ORIGINAL ARTICLE
Radiofrequency Ablation Versus Hepatic Resection for Small Hepatocellular Carcinoma A Meta-analysis of Randomized Controlled Trials Xingshun Qi, MD,*w Yulong Tang, MD,z Dan An, MD,y Ming Bai, MD,* Xiaolei Shi, MD,*8 Juan Wang, MD,w Guohong Han, MD,* and Daiming Fan, MD, PhD*
Background and Goals: Whether radiofrequency ablation or hepatic resection is superior for improving the survival in patients with small hepatocellular carcinoma (HCC) remains controversial. A meta-analysis of randomized controlled trials was performed to examine this issue. Methods: PubMed, EMBASE, and Cochrane Library databases were used to identify all randomized controlled trials comparing the survival between small HCC patients receiving radiofrequency ablation and hepatic resection. The hazard ratio (HR) was pooled to compare the overall survival and recurrence-free survival rates. The odds ratio was pooled to compare the incidence of treatment-related complications. The mean difference was pooled to compare the hospitalization duration. Heterogeneity among studies was assessed. Results: Three randomized controlled trials were included in this meta-analysis. All patients met the Milan criteria. Hepatic resection was superior to radiofrequency ablation for the improvement of overall survival [HR = 1.41; 95% confidence interval (CI), 1.06-1.89; P = 0.02] and recurrence-free survival (HR = 1.41; 95% CI, 1.141.74; P = 0.001). Heterogeneity among studies was not significant (overall survival: P = 0.14; recurrence-free survival: P = 0.28). Patients treated with hepatic resection had a significantly higher incidence of treatment-related complications (odds ratio = 0.12; 95% CI, 0.03-0.47; P = 0.002) and a significantly longer hospitalization duration (mean difference: 8.77; 95% CI, 10.36 to 7.18; P < 0.00001) than those treated with radiofrequency ablation. Heterogeneity among studies was significant (treatment-related complications: P = 0.006; hospitalization duration: P = 0.003). No hospital death occurred in the 2 groups. Conclusions: Evidence from the meta-analysis of randomized controlled trials suggested that hepatic resection might improve the overall survival and recurrence-free survival in small HCC patients, whereas increase the complications and hospitalization duration. However, this conclusion should be explained with caution, due to
Received for publication March 4, 2013; accepted September 16, 2013. From the *Xijing Hospital of Digestive Diseases & State Key Laboratory of Cancer Biology; 8Department of Gastrointestinal Surgery, Xijing Hospital of Digestive Diseases; zDepartment of Implant Dentistry, School of Stomatology, Fourth Military Medical University; yDepartment of Hepatobiliary Surgery, The First Affiliated Hospital of Medical College, Xi’an Jiaotong University, Xi’an; and wDepartment of Gastroenterology, No. 463 Hospital of Chinese PLA, Shenyang, China. The authors declare that they have nothing to disclose. X.Q. and Y.T. contributed equally. Reprints: Daiming Fan, MD, PhD and Guohong Han, MD, Xijing Hospital of Digestive Diseases, Fourth Military Medical University, 27 West Changle Road, Xi’an 710032, China (e-mails: [email protected]; [email protected]). Supplemental Digital Content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Website, www.jcge.com. Copyright r 2013 by Lippincott Williams & Wilkins
450 | www.jcge.com
the absence of further subgroup analysis with respect to the outcome in patients with different tumor size (< 3 and 3 to 5 cm). Key Words: hepatic resection, hepatocelluar carcinoma, metaanalysis, radiofrequency ablation, randomized trial
(J Clin Gastroenterol 2014;48:450–457)
A
ccording to the reports about estimates of worldwide burden of cancer in 2008,1 hepatocellular carcinoma (HCC) is the sixth most common cancer and the third most frequent cause of cancer death.2 Recently, an updated systematic analysis of global burden disease studies demonstrates that the number of global death from liver cancer is increased from 463,000 in 1990 to 752,100 in 2010.3 Its global death rank also rises from 24th in 1990 to 16th in 2010.3 In the Asia-Pacific region, the ranking is remarkably raised to the top 10. Early diagnosis and treatment of HCC are the only options to decrease these patients’ mortality.2 Currently, radiofrequency ablation and hepatic resection have been recommended to improve the survival in patients with earlystage HCC (single or 3 nodules 1210: 8
Mean (SD): Mean (SD): 41.3 Median (range): 262.8 38.6 (6.2) g/L (7.8) IU/L (1.7-10,220) ng/mL
Mean (SD): Mean (SD): 36.7 Median (range): 215.5 40.1 (4.6) g/L (10.2) IU/L (0.5-8530) ng/mL
ALB
Solitary, r3 cm: 37 Solitary, 3.15 cm: 34 Solitary, r3 cm: 42 Solitary, 3.15 cm: 48
Solitary, >3 cm: 27 Solitary, 3-5 cm: 57 Multifocal (2-3 tumors): 31 Solitary, >3 cm: 44 Solitary, 3-5 cm: 45 Multifocal (2-3 tumors): 26
r2 cm: 31 2-4 cm: 53 1 tumor: 48 2 tumors: 36 r2 cm: 25 2-4 cm: 59 1 tumor: 52 2 tumors: 32
Tumor Size and Number
r
*90 patients were randomized to the RFA group. Among them, 71 received RFA and 19 withdrew their consent after randomization and received surgical resection. The authors reported the baseline characteristics of 71 patients but not those of 90 patients randomized. Accordingly, the data presented in this table were the baseline characteristics of 71 patients. In 2 studies by Huang and Chen, the units of variables were not provided. AFP indicates a fetoprotein; ALB, albumin; ALT; alanine aminotransferase; BIL, bilirubin; ICG-R15, indocyanine green retention at 15 minutes; NA, not available; RFA, radiofrequency ablation.
90
56/15
85/30
79/36
75/9
79/5
Male/ Female
J Clin Gastroenterol
Surgical resection
Chen et al (Ann Surg, 2006)21 RFA 71* 51.9 (11.2)
Surgical 115 resection
Huang et al (Ann Surg, 2010)22 RFA 115
Surgical resection
Feng et al (J Hepatol, 2012)32 RFA 84
Authors (Journal, Year)
TABLE 2. Baseline Characteristics of the Included Randomized Controlled Trials
Qi et al Volume 48, Number 5, May/June 2014
2013 Lippincott Williams & Wilkins
J Clin Gastroenterol
Volume 48, Number 5, May/June 2014
RFA Versus Resection
FIGURE 2. The forest plot of overall survival between the patients treated with resection and ablation. Studies are arranged by publication year. In a study by Chen et al,21 90 patients were randomized to the ablation group, but 71 of them received ablation and 19 withdrew their consent after randomization and received surgical resection. The authors reported the overall survival of 90 patients according to the intention-to-treat analysis. CI indicates confidence interval.
the cumulative recurrence-free survival rate was significantly better in the resection group than in the ablation group.22 The heterogeneity among studies was not significant for the recurrence-free survival (I2 = 22%; P = 0.28). Using a fixed-effects model, the pooled HR was found to be significant (HR = 1.41; 95% CI, 1.14-1.74; P = 0.001; Fig. 3), suggesting that resection was superior to ablation for the improvement of recurrence-free survival.
Treatment-related Complications All included studies showed that the incidence of treatment-related complications was significantly higher in the resection group than in the ablation group.21,22,32 The heterogeneity among studies was significant for the treatment-related complications (I2 = 80%; P = 0.006). Using a random-effects model, the pooled OR was found to be significant (OR = 0.12; 95% CI, 0.03-0.47; P = 0.002; Fig. 4), suggesting that the resection group had a higher incidence of treatment-related complications compared with the ablation group. All included studies demonstrated that hospital mortality due to complications was 0% in both groups.
Hospitalization Duration All included studies showed that the hospitalization duration was significantly longer in the resection group than in the ablation group.21,22,31,32 The heterogeneity among studies was significant for the hospitalization duration (I2 = 83%; P = 0.003). Using a random-effects model, the pooled mean difference was found to be significant (mean difference: 8.77; 95% CI, 10.36 to 7.18; P < 0.00001; Fig. 5), suggesting that the resection group had a longer hospitalization duration compared with the ablation group.
DISCUSSION The major finding of the present meta-analysis was that the overall survival and recurrence-free survival rates were significantly higher in small HCC patients treated with hepatic resection than in those treated with radiofrequency ablation. However, it should be noted that only 1 of 3 studies showed significant effect for overall survival as well as recurrence-free survival. Our meta-analysis also demonstrated that small HCC patients treated with hepatic resection had a higher incidence of treatment-related complications and longer hospitalization duration than those treated with radiofrequency ablation. These findings suggested the benefits of radiofrequency ablation in terms of complications and hospitalization duration. Notably, no patient died from treatment-related complications after either hepatic resection or radiofrequency ablation in these included randomized controlled trials. Thus, our findings basically accord with the current recommendation from the Barcelona Clinic Liver Cancer (BCLC) staging system and the European Association for the Study of the Liver— European Organisation for the Research and Treatment of Cancer clinical practice guidelines that hepatic resection should be the first-line treatment option for patients with BCLC very early or early-stage HCC, and radiofrequency ablation is considered the standard of care for such patients who are not suitable for surgery.4,5 Six previous meta-analyses have explored whether radiofrequency ablation or hepatic resection is superior for improving the survival in patients with HCC.15–20 However, results of these meta-analyses were not consistent. Three of them concluded that the resection group had a significantly higher survival rate than the radiofrequency ablation group,16,18,19 and 3 other studies demonstrated a similar survival rate between the resection and radiofrequency ablation groups.15,17,20 This unexpected phenomenon was primarily because a majority of included studies were
FIGURE 3. The forest plot of recurrence-free survival between the patients treated with resection and ablation. Studies are arranged by publication year. In a study by Chen et al,21 90 patients were randomized to the ablation group, but 71 of them received ablation and 19 withdrew their consent after randomization and received surgical resection. The authors reported the recurrence-free survival in 90 patients according to the intention-to-treat analysis. CI indicates confidence interval. r
2013 Lippincott Williams & Wilkins
www.jcge.com |
455
Qi et al
J Clin Gastroenterol
Volume 48, Number 5, May/June 2014
FIGURE 4. The forest plot of treatment-related complications between the patients treated with resection and ablation. Studies are arranged by publication year. In a study by Chen et al,21 90 patients were randomized to the ablation group, but 71 of them received ablation and 19 withdrew their consent after randomization and received surgical resection. The authors did not report the treatmentrelated complications in 90 patients randomized but only 71 patients treated with ablation. CI indicates confidence interval.
retrospective studies with the lower level of evidence. Accordingly, these meta-analyses showing a statistically significant difference between the 2 groups unanimously emphasized that their findings should be interpreted carefully.16,18,19 Compared with them, our meta-analysis has several strengths, as follows. First, we just included randomized controlled trials into the present meta-analysis but not low-quality retrospective comparative studies. Randomized controlled trials can provide the high-level evidence by evaluating the “hard” clinical endpoints and using the most efficient and reliable way.33 In contrast, retrospective studies provide the relatively low-level clinical evidence due to a potential selection bias. Indeed, the fact that only 1 or 2 randomized controlled trial was included in previous meta-analyses may be the major reason for their contradictory conclusions. Second, we extracted the HRs, but not the ORs or relative risks, to compare the overall survival and recurrence-free survival after hepatic resection and radiofrequency ablation for HCC. Cumulative overall survival or recurrence-free survival rates are time-to-event outcomes. In addition, HRs are the most appropriate parameters to measure the time-dependent outcomes.26 On comparison, ORs or relative risks are found to be appropriate for measuring the dichotomous outcomes, because they just measured the number of events without any consideration of the time when the events developed. Third, publication language was not restricted in our study. Thus, we might have a lower bias of study selection. All included patients met the Milan criteria (ie, solitary HCC nodule r5 cm or up to 3 nodules, each
Radiofrequency Ablation Versus Hepatic Resection for Small Hepatocellular Carcinoma A Meta-analysis of Randomized Controlled Trials Xingshun Qi, MD,*w Yulong Tang, MD,z Dan An, MD,y Ming Bai, MD,* Xiaolei Shi, MD,*8 Juan Wang, MD,w Guohong Han, MD,* and Daiming Fan, MD, PhD*
Background and Goals: Whether radiofrequency ablation or hepatic resection is superior for improving the survival in patients with small hepatocellular carcinoma (HCC) remains controversial. A meta-analysis of randomized controlled trials was performed to examine this issue. Methods: PubMed, EMBASE, and Cochrane Library databases were used to identify all randomized controlled trials comparing the survival between small HCC patients receiving radiofrequency ablation and hepatic resection. The hazard ratio (HR) was pooled to compare the overall survival and recurrence-free survival rates. The odds ratio was pooled to compare the incidence of treatment-related complications. The mean difference was pooled to compare the hospitalization duration. Heterogeneity among studies was assessed. Results: Three randomized controlled trials were included in this meta-analysis. All patients met the Milan criteria. Hepatic resection was superior to radiofrequency ablation for the improvement of overall survival [HR = 1.41; 95% confidence interval (CI), 1.06-1.89; P = 0.02] and recurrence-free survival (HR = 1.41; 95% CI, 1.141.74; P = 0.001). Heterogeneity among studies was not significant (overall survival: P = 0.14; recurrence-free survival: P = 0.28). Patients treated with hepatic resection had a significantly higher incidence of treatment-related complications (odds ratio = 0.12; 95% CI, 0.03-0.47; P = 0.002) and a significantly longer hospitalization duration (mean difference: 8.77; 95% CI, 10.36 to 7.18; P < 0.00001) than those treated with radiofrequency ablation. Heterogeneity among studies was significant (treatment-related complications: P = 0.006; hospitalization duration: P = 0.003). No hospital death occurred in the 2 groups. Conclusions: Evidence from the meta-analysis of randomized controlled trials suggested that hepatic resection might improve the overall survival and recurrence-free survival in small HCC patients, whereas increase the complications and hospitalization duration. However, this conclusion should be explained with caution, due to
Received for publication March 4, 2013; accepted September 16, 2013. From the *Xijing Hospital of Digestive Diseases & State Key Laboratory of Cancer Biology; 8Department of Gastrointestinal Surgery, Xijing Hospital of Digestive Diseases; zDepartment of Implant Dentistry, School of Stomatology, Fourth Military Medical University; yDepartment of Hepatobiliary Surgery, The First Affiliated Hospital of Medical College, Xi’an Jiaotong University, Xi’an; and wDepartment of Gastroenterology, No. 463 Hospital of Chinese PLA, Shenyang, China. The authors declare that they have nothing to disclose. X.Q. and Y.T. contributed equally. Reprints: Daiming Fan, MD, PhD and Guohong Han, MD, Xijing Hospital of Digestive Diseases, Fourth Military Medical University, 27 West Changle Road, Xi’an 710032, China (e-mails: [email protected]; [email protected]). Supplemental Digital Content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Website, www.jcge.com. Copyright r 2013 by Lippincott Williams & Wilkins
450 | www.jcge.com
the absence of further subgroup analysis with respect to the outcome in patients with different tumor size (< 3 and 3 to 5 cm). Key Words: hepatic resection, hepatocelluar carcinoma, metaanalysis, radiofrequency ablation, randomized trial
(J Clin Gastroenterol 2014;48:450–457)
A
ccording to the reports about estimates of worldwide burden of cancer in 2008,1 hepatocellular carcinoma (HCC) is the sixth most common cancer and the third most frequent cause of cancer death.2 Recently, an updated systematic analysis of global burden disease studies demonstrates that the number of global death from liver cancer is increased from 463,000 in 1990 to 752,100 in 2010.3 Its global death rank also rises from 24th in 1990 to 16th in 2010.3 In the Asia-Pacific region, the ranking is remarkably raised to the top 10. Early diagnosis and treatment of HCC are the only options to decrease these patients’ mortality.2 Currently, radiofrequency ablation and hepatic resection have been recommended to improve the survival in patients with earlystage HCC (single or 3 nodules 1210: 8
Mean (SD): Mean (SD): 41.3 Median (range): 262.8 38.6 (6.2) g/L (7.8) IU/L (1.7-10,220) ng/mL
Mean (SD): Mean (SD): 36.7 Median (range): 215.5 40.1 (4.6) g/L (10.2) IU/L (0.5-8530) ng/mL
ALB
Solitary, r3 cm: 37 Solitary, 3.15 cm: 34 Solitary, r3 cm: 42 Solitary, 3.15 cm: 48
Solitary, >3 cm: 27 Solitary, 3-5 cm: 57 Multifocal (2-3 tumors): 31 Solitary, >3 cm: 44 Solitary, 3-5 cm: 45 Multifocal (2-3 tumors): 26
r2 cm: 31 2-4 cm: 53 1 tumor: 48 2 tumors: 36 r2 cm: 25 2-4 cm: 59 1 tumor: 52 2 tumors: 32
Tumor Size and Number
r
*90 patients were randomized to the RFA group. Among them, 71 received RFA and 19 withdrew their consent after randomization and received surgical resection. The authors reported the baseline characteristics of 71 patients but not those of 90 patients randomized. Accordingly, the data presented in this table were the baseline characteristics of 71 patients. In 2 studies by Huang and Chen, the units of variables were not provided. AFP indicates a fetoprotein; ALB, albumin; ALT; alanine aminotransferase; BIL, bilirubin; ICG-R15, indocyanine green retention at 15 minutes; NA, not available; RFA, radiofrequency ablation.
90
56/15
85/30
79/36
75/9
79/5
Male/ Female
J Clin Gastroenterol
Surgical resection
Chen et al (Ann Surg, 2006)21 RFA 71* 51.9 (11.2)
Surgical 115 resection
Huang et al (Ann Surg, 2010)22 RFA 115
Surgical resection
Feng et al (J Hepatol, 2012)32 RFA 84
Authors (Journal, Year)
TABLE 2. Baseline Characteristics of the Included Randomized Controlled Trials
Qi et al Volume 48, Number 5, May/June 2014
2013 Lippincott Williams & Wilkins
J Clin Gastroenterol
Volume 48, Number 5, May/June 2014
RFA Versus Resection
FIGURE 2. The forest plot of overall survival between the patients treated with resection and ablation. Studies are arranged by publication year. In a study by Chen et al,21 90 patients were randomized to the ablation group, but 71 of them received ablation and 19 withdrew their consent after randomization and received surgical resection. The authors reported the overall survival of 90 patients according to the intention-to-treat analysis. CI indicates confidence interval.
the cumulative recurrence-free survival rate was significantly better in the resection group than in the ablation group.22 The heterogeneity among studies was not significant for the recurrence-free survival (I2 = 22%; P = 0.28). Using a fixed-effects model, the pooled HR was found to be significant (HR = 1.41; 95% CI, 1.14-1.74; P = 0.001; Fig. 3), suggesting that resection was superior to ablation for the improvement of recurrence-free survival.
Treatment-related Complications All included studies showed that the incidence of treatment-related complications was significantly higher in the resection group than in the ablation group.21,22,32 The heterogeneity among studies was significant for the treatment-related complications (I2 = 80%; P = 0.006). Using a random-effects model, the pooled OR was found to be significant (OR = 0.12; 95% CI, 0.03-0.47; P = 0.002; Fig. 4), suggesting that the resection group had a higher incidence of treatment-related complications compared with the ablation group. All included studies demonstrated that hospital mortality due to complications was 0% in both groups.
Hospitalization Duration All included studies showed that the hospitalization duration was significantly longer in the resection group than in the ablation group.21,22,31,32 The heterogeneity among studies was significant for the hospitalization duration (I2 = 83%; P = 0.003). Using a random-effects model, the pooled mean difference was found to be significant (mean difference: 8.77; 95% CI, 10.36 to 7.18; P < 0.00001; Fig. 5), suggesting that the resection group had a longer hospitalization duration compared with the ablation group.
DISCUSSION The major finding of the present meta-analysis was that the overall survival and recurrence-free survival rates were significantly higher in small HCC patients treated with hepatic resection than in those treated with radiofrequency ablation. However, it should be noted that only 1 of 3 studies showed significant effect for overall survival as well as recurrence-free survival. Our meta-analysis also demonstrated that small HCC patients treated with hepatic resection had a higher incidence of treatment-related complications and longer hospitalization duration than those treated with radiofrequency ablation. These findings suggested the benefits of radiofrequency ablation in terms of complications and hospitalization duration. Notably, no patient died from treatment-related complications after either hepatic resection or radiofrequency ablation in these included randomized controlled trials. Thus, our findings basically accord with the current recommendation from the Barcelona Clinic Liver Cancer (BCLC) staging system and the European Association for the Study of the Liver— European Organisation for the Research and Treatment of Cancer clinical practice guidelines that hepatic resection should be the first-line treatment option for patients with BCLC very early or early-stage HCC, and radiofrequency ablation is considered the standard of care for such patients who are not suitable for surgery.4,5 Six previous meta-analyses have explored whether radiofrequency ablation or hepatic resection is superior for improving the survival in patients with HCC.15–20 However, results of these meta-analyses were not consistent. Three of them concluded that the resection group had a significantly higher survival rate than the radiofrequency ablation group,16,18,19 and 3 other studies demonstrated a similar survival rate between the resection and radiofrequency ablation groups.15,17,20 This unexpected phenomenon was primarily because a majority of included studies were
FIGURE 3. The forest plot of recurrence-free survival between the patients treated with resection and ablation. Studies are arranged by publication year. In a study by Chen et al,21 90 patients were randomized to the ablation group, but 71 of them received ablation and 19 withdrew their consent after randomization and received surgical resection. The authors reported the recurrence-free survival in 90 patients according to the intention-to-treat analysis. CI indicates confidence interval. r
2013 Lippincott Williams & Wilkins
www.jcge.com |
455
Qi et al
J Clin Gastroenterol
Volume 48, Number 5, May/June 2014
FIGURE 4. The forest plot of treatment-related complications between the patients treated with resection and ablation. Studies are arranged by publication year. In a study by Chen et al,21 90 patients were randomized to the ablation group, but 71 of them received ablation and 19 withdrew their consent after randomization and received surgical resection. The authors did not report the treatmentrelated complications in 90 patients randomized but only 71 patients treated with ablation. CI indicates confidence interval.
retrospective studies with the lower level of evidence. Accordingly, these meta-analyses showing a statistically significant difference between the 2 groups unanimously emphasized that their findings should be interpreted carefully.16,18,19 Compared with them, our meta-analysis has several strengths, as follows. First, we just included randomized controlled trials into the present meta-analysis but not low-quality retrospective comparative studies. Randomized controlled trials can provide the high-level evidence by evaluating the “hard” clinical endpoints and using the most efficient and reliable way.33 In contrast, retrospective studies provide the relatively low-level clinical evidence due to a potential selection bias. Indeed, the fact that only 1 or 2 randomized controlled trial was included in previous meta-analyses may be the major reason for their contradictory conclusions. Second, we extracted the HRs, but not the ORs or relative risks, to compare the overall survival and recurrence-free survival after hepatic resection and radiofrequency ablation for HCC. Cumulative overall survival or recurrence-free survival rates are time-to-event outcomes. In addition, HRs are the most appropriate parameters to measure the time-dependent outcomes.26 On comparison, ORs or relative risks are found to be appropriate for measuring the dichotomous outcomes, because they just measured the number of events without any consideration of the time when the events developed. Third, publication language was not restricted in our study. Thus, we might have a lower bias of study selection. All included patients met the Milan criteria (ie, solitary HCC nodule r5 cm or up to 3 nodules, each
