Hepatic arterial embolization for unresectable hepatocellular carcinomas: do technical factors affect prognosis?

  • Published on
    08-Dec-2016

  • View
    212

  • Download
    0

Embed Size (px)

Transcript

<ul><li><p>ORIGINAL ARTICLE</p><p>Hepatic arterial embolization for unresectable hepatocellularcarcinomas: do technical factors affect prognosis?</p><p>Koichiro Yamakado Shiro Miyayama Shozo Hirota Kimiyoshi Mizunuma </p><p>Kenji Nakamura Yoshitaka Inaba Akihiro Maeda Kunihiro Matsuo Norifumi Nishida </p><p>Takeshi Aramaki Hiroshi Anai Shinichi Koura Shigeo Oikawa Ken Watanabe </p><p>Taku Yasumoto Kinya Furuichi Masato Yamaguchi</p><p>Received: 4 April 2012 / Accepted: 10 May 2012 / Published online: 30 May 2012</p><p> Japan Radiological Society 2012</p><p>Abstract</p><p>Purpose To evaluate retrospectively whether technical</p><p>factors of hepatic arterial embolization affect the prognosis</p><p>of patients with hepatocellular carcinoma (HCC).</p><p>Materials and methods Inclusion criteria of this study</p><p>were the following: (1) patients received embolization as</p><p>the initial treatment during 20032004, (2) Child A or B</p><p>liver profile, (3) five or fewer HCCs with maximum</p><p>diameter of 7 cm or smaller, and (4) no extrahepatic</p><p>metastasis. Patient data were gathered from 43 centers.</p><p>Prognostic factors were evaluated using univariate and</p><p>multivariate analyses.</p><p>Results Eight hundred fifteen patients were enrolled. The</p><p>1-, 3-, 5-, and 7-year overall survival rates were 92.0 %</p><p>(95 % CI 90.193.9), 62.9 % (95 % CI 59.366.6), 39.0 %</p><p>(95 % CI 35.143.0), and 26.7 % (95 % CI 22.630.8) in all</p><p>patients. Univariate analysis showed a Child-Pugh class-A,</p><p>alpha-fetoprotein level lower than 100 ng/ml, tumor size of</p><p>On behalf of the Clinical Research Group of the Japanese Society for</p><p>Transcatheter Hepatic Arterial Embolization and Japanese Society of</p><p>Interventional Radiology.</p><p>K. Yamakado (&amp;)Department of Interventional Radiology, Mie University School</p><p>of Medicine, 2-174 Edobashi, Tsu 514-8507, Japan</p><p>e-mail: yama@clin.medic.mie-u.ac.jp</p><p>S. Miyayama</p><p>Department of Diagnostic Radiology, Fukui-ken Saiseikai</p><p>Hospital, Fukui, Japan</p><p>S. Hirota A. MaedaDepartment of Radiology, Hyogo College of Medicine,</p><p>Nishinomiya, Japan</p><p>K. Mizunuma</p><p>Department of Radiology, Ohtawara Red Cross Hospital,</p><p>Ohtawara, Japan</p><p>K. Nakamura</p><p>Department of Radiology, Daito Central Hospital, Daito, Japan</p><p>Y. Inaba</p><p>Department of Diagnostic and Interventional Radiology,</p><p>Aichi Cancer Center Hospital, Nagoya, Japan</p><p>K. Matsuo</p><p>Department of Radiology, Narumi Hospital, Hirosaki, Japan</p><p>N. Nishida</p><p>Department of Radiology, Osaka City University, Osaka, Japan</p><p>T. Aramaki</p><p>Department of Diagnostic Radiology, Shizuoka Cancer Center,</p><p>Shizuoka, Japan</p><p>H. Anai</p><p>Department of Radiology, Nara Medical University,</p><p>Kashihara, Japan</p><p>S. Koura</p><p>Department of Radiology, Fukuoka University,</p><p>Fukuoka, Japan</p><p>S. Oikawa</p><p>Department of Radiology, Iwate Prefectural Central Hospital,</p><p>Morioka, Japan</p><p>K. Watanabe</p><p>Department of Radiology, Jikei University, Tokyo, Japan</p><p>T. Yasumoto</p><p>Department of Radiology, Toyonaka Municipal Hospital,</p><p>Toyonaka, Japan</p><p>K. Furuichi</p><p>Department of Radiology, Higashiosaka City General</p><p>Hospital, Higashiosaka, Japan</p><p>M. Yamaguchi</p><p>Department of Radiology, Kobe University, Kobe, Japan</p><p>123</p><p>Jpn J Radiol (2012) 30:560566</p><p>DOI 10.1007/s11604-012-0088-1</p></li><li><p>3 cm or smaller, tumor number of 3 or fewer, one-lobe tumor</p><p>distribution, nodular tumor type, within the Milan criteria,</p><p>stage I or II, no portal venous invasion, use of iodized oil, and</p><p>selective embolization were significantly better prognostic</p><p>factors. In the multivariate Cox model, the benefit to survival</p><p>of selective embolization remained significant (hazard ratio</p><p>0.68; 95 % CI 0.480.97; p = 0.033).</p><p>Conclusion Selective embolization contributes to sur-</p><p>vival in patients with HCCs.</p><p>Keywords Hepatocellular carcinoma Arterialembolization Prognosis</p><p>Introduction</p><p>The incidence of hepatocellular carcinoma (HCC), the fifth</p><p>most common cancer in the world, is increasing worldwide</p><p>[1]. Curative therapies including resection, liver transplan-</p><p>tation, and percutaneous ablation such as percutaneous eth-</p><p>anol injection (PEI) and radiofrequency (RF) ablation are</p><p>applicable in only 3040 % of patients with HCC [1]. Other</p><p>HCC patients are still not eligible for curative treatment</p><p>because of an advanced tumor stage or poor hepatic func-</p><p>tional reserve. Therefore, a continuing need persists for</p><p>effective palliative treatments. Recently, the benefit to sur-</p><p>vival of undergoing chemoembolization has been shown</p><p>compared with the best supportive care in meta-analyses of</p><p>randomized trials and in two individual trials [25].</p><p>It is important to achieve complete tumor necrosis to</p><p>prolong patient survival [6, 7]. Therefore, some techniques</p><p>have been developed in an attempt to reinforce anticancer</p><p>effects on HCC. These techniques include the use of che-</p><p>motherapeutic agents and iodized oil, and the introduction</p><p>of selective embolization [820]. Although some reports</p><p>have described a benefit to survival of using iodized oil and</p><p>anticancer drugs [11, 12], others have not [810]. Despite</p><p>the accumulation of evidence indicating that selective</p><p>embolization achieves better anticancer effects than non-</p><p>selective embolization, data demonstrating a benefit to</p><p>survival of this technique are lacking [1320].</p><p>Therefore, we conducted this retrospective study to eval-</p><p>uate whether technical factors of transarterial embolization</p><p>have impacts on survival in patients with unresectable HCCs.</p><p>Materials and methods</p><p>Study design</p><p>The Clinical Research Group of the Japanese Society of</p><p>Transcatheter Hepatic Arterial Embolization asked 255</p><p>training centers accredited by the Japanese Society of</p><p>Interventional Radiology to take part in this study. Ques-</p><p>tionnaire sheets were sent to them. Patient data were gathered</p><p>from the 43 institutions (16.9 %, 43/255) that agreed to</p><p>participate. At each institution, IRB approval was obtained</p><p>for this study. Because of the retrospective nature of this</p><p>study, the requirement of obtaining informed consent to take</p><p>part in this study was waived at all but two institutions, where</p><p>informed consent was obtained from living patients.</p><p>Inclusion criteria of this study were the following: (1)</p><p>patients received embolization as the initial treatment</p><p>during 20032004 and followed at least 3 months, (2)</p><p>Child A or B liver profile, (3) five or fewer HCCs with a</p><p>maximum diameter of 7 cm or smaller, and (4) no extra-</p><p>hepatic metastasis.</p><p>Patients</p><p>In 2003 and 2004, 1290 patients received transarterial</p><p>embolization as the initial treatment of unresectable HCCs at</p><p>the 43 institutions. Of them, 815 patients (63.2 %, 815/1290)</p><p>met the inclusion criteria and were enrolled in this study.</p><p>The diagnosis of HCC was made mainly based on</p><p>imaging modalities using ultrasonography, contrast-</p><p>enhanced computed tomography (CT), magnetic resonance</p><p>(MR) imaging, and angiography, in addition to elevation of</p><p>tumor markers such as a-fetoprotein and des-c-carboxylprothrombin. The typical HCC was depicted as an</p><p>enhanced tumor in the arterial phase and washout in the</p><p>delayed phase in contrast-enhanced CT and MRI, and as a</p><p>hypervascular tumor in digital subtraction angiography</p><p>[21]. Alpha-fetoprotein was positive ([20 ng/ml) in 460patients (60.6 %, 460/759), as was des-c-carboxyl pro-thrombin ([40 mAU/ml) in 371 patients (48.8 %). Biopsywas done in seven patients (0.9 %).</p><p>The decision for unresectable tumor was made by sur-</p><p>geons in each institution taking into account liver function,</p><p>tumor number and location, and patients status, such as</p><p>their age and performance status.</p><p>Patient and tumor characteristics are presented in</p><p>Table 1. The 535 men (65.6 %, 535/815) and 280 women</p><p>(34.4%, 280/815) had a mean age of 69.0 8.4 years</p><p>(standard deviation) (range 4091 years).</p><p>The mean maximum tumor diameter was 3.1 1.5 cm</p><p>(range 0.57.0 cm) and the mean tumor number 1.8 1.1</p><p>(range 15). Based on the Liver Cancer Study Group of Japan</p><p>(LCSGJ) tumor-node-metastasis (TNM) staging system,</p><p>65.6 % (535/815) of patients had stage I or II disease, and</p><p>34.4 % (280/815) of patients had stage III or IVA [22].</p><p>Transarterial embolization</p><p>Transarterial embolization was done using a gelatin sponge</p><p>in all patients (Table 1). Iodized oil was used in 98 % of</p><p>Jpn J Radiol (2012) 30:560566 561</p><p>123</p></li><li><p>patients (799/815) and anticancer drugs in 98.2 % of patients</p><p>(800/815). The iodized oil dose was 115 ml, with a mean</p><p>dose of 3.9 2.1 ml. Anticancer drugs were epirubicin,</p><p>used in 76.9 % of patients (615/800), epirubicin and mito-</p><p>mycin in 17.8 % of patients (142/800), and others in 5.4 % of</p><p>patients (43/800). The epirubicin dose was 590 mg with a</p><p>mean dose of 32.5 14.2 mg. That of mitomycin was</p><p>212 mg with a mean dose of 6.2 2.3 mg.</p><p>The definition of selective embolization was the fol-</p><p>lowing: transarterial embolization performed in the seg-</p><p>mental artery or more peripherally. Even when selective</p><p>embolization was performed at two or more different sites,</p><p>the technique was defined as selective embolization. When</p><p>selective embolization was combined with lobar or whole</p><p>liver embolization, the procedure was not regarded as</p><p>selective embolization. Selective embolization was done in</p><p>86.6 % of patients (706/815). Subsegmental embolization</p><p>consisted of 28.3 % in selective embolization (200/706),</p><p>segmental embolization of 30.3 % (214/706), and multiple</p><p>sessions of both of 41.4 % (292/706).</p><p>Assessments</p><p>Overall and recurrence-free survival was evaluated. Overall</p><p>and recurrence-free survival was defined as the time from the</p><p>initial transarterial embolization to death or last patient</p><p>contact, and recurrence-free survival to death, last patient</p><p>contact, or detection of disease progression. Disease pro-</p><p>gression was divided into three categories: local tumor pro-</p><p>gression in the treated HCC lesion, a new HCC lesion that</p><p>appeared in the untreated liver, and extrahepatic metastasis.</p><p>Complications related to transarterial embolization were</p><p>evaluated using clinical records.</p><p>Statistical analysis</p><p>The 16 variables presented in Table 1 were analyzed via</p><p>univariate analysis to identify factors affecting overall and</p><p>recurrence-free survival. The multivariate analysis was</p><p>performed using the Cox proportional hazard model. The</p><p>overall and recurrence-free survival rates were obtained</p><p>using the Kaplan-Meier method and compared using the log-</p><p>rank test. All variables with a p value of\0.05 by univariateanalysis and sex and age were subjected to multivariate</p><p>analysis. All significance tests were two-tailed, and a p value</p><p>\0.05 was regarded as statistically significant. All statisticalanalyses were performed using the Statistical Analysis</p><p>System (SAS version 8.02; SAS Inc., Cary, NC, USA).</p><p>Results</p><p>Overall survival</p><p>The 1-, 3-, 5-, and 7-year overall survival rates were 92.0 %</p><p>(95 % CI 90.193.9), 62.9 % (95 % CI 59.366.6), 39.0 %</p><p>(95 % CI 35.143.0), and 26.7 % (95 % CI 22.630.8) in all</p><p>patients (Fig. 1). The univariate analysis showed that a</p><p>Child-Pugh class-A, alpha-fetoprotein level lower than</p><p>100 ng/ml, tumor size of 3 cm or less, tumor number of three</p><p>or fewer, one-lobe tumor distribution, nodular tumor type,</p><p>within the Milan criteria, stage I or II, no portal venous</p><p>invasion, use of iodized oil, and selective embolization are</p><p>significantly better prognostic factors (Fig. 2; Table 2). In</p><p>the multivariate Cox model, the benefit to survival of</p><p>selective embolization remained significant (hazard ratio</p><p>0.68; 95 % CI 0.480.97, p = 0.033) (Table 3).</p><p>During the mean follow-up of 39.6 25.8 months</p><p>(range 3.098.0 months), 447 patients (54.8 %, 447/815)</p><p>died. At the end of the follow-up, 124 patients were still</p><p>Table 1 Patient and tumor backgrounds</p><p>Patient</p><p>Number 815</p><p>Male/female 535 (65.6)/280 (34.4)</p><p>Age (years) (range) 69.0 8.4 (4091)</p><p>B70/[70</p><p>Hepatitis B/C/others/unknown 71 (8.7)/619 (76.0)/85</p><p>(10.4)/40 (4.9)</p><p>Child-Pugh class A/B 585 (71.8)/230 (28.2)</p><p>Tumor</p><p>Maximum diameter</p><p>(cm) (range)</p><p>3.1 1.5 (0.57.0)</p><p>B3/3.17 499 (61.2)/316 (38.8)</p><p>Number (range) 1.8 1.1 (15)</p><p>13/45 731 (89.7)/84 (10.3)</p><p>Type, nodular/infiltrating 773 (94.85)/33 (4.05)/9 (1.10)</p><p>Distribution, hemilobe/bilateral</p><p>lobes/unknown</p><p>585 (71.78)/170 (20.86)/60 (7.36)</p><p>Portal venous invasion,</p><p>no/yes/unknown</p><p>778 (95.5)/37 (4.5)</p><p>Alpha-fetoprotein</p><p>(ng/ml) (range)</p><p>1263.3 12412.5 (1.0302200.0)</p><p>B100/[100/unknown 530 (65.0)/230 (28.2)/55 (6.8)</p><p>Milan criteria, within/</p><p>beyond/unknown</p><p>543 (66.6)/263 (32.3)/9 (1.1)</p><p>Stage I or II/III or IVA 535 (65.6)/280 (34.4)</p><p>Hepatic arterial embolization</p><p>Anticancer drug, used/not used 800 (98.2)/15 (1.8)</p><p>Epi/Epi ? MMC/others 615 (76.875)/142 (17.75)/43</p><p>(5.375)</p><p>Iodized oil, used/not used 799 (98.0)/16 (2.0)</p><p>Selective embolization, yes/no 706 (86.6)/109 (13.4)</p><p>Parentheses, percentage</p><p>Epi epirubicin, MMC mitomycin C</p><p>562 Jpn J Radiol (2012) 30:560566</p><p>123</p></li><li><p>alive (15.2 %). Cancer progression was the most frequent</p><p>cause of death, followed by liver failure, other disease,</p><p>gastrointestinal hemorrhage, unknown cause, and tumor</p><p>rupture (Table 4).</p><p>Fig. 1 Overall survival curve. The 1-, 3-, 5-, and 7-year overallsurvival rates were 92.0 % (95 % CI 90.193.9), 62.9 % (95 % CI</p><p>59.366.6), 39.0 % (95 % CI 35.143.0), and 26.7 % (95 % CI</p><p>22.630.8) in all 815 patients</p><p>Fig. 2 Overall survival curves based on a selective embolizationtechnique. A significant difference was found in survival rates</p><p>between patients who underwent selective embolization and those</p><p>who underwent non-selective embolization (p = 0.0034). The respec-tive 1-, 3-, 5-, and 7-year survival rates were 92.7 % (95 % CI,</p><p>90.894.7), 64.5 % (95 % CI 60.968.6), 40.8 % (95 % CI</p><p>36.645.1), and 28.4 % (95 % CI 24.032.7) in the patient group</p><p>that received selective embolization, and 87.0 % (95 % CI</p><p>80.493.6), 51.3 % (95 % CI 40.462.2), 25.7 % (14.936.5 %),</p><p>and 12.9 % (95 % CI 2.023.7) in the group that received non-</p><p>selective embolization</p><p>Table 2 Results of univariate analysis</p><p>Variable Survival (%) p value</p><p>1 year 3 year 5 year 7 year</p><p>Gender</p><p>Male 92.8 63.9 39.1 25.5 0.89</p><p>Female 90.4 61.1 38.9 29</p><p>Age</p><p>B70 91.4 61.2 40.1 27.3 0.42</p><p>[70 92.3 61.1 34.7 24.1</p><p>Hepatitis</p><p>C 91.5 62.6 37.3 24.3 0.11</p><p>Others 93.5 64 45.2 34.6</p><p>Child-Pugh</p><p>Class A 94.7 68 44.1 31.3 \0.0001</p><p>Class B 85.1 48.8 24.5 12.7</p><p>Maximum tumor size (cm)</p><p>B3 94.9 67.8 43.6 29.1 0.0001</p><p>\3 87.3 54.6 30.8 22.9</p><p>Tumor number</p><p>One to three 92.4 64.4 40.2 27.8 0.011</p><p>Four or five 88.7 49.9 29 17.4</p><p>Tumor type</p><p>Nodular 92.9 64.3 39.5 26.7 0.0049</p><p>Infiltrating 74.7 36 28 23.3</p><p>Tumor distribution</p><p>One lobe 92.8 64.9 41.8 30.2 \0.0001</p><p>Both lobes 89.4 50.9 27.1 8.3</p><p>Portal venous invasion</p><p>No 93.4 63.9 40 27.9 \0.0001</p><p>Yes 58.3 40.6 17.7 0.4</p><p>Alpha-fetoprotein (ng/ml)</p><p>B100 94 67.5 42.6 30.1 \0.0001</p><p>[100 87.4 47.6 27.8 16.9</p><p>Milan Criteria</p><p>Within 94.1 66.8 42.7 29.3 0.0006</p><p>Beyond 87.9 55.2 31 21</p><p>Tumor stage</p><p>I or II 94.2 67 44 31.2 \0.0001</p><p>III or IVA 87.5 54.5 28.2 16.8</p><p>Anticancer drug</p><p>Used 92.2 63.3 39.2 26.6 0.33</p><p>Not used 79.4 43.3 32.5 32.5</p><p>Anticancer drug</p><p>Epi 91.9 63.2 39.6 26.9 0.18</p><p>Epi ? MMC 94.2 64.8 42.7 30</p><p>Others 93 62.2 24.3 15.7</p><p>Iodized oil</p><p>Used 91.9 62.1 38.8 26.6 0.018</p><p>Not used 80.2 33.3 16.7 16.7</p><p>How to embolize</p><p>Selective 92.7 64.5 40.8 28.4 0.0034</p><p>Non-selective 87 51.3 25.7 12.9</p><p>Total 92 62.9 39 26.7</p><p>Epi epirubicin, MMC mitomycin C</p><p>Jpn J Radiol (2012) 30:560566 563</p><p>123</p></li><li><p>Recurrence-free survival</p><p>Recurr...</p></li></ul>