 2007; 18: