is based on the fact that HCCs are exclusively supplied by the hepatic artery. When a tumor is advanced in stage

Transcription

1 Hepatocellular Carcinoma with Internal Mammary Artery Supply: Feasibility and Efficacy of Transarterial Chemoembolization and Factors Affecting Patient Prognosis Hyo-Cheol Kim, MD, Jin Wook Chung, MD, Seung Hong Choi, MD, Jung-Hwan Yoon, MD, Hyo-Suk Lee, MD, Hwan Jun Jae, MD, Whal Lee, MD, and Jae Hyung Park, MD PURPOSE: To determine technical feasibility, therapeutic efficacies, and identify prognostic factors in patients with hepatocellular carcinoma (HCC) treated by transarterial chemoembolization via the internal mammary artery. MATERIALS AND METHODS: From August 1996 to July 2005, the authors identified the internal mammary arteries supplying HCCs in 97 (2.2%) of 4,438 HCC patients. Computed tomography scans and digital subtraction angiography images of these 97 patients were retrospectively reviewed by consensus between two of the authors regarding technical success and clinical outcome. The technical success of internal mammary artery chemoembolization was defined as achievement of catheterization into feeding vessels, delivery of drugs via those vessels, and no residual tumor staining fed by the internal mammary artery. Multivariate Cox proportional hazard regression analysis was performed to enable evaluation of prognostic factors for survival. RESULTS: Technical success by selective chemoembolization via internal mammary artery was achieved in 53 (55%) of the 97 patients. The clinical responses of the 97 patients were complete remission (n 3), partial remission (n 19), no response (n 65), and no available follow-up image (n 10). Overall cumulative survival rates, calculated from times of internal mammary artery chemoembolization, were 55.9% (at 6 months), 32.5% (at 1 year), 15.4% (at 2 years), and 8.2% (at 3 years). Multivariate analysis showed that tumor multiplicity (single vs multiple nodular or diffuse; P.013), portal vein thrombosis (P.004), and the technical success of internal mammary artery chemoembolization (P.02) each significantly affected survival. CONCLUSIONS: Although patients with HCC supplied by the internal mammary artery had poor prognosis because of their advanced stage of disease, chemoembolization via the internal mammary artery is possible and worth performing in selective situations. J Vasc Interv Radiol 2007; 18: Abbreviations: HCC hepatocellular carcinoma, TACE transarterial chemoembolization From the Department of Radiology, Seoul National University College of Medicine, Institute of Radiation Medicine, Seoul National University Medical Research Center, and Clinical Research Institute, Seoul National University Hospital, No. 28 Yongon-dong, Chongnogu, Seoul, , Republic of Korea (H-C.K., J.W.C., S.H.C., H.J.J., W.L., J.H.P.); and Department of Internal Medicine and Liver Research Institute, Seoul National University College of Medicine, Seoul, Republic of Korea (J-H.W., H-S.L.). Received July 15, TRANSARTERIAL chemoembolization (TACE) is widely performed to treat hepatocellular carcinomas (HCCs) and particularly for the palliation of unresectable HCCs (1,2). Chemoembolization is based on the fact that HCCs are exclusively supplied by the hepatic artery. When a tumor is advanced in stage 2006; final revision received February 5, 2007; accepted February 19, Address correspondence to J.W.C.; chungjw@radcom.snu.ac.kr None of the authors have identified a conflict of interest. SIR, 2007 DOI: /j.jvir or is located adjacent to the hepatic ligaments or bare area, however, various extrahepatic collateral pathways can develop (3 10). Because extrahepatic collateral supplies can prohibit the effective control of the tumor by TACE, we should treat the tumor via extrahepatic collateral pathway or choose other available treatment modalities such as radiofrequency ablation. The internal mammary artery is one of the extrahepatic collateral vessels that supply the liver and may serve as a feeding artery when a HCC is located in ventral hepatic areas, abutting the diaphragm and anterior abdomi- 611

2 612 Hepatocellular Carcinoma with Internal Mammary Artery Supply May 2007 JVIR nal wall. Since an initial case report was issued by Kim et al (11) in 1995, several reports have been issued on internal mammary artery involvement in HCC (12 15). However, these are either case reports or a single study involving only 10 patients. During the past 9 years, we have encountered 97 patients with a HCC fed by the internal mammary artery and have attempted to treat using TACE via the internal mammary artery. In this study, we determined the technical success rates and therapeutic efficacies of TACE and attempted to identify prognostic factors in patients treated with TACE via the internal mammary artery. MATERIALS AND METHODS This retrospective study was approved by our institutional review board, and the need to obtain informed consent was waived. Patients From August 1996 to July 2005, 15,295 sessions of TACE were performed in 4,438 HCC patients at our institution. In 97 (2.2%) of these 4,438 patients, we found that the internal mammary arteries supplied the HCC. These 97 patients included 76 men and 21 women with ages ranging from 19 to 79 years (mean, 55 years). Methods of Chemoembolization and Follow-up All patients received enhanced biphasic helical computed tomography (CT) scans of the liver before TACE. All TACE procedures were performed by one of two authors (one, 14 years experience; the other, 25 years experience). The TACE method used has been described in detail elsewhere (4,5,16). Briefly, the procedure was commenced by infusing emulsion of 2 12 ml of iodized oil (Lipiodol; Andre Gurbet, Aulnay-sous-Bois, France) and mg of doxorubicin hydrochloride (Adriamycin RDF; Ildong Pharmaceutical, Seoul, Korea) until arterial flow stasis was achieved and/or iodized oil appeared in portal branches. If a patient had a tumor supplied by both the hepatic and extrahepatic collateral arteries, the chemotherapeutic agent was divided between the two arterial distributions. If initial blockade of the hepatic artery was insufficient because of a large mass size or arterioportal shunting, embolization was performed using absorbable gelatin sponge particles (Gelfoam; Upjohn, Kalamazoo, Mich) 1 2 mm in diameter soaked in a mixture of 4 6 mg of crystalline mitomycin (Mitomycin-10; Kyowa Hakko Kogyo, Tokyo, Japan) and 10 ml of nonionic contrast medium. When CT and angiographic findings suggestive of an extrahepatic arterial supply to the tumor, described in detail elsewhere (4,5), were present, we determined the potential extrahepatic collateral arteries based on tumor location and performed selective angiography of all collateral vessels suspected of supplying the tumor. We catheterized the internal mammary artery at the following situation: first, when a HCC was located in the ventral hepatic area and its staining on hepatic angiogram had a focal defect; second, when iodized oil infused at a previous TACE session had not accumulated in the ventral portion of the tumor on CT scan; third, even though a viable tumor abutting the diaphragm was located in dorsal hepatic portion on CT scan, its correspondent tumor staining was not found by the hepatic and inferior phrenic angiography. Selective angiography of internal mammary arteries was performed using 5-F catheter to visualize entire artery anatomies. When superselective catheterization was achieved by placing a 3-F microcatheter tip as close as possible to the specific branch or branches supplying a tumor, iodized oil and doxorubicin hydrochloride emulsion was infused. When superselective catheterization of the tumor feeder was not possible, chemoembolization was not performed in most cases, but embolization using gelatin sponge (Gelfoam) or polyvinyl alcohol particle (Contour; Boston Scientific, Natick, Mass) was performed in selective cases (Fig 1). Unenhanced CT through entire livers was performed 1 3 weeks after TACE to assess livers for traces of iodized oil, and enhanced biphasic helical CT was performed 1 2 weeks before a next scheduled TACE session to assess the efficacy of previous treatment and to detect residual viable or recurrent tumors. Although a 3-month interval between TACE sessions is our routine protocol, these repeat sessions were tailored to patient tolerance to the procedure and tumor response. Study Definitions Data were recorded in an electronic database (Microsoft Access; Microsoft, Redmond, Wash) immediately after angiography by one of two of the authors (J.W.C., J.H.P.): these records included clinical response since a previous TACE session, the presence of an extrahepatic collateral artery supplying a tumor, an anatomical description of the hepatic artery and extrahepatic collateral artery, the presence of hepatic artery stenosis or occlusion caused by repeated TACE, the amount of iodized oil and anticancer agents used, the identity of the artery infused with iodized oil and anticancer agent, complications such as skin rash, residual tumor staining after TACE, and the presence of portal or hepatic vein invasion by a tumor. When chemoembolization was not performed despite the presence of tumor staining by hepatic angiography or selective angiography of extrahepatic collateral arteries, the reason why we did not perform chemoembolization was also recorded. CT scans and digital subtraction angiographs of 97 patients were retrospectively reviewed by consensus between the two interventional radiologists who had performed all TACE procedures. Tumors were classified as single nodular, multinodular, or diffuse. Tumor size was defined as the largest tumor diameter on transverse CT scans. Tumors were assigned to liver segments according to the Couinaud classification (17). If a tumor occupied two or more segments, its location was assigned to the dominant segment. For multinodular tumors, the location and size of tumor that was supplied by the internal mammary artery were recorded. Tumor feeders were classified as pericardiacophrenic artery, musculophrenic artery, superior epigastric artery, anterior intercostal artery, ensiform artery, and phrenic branch. The proportion of blood supplied by the internal mammary artery was classified as exclusive supply (100% of tumor staining supplied by the internal mammary artery) (Fig 2), dominant (above 50%), and supplementary (below 50%). In 29 pa-

3 Volume 18 Number 5 Kim et al 613 Figure 1. A 62-year-old man with hepatocellular carcinoma supplied by the right internal mammary artery. (a) Transverse CT scan shows ill-defined enhancing lesion (arrowhead) in liver with previously uptaken iodized oil. (b) Posteroanterior internal mammary arteriogram shows ill-defined tumor staining supplied by the fine branches of the musculophrenic artery (arrows). Note the tip (arrowhead) of the microcatheter. Embolization using polyvinyl alcohol particle was performed, and no cutaneous complication developed. tients who received repeated TACE via the internal mammary artery, we included the data obtained when the internal mammary artery feeding a tumor was first detected. We defined the technical success of internal mammary artery chemoembolization as the successful catheterization of all feeding vessels, the successful delivery of treatment drugs, and no residual tumor staining from the internal mammary artery on angiography. Technical failure was defined as either failure of vessel catheterization, presence of residual tumor staining, or no attempt to catheterize the internal mammary artery due to multiple fine tumor feeders. Clinical response, evaluated by imaging, was assessed at subsequent enhanced CT and angiographic follow-ups and was classified as complete remission (tumor disappearance), partial remission (indicating greater than a 50% reduction in the product of the two maximum perpendicular dimensions of a viable tumor on CT scan and concomitant reduction in tumor vascularity by hepatic angiography), or no response (a stationary or aggravated state). For patients with multiple lesions, the cross-product of several indicator lesions was added to categorize the patient s tumor response. Long-term clinical outcomes for all patients were obtained from medical records and by telephone interviews. Statistical Analysis Cumulative survival rates were estimated using the Kaplan-Meier method and were compared using the log rank test. The following factors were included in univariate Cox regression analyses of survival: age, sex, tumor size, tumor multiplicity, proportion of blood supplied by the internal mammary artery, hepatic artery stenosis or occlusion, portal vein thrombosis, hepatic vein invasion, metastasis, other concurrent extrahepatic collateral arteries, previous hepatic resection, Child-Pugh class, technical success of internal mammary artery chemoembolization, and the presence of any residual tumor staining after TACE. In these analyses, patient age and tumor size were treated as continuous variables and all others as categorical variables. Multivariate Cox proportional hazard regression analyses using a backward likelihood ratio method were used to identify multivariate predictors of patient survival. We included factors that were significant or showed a trend toward statistical significance (P.1) in the univariate analysis. A two-tailed P value of.05 or less was considered to indicate statistical significance, and statistical computer software (SPSS 10.0; SPSS, Chicago, Ill) was used. RESULTS Tumor sizes ranged from 1.5 cm to 20 cm (mean, 8.3 cm 4.5). Tumor multiplicity was of the single nodular (n 25), multinodular (n 50), or diffuse (n 22). Tumors were located in S4 (n 54), S8 (n 12), S3-2 (n 24), S7 (n 1), a pericardial lymph node (n 3), abdominal wall (metastasis, n 2), and diaphragm (metastasis, n 1). Involvement of the internal mammary artery was detected at the 1st to 26th (mean, 5.4; median, 4) TACE session, but during the 1st session in 16 (16%) patients. Time from initial TACE to internal mammary arteriogram ranged from 0 to 162 months (mean, 23.8 months; median, 10 months). The proportion of blood supplied by the internal mammary artery was classified as exclusive supply (n 16), dominant (n 19), and supplementary (n 62). Celiac stenosis or occlusion was present in six (6%) patients, but the hepatic artery was widely patent at initial TACE in all 97 patients. In 72 (74%) patients, the hepatic artery was still widely patent when the internal mammary arteriogram was obtained.

4 614 Hepatocellular Carcinoma with Internal Mammary Artery Supply May 2007 JVIR Figure 2. A 64-year-old man with hepatocellular carcinoma supplied by the right internal mammary artery. (a) Transverse CT scan shows 6-cm-sized enhancing mass (arrowheads) in liver S8. (b) Posteroanterior celiac angiogram shows no tumor staining (arrowheads) supplied by the hepatic artery. (c) Posteroanterior internal mammary arteriogram shows exclusive blood supply from the internal mammary artery (arrow). Note lateral costal branch (arrowhead) and no demonstrable pericardiacophrenic artery.

5 Volume 18 Number 5 Kim et al 615 Table 1 Technical Aspect of Chemoembolization via the Internal Mammary Artery in 97 Patients with Hepatocellular Carcinoma Results Number of Patients Attempt to chemoembolize tumors via the internal mammary artery 79 Technical success 53 Technical failure 26 Residual tumor staining supplied by the tiny branches of internal mammary artery, despite selective 14 embolization of main feeders Incomplete chemoembolization due to systemic arterial-to-pulmonary venous shunt 5 Embolization by gelatin sponge particle of multiple fine feeders 2 Embolization by polyvinyl alcohol particle of multiple fine feeders 2 Unselective chemoembolization 1 Reverse blood flow caused by arterial spasm 1 Failure of tumor feeder selection 1 No attempt to chemoembolize tumors via the internal mammary artery 18 Minimal tumor staining supplied by multiple fine tumor feeders 17 Severe systemic arterial-to-pulmonary venous shunt 1 Note. Technical success means successful catheterization into feeding vessels, successful delivery of treatment drugs, and no residual tumor staining supplied by the internal mammary artery. Table 2 Clinical Response in 97 Patients with Hepatocellular Carcinoma Clinical Response Complete Remission Partial Remission No Response Follow-up Loss Technical aspect Technical success Technical failure or no attempt Blood supply Exclusive supply Dominant supply Supplementary supply Total Total In six (6%) patients, the right hepatic (n 3) or left hepatic (n 3) arteries were occluded by repeated TACE. In 19 (20%) patients, the severe stenosis or occlusion of segmental hepatic artery was noted when the internal mammary arteriogram was obtained. In 33 (34%) patients, other extrahepatic collateral arteries were observed to supply HCCs at this TACE session: right inferior phrenic artery (n 21), left inferior phrenic artery (n 2), intercostal artery (n 8), omental branch (n 4), adrenal artery (n 2), right gastric artery (n 4), and left gastric artery (n 3). All 97 patients underwent TACE via the hepatic artery and/or internal mammary artery. Technical success with selective chemoembolization via the internal mammary artery was achieved in 53 (55%) patients. The detailed technical aspects of TACE via the internal mammary artery are summarized in Table 1. Cutaneous complications such as a skin rash occurred in seven (9%) patients. The embolized vessels in these seven patients were superior epigastric artery (n 3), anterior intercostal artery (n 2), and phrenic branch (n 2). Seven patients with a skin rash were treated by observation, and no patient experienced skin necrosis. The clinical responses of the 97 patients were complete remission (n 3), partial remission (n 19), no response (n 65), and no available follow-up image (n 10). The clinical responses were classified according to the technical success and blood supply in Table 2. Follow-up period ranges from 1 to 125 months (mean, 10.8 months; median, 6 months). In November 2005, 12 patients were alive and being followed up. Two patients had no evidence of a viable tumor on CT and maintained normal -fetoprotein levels at 43- and 127-month follow-up, respectively (Fig 3). Overall cumulative survival rates, calculated from times of internal mammary artery chemoembolization, were 55.9% (at 6 months), 32.5% (at 1 year), 15.4% (at 2 years), and 8.2% (at 3 years). Influences of patient- and tumorrelated factors on survival are shown in Table 3 and Table 4. Multivariate analysis was performed using the following variables: tumor size, tumor multiplicity, proportion of blood supplied by the internal mammary artery, portal vein thrombosis, technical success of internal mammary artery chemoembolization, and any residual tumor staining after TACE. Multivariate analysis showed that survival rates were dependent on tumor multiplicity (single vs multiple nodular or diffuse; P.013), portal vein thrombosis (P.004), and the technical success of in-

6 616 Hepatocellular Carcinoma with Internal Mammary Artery Supply May 2007 JVIR Figure 3. A 34-year-old woman with hepatocellular carcinoma supplied by the right internal mammary artery. (a) Transverse CT scan shows 8-cm-sized mass lesion (arrowheads) in the ventral liver. (b) Posteroanterior celiac angiogram shows tumor staining (arrowheads) supplied by the hepatic artery. (c) Posteroanterior internal mammary arteriogram shows ill-defined tumor staining supplied by the phrenic branch (arrow). (d) Transverse CT scan after 120 months shows shrunk tumor with iodized oil retention (arrowheads). Stomach (S) filled with oral contrast media. ternal mammary artery chemoembolization (P.02). DISCUSSION Among 4,438 patients with HCC, this study enrolled 97 (2.2%) patients whose tumors were supplied by the internal mammary artery. In 72 (74%) patients, the hepatic artery was still widely patent when the internal mammary arteriogram was obtained. Technically successful chemoembolization via the internal mammary artery was possible in 55% and clinical response (complete or partial remission) at a subsequent TACE session was achieved in 23%. Three variables (tumor multiplicity, portal vein thrombosis, and

7 Volume 18 Number 5 Kim et al 617 Table 3 Patients and TACE Characteristics: Hazards Ratios Based on Univariate Cox Models of Patient Survival Characteristics Number Hazard Ratio 95% Confidence Interval P Value Sex 0.870, Male Female Age , Tumor size , Tumor multiplicity 0.363, Single nodular Multinodular or diffuse Grade of blood supply by the internal mammary artery 0.397, Minimal supply Dominant or exclusive supply Hepatic artery stenosis or occlusion 0.549, Absent Present Portal vein thrombosis 1.266, Absent Present Hepatic vein invasion 0.519, Absent Present Metastasis 0.851, Absent Present Concurrent other extrahepatic collateral arteries 0.768, Absent Present Previous hepatic resection 0.495, Absent Present Child-Pugh class 0.670, A B or C Technical success of internal mammary artery chemoembolization 0.374, Failure Success Any residual tumor staining after TACE 0.279, Absent Present Table 4 Multivariate Analysis of Prognostic Factors with Cox Proportional Hazards Model Variable Hazard Ratio 95% Confidence Interval P Value Tumor type , Portal vein invasion , Technical success of internal mammary artery chemoembolization , technical success of internal mammary artery chemoembolization) were found to be independent prognostic indicators at multivariate analysis. In this study, the mean number of previous TACEs was 4.4, and the median time from initial TACE to internal mammary angiography was 10 months. Nakai et al (15) reported a mean number of previous TACEs of 5.8, and a median time from initial TACE to internal mammary angiography of 31.5 months. In this study, 25 (26%) of 97 patients had occlusion or severe stenosis of the hepatic artery caused by repeated TACE. However, Nakai et al (15) reported occlusion or severe stenosis of the hepatic artery in nine (90%) of 10 patients. Miyayama et al (8) found that the main cause of blood supply development from the omental branch was attenuation or occlusion of the hepatic artery due to repeated TACE, which was observed in 80% of patients. This indicates that the internal mammary artery supplying a tumor was detected at an early

8 618 Hepatocellular Carcinoma with Internal Mammary Artery Supply May 2007 JVIR stage in our study, probably because we performed selective angiography of all suspected collateral vessels. We believe that hepatic artery occlusion or severe stenosis can increase the likelihood of a collateral blood supply to a tumor by the internal mammary artery, but it is evident that tumor location is a more important factor of the collateral blood supply development because the hepatic artery was widely patent in 74% of patients in the current study. Nakai et al (15) advocated that the internal mammary artery more frequently supplies tumors in patients who survive for a prolonged period and who have undergone a large number of TACEs. We concur, but we also detected an internal mammary artery supply at the first TACE session in 16 (16%) patients. Thus, the possibility of an extrahepatic collateral supply should be borne in mind during even initial TACE in patients with a patent hepatic artery. The well-known complication of internal mammary artery embolization is a precordial skin rash or necrosis. These are probably caused by insufficient distal catheterization of the artery or the inadvertent regurgitation of embolizing material into other arteries (18,19). In our series, a skin rash occurred in seven patients. Anterior intercostal or superior epigastric arteries were embolized in five of these seven, both of which possessed branches that fed the skin, indicating that selective embolization is needed in such cases. Skin rashes occurred in two patients in whom phrenic branches were embolized, and these were attributed to chemotherapeutic agent reflux into main artery. Thus, selection of a tumor feeder and incremental agent injection are mandatory to prevent cutaneous complications. Advances in catheter and guidewire technologies now enable a microcatheter to be introduced into the small branches of extrahepatic arteries. In cases where the safety of chemoembolization is equivocal (ie, tumors are supplied by multiple tiny feeders that cannot be catheterized by microcatheter), bland embolization without the use of chemotherapeutic drugs may be a less risky alternative. In the current study, multiple small feeders in four patients were embolized with polyvinyl alcohol or gelatin sponge particles without cutaneous complication. In another patient, small feeders were embolized nonselectively, and this resulted in a skin rash. Recently, drug-eluting beads are available and could be used without cutaneous complication in cases with multiple tiny feeders. We did not attempt chemoembolization via the internal mammary artery in 18 patients, in whom we observed severe shunt or small tumor staining supplied by the internal mammary artery despite huge viable tumor supplied by the hepatic artery. In these 18 patients, we performed TACE via only the hepatic artery, because we thought that additional TACE through the internal mammary artery had no benefit for tumor control, and selective TACE was difficult or impossible due to multiple tiny feeders. Blood supplied by the internal mammary artery was exclusive in 16 patients. TACE in these patients must be performed via that vessel to expect any response, particularly in single nodular tumor. Exclusive supply to tumors, which are small recurrent ones at the periphery of the liver, usually occurs when multiple sessions of TACE via the hepatic artery cause significant stenosis or occlusion of the hepatic artery. In these cases, hepatic angiography does not reveal tumor staining that is present on CT scan. In case of multinodular tumor, only tumors supplied by the hepatic artery could be chemoembolized and ones supplied by the internal mammary artery could be missed, unless careful review of CT scan and angiography was taken. Thus, correlation between CT scan and hepatic angiography is essential to detection of extrahepatic collateral supply. Clinical response at subsequent TACE sessions was complete or partial remission in only 23%. The relatively poor outcome of these patients can be explained by the following: first, many patients had multinodular or advancedstage tumors; second, other extrahepatic collateral arteries were observed to supply HCCs in 33 patients; third, multiple tumors recurred at other liver sites. Although most patients had poor prognosis because of advanced stage at the time of detection of internal mammary artery blood supply, there are two patients free of disease by TACE. The prognosis of HCC after TACE seems to depend on factors related to both tumors and patients, although results presented in the literature are often disputed. Nevertheless, the important negative prognostic factors are the tumor size and multiplicity, portal vein thrombosis, and an advanced cirrhosis stage (20 23). In our study, tumor multiplicity and portal vein thrombosis were identified as independent negative prognostic factors by multivariate analysis. We also found that the technical success of internal mammary artery chemoembolization is an independent prognostic factor. Thus, efforts should be made to identify the internal mammary artery feeding a HCC and to assiduously perform selective chemoembolization via that artery. Some limitations of the current study should be mentioned. First, the frequency of 2.2% for an internal mammary artery supply may be somewhat underestimated. We may have missed unsuspected arterial feeders from the internal mammary artery in some patients. In our clinical practice, internal mammary angiography was not performed in some patients at a very advanced stage of HCC despite suspicion of collateral blood supply. Second, four patients received embolization with gelatin sponge or polyvinyl alcohol particles because of possible cutaneous complication. Chemoembolization via the internal mammary artery was not attempted in 18 patients. Moreover, 33 patients had tumors supplied by other extrahepatic collateral arteries, but the chemoembolization via other extrahepatic collateral arteries was not analyzed. These factors were not considered in survival analysis. Third, in patients with multinodular tumor, tumor size was recorded as the size of the tumor supplied by the internal mammary artery. So, tumor size did not represent true tumor burden in these patients. Fourth, this study was not a prospective trial and lacked a control group. Although blood supply to a HCC from the internal mammary artery is a rare condition, further randomized controlled studies regarding TACE via the internal mammary artery are warranted. In conclusion, although patients with HCC supplied by the internal

9 Volume 18 Number 5 Kim et al 619 mammary artery had poor prognosis because of their advanced stage of disease, chemoembolization via the internal mammary artery is possible and worth performing in selective situations. Acknowledgment: This study was supported by a grant of the National R&D Program for Cancer Control, Ministry of Health & Welfare, Republic of Korea ( ). References 1. Geschwind JF. Chemoembolization for hepatocellular carcinoma: where does the truth lie? J Vasc Interv Radiol 2002; 13: Llovet JM, Bruix J. Systematic review of randomized trials for unresectable hepatocellular carcinoma: chemoembolization improves survival. Hepatology 2003; 37: Miyayama S, Matsui O, Taki K, et al. Extrahepatic blood supply to hepatocellular carcinoma: angiographic demonstration and transcatheter arterial chemoembolization. Cardiovasc Intervent Radiol 2006; 29: Chung JW, Park JH, Han JK, Choi BI, Kim TK, Han MC. Transcatheter oily chemoembolization of the inferior phrenic artery in hepatocellular carcinoma: the safety and potential therapeutic role. J Vasc Interv Radiol 1998; 9: Kim HC, Chung JW, Lee W, Jae HJ, Park JH. Recognizing extrahepatic collateral vessels that supply hepatocellular carcinoma to avoid complications of transcatheter arterial chemoembolization. Radiographics 2005; 25: S25 S Kim HC, Chung JW, Jae HJ, Lee W, So YH, Park JH. Hepatocellular carcinoma: transcatheter arterial chemoembolization of the gonadal artery. J Vasc Interv Radiol 2006; 17: Chung JW, Kim HC, Yoon JH, et al. Transcatheter arterial chemoembolization of hepatocellular carcinoma: prevalence and causative factors of extrahepatic collateral vessels in 479 patients. Korean J Radiol 2006; 7: Miyayama S, Matsui O, Akakura Y, et al. Hepatocellular carcinoma with blood supply from omental branches: treatment with transcatheter arterial embolization. J Vasc Interv Radiol 2001; 12: Park SI, Lee DY, Won JY, Lee JT. Extrahepatic collateral supply of hepatocellular carcinoma by the intercostal arteries. J Vasc Interv Radiol 2003; 14: Suh SH, Won JY, Lee DY, Lee JT, Lee KH. Chemoembolization of the left inferior phrenic artery in patients with hepatocellular carcinoma: radiographic findings and clinical outcome. J Vasc Interv Radiol 2005;16: Kim JH, Chung JW, Han JK, Park JH, Choi BI, Han MC. Transcatheter arterial embolization of the internal mammary artery in hepatocellular carcinoma. J Vasc Interv Radiol 1995; 6: Kanetsuki I, Hori A, Ohshiro K, et al. Left lobe recurrent hepatocellular carcinoma treated with lipiodol-tae via the left internal mammary artery. Cardiovasc Intervent Radiol 1997; 20: Memis A, Oran I, Calli C, Yuzer Y, Mentes A. Transcatheter arterial embolization of internal mammary artery in a hepatocellular carcinoma with abdominal wall invasion. Eur J Radiol 1998; 28: Nakai M, Sato M, Kawai N, et al. Analysis of hepatocellular carcinoma fed by internal thoracic artery. Jpn J Intervent Radiol 1999; 14: Nakai M, Sato M, Kawai N, et al. Hepatocellular carcinoma: involvement of the internal mammary artery. Radiology 2001; 219: Chung JW, Park JH, Han JK, et al. Hepatic tumors: predisposing factors for complications of transcatheter oily chemoembolization. Radiology 1996; 198: Couinaud C. Le foie: études anatomiques et chirurgicales. Paris: Masson, 1957: Arora R, Soulen MC, Haskal ZJ. Cutaneous complications of hepatic chemoembolization via extrahepatic collaterals. J Vasc Interv Radiol 1999; 10: Lee JH, Chon CY, Ahn SH, et al. An ischemic skin lesion after chemoembolization of the right internal mammary artery in a patient with hepatocellular carcinoma. Yonsei Med J 2001; 42: Yamamoto K, Masuzawa M, Kato M, Okuyama T, Tamura K. Analysis of prognostic factors in patients with hepatocellular carcinoma treated by transcatheter arterial embolization. CancerChemotherPharmacol1992; 31S: S77 S Mondazzi L, Bottelli R, Brambilla G, et al. Transarterial oily chemoembolization for the treatment of hepatocellular carcinoma: a multivariate analysis of prognostic factors. Hepatology 1994; 19: Taniguchi K, Nakata K, Kato Y, et al. Treatment of hepatocellular carcinoma with transcatheter arterial embolization. Analysis of prognostic factors. Cancer 1994; 73: Llado L, Virgili J, Figueras J, et al. A prognostic index of the survival of patients with unresectable hepatocellular carcinoma after transcatheter arterial chemoembolization. Cancer 2000; 88: