Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.
Original Research n Vascular
Hyo-Joon Yang, MD Jeong-Hoon Lee, MD Dong Hyeon Lee, MD Su Jong Yu, MD Yoon Jun Kim, MD Jung-Hwan Yoon, MD Hyo-Cheol Kim, MD Jeong Min Lee, MD Jin Wook Chung, MD Nam-Joon Yi, MD Kwang-Woong Lee, MD Kyung-Suk Suh, MD Hyo-Suk Lee, MD
1
From the Department of Internal Medicine and Liver Research Institute (H.J.Y., J.H.L., D.H.L., S.J.Y., Y.J.K., J.H.Y., H.S.L.), Department of Radiology (H.C.K., J.M.L., J.W.C.), and Department of Surgery (N.J.Y., K.W.L., K.S.S.), Seoul National University College of Medicine, 103 Daehak-ro, Jongno-gu, Seoul 110-744, Korea. Received August 7, 2013; revision requested September 19; revision received November 7; accepted November 21; final version accepted November 22. Supported by the Seoul National University Hospital Research Fund (grant 04-2013-3130). Address correspondence to J.H.L. (e-mail: [email protected]).
Purpose:
To compare the effectiveness of transarterial chemoembolization (TACE), radiofrequency ablation (RFA), and hepatic resection (HR) in patients with small single-nodule hepatocellular carcinoma by using inverse probability weighting to control selection bias.
Materials and Methods:
This retrospective cohort study was approved by the institutional review board, and the requirement to obtain informed consent was waived. The study included 197 consecutive patients (146 men and 51 women; mean age 6 standard deviation, 57.4 years 6 9.7) with single-nodule hepatocellular carcinomas measuring 3 cm or smaller and no vascular invasion who were treated initially with HR (n = 52), RFA (n = 79), or TACE (n = 66) from January 2005 to December 2006 at a single tertiary hospital. The primary endpoint was overall survival.
Results:
The baseline liver status of the groups differed significantly and was most favorable for the HR group, followed by the RFA group, and then the TACE group. The 5-year overall survival rates were 93.6% in the HR group, 86.6% in the RFA group, and 74.2% in the TACE group (P = .023). However, after inverse probability weighting, weighted survival rates among the three groups were similar (5-year overall survival: 85.6% with HR, 87.6% with RFA, and 80.7% with TACE; P = .834). In multivariate Cox regression analysis, TACE showed a hazard ratio of 0.978 (95% confidence interval: 0.407, 2.347; P = .960) compared with HR and of 1.335 (95% confidence interval: 0.619, 2.879; P = .461) compared with RFA.
Conclusion:
TACE is an effective treatment that allows achievement of long-term survival rates comparable to those with HR and RFA in patients with small single-nodule hepatocellular carcinoma. RSNA, 2014
q
Online supplemental material is available for this article.
RSNA, 2014
q
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
909
and Interventional Radiology
Small Single-Nodule Hepatocellular Carcinoma: Comparison of Transarterial Chemoembolization, Radiofrequency Ablation, and Hepatic Resection by Using Inverse Probability Weighting1
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
H
epatocellular carcinoma (HCC) is the sixth most common cancer in the world and the third most common cause of cancer-related death (1). With improved surveillance of patients with chronic liver disease and advances in imaging, more patients are diagnosed with early-stage HCC (2). For the treatment of earlystage HCC, curative therapies including liver transplantation, hepatic resection (HR), and radiofrequency ablation (RFA) are recommended (3–5). Although liver transplantation is generally limited by the shortage of donor allografts, HR is widely used as the main choice of treatment for resectable HCC. However, the risk of postoperative hepatic dysfunction often precludes HR (6). RFA has been shown to be as effective as HR in the treatment of small single-nodule HCC (7,8). However, RFA of lesions located close to major organs or the
Advances in Knowledge nn Transarterial chemoembolization (TACE) showed a 5-year overall survival rate of 80.7%, which was comparable to those of hepatic resection (85.6%) and radiofrequency ablation (87.6%), in patients with a single-nodule hepatocellular carcinoma of less than or equal to 3 cm without vascular invasion when the underlying liver status was balanced among the patients receiving each treatment by using inverse probability weighting. nn Most (97.0%) of the patients treated with TACE achieved a complete response, and the achievement of complete response was one of the independent prognostic factors of survival (P , .001). nn TACE showed a significantly reduced risk of adverse events compared with radiofrequency ablation and hepatic resection (P = .007), but also had significantly higher frequency of tumor recurrence (P , .001). 910
Yang et al
liver capsule is often contraindicated (9). Consequently, transarterial chemoembolization (TACE) is also used frequently for the treatment of small single-nodule HCC when other curative treatments are not possible. On the basis of the data from randomized controlled trials and systematic review, TACE is currently recommended only for large or multinodular HCC (10–12). In comparison, there is a paucity of long-term data comparing TACE with HR or RFA for the treatment of small HCC (13,14). In the past few years, authors of some retrospective cohort studies (15,16) reported the efficacy of TACE compared with that of HR and with RFA in early-stage HCC. Because HR, RFA, and TACE are the major treatment modalities for small (3 cm) single-nodule HCC, there is a considerable need for an evaluation of the therapeutic efficacy of the three treatments in a single cohort of patients. To obtain confirmative evidence, a well-designed randomized controlled trial is necessary, but this is not easy to conduct because of the current practice guidelines and ethical concerns. A cohort study with inverse probability weighting (IPW) and a regression method was an option to balance underlying liver function among the treatment groups. The purpose of our study was to compare the effectiveness of TACE, RFA, and HR in patients with small single-nodule HCC by using IPW to control selection bias.
Implications for Patient Care nn TACE is a safe and effective alternative to hepatic resection and radiofrequency ablation for single-nodule hepatocellular carcinomas of 3 cm or smaller without vascular invasion, especially for patients who are vulnerable to adverse events of treatments. nn To attain long-term survival in patients treated with TACE, special care should be taken to obtain a complete response and to undertake surveillance for tumor recurrence.
Materials and Methods Patients This retrospective cohort study was approved by the institutional review board of Seoul National University Hospital (Seoul, Korea) and was exempt from the requirement to obtain informed consent. We included consecutive patients (a) in whom a single-nodule HCC smaller than or equal to 3 cm without vascular invasion was diagnosed from January 2005 and December 2006 at Seoul National University Hospital and (b) who were treated initially with HR, RFA, or TACE. The diagnosis of HCC was based on the recommendations of international practice guidelines (3–5). The presence of vascular invasion was confirmed with the demonstration of an enhancing expansile low-attenuation intraluminal mass or filling defect in the portal or hepatic vessels at dynamic computed tomography (CT) and/ or magnetic resonance (MR) imaging (17,18). Patients were excluded if they had (a) Child-Pugh class C liver function, (b) concomitant serious medical illness including malignant tumors other than HCC, or (c) received multimodality treatment for the single-nodule HCC. Published online before print 10.1148/radiol.13131760 Content codes: Radiology 2014; 271:909–918 Abbreviations: CI = confidence interval HCC = hepatocellular carcinoma HR = hepatic resection IPW = inverse probability weighting RFA = radiofrequency ablation TACE = transarterial chemoembolization Author contributions: Guarantors of integrity of entire study, H.J.Y., J.H.L., D.H.L., S.J.Y.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, H.J.Y., J.H.L., D.H.L., N.J.Y.; clinical studies, all authors; statistical analysis, H.J.Y., J.H.L., D.H.L.; and manuscript editing, H.J.Y., J.H.L., D.H.L., S.J.Y., Y.J.K., J.H.Y., H.C.K., J.M.L., J.W.C., K.W.L., K.S.S., H.S.L. Conflicts of interest are listed at the end of this article. See also the article by Halpern in this issue.
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
Patients were divided into three groups according to the treatment they had received: (a) HR group, (b) RFA group, and (c) TACE group. The decision of treatment modality was based mainly on practice guidelines (3–5). For the patients with wellpreserved liver function, HR was recommended initially. If the patients were not eligible for resection, RFA was recommended before other treatments. Among the patients who could undergo surgery, some received RFA because of their individual preferences. When curative therapies including HR and RFA were not possible, TACE was recommended.
Procedures HR was performed by three specialized surgeons (including K.S.S. and N.J.Y., with more than 20 and 10 years of experience in hepatic surgery, respectively). The decision about the type and extent of resection was based on the tumor location, underlying liver status, and the patients’ other medical conditions. If necessary, intraoperative ultrasonography was used to identify tumor margins and hepatic vasculature. RFA was performed by three specialized interventional radiologists (including J.M.L., with more than 15 years of experience) by using an electrosurgical unit (RF-2000; RadioTherapeutics, Mountain View, Calif) or a 200-watt radiofrequency generator (Radionics, Burlington, Mass) with a 17-gauge single cooled-tip needle electrode for HCCs smaller than 2 cm or a triple-cluster needle electrode for HCCs of 2 cm or larger. The completeness of the ablation was evaluated by using dynamic CT immediately after the procedure, and an additional session of RFA was performed for an incomplete ablation. TACE was conducted by three specialized interventional radiologists (including J.W.C., with more than 20 years of experience) initially by selecting segmental or subsegmental hepatic arteries by using a microcatheter (Microferret [Cook, Bloomington, Ind] or Progreat [Terumo, Tokyo,
Japan]). An emulsion of iodized oil (Lipiodol; Andre Guerbet, Aulnaysous-Bois, France) and doxorubicin hydrochloride (Adriamycin RDF; Ildong Pharmaceutical, Seoul, Korea) was infused through the feeder vessels. Then, by using a mixture of gelatin sponge particles (Gelfoam; Upjohn, Kalamazoo, Mich), mitomycin (Mitomycin-10; Kyowa Hakko Kogyo, Tokyo, Japan), and contrast material, embolization was performed. Iodized oil labeling of the tumor was evaluated by using unenhanced CT 3 weeks after the procedure. All the patients were followed up at months 1, 3, 6, 9, and 12 before 1 year, and then every 3–6 months when necessary. At every visit, radiologic studies, tumor markers, and biochemical liver function tests were performed.
Assessment Data on clinical and laboratory findings were collected retrospectively from all patients by reviewing the electronic medical records, and ChildPugh class, model for end-stage liver disease score (19), and model to estimate survival in ambulatory HCC patients score (20) were determined by two authors (H.J.Y. and D.H.L., with 9 and 6 years of experience, respectively) in consensus. One radiologist, (J.W.C.) independently reviewed all radiologic images, blinded to survival data, to determine singularity and size of tumor, absence of vascular invasion, treatment response, and recurrence. Study Endpoints and Definitions The primary endpoint was overall survival, defined as the time from the date of diagnosis until the date of death from any cause. The secondary endpoints included complete response, adverse events, and recurrence. Complete response was defined according to the modified Response Evaluation Criteria in Solid Tumors (4) and subcategorized into initial complete response, defined as complete response after initial treatment and overall complete response,
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Yang et al
defined as complete response after a single session or repeated sessions of a single-treatment modality. Adverse events that were grade 2 or higher in severity defined according to the National Cancer Institute Common Terminology Criteria (version 4.0) (21) were evaluated. Grade 1 events were not assessed because they were common after the three treatments and were not recorded separately in practice.
Statistical Analysis Continuous data were compared by using the analysis of variance or Kruskal-Wallis test, and categorical data were compared by using a x2 or linear-by-linear association test. Survival and recurrence were plotted by means of the Kaplan-Meier method and compared with the log-rank test. To identify independent prognostic factors and compare the treatment effects adjusted for those factors, the Cox proportional hazards model was fitted. The clinically or statistically significant variables in univariate analysis (P , .1) were included in multivariate analysis. To minimize selection bias and describe the treatment effect of different modalities better in our observational study, we used IPW (22,23). Propensity scores were calculated by generating a logistic regression model to predict the probability of each patient receiving a particular treatment on the basis of the 18 pretreatment variables described in the baseline characteristics. After the inverse probabilities of treatment weight (propensity score weight) were created, the three groups were then balanced by means of IPW. A post hoc power analysis was performed for the results of multiple Cox regression of overall survival and IPW followed by Cox regression. All statistical analyses were performed by using SPSS software (SPSS version 19.0; SPSS, Chicago, Ill) and the R statistical programming environment (version 2.13.0; R development Core Team, Vienna, Austria, http://www.R-project.org). A P value less than .05 was considered to indicate a significant difference. 911
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
Yang et al
TACE (three of 66, 4.5%; P = .007) and so did adverse events greater in severity than grade 3 (P , .001) (Table 2). The details of the adverse events are described in Table E1 (online).
Figure 1
Figure 1: Flow diagram shows the screening, enrollment, and treatment allocation of patients.
Results Study Population The study profile is shown in Figure 1. Between January 2005 and December 2006, a total of 1184 consecutive patients received a diagnosis of HCC. Among these, 230 patients without vascular invasion underwent HR, RFA, or TACE for single-nodule HCC of 3 cm or smaller. Thirty-three patients were excluded owing to multimodality treatment (n = 21), Child-Pugh class C liver function (n = 7), or concomitant serious medical illness (n = 5), thus leaving 197 patients for the analysis (HR, 52; RFA, 79; and TACE, 66). The baseline characteristics are summarized in Table 1. There were no significant differences in demographic and tumor-related characteristics among the three groups. The mean patient age was 57.4 years 6 9.7; 74.1% (146 of 197) were male, 69.0% (136 of 197) were positive for hepatitis B surface antigen, and the mean tumor size was 2.1 cm 6 0.6. However, there were significant differences in the variables reflecting underlying liver status, including model for end-stage liver disease and model 912
to estimate survival in ambulatory HCC patients score. They were most favorable in the HR group, followed by those in the RFA group, and then those in the TACE group.
Treatment Response and Adverse Events The overall complete response rates among the groups were similar; 98.1% in the HR group, 97.5% in the RFA group, and 97.0% in the TACE group (P = .705) (Table 2). However, the initial complete response rate was significantly inferior in the TACE group compared with those of the other groups (P , .001). After the initial treatment, there was one treatment-related death (1.9%) in the HR group; the patient was a 72-year-old man with chronic hepatitis C–related liver cirrhosis and Child-Pugh class A status. He underwent peripheral segmentectomy and died of postoperative pneumonia (ie, grade 5 lung infection) on postoperative day 21. There were no deaths within 90 days of treatment in the other groups. Adverse events greater in severity than grade 2 developed significantly more often after HR (12 of 52, 23.1%) than after RFA (eight of 79, 10.1%) or
Survival and Tumor Recurrence Analysis During follow-up, four of 52 (7.7%) patients in the HR group, 12 of 79 (15.2%) in the RFA group, and 18 of 66 (27.3%) in the TACE group died: two of four (50.0%), eight of 12 (66.7%), and 12 of 18 (66.7%) patients died of progression of HCC, worsening cirrhosis, or both, respectively. The cumulative probabilities of overall survival at 1, 3, and 5 years were as follows: 98.0%, 96.0%, and 93.6% in the HR group; 94.8%, 90.9%, and 86.6% in the RFA group; and 96.9%, 82.9%, and 74.2% in the TACE group, respectively (Fig 2, A). The survival rates among the groups were significantly different (P = .023). Recurrence was diagnosed in 23 of 51 (45.1%) patients in the HR group, 57 of 77 (74.0%) in the RFA group, and 54 of 64 (84.4%) in the TACE group; 5-year cumulative probabilities were 44.6%, 73.2%, and 82.5%, respectively (P , .001) (Fig 2, B). In multivariate Cox regression analysis, the treatment modality was not significantly associated with survival. Failure to achieve an overall complete response (hazard ratio, 16.750; 95% confidence interval [CI]: 4.525, 62.006; P , .001), presence of ascites (hazard ratio, 3.440; 95% CI: 1.384, 8.547; P = .008), and high model to estimate survival in ambulatory HCC patients score (hazard ratio, 2.241; 95% CI: 1.282, 3.919; P = .005) were independent predictive factors associated with poor survival (Table 3). With adjustment for these predictors, the hazard ratios between treatments were considerably reduced compared with the unadjusted hazard ratios: from 3.774 to 2.191 for TACE versus HR and from 1.858 to 1.295 for TACE versus RFA. A post hoc power analysis resulted in a power of 0.60 for TACE versus HR and of 0.11 for TACE versus RFA. The estimated hazard ratios detectable with a power
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
Table 1 Baseline Characteristics of the Study Population Parameter Age (y) Sex M F Viral marker Hepatitis B surface antigen Anti-hepatitis C virus Platelet (103/mm3)‡ Serum albumin (g/dL)§ Total bilirubin (mg/dL)|| Creatinine (mg/dL)# Prothrombin time (INR) Cirrhosis Absent Present Ascites Absent Present Esophageal varix Absent Present Child class A B Model for end-stage liver disease score Eastern Cooperative Oncology Group performance score 0 1 Tumor size 2 cm .2 cm a-fetoprotein** 200 ng/mL .200 ng/mL Model to estimate survival in ambulatory HCC patients score
HR Group (n = 52)
RFA Group (n = 79)
TACE Group (n = 66)
55.7 6 10.6*
57.2 6 9.2*
59.0 6 9.5*
38 (73.1) 14 (26.9)
59 (74.7) 20 (25.3)
49 (74.2) 17 (25.8)
34 (65.4) 6 (11.5) 153 6 51* 4.04 6 0.43* 1.02 6 0.45* 1.03 6 0.15* 1.10 6 0.09*
56 (70.9) 18 (22.8) 112 6 52* 3.71 6 0.51* 1.27 6 0.71* 1.00 6 0.16* 1.18 6 0.17*
46 (69.7) 13 (19.7) 113 6 62* 3.66 6 0.51* 1.14 6 0.54* 1.17 6 1.15* 1.22 6 0.18*
23 (44.2) 29 (55.8)
15 (19.0) 64 (81.0)
6 (9.1) 60 (90.9)
P Value .190 .979
.793 .265 ,.001† ,.001† .229 .188 ,.001† ,.001†
.031† 51 (98.1) 1 (1.9)
74 (93.7) 5 (6.3)
58 (87.9) 8 (12.1) .006†
51 (98.1) 1 (1.9)
62 (78.5) 17 (21.5)
53 (80.3) 13 (19.7)
50 (96.2) 2 (3.8) 8.3 6 2.2*
68 (86.1) 11 (13.9) 9.4 6 2.9*
55 (83.3) 11 (16.7) 10.0 6 3.9*
.040†
.016† .944
46 (88.5) 6 (11.5)
69 (87.3) 10 (12.7)
57 (86.4) 9 (13.6)
21 (40.4) 31 (59.6)
36 (45.6) 43 (54.4)
40 (60.6) 26 (39.4)
40 (76.9) 12 (23.1) 3.3 6 0.39*
63 (79.7) 16 (20.3) 3.5 6 0.41*
53 (80.3) 13 (19.7) 3.5 6 0.58*
.065
.893
.045†
Note.—Except where indicated, data are numbers of patients, with percentages in parentheses. * Data are means 6 standard deviation. †
Statistically significant.
‡
To convert to Système international (SI) units (109 per liter), multiply by 1.
§
To convert to SI units (grams per liter), multiply by 10.
||
To convert to SI units (micromoles per liter), multiply by 17.104.
#
To convert to SI units (micromoles per liter), multiply by 76.25.
** To convert to SI units (micrograms per liter), multiply by 1.
of 0.80 were 2.716 for TACE versus HR and 2.643 for TACE versus RFA (Table E2 [online]). Meanwhile, the
treatment modality (overall P , .001), total bilirubin (hazard ratio, 1.545; 95% CI: 1.152, 2.072; P = .004), and
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Yang et al
age (hazard ratio, 1.024; 95% CI: 1.005, 1.044; P = .013) were independent predictive factors of recurrence (Table E3 [online]).
Survival and Tumor Recurrence Analysis by Using Inverse Probability Weighting After IPW, the baseline characteristics, including liver status–related variables, became more balanced than they were before weighting and did not differ significantly among the groups (data were not shown). The weighted cumulative probabilities of survival among the three groups became comparable; 5-year overall survival rates were 85.6% with HR, 87.6% with RFA, and 80.7% with TACE (P = .834) (Fig 3, A). In multivariate analysis, the hazard ratios for survival between TACE and HR and between TACE and RFA were 0.978 (95% CI: 0.407, 2.347; P = .960) and 1.335 (95% CI: 0.619, 2.879; P = .461), respectively (Table E4 [online]). A post hoc power analysis showed a power of 0.03 between TACE and HR and of 0.13 between TACE and RFA. The estimated hazard ratios detectable with a power of 0.80 were 2.790 between TACE and HR and 2.699 between TACE and RFA. On the other hand, treatment with TACE versus that with HR and RFA was still an independent risk factor of recurrence even after the weighting (overall, P , .001) (Fig 3, B and Table E5 [online]). Discussion In our study, we compared the treatment effects of HR, RFA, and TACE on long-term survival by using IPW and regression methods. We found that 5-year overall survival with TACE was similar to that with HR and RFA in patients with single-nodule HCC of 3 cm or smaller without vascular invasion when the underlying liver status was balanced among the patients receiving each treatment by means of IPW. In addition, our results showed that most of the patients initially treated with TACE achieved a complete response, which was one of the independent prognostic factors of survival. 913
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
Yang et al
Table 2 Summary of Treatment Response and Adverse Events Treatment Response and Adverse Event Initial complete response rate Overall complete response rate Number of treatments One Two Three Four Five Adverse events after initial treatment Grade 2 Patients Events‡ Grade 3 Patients Events‡
HR Group (n = 52)
RFA Group (n = 79)
TACE Group (n = 66)
P Value
50 (96.2) 51 (98.1)
76 (96.2) 77 (97.5)
53 (80.3) 64 (97.0)
,.001* .705
50 (96.2) 1 (1.9) 0 (0.0) 0 (0.0) 0 (0.0)
76 (96.2) 1 (1.3) 0 (0.0) 0 (0.0) 0 (0.0)
53 (80.3) 6 (9.2) 2 (3.0) 2 (3.0) 1 (1.5)
12 (23.1) 13
8 (10.1) 12
3 (4.5) 5
.007*
7 (13.5) 7
1 (1.3) 1
0 (0.0) 0
,.001*
Note.—Except where indicated, data are numbers of patients, with percentages in parentheses. Complete response is defined according to the modified Response Evaluation Criteria in Solid Tumors. Grades are defined by the National Cancer Institute Common Terminology Criteria (version 4.0). * Statistically significant. †
Data are numbers of events.
The survival of patients with HCC is not only strongly associated with the extent of the tumor but also with the severity of underlying liver dysfunction. However, the degree of hepatic functional reserve of patients affects the decision on the choice of treatment modality. TACE, for example, is usually considered when HR or RFA is not feasible. This selection bias is the major obstacle that hampers the evaluation of effectiveness of different treatments for HCC in a cohort study. Although a randomized controlled trial would allow an unbiased comparison of treatment effectiveness, such an approach would raise an ethical problem in that the current practice guidelines recommend treating small single-nodule HCC primarily with HR and alternatively with liver transplantation or RFA rather than TACE (3–5). IPW, as one of the propensity score methods, can be a useful alternative strategy to address potential selection bias in an observational study (22,23). In our study, we 914
hypothesized that the patients’ baseline characteristics, including underlying liver function, influenced treatment selection; and we evaluated the probability of patients being included in the three treatment groups on the basis of the baseline variables, and we weighted the data from each patient inversely to each probability. This balanced the baseline characteristics, of which liver function was significantly different, to reduce selection bias and better describe treatment effects on patient survival. Our results showed that there was a significant difference in overall survival rates according to the choice of initial treatments. However, when adjusted for the covariates in the multivariate regression analysis, the treatment modality was not an independent prognostic factor of overall survival. Survival analysis by using IPW revealed comparable overall survival rates among the three treatment groups; the estimated 5-year overall survival rates were shown to differ by only 5%–7%, and
the hazard ratios of death in the TACE group versus the HR and RFA groups were remarkably decreased compared with those before IPW. These results are consistent with those of cohort studies in which multiple regression (15) and propensity score matching (16) were used; they demonstrated that TACE provided overall survival similar to HR even after adjustment for baseline variables (15) and to RFA when populations were matched by using propensity score (16). A probable explanation for this is that the survival differences observed in our population before weighting might be determined by the degree of baseline liver dysfunction, on which the treatment selection was based, rather than by the specific treatment modality itself. Therefore, TACE might offer long-term survival rates comparable to those of HR and RFA for small single-nodule HCC if underlying liver function was similar among the patients receiving each treatment. The major disadvantages of TACE were the lower rate of initial complete response (80.3%) and the higher risk of recurrence (82.5% after 5 years from complete response). Nevertheless, the TACE group showed a similar overall survival rate compared with that of the other groups after IPW. From our analysis, we can surmise the reasons for this. First, the overall complete response, rather than the initial complete response, did affect survival significantly, and the TACE group showed a high overall complete response rate similar to those of the other groups. Second, TACE showed the lowest frequency and severity of adverse events, and this may have affected survival rates of the patients in the TACE group favorably. Third, recurrence was not an independent prognostic factor of overall survival, implying that the higher risk of recurrence after TACE may not increase the risk of death in such patients with early-stage HCC. Considering these drawbacks, TACE would not be the first-choice treatment for small singlenodule HCC. However, most of those patients treated initially with TACE
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
Figure 2
Figure 2: Graphs show cumulative probability of, A, overall survival and, B, tumor recurrence according to treatment groups. Data were obtained with Kaplan-Meier method. There were significant differences among treatment groups both in overall survival (P = .023, log-rank test) and in tumor recurrence (P , .001, log-rank test).
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Yang et al
eventually can achieve a complete response, although some should receive repeated treatments, and once a complete response is achieved, long-term survival similar to that with HR and RFA would be expected. Therefore, TACE can be a safe and effective alternative treatment option, especially for those who are vulnerable to the adverse events of treatments. Although angiography has lower sensitivity for detection of small HCC than does CT or MR imaging (24), there was no patient in our cohort in whom angiography was unable to show the lesion that was already found at CT or MR imaging. Thus, the low sensitivity of angiography as a diagnostic tool might not be a substantial weakness of TACE as a therapeutic modality. The limitations of our study were the retrospective observational design and the relatively small sample size. In the initial comparison of the cumulative survival probabilities, the sample size was large enough to reveal the significant survival differences among the treatment groups. After the multiple regression and IPW analysis, however, decreased survival differences among the treatment groups resulted in low statistical power. Results of a post hoc power analysis also indicated that, under the conditions of our study, a treatment difference smaller than a 2.70 of hazard ratio of survival might exist without being detected. Thus, the failure to reach statistical significance between the treatment groups appears to be attributable not only to the lack of clinically significant differences but also to the small sample size. Therefore, a further study with a larger sample size is needed to elucidate whether the observed small differences in survival among the treatment options truly exist. Another limitation is that we included up to six covariates in the multiple Cox regression models for overall survival, which might have caused a problem of overfitting, considering the 34 deaths in our cohort. This should be confirmed in a further study with more events. Although we showed that TACE was 915
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
Yang et al
Table 3 Univariate and Multivariate Analyses of Prognostic Factors for Overall Survival Univariate Analysis Variables* Age (y) Sex (M vs F) hepatitis B surface antigen (+ vs 2) Anti-hepatitis C virus (+ vs 2) Platelet (103/mm3)† Serum albumin (g/dL)‡ Total bilirubin (mg/dL)§ Creatinine (mg/dL)|| Prothrombin time (INR) Cirrhosis (+ vs 2) Ascites (+ vs 2) Esophageal varix (+ vs 2) Child class (B vs A) Model for end-stage liver disease score ECOG PS (1 vs 0) Tumor size (.2 cm vs 2 cm) a-fetoprotein (.200 ng/mL vs 200 ng/mL)# Model to estimate survival in ambulatory HCC patients score Treatment modality (RFA vs HR) (TACE vs HR) (TACE vs RFA) Initial complete response (2 vs +) Overall complete response (2 vs +) HCC recurrence (+ vs 2)
Multivariate Analysis
Hazard Ratio
95% CI
P Value
Hazard Ratio
95% CI
P Value
1.014 1.028 1.382 1.183 0.999 0.323 1.554 1.311 11.297 3.175 5.051 1.505 3.767 1.143 1.634 1.103 1.416 3.136
0.979, 1.051 0.480, 2.201 0.626, 3.054 0.514, 2.720 0.993, 1.005 0.178, 0.588 0.918, 2.631 1.057, 1.626 1.996, 63.929 0.971, 10.385 2.281, 11.186 0.655, 3.458 1.800, 7.884 1.066, 1.225 0.676, 3.948 0.563, 2.163 0.661, 3.034 1.836, 5.356
... ... ... ... ... ... ... ... ... ... 3.440 ... ... ... ... ... ... 2.241
... ... ... ... ... ... ... ... ... ... 1.384, 8.547 ... ... ... ... ... ... 1.282, 3.919
2.032 3.774 1.858 2.835 13.507 1.628
0.655, 6.300 1.277, 11.158 0.894, 3.858 1.170, 2.835 3.920, 46.537 0.667, 3.973
.433 .944 .424 .693 .769 ,.001 .101 .014 .006 .056 ,.001 .335 ,.001 ,.001 .275 .774 .371 ,.001 .032 .220 .016 .097 .021 ,.001 .284
1.692 2.191 1.295 ... 16.750 ...
0.542, 5.279 0.705, 6.807 0.598, 2.807 ... 4.525, 62.006 ...
... ... ... ... ... ... ... ... ... ... .008 ... ... ... ... ... ... .005 .390 .365 .175 .512 ... ,.001 ...
Note.—Variables that were not significant in the univariate analysis were not included in the multivariate analysis. ECOG PS = Eastern Cooperative Oncology Group performance score. * For categorical variables with the categories in the parentheses, the former was compared with the latter (the reference) in calculating hazard ratios and 95% confidence intervals by using the Cox proportional hazards model. †
To convert to Système international (SI) units (109 per liter), multiply by 1.
‡
To convert to SI units (grams per liter), multiply by 10.
§
To convert to SI units (micromoles per liter), multiply by 17.104.
||
To convert to SI units (micromoles per liter), multiply by 76.25.
#
To convert to SI units (micrograms per liter), multiply by 1.
safer than RFA in patients with small single-nodule HCC, RFA is generally considered to be a safe procedure that causes fewer adverse events than TACE does (3,4). Therefore, caution should be exercised when applying our safety results to other patients treated with either of the modalities for lesions of different sizes or numbers. In conclusion, the results of our study show that TACE is an effective and safe treatment for small single-nodule HCC, achieving long-term survival rates comparable to those of HR and RFA, when the underlying liver dysfunction 916
is similar among the patients receiving each treatment. However, when TACE is used as an initial treatment, special care should be taken to obtain a complete response, and surveillance for tumor recurrence should be undertaken. This study may warrant a large randomized controlled trial to reassess the role of TACE in patients with small single-nodule HCC. Acknowledgments: We appreciate statistical consultation and power analysis provided by the Medical Research Collaborating Center at Seoul National University College of Medicine/ Seoul National University Hospital.
Disclosures of Conflicts of Interest: H.J.Y. No relevant conflicts of interest to disclose. J.H.L. No relevant conflicts of interest to disclose. D.H.L. No relevant conflicts of interest to disclose. S.J.Y. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: Payment for lectures from Bayer Healthcare Pharmaceuticals. Other relationships: none to disclose. Y.J.K. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: grants/grants pending from BristolMyers Squibb, Roche, JW Creagene, Bukwang Pharmaceuticals, Handok Pharmaceuticals, Hanmi Pharmaceuticals, Yuhan Pharmaceuticals, and Pharmaking; payment for lectures from Bayer, Bukwang, CJ Pharmaceuticals, Gilead Science, Handok, MSD Korea, Yuhan,
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
Figure 3
Yang et al
and Samil Pharmaceuticals. Other relationships: none to disclose. J.H.Y. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: grants/grants pending from Bayer. Other relationships: none to disclose. H.C.K. No relevant conflicts of interest to disclose. J.M.L. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: board membership for Bayer; consultancy for Bayer and Siemens; grants/grants pending from GE Healthcare, Dongseo Medical, CMS, Acuzen, Starmed, and RF Medical; payment for lectures from Bayer and Supersonic. Other relationships: none to disclose. J.W.C. No relevant conflicts of interest to disclose. N.J.Y. No relevant conflicts of interest to disclose. K.W.L. No relevant conflicts of interest to disclose. K.S.S. No relevant conflicts of interest to disclose. H.S.L. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: grants/grants pending from Dong-A Pharmaceuticals and Bukwang. Other relationships: none to disclose.
References 1. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer 2010;127(12):2893–2917. 2. Zhang BH, Yang BH, Tang ZY. Randomized controlled trial of screening for hepatocellular carcinoma. J Cancer Res Clin Oncol 2004;130(7):417–422. 3. Bruix J, Sherman M; American Association for the Study of Liver Diseases. Management of hepatocellular carcinoma: an update. Hepatology 2011;53(3):1020–1022. 4. European Association For The Study Of The Liver; European Organisation For Research And Treatment Of Cancer. EASL-EORTC clinical practice guidelines: management of hepatocellular carcinoma. J Hepatol 2012;56(4):908–943. [Published correction appears in J Hepatol 2012;56(6):1430.] 5. Benson AB 3rd, Abrams TA, Ben-Josef E, et al. NCCN clinical practice guidelines in oncology: hepatobiliary cancers. J Natl Compr Canc Netw 2009;7(4):350–391.
Figure 3: Graphs show weighted cumulative probability of, A, overall survival and, B, tumor recurrence according to treatment groups. Data were obtained with weighted Kaplan-Meier method. Weighted probabilities of overall survival were comparable (P = .834, univariate Cox regression analysis) among treatment groups, while weighted risks of tumor recurrence were still significantly different (P , .001, univariate Cox regression analysis).
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
6. Bruix J, Castells A, Bosch J, et al. Surgical resection of hepatocellular carcinoma in cirrhotic patients: prognostic value of preoperative portal pressure. Gastroenterology 1996;111(4):1018–1022. 7. Chen MS, Li JQ, Zheng Y, et al. A prospective randomized trial comparing percutaneous local ablative therapy and partial hepatectomy for small hepatocellular carcinoma. Ann Surg 2006;243(3):321–328.
917
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
8. Livraghi T, Meloni F, Di Stasi M, et al. Sustained complete response and complications rates after radiofrequency ablation of very early hepatocellular carcinoma in cirrhosis: Is resection still the treatment of choice? Hepatology 2008;47(1):82–89. 9. Livraghi T, Solbiati L, Meloni MF, Gazelle GS, Halpern EF, Goldberg SN. Treatment of focal liver tumors with percutaneous radio-frequency ablation: complications encountered in a multicenter study. Radiology 2003;226(2):441–451. 10. Llovet JM, Real MI, Montaña X, et al. Arterial embolisation or chemoembolisation versus symptomatic treatment in patients with unresectable hepatocellular carcinoma: a randomised controlled trial. Lancet 2002;359(9319):1734–1739. 11. Lo CM, Ngan H, Tso WK, et al. Randomized controlled trial of transarterial lipiodol chemoembolization for unresectable hepatocellular carcinoma. Hepatology 2002;35(5):1164–1171. 12. Llovet JM, Bruix J. Systematic review of randomized trials for unresectable hepatocellular carcinoma: Chemoembolization improves survival. Hepatology 2003;37(2):429–442. 13. Huang YH, Chen CH, Chang TT, et al. The role of transcatheter arterial embolization in patients with resectable hepatocellular carcinoma: a nation-wide, multicenter study. Liver Int 2004;24(5):419–424.
918
Yang et al
14. Lee HS, Kim KM, Yoon JH, et al. Therapeutic efficacy of transcatheter arterial chemoembolization as compared with hepatic resection in hepatocellular carcinoma patients with compensated liver function in a hepatitis B virus-endemic area: a prospective cohort study. J Clin Oncol 2002;20(22):4459–4465. 15. Hsu KF, Chu CH, Chan DC, et al. Superselective transarterial chemoembolization vs hepatic resection for resectable early-stage hepatocellular carcinoma in patients with Child-Pugh class a liver function. Eur J Radiol 2012;81(3):466–471. 16. Hsu CY, Huang YH, Chiou YY, et al. Comparison of radiofrequency ablation and transarterial chemoembolization for hepatocellular carcinoma within the Milan criteria: a propensity score analysis. Liver Transpl 2011;17(5):556–566.
with end-stage liver disease. Hepatology 2001;33(2):464–470. 20. Yang JD, Kim WR, Park KW, et al. Model to estimate survival in ambulatory patients with hepatocellular carcinoma. Hepatology 2012;56(2):614–621. 21. Common Terminology Criteria for Adverse Events v.4.0 (CTCAE). NIH publication #09-7473. National Cancer Institute. http:// ctep.cancer.gov/protocolDevelopment/electronic_applications/ctc.htm. Published May 29, 2009. Accessed March 1, 2013. 22. Park KW, Lim WH, Kim JH, et al. Comparison between zotarolimus-eluting stents and first generation drug-eluting stents in the treatment of patients with acute ST-segment elevation myocardial infarction. Int J Cardiol 2013;166(1):118–125.
17. Chung JW, Park JH, Han JK, Choi BI, Han MC. Hepatocellular carcinoma and portal vein invasion: results of treatment with transcatheter oily chemoembolization. AJR Am J Roentgenol 1995;165(2):315–321.
23. Kurth T, Walker AM, Glynn RJ, et al. Results of multivariable logistic regression, propensity matching, propensity adjustment, and propensity-based weighting under conditions of nonuniform effect. Am J Epidemiol 2006;163(3):262–270.
18. Shah ZK, McKernan MG, Hahn PF, Sahani DV. Enhancing and expansile portal vein thrombosis: value in the diagnosis of hepatocellular carcinoma in patients with multiple hepatic lesions. AJR Am J Roentgenol 2007;188(5):1320–1323.
24. Bizollon T, Rode A, Bancel B, et al. Di agnostic value and tolerance of Lipiodolcomputed tomography for the detection of small hepatocellular carcinoma: correlation with pathologic examination of explanted livers. J Hepatol 1998;28(3):491–496.
19. Kamath PS, Wiesner RH, Malinchoc M, et al. A model to predict survival in patients
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
Original Research n Vascular
Hyo-Joon Yang, MD Jeong-Hoon Lee, MD Dong Hyeon Lee, MD Su Jong Yu, MD Yoon Jun Kim, MD Jung-Hwan Yoon, MD Hyo-Cheol Kim, MD Jeong Min Lee, MD Jin Wook Chung, MD Nam-Joon Yi, MD Kwang-Woong Lee, MD Kyung-Suk Suh, MD Hyo-Suk Lee, MD
1
From the Department of Internal Medicine and Liver Research Institute (H.J.Y., J.H.L., D.H.L., S.J.Y., Y.J.K., J.H.Y., H.S.L.), Department of Radiology (H.C.K., J.M.L., J.W.C.), and Department of Surgery (N.J.Y., K.W.L., K.S.S.), Seoul National University College of Medicine, 103 Daehak-ro, Jongno-gu, Seoul 110-744, Korea. Received August 7, 2013; revision requested September 19; revision received November 7; accepted November 21; final version accepted November 22. Supported by the Seoul National University Hospital Research Fund (grant 04-2013-3130). Address correspondence to J.H.L. (e-mail: [email protected]).
Purpose:
To compare the effectiveness of transarterial chemoembolization (TACE), radiofrequency ablation (RFA), and hepatic resection (HR) in patients with small single-nodule hepatocellular carcinoma by using inverse probability weighting to control selection bias.
Materials and Methods:
This retrospective cohort study was approved by the institutional review board, and the requirement to obtain informed consent was waived. The study included 197 consecutive patients (146 men and 51 women; mean age 6 standard deviation, 57.4 years 6 9.7) with single-nodule hepatocellular carcinomas measuring 3 cm or smaller and no vascular invasion who were treated initially with HR (n = 52), RFA (n = 79), or TACE (n = 66) from January 2005 to December 2006 at a single tertiary hospital. The primary endpoint was overall survival.
Results:
The baseline liver status of the groups differed significantly and was most favorable for the HR group, followed by the RFA group, and then the TACE group. The 5-year overall survival rates were 93.6% in the HR group, 86.6% in the RFA group, and 74.2% in the TACE group (P = .023). However, after inverse probability weighting, weighted survival rates among the three groups were similar (5-year overall survival: 85.6% with HR, 87.6% with RFA, and 80.7% with TACE; P = .834). In multivariate Cox regression analysis, TACE showed a hazard ratio of 0.978 (95% confidence interval: 0.407, 2.347; P = .960) compared with HR and of 1.335 (95% confidence interval: 0.619, 2.879; P = .461) compared with RFA.
Conclusion:
TACE is an effective treatment that allows achievement of long-term survival rates comparable to those with HR and RFA in patients with small single-nodule hepatocellular carcinoma. RSNA, 2014
q
Online supplemental material is available for this article.
RSNA, 2014
q
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
909
and Interventional Radiology
Small Single-Nodule Hepatocellular Carcinoma: Comparison of Transarterial Chemoembolization, Radiofrequency Ablation, and Hepatic Resection by Using Inverse Probability Weighting1
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
H
epatocellular carcinoma (HCC) is the sixth most common cancer in the world and the third most common cause of cancer-related death (1). With improved surveillance of patients with chronic liver disease and advances in imaging, more patients are diagnosed with early-stage HCC (2). For the treatment of earlystage HCC, curative therapies including liver transplantation, hepatic resection (HR), and radiofrequency ablation (RFA) are recommended (3–5). Although liver transplantation is generally limited by the shortage of donor allografts, HR is widely used as the main choice of treatment for resectable HCC. However, the risk of postoperative hepatic dysfunction often precludes HR (6). RFA has been shown to be as effective as HR in the treatment of small single-nodule HCC (7,8). However, RFA of lesions located close to major organs or the
Advances in Knowledge nn Transarterial chemoembolization (TACE) showed a 5-year overall survival rate of 80.7%, which was comparable to those of hepatic resection (85.6%) and radiofrequency ablation (87.6%), in patients with a single-nodule hepatocellular carcinoma of less than or equal to 3 cm without vascular invasion when the underlying liver status was balanced among the patients receiving each treatment by using inverse probability weighting. nn Most (97.0%) of the patients treated with TACE achieved a complete response, and the achievement of complete response was one of the independent prognostic factors of survival (P , .001). nn TACE showed a significantly reduced risk of adverse events compared with radiofrequency ablation and hepatic resection (P = .007), but also had significantly higher frequency of tumor recurrence (P , .001). 910
Yang et al
liver capsule is often contraindicated (9). Consequently, transarterial chemoembolization (TACE) is also used frequently for the treatment of small single-nodule HCC when other curative treatments are not possible. On the basis of the data from randomized controlled trials and systematic review, TACE is currently recommended only for large or multinodular HCC (10–12). In comparison, there is a paucity of long-term data comparing TACE with HR or RFA for the treatment of small HCC (13,14). In the past few years, authors of some retrospective cohort studies (15,16) reported the efficacy of TACE compared with that of HR and with RFA in early-stage HCC. Because HR, RFA, and TACE are the major treatment modalities for small (3 cm) single-nodule HCC, there is a considerable need for an evaluation of the therapeutic efficacy of the three treatments in a single cohort of patients. To obtain confirmative evidence, a well-designed randomized controlled trial is necessary, but this is not easy to conduct because of the current practice guidelines and ethical concerns. A cohort study with inverse probability weighting (IPW) and a regression method was an option to balance underlying liver function among the treatment groups. The purpose of our study was to compare the effectiveness of TACE, RFA, and HR in patients with small single-nodule HCC by using IPW to control selection bias.
Implications for Patient Care nn TACE is a safe and effective alternative to hepatic resection and radiofrequency ablation for single-nodule hepatocellular carcinomas of 3 cm or smaller without vascular invasion, especially for patients who are vulnerable to adverse events of treatments. nn To attain long-term survival in patients treated with TACE, special care should be taken to obtain a complete response and to undertake surveillance for tumor recurrence.
Materials and Methods Patients This retrospective cohort study was approved by the institutional review board of Seoul National University Hospital (Seoul, Korea) and was exempt from the requirement to obtain informed consent. We included consecutive patients (a) in whom a single-nodule HCC smaller than or equal to 3 cm without vascular invasion was diagnosed from January 2005 and December 2006 at Seoul National University Hospital and (b) who were treated initially with HR, RFA, or TACE. The diagnosis of HCC was based on the recommendations of international practice guidelines (3–5). The presence of vascular invasion was confirmed with the demonstration of an enhancing expansile low-attenuation intraluminal mass or filling defect in the portal or hepatic vessels at dynamic computed tomography (CT) and/ or magnetic resonance (MR) imaging (17,18). Patients were excluded if they had (a) Child-Pugh class C liver function, (b) concomitant serious medical illness including malignant tumors other than HCC, or (c) received multimodality treatment for the single-nodule HCC. Published online before print 10.1148/radiol.13131760 Content codes: Radiology 2014; 271:909–918 Abbreviations: CI = confidence interval HCC = hepatocellular carcinoma HR = hepatic resection IPW = inverse probability weighting RFA = radiofrequency ablation TACE = transarterial chemoembolization Author contributions: Guarantors of integrity of entire study, H.J.Y., J.H.L., D.H.L., S.J.Y.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, H.J.Y., J.H.L., D.H.L., N.J.Y.; clinical studies, all authors; statistical analysis, H.J.Y., J.H.L., D.H.L.; and manuscript editing, H.J.Y., J.H.L., D.H.L., S.J.Y., Y.J.K., J.H.Y., H.C.K., J.M.L., J.W.C., K.W.L., K.S.S., H.S.L. Conflicts of interest are listed at the end of this article. See also the article by Halpern in this issue.
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
Patients were divided into three groups according to the treatment they had received: (a) HR group, (b) RFA group, and (c) TACE group. The decision of treatment modality was based mainly on practice guidelines (3–5). For the patients with wellpreserved liver function, HR was recommended initially. If the patients were not eligible for resection, RFA was recommended before other treatments. Among the patients who could undergo surgery, some received RFA because of their individual preferences. When curative therapies including HR and RFA were not possible, TACE was recommended.
Procedures HR was performed by three specialized surgeons (including K.S.S. and N.J.Y., with more than 20 and 10 years of experience in hepatic surgery, respectively). The decision about the type and extent of resection was based on the tumor location, underlying liver status, and the patients’ other medical conditions. If necessary, intraoperative ultrasonography was used to identify tumor margins and hepatic vasculature. RFA was performed by three specialized interventional radiologists (including J.M.L., with more than 15 years of experience) by using an electrosurgical unit (RF-2000; RadioTherapeutics, Mountain View, Calif) or a 200-watt radiofrequency generator (Radionics, Burlington, Mass) with a 17-gauge single cooled-tip needle electrode for HCCs smaller than 2 cm or a triple-cluster needle electrode for HCCs of 2 cm or larger. The completeness of the ablation was evaluated by using dynamic CT immediately after the procedure, and an additional session of RFA was performed for an incomplete ablation. TACE was conducted by three specialized interventional radiologists (including J.W.C., with more than 20 years of experience) initially by selecting segmental or subsegmental hepatic arteries by using a microcatheter (Microferret [Cook, Bloomington, Ind] or Progreat [Terumo, Tokyo,
Japan]). An emulsion of iodized oil (Lipiodol; Andre Guerbet, Aulnaysous-Bois, France) and doxorubicin hydrochloride (Adriamycin RDF; Ildong Pharmaceutical, Seoul, Korea) was infused through the feeder vessels. Then, by using a mixture of gelatin sponge particles (Gelfoam; Upjohn, Kalamazoo, Mich), mitomycin (Mitomycin-10; Kyowa Hakko Kogyo, Tokyo, Japan), and contrast material, embolization was performed. Iodized oil labeling of the tumor was evaluated by using unenhanced CT 3 weeks after the procedure. All the patients were followed up at months 1, 3, 6, 9, and 12 before 1 year, and then every 3–6 months when necessary. At every visit, radiologic studies, tumor markers, and biochemical liver function tests were performed.
Assessment Data on clinical and laboratory findings were collected retrospectively from all patients by reviewing the electronic medical records, and ChildPugh class, model for end-stage liver disease score (19), and model to estimate survival in ambulatory HCC patients score (20) were determined by two authors (H.J.Y. and D.H.L., with 9 and 6 years of experience, respectively) in consensus. One radiologist, (J.W.C.) independently reviewed all radiologic images, blinded to survival data, to determine singularity and size of tumor, absence of vascular invasion, treatment response, and recurrence. Study Endpoints and Definitions The primary endpoint was overall survival, defined as the time from the date of diagnosis until the date of death from any cause. The secondary endpoints included complete response, adverse events, and recurrence. Complete response was defined according to the modified Response Evaluation Criteria in Solid Tumors (4) and subcategorized into initial complete response, defined as complete response after initial treatment and overall complete response,
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Yang et al
defined as complete response after a single session or repeated sessions of a single-treatment modality. Adverse events that were grade 2 or higher in severity defined according to the National Cancer Institute Common Terminology Criteria (version 4.0) (21) were evaluated. Grade 1 events were not assessed because they were common after the three treatments and were not recorded separately in practice.
Statistical Analysis Continuous data were compared by using the analysis of variance or Kruskal-Wallis test, and categorical data were compared by using a x2 or linear-by-linear association test. Survival and recurrence were plotted by means of the Kaplan-Meier method and compared with the log-rank test. To identify independent prognostic factors and compare the treatment effects adjusted for those factors, the Cox proportional hazards model was fitted. The clinically or statistically significant variables in univariate analysis (P , .1) were included in multivariate analysis. To minimize selection bias and describe the treatment effect of different modalities better in our observational study, we used IPW (22,23). Propensity scores were calculated by generating a logistic regression model to predict the probability of each patient receiving a particular treatment on the basis of the 18 pretreatment variables described in the baseline characteristics. After the inverse probabilities of treatment weight (propensity score weight) were created, the three groups were then balanced by means of IPW. A post hoc power analysis was performed for the results of multiple Cox regression of overall survival and IPW followed by Cox regression. All statistical analyses were performed by using SPSS software (SPSS version 19.0; SPSS, Chicago, Ill) and the R statistical programming environment (version 2.13.0; R development Core Team, Vienna, Austria, http://www.R-project.org). A P value less than .05 was considered to indicate a significant difference. 911
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
Yang et al
TACE (three of 66, 4.5%; P = .007) and so did adverse events greater in severity than grade 3 (P , .001) (Table 2). The details of the adverse events are described in Table E1 (online).
Figure 1
Figure 1: Flow diagram shows the screening, enrollment, and treatment allocation of patients.
Results Study Population The study profile is shown in Figure 1. Between January 2005 and December 2006, a total of 1184 consecutive patients received a diagnosis of HCC. Among these, 230 patients without vascular invasion underwent HR, RFA, or TACE for single-nodule HCC of 3 cm or smaller. Thirty-three patients were excluded owing to multimodality treatment (n = 21), Child-Pugh class C liver function (n = 7), or concomitant serious medical illness (n = 5), thus leaving 197 patients for the analysis (HR, 52; RFA, 79; and TACE, 66). The baseline characteristics are summarized in Table 1. There were no significant differences in demographic and tumor-related characteristics among the three groups. The mean patient age was 57.4 years 6 9.7; 74.1% (146 of 197) were male, 69.0% (136 of 197) were positive for hepatitis B surface antigen, and the mean tumor size was 2.1 cm 6 0.6. However, there were significant differences in the variables reflecting underlying liver status, including model for end-stage liver disease and model 912
to estimate survival in ambulatory HCC patients score. They were most favorable in the HR group, followed by those in the RFA group, and then those in the TACE group.
Treatment Response and Adverse Events The overall complete response rates among the groups were similar; 98.1% in the HR group, 97.5% in the RFA group, and 97.0% in the TACE group (P = .705) (Table 2). However, the initial complete response rate was significantly inferior in the TACE group compared with those of the other groups (P , .001). After the initial treatment, there was one treatment-related death (1.9%) in the HR group; the patient was a 72-year-old man with chronic hepatitis C–related liver cirrhosis and Child-Pugh class A status. He underwent peripheral segmentectomy and died of postoperative pneumonia (ie, grade 5 lung infection) on postoperative day 21. There were no deaths within 90 days of treatment in the other groups. Adverse events greater in severity than grade 2 developed significantly more often after HR (12 of 52, 23.1%) than after RFA (eight of 79, 10.1%) or
Survival and Tumor Recurrence Analysis During follow-up, four of 52 (7.7%) patients in the HR group, 12 of 79 (15.2%) in the RFA group, and 18 of 66 (27.3%) in the TACE group died: two of four (50.0%), eight of 12 (66.7%), and 12 of 18 (66.7%) patients died of progression of HCC, worsening cirrhosis, or both, respectively. The cumulative probabilities of overall survival at 1, 3, and 5 years were as follows: 98.0%, 96.0%, and 93.6% in the HR group; 94.8%, 90.9%, and 86.6% in the RFA group; and 96.9%, 82.9%, and 74.2% in the TACE group, respectively (Fig 2, A). The survival rates among the groups were significantly different (P = .023). Recurrence was diagnosed in 23 of 51 (45.1%) patients in the HR group, 57 of 77 (74.0%) in the RFA group, and 54 of 64 (84.4%) in the TACE group; 5-year cumulative probabilities were 44.6%, 73.2%, and 82.5%, respectively (P , .001) (Fig 2, B). In multivariate Cox regression analysis, the treatment modality was not significantly associated with survival. Failure to achieve an overall complete response (hazard ratio, 16.750; 95% confidence interval [CI]: 4.525, 62.006; P , .001), presence of ascites (hazard ratio, 3.440; 95% CI: 1.384, 8.547; P = .008), and high model to estimate survival in ambulatory HCC patients score (hazard ratio, 2.241; 95% CI: 1.282, 3.919; P = .005) were independent predictive factors associated with poor survival (Table 3). With adjustment for these predictors, the hazard ratios between treatments were considerably reduced compared with the unadjusted hazard ratios: from 3.774 to 2.191 for TACE versus HR and from 1.858 to 1.295 for TACE versus RFA. A post hoc power analysis resulted in a power of 0.60 for TACE versus HR and of 0.11 for TACE versus RFA. The estimated hazard ratios detectable with a power
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
Table 1 Baseline Characteristics of the Study Population Parameter Age (y) Sex M F Viral marker Hepatitis B surface antigen Anti-hepatitis C virus Platelet (103/mm3)‡ Serum albumin (g/dL)§ Total bilirubin (mg/dL)|| Creatinine (mg/dL)# Prothrombin time (INR) Cirrhosis Absent Present Ascites Absent Present Esophageal varix Absent Present Child class A B Model for end-stage liver disease score Eastern Cooperative Oncology Group performance score 0 1 Tumor size 2 cm .2 cm a-fetoprotein** 200 ng/mL .200 ng/mL Model to estimate survival in ambulatory HCC patients score
HR Group (n = 52)
RFA Group (n = 79)
TACE Group (n = 66)
55.7 6 10.6*
57.2 6 9.2*
59.0 6 9.5*
38 (73.1) 14 (26.9)
59 (74.7) 20 (25.3)
49 (74.2) 17 (25.8)
34 (65.4) 6 (11.5) 153 6 51* 4.04 6 0.43* 1.02 6 0.45* 1.03 6 0.15* 1.10 6 0.09*
56 (70.9) 18 (22.8) 112 6 52* 3.71 6 0.51* 1.27 6 0.71* 1.00 6 0.16* 1.18 6 0.17*
46 (69.7) 13 (19.7) 113 6 62* 3.66 6 0.51* 1.14 6 0.54* 1.17 6 1.15* 1.22 6 0.18*
23 (44.2) 29 (55.8)
15 (19.0) 64 (81.0)
6 (9.1) 60 (90.9)
P Value .190 .979
.793 .265 ,.001† ,.001† .229 .188 ,.001† ,.001†
.031† 51 (98.1) 1 (1.9)
74 (93.7) 5 (6.3)
58 (87.9) 8 (12.1) .006†
51 (98.1) 1 (1.9)
62 (78.5) 17 (21.5)
53 (80.3) 13 (19.7)
50 (96.2) 2 (3.8) 8.3 6 2.2*
68 (86.1) 11 (13.9) 9.4 6 2.9*
55 (83.3) 11 (16.7) 10.0 6 3.9*
.040†
.016† .944
46 (88.5) 6 (11.5)
69 (87.3) 10 (12.7)
57 (86.4) 9 (13.6)
21 (40.4) 31 (59.6)
36 (45.6) 43 (54.4)
40 (60.6) 26 (39.4)
40 (76.9) 12 (23.1) 3.3 6 0.39*
63 (79.7) 16 (20.3) 3.5 6 0.41*
53 (80.3) 13 (19.7) 3.5 6 0.58*
.065
.893
.045†
Note.—Except where indicated, data are numbers of patients, with percentages in parentheses. * Data are means 6 standard deviation. †
Statistically significant.
‡
To convert to Système international (SI) units (109 per liter), multiply by 1.
§
To convert to SI units (grams per liter), multiply by 10.
||
To convert to SI units (micromoles per liter), multiply by 17.104.
#
To convert to SI units (micromoles per liter), multiply by 76.25.
** To convert to SI units (micrograms per liter), multiply by 1.
of 0.80 were 2.716 for TACE versus HR and 2.643 for TACE versus RFA (Table E2 [online]). Meanwhile, the
treatment modality (overall P , .001), total bilirubin (hazard ratio, 1.545; 95% CI: 1.152, 2.072; P = .004), and
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Yang et al
age (hazard ratio, 1.024; 95% CI: 1.005, 1.044; P = .013) were independent predictive factors of recurrence (Table E3 [online]).
Survival and Tumor Recurrence Analysis by Using Inverse Probability Weighting After IPW, the baseline characteristics, including liver status–related variables, became more balanced than they were before weighting and did not differ significantly among the groups (data were not shown). The weighted cumulative probabilities of survival among the three groups became comparable; 5-year overall survival rates were 85.6% with HR, 87.6% with RFA, and 80.7% with TACE (P = .834) (Fig 3, A). In multivariate analysis, the hazard ratios for survival between TACE and HR and between TACE and RFA were 0.978 (95% CI: 0.407, 2.347; P = .960) and 1.335 (95% CI: 0.619, 2.879; P = .461), respectively (Table E4 [online]). A post hoc power analysis showed a power of 0.03 between TACE and HR and of 0.13 between TACE and RFA. The estimated hazard ratios detectable with a power of 0.80 were 2.790 between TACE and HR and 2.699 between TACE and RFA. On the other hand, treatment with TACE versus that with HR and RFA was still an independent risk factor of recurrence even after the weighting (overall, P , .001) (Fig 3, B and Table E5 [online]). Discussion In our study, we compared the treatment effects of HR, RFA, and TACE on long-term survival by using IPW and regression methods. We found that 5-year overall survival with TACE was similar to that with HR and RFA in patients with single-nodule HCC of 3 cm or smaller without vascular invasion when the underlying liver status was balanced among the patients receiving each treatment by means of IPW. In addition, our results showed that most of the patients initially treated with TACE achieved a complete response, which was one of the independent prognostic factors of survival. 913
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
Yang et al
Table 2 Summary of Treatment Response and Adverse Events Treatment Response and Adverse Event Initial complete response rate Overall complete response rate Number of treatments One Two Three Four Five Adverse events after initial treatment Grade 2 Patients Events‡ Grade 3 Patients Events‡
HR Group (n = 52)
RFA Group (n = 79)
TACE Group (n = 66)
P Value
50 (96.2) 51 (98.1)
76 (96.2) 77 (97.5)
53 (80.3) 64 (97.0)
,.001* .705
50 (96.2) 1 (1.9) 0 (0.0) 0 (0.0) 0 (0.0)
76 (96.2) 1 (1.3) 0 (0.0) 0 (0.0) 0 (0.0)
53 (80.3) 6 (9.2) 2 (3.0) 2 (3.0) 1 (1.5)
12 (23.1) 13
8 (10.1) 12
3 (4.5) 5
.007*
7 (13.5) 7
1 (1.3) 1
0 (0.0) 0
,.001*
Note.—Except where indicated, data are numbers of patients, with percentages in parentheses. Complete response is defined according to the modified Response Evaluation Criteria in Solid Tumors. Grades are defined by the National Cancer Institute Common Terminology Criteria (version 4.0). * Statistically significant. †
Data are numbers of events.
The survival of patients with HCC is not only strongly associated with the extent of the tumor but also with the severity of underlying liver dysfunction. However, the degree of hepatic functional reserve of patients affects the decision on the choice of treatment modality. TACE, for example, is usually considered when HR or RFA is not feasible. This selection bias is the major obstacle that hampers the evaluation of effectiveness of different treatments for HCC in a cohort study. Although a randomized controlled trial would allow an unbiased comparison of treatment effectiveness, such an approach would raise an ethical problem in that the current practice guidelines recommend treating small single-nodule HCC primarily with HR and alternatively with liver transplantation or RFA rather than TACE (3–5). IPW, as one of the propensity score methods, can be a useful alternative strategy to address potential selection bias in an observational study (22,23). In our study, we 914
hypothesized that the patients’ baseline characteristics, including underlying liver function, influenced treatment selection; and we evaluated the probability of patients being included in the three treatment groups on the basis of the baseline variables, and we weighted the data from each patient inversely to each probability. This balanced the baseline characteristics, of which liver function was significantly different, to reduce selection bias and better describe treatment effects on patient survival. Our results showed that there was a significant difference in overall survival rates according to the choice of initial treatments. However, when adjusted for the covariates in the multivariate regression analysis, the treatment modality was not an independent prognostic factor of overall survival. Survival analysis by using IPW revealed comparable overall survival rates among the three treatment groups; the estimated 5-year overall survival rates were shown to differ by only 5%–7%, and
the hazard ratios of death in the TACE group versus the HR and RFA groups were remarkably decreased compared with those before IPW. These results are consistent with those of cohort studies in which multiple regression (15) and propensity score matching (16) were used; they demonstrated that TACE provided overall survival similar to HR even after adjustment for baseline variables (15) and to RFA when populations were matched by using propensity score (16). A probable explanation for this is that the survival differences observed in our population before weighting might be determined by the degree of baseline liver dysfunction, on which the treatment selection was based, rather than by the specific treatment modality itself. Therefore, TACE might offer long-term survival rates comparable to those of HR and RFA for small single-nodule HCC if underlying liver function was similar among the patients receiving each treatment. The major disadvantages of TACE were the lower rate of initial complete response (80.3%) and the higher risk of recurrence (82.5% after 5 years from complete response). Nevertheless, the TACE group showed a similar overall survival rate compared with that of the other groups after IPW. From our analysis, we can surmise the reasons for this. First, the overall complete response, rather than the initial complete response, did affect survival significantly, and the TACE group showed a high overall complete response rate similar to those of the other groups. Second, TACE showed the lowest frequency and severity of adverse events, and this may have affected survival rates of the patients in the TACE group favorably. Third, recurrence was not an independent prognostic factor of overall survival, implying that the higher risk of recurrence after TACE may not increase the risk of death in such patients with early-stage HCC. Considering these drawbacks, TACE would not be the first-choice treatment for small singlenodule HCC. However, most of those patients treated initially with TACE
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
Figure 2
Figure 2: Graphs show cumulative probability of, A, overall survival and, B, tumor recurrence according to treatment groups. Data were obtained with Kaplan-Meier method. There were significant differences among treatment groups both in overall survival (P = .023, log-rank test) and in tumor recurrence (P , .001, log-rank test).
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Yang et al
eventually can achieve a complete response, although some should receive repeated treatments, and once a complete response is achieved, long-term survival similar to that with HR and RFA would be expected. Therefore, TACE can be a safe and effective alternative treatment option, especially for those who are vulnerable to the adverse events of treatments. Although angiography has lower sensitivity for detection of small HCC than does CT or MR imaging (24), there was no patient in our cohort in whom angiography was unable to show the lesion that was already found at CT or MR imaging. Thus, the low sensitivity of angiography as a diagnostic tool might not be a substantial weakness of TACE as a therapeutic modality. The limitations of our study were the retrospective observational design and the relatively small sample size. In the initial comparison of the cumulative survival probabilities, the sample size was large enough to reveal the significant survival differences among the treatment groups. After the multiple regression and IPW analysis, however, decreased survival differences among the treatment groups resulted in low statistical power. Results of a post hoc power analysis also indicated that, under the conditions of our study, a treatment difference smaller than a 2.70 of hazard ratio of survival might exist without being detected. Thus, the failure to reach statistical significance between the treatment groups appears to be attributable not only to the lack of clinically significant differences but also to the small sample size. Therefore, a further study with a larger sample size is needed to elucidate whether the observed small differences in survival among the treatment options truly exist. Another limitation is that we included up to six covariates in the multiple Cox regression models for overall survival, which might have caused a problem of overfitting, considering the 34 deaths in our cohort. This should be confirmed in a further study with more events. Although we showed that TACE was 915
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
Yang et al
Table 3 Univariate and Multivariate Analyses of Prognostic Factors for Overall Survival Univariate Analysis Variables* Age (y) Sex (M vs F) hepatitis B surface antigen (+ vs 2) Anti-hepatitis C virus (+ vs 2) Platelet (103/mm3)† Serum albumin (g/dL)‡ Total bilirubin (mg/dL)§ Creatinine (mg/dL)|| Prothrombin time (INR) Cirrhosis (+ vs 2) Ascites (+ vs 2) Esophageal varix (+ vs 2) Child class (B vs A) Model for end-stage liver disease score ECOG PS (1 vs 0) Tumor size (.2 cm vs 2 cm) a-fetoprotein (.200 ng/mL vs 200 ng/mL)# Model to estimate survival in ambulatory HCC patients score Treatment modality (RFA vs HR) (TACE vs HR) (TACE vs RFA) Initial complete response (2 vs +) Overall complete response (2 vs +) HCC recurrence (+ vs 2)
Multivariate Analysis
Hazard Ratio
95% CI
P Value
Hazard Ratio
95% CI
P Value
1.014 1.028 1.382 1.183 0.999 0.323 1.554 1.311 11.297 3.175 5.051 1.505 3.767 1.143 1.634 1.103 1.416 3.136
0.979, 1.051 0.480, 2.201 0.626, 3.054 0.514, 2.720 0.993, 1.005 0.178, 0.588 0.918, 2.631 1.057, 1.626 1.996, 63.929 0.971, 10.385 2.281, 11.186 0.655, 3.458 1.800, 7.884 1.066, 1.225 0.676, 3.948 0.563, 2.163 0.661, 3.034 1.836, 5.356
... ... ... ... ... ... ... ... ... ... 3.440 ... ... ... ... ... ... 2.241
... ... ... ... ... ... ... ... ... ... 1.384, 8.547 ... ... ... ... ... ... 1.282, 3.919
2.032 3.774 1.858 2.835 13.507 1.628
0.655, 6.300 1.277, 11.158 0.894, 3.858 1.170, 2.835 3.920, 46.537 0.667, 3.973
.433 .944 .424 .693 .769 ,.001 .101 .014 .006 .056 ,.001 .335 ,.001 ,.001 .275 .774 .371 ,.001 .032 .220 .016 .097 .021 ,.001 .284
1.692 2.191 1.295 ... 16.750 ...
0.542, 5.279 0.705, 6.807 0.598, 2.807 ... 4.525, 62.006 ...
... ... ... ... ... ... ... ... ... ... .008 ... ... ... ... ... ... .005 .390 .365 .175 .512 ... ,.001 ...
Note.—Variables that were not significant in the univariate analysis were not included in the multivariate analysis. ECOG PS = Eastern Cooperative Oncology Group performance score. * For categorical variables with the categories in the parentheses, the former was compared with the latter (the reference) in calculating hazard ratios and 95% confidence intervals by using the Cox proportional hazards model. †
To convert to Système international (SI) units (109 per liter), multiply by 1.
‡
To convert to SI units (grams per liter), multiply by 10.
§
To convert to SI units (micromoles per liter), multiply by 17.104.
||
To convert to SI units (micromoles per liter), multiply by 76.25.
#
To convert to SI units (micrograms per liter), multiply by 1.
safer than RFA in patients with small single-nodule HCC, RFA is generally considered to be a safe procedure that causes fewer adverse events than TACE does (3,4). Therefore, caution should be exercised when applying our safety results to other patients treated with either of the modalities for lesions of different sizes or numbers. In conclusion, the results of our study show that TACE is an effective and safe treatment for small single-nodule HCC, achieving long-term survival rates comparable to those of HR and RFA, when the underlying liver dysfunction 916
is similar among the patients receiving each treatment. However, when TACE is used as an initial treatment, special care should be taken to obtain a complete response, and surveillance for tumor recurrence should be undertaken. This study may warrant a large randomized controlled trial to reassess the role of TACE in patients with small single-nodule HCC. Acknowledgments: We appreciate statistical consultation and power analysis provided by the Medical Research Collaborating Center at Seoul National University College of Medicine/ Seoul National University Hospital.
Disclosures of Conflicts of Interest: H.J.Y. No relevant conflicts of interest to disclose. J.H.L. No relevant conflicts of interest to disclose. D.H.L. No relevant conflicts of interest to disclose. S.J.Y. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: Payment for lectures from Bayer Healthcare Pharmaceuticals. Other relationships: none to disclose. Y.J.K. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: grants/grants pending from BristolMyers Squibb, Roche, JW Creagene, Bukwang Pharmaceuticals, Handok Pharmaceuticals, Hanmi Pharmaceuticals, Yuhan Pharmaceuticals, and Pharmaking; payment for lectures from Bayer, Bukwang, CJ Pharmaceuticals, Gilead Science, Handok, MSD Korea, Yuhan,
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
Figure 3
Yang et al
and Samil Pharmaceuticals. Other relationships: none to disclose. J.H.Y. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: grants/grants pending from Bayer. Other relationships: none to disclose. H.C.K. No relevant conflicts of interest to disclose. J.M.L. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: board membership for Bayer; consultancy for Bayer and Siemens; grants/grants pending from GE Healthcare, Dongseo Medical, CMS, Acuzen, Starmed, and RF Medical; payment for lectures from Bayer and Supersonic. Other relationships: none to disclose. J.W.C. No relevant conflicts of interest to disclose. N.J.Y. No relevant conflicts of interest to disclose. K.W.L. No relevant conflicts of interest to disclose. K.S.S. No relevant conflicts of interest to disclose. H.S.L. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: grants/grants pending from Dong-A Pharmaceuticals and Bukwang. Other relationships: none to disclose.
References 1. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer 2010;127(12):2893–2917. 2. Zhang BH, Yang BH, Tang ZY. Randomized controlled trial of screening for hepatocellular carcinoma. J Cancer Res Clin Oncol 2004;130(7):417–422. 3. Bruix J, Sherman M; American Association for the Study of Liver Diseases. Management of hepatocellular carcinoma: an update. Hepatology 2011;53(3):1020–1022. 4. European Association For The Study Of The Liver; European Organisation For Research And Treatment Of Cancer. EASL-EORTC clinical practice guidelines: management of hepatocellular carcinoma. J Hepatol 2012;56(4):908–943. [Published correction appears in J Hepatol 2012;56(6):1430.] 5. Benson AB 3rd, Abrams TA, Ben-Josef E, et al. NCCN clinical practice guidelines in oncology: hepatobiliary cancers. J Natl Compr Canc Netw 2009;7(4):350–391.
Figure 3: Graphs show weighted cumulative probability of, A, overall survival and, B, tumor recurrence according to treatment groups. Data were obtained with weighted Kaplan-Meier method. Weighted probabilities of overall survival were comparable (P = .834, univariate Cox regression analysis) among treatment groups, while weighted risks of tumor recurrence were still significantly different (P , .001, univariate Cox regression analysis).
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
6. Bruix J, Castells A, Bosch J, et al. Surgical resection of hepatocellular carcinoma in cirrhotic patients: prognostic value of preoperative portal pressure. Gastroenterology 1996;111(4):1018–1022. 7. Chen MS, Li JQ, Zheng Y, et al. A prospective randomized trial comparing percutaneous local ablative therapy and partial hepatectomy for small hepatocellular carcinoma. Ann Surg 2006;243(3):321–328.
917
VASCULAR AND INTERVENTIONAL RADIOLOGY: Small Single-Nodule Hepatocellular Carcinoma
8. Livraghi T, Meloni F, Di Stasi M, et al. Sustained complete response and complications rates after radiofrequency ablation of very early hepatocellular carcinoma in cirrhosis: Is resection still the treatment of choice? Hepatology 2008;47(1):82–89. 9. Livraghi T, Solbiati L, Meloni MF, Gazelle GS, Halpern EF, Goldberg SN. Treatment of focal liver tumors with percutaneous radio-frequency ablation: complications encountered in a multicenter study. Radiology 2003;226(2):441–451. 10. Llovet JM, Real MI, Montaña X, et al. Arterial embolisation or chemoembolisation versus symptomatic treatment in patients with unresectable hepatocellular carcinoma: a randomised controlled trial. Lancet 2002;359(9319):1734–1739. 11. Lo CM, Ngan H, Tso WK, et al. Randomized controlled trial of transarterial lipiodol chemoembolization for unresectable hepatocellular carcinoma. Hepatology 2002;35(5):1164–1171. 12. Llovet JM, Bruix J. Systematic review of randomized trials for unresectable hepatocellular carcinoma: Chemoembolization improves survival. Hepatology 2003;37(2):429–442. 13. Huang YH, Chen CH, Chang TT, et al. The role of transcatheter arterial embolization in patients with resectable hepatocellular carcinoma: a nation-wide, multicenter study. Liver Int 2004;24(5):419–424.
918
Yang et al
14. Lee HS, Kim KM, Yoon JH, et al. Therapeutic efficacy of transcatheter arterial chemoembolization as compared with hepatic resection in hepatocellular carcinoma patients with compensated liver function in a hepatitis B virus-endemic area: a prospective cohort study. J Clin Oncol 2002;20(22):4459–4465. 15. Hsu KF, Chu CH, Chan DC, et al. Superselective transarterial chemoembolization vs hepatic resection for resectable early-stage hepatocellular carcinoma in patients with Child-Pugh class a liver function. Eur J Radiol 2012;81(3):466–471. 16. Hsu CY, Huang YH, Chiou YY, et al. Comparison of radiofrequency ablation and transarterial chemoembolization for hepatocellular carcinoma within the Milan criteria: a propensity score analysis. Liver Transpl 2011;17(5):556–566.
with end-stage liver disease. Hepatology 2001;33(2):464–470. 20. Yang JD, Kim WR, Park KW, et al. Model to estimate survival in ambulatory patients with hepatocellular carcinoma. Hepatology 2012;56(2):614–621. 21. Common Terminology Criteria for Adverse Events v.4.0 (CTCAE). NIH publication #09-7473. National Cancer Institute. http:// ctep.cancer.gov/protocolDevelopment/electronic_applications/ctc.htm. Published May 29, 2009. Accessed March 1, 2013. 22. Park KW, Lim WH, Kim JH, et al. Comparison between zotarolimus-eluting stents and first generation drug-eluting stents in the treatment of patients with acute ST-segment elevation myocardial infarction. Int J Cardiol 2013;166(1):118–125.
17. Chung JW, Park JH, Han JK, Choi BI, Han MC. Hepatocellular carcinoma and portal vein invasion: results of treatment with transcatheter oily chemoembolization. AJR Am J Roentgenol 1995;165(2):315–321.
23. Kurth T, Walker AM, Glynn RJ, et al. Results of multivariable logistic regression, propensity matching, propensity adjustment, and propensity-based weighting under conditions of nonuniform effect. Am J Epidemiol 2006;163(3):262–270.
18. Shah ZK, McKernan MG, Hahn PF, Sahani DV. Enhancing and expansile portal vein thrombosis: value in the diagnosis of hepatocellular carcinoma in patients with multiple hepatic lesions. AJR Am J Roentgenol 2007;188(5):1320–1323.
24. Bizollon T, Rode A, Bancel B, et al. Di agnostic value and tolerance of Lipiodolcomputed tomography for the detection of small hepatocellular carcinoma: correlation with pathologic examination of explanted livers. J Hepatol 1998;28(3):491–496.
19. Kamath PS, Wiesner RH, Malinchoc M, et al. A model to predict survival in patients
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
