ORIGINAL

ARTICLE

Comparative Efficacy of Radiofrequency and Laser Ablation for the Treatment of Benign Thyroid Nodules: Systematic Review Including Traditional Pooling and Bayesian Network Meta-analysis Eun Ju Ha, Jung Hwan Baek, Kyung Won Kim, Junhee Pyo, Jeong Hyun Lee, Seung Hee Baek, Helle Døssing, and Laszlo Hegedu¨s Department of Radiology (E.J.H.), Ajou University School of Medicine, Suwon 443–380, Korea; Department of Radiology and Research Institute of Radiology (J.H.B., K.W.K., J.H.L.), University of Ulsan College of Medicine, Asan Medical Center, Seoul 138-736, Korea; WHO Collaborating Centre for Pharmaceutical Policy and Regulation, Department of Pharmaceutical Science (J.P.), Utrecht University, Netherlands; Department of Clinical Epidemiology and Biostatistics (S.H.B.), University of Ulsan College of Medicine, Asan Medical Center, Seoul 138-736, Korea; and Departments of Otorhinolaryngology (H.D.) and Endocrinology (L.H.), Odense University Hospital, DK-5000 Odense C, Denmark

Purpose: To compare the efficacy of radiofrequency ablation (RFA) and laser ablation (LA) for treatment of benign solid thyroid nodules, using a systematic review including traditional pooling and Bayesian network meta-analysis. Materials and Methods: A comprehensive literature search in PubMed-MEDLINE, EMBASE, and the Cochrane Library databases identified prospective studies evaluating the percentage mean change [absolute mean change (mL)] in nodule volume after RFA or LA. Studies from January 1, 2000, to November 1, 2013, were included. Review of 128 potential papers, including a full-text review of 33, identified 10 eligible papers covering a total of 184 patients for meta-analysis. The percentage mean change [absolute mean change] in nodule volume over a 6-month follow-up was compared between RFA and LA. Results: Based on the traditional frequentist approach, the pooled percentage mean changes (95% confidence interval) of RFA and LA were 76.1% (70.1– 82.1) and 49.9% (41.4 –58.5), respectively, and the pooled absolute mean changes (95% confidence interval) of RFA and LA were 8.9 mL (6.6 –11.2) and 5.2 mL (4.3– 6.1), respectively. Based on the Bayesian network meta-analysis, RFA achieved a larger pooled percentage mean change (95% credible interval) and absolute mean change (95% credible interval) compared to LA [77.8% (67.7– 88.0) vs 49.5% (26.7–72.4), and 9.2 mL (5.8 –11.9) vs 5.3 mL (2.1– 8.5), respectively]. The RFA group has the highest probability of having the most efficacious treatment (98.7%). There were no major complications after either RFA or LA. Conclusions: RFA appears to be superior to LA in reducing benign solid thyroid nodule volume, despite the smaller number of treatment sessions without major side effects. (J Clin Endocrinol Metab 100: 1903–1911, 2015)

odular thyroid disease is common in adult populations (1). Although most thyroid nodules are benign and treatment is often not warranted, some patients require treatment due to pressure symptoms or cosmetic

N

complaints (2, 3). In that case, surgery, radioiodine therapy, and TSH suppression have been used; however, all these treatment options have several drawbacks (4). Therefore, percutaneous ablation therapies, such as eth-

ISSN Print 0021-972X ISSN Online 1945-7197 Printed in U.S.A. Copyright © 2015 by the Endocrine Society Received November 12, 2014. Accepted February 13, 2015. First Published Online February 19, 2015

Abbreviations: CI, confidence interval; CrI, credible interval; DIC, deviance information criterion; LA, laser ablation; RCT, randomized controlled trial; RFA, radiofrequency ablation.

doi: 10.1210/jc.2014-4077

J Clin Endocrinol Metab, May 2015, 100(5):1903–1911

jcem.endojournals.org

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 May 2015. at 04:06 For personal use only. No other uses without permission. . All rights reserved.

1903

1904

Ha et al

Comparative Efficacy of RFA and LA

anol ablation, radiofrequency ablation (RFA), and laser ablation (LA), have been introduced as satisfactory alternatives (2, 5). Ethanol ablation has been suggested as a first-line treatment modality for cystic and predominantly cystic thyroid nodules (6 –13). However, limitations related to seepage of ethanol causing serious side effects have limited its utility in solid nodules (14, 15). Thermal ablation techniques, such as LA and RFA, have been introduced and have proven effective in inducing necrosis and subsequent nodule shrinkage in solid thyroid nodules (16 –20). However, importantly, no group has accumulated sufficient clinical experience using both techniques, and a direct comparison in prospective randomized controlled trials (RCTs) is not available (2). Therefore, controversy still exists concerning whether one or the other of the two treatment modalities is superior in the therapy of benign solid thyroid nodules. Systematic review, using meta-analysis, is a well-accepted basis for providing evidence-based treatment guidelines. Bayesian network meta-analysis combines both direct and indirect evidence for treatment comparisons and estimates the inter-relations across the treatments (21, 22). Therefore, we conducted a systematic review, including traditional pooling and Bayesian network meta-analysis, for comparison of the efficacy of the two predominantly used thermal ablation techniques, RFA and LA, focusing on their efficacy and side effects in the treatment of benign solid thyroid nodules.

Materials and Methods Study objectives Our primary aim was to identify the percentage mean change [absolute mean change (mL)] in benign solid thyroid nodule volume during 6- to 12-month follow-up after RFA or LA. The secondary aim was to identify complications to these therapies.

Data sources and searches Relevant studies were identified by searching MEDLINE on PubMed, Embase, and the Cochrane Central Register of Controlled Trials using “thyroid nodule” or “thyroid tumor” as common text words combined with “percutaneous ablation,” “radiofrequency ablation,” “radio-frequency ablation,” “laser ablation,” “interstitial photocoagulation,” and “laser photocoagulation.” This was supplemented by a manual search and review of the reference lists. Two authors (J.H.B. and E.J.H.) independently performed the literature search, and disagreement regarding inclusion of the studies was resolved by discussion until unanimity was reached among all the authors. We were not blinded to authors, institutions, journals, or interventions while selecting studies or extracting the data.

J Clin Endocrinol Metab, May 2015, 100(5):1903–1911

Study selection Prospective studies published as abstracts or complete manuscripts in English between January 1, 2000, and November 1, 2013, in peer-reviewed journals were included. Only studies where RFA or LA was used as the primary treatment for benign solitary thyroid nodules were included. Studies were eligible if the initial nodule volume, the proportion of a solid component after therapy, and complications were recorded.

Data extraction and quality assessment Two authors (J.H.B. and E.J.H.) independently extracted the following information and entered it into a data extraction database: 1) study characteristics including authors, year of publication, country, study design, interventions, and number of patients; 2) demographic characteristics including patient age and sex; 3) nodule characteristics and treatment strategies including initial nodule volume, the proportion of a solid component, the number of treatment sessions, and the follow-up period; and 4) outcomes, including percentage mean change [absolute mean change (mL)] in nodule volume and side effects. Previous studies of nonsurgical treatments have suggested that variables such as the initial nodule volume (23, 24), the proportion of the solid component (14, 23, 25–27), the number of treatment sessions (23, 24, 28), and the follow-up period (23, 25, 29, 30) are related to outcome. To control for this confounding bias, which may affect the outcomes between RFA and LA therapy, we matched the propensity of both treatment groups based on the following characteristics: a prospective single treatment session, initial nodule volume between 5 and 15 mL, proportion of a solid component greater than 50%, and a follow-up period of 6 –12 months. The percentage mean change [absolute mean change] in nodule volume was preferentially selected at the 6- to 12-month follow-up. When relevant information regarding the study design or outcomes was unclear or if there was doubt regarding duplicate publications, we contacted the original authors for clarification. The methodological quality and the risk of bias among the included studies were assessed using the Cochrane Collaboration Quality Assessment Checklist in consensus between two authors (J.H.B. and E.J.H.). The Cochrane Checklist was assessed and descriptively presented in this review using the seven domains of “random sequence,” “allocation concealment,” “blinding to participants and personnel,” “blinding to outcome assessment,” “incomplete outcome data,” “selective reporting,” and “other bias.” We followed the PRISMA reporting guideline when conducting this systemic review and meta-analysis.

Data synthesis and analysis The percentage mean change [absolute mean change] in nodule volume during 6 to 12 months of follow-up was used as the main efficacy index for this meta-analysis. Two types of metaanalyses were carried out, ie, traditional frequentist meta-analysis of pooling, and Bayesian network meta-analysis. Based on the mean volume and SD value at baseline and follow-up, the percentage mean change [absolute mean change], SE, and its confidence intervals (CIs) were calculated for each study. Given the SE, we weighted and used the inverse-variance method of meta-analysis to combine the percentage mean change [absolute mean change] and its CI. For this we employed Stata software (version 13.0; StataCorp). Using both the fixed- and random-effects models (31, 32), we calculated the pooled estimates

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 May 2015. at 04:06 For personal use only. No other uses without permission. . All rights reserved.

doi: 10.1210/jc.2014-4077

Table 1.

jcem.endojournals.org

1905

Summary of the Included Studies and Patient Characteristics

First Author (Ref.)

Year

Country

Study Design

Interventions

No. of Patients

Age, y (Mean)

No. of Females: Males

Døssing (16) Døssing (35) Døssing (28)

2002 2005 2006

Denmark Denmark Denmark

Prospective OS RCT RCT

Døssing (36) Gambelunghe (37) Døssing (38) Papini (39) Baek (40) Huh (24)

2006 2006 2007 2007 2010 2012

Denmark Italy Denmark Italy Korea Korea

Prospective OS RCT RCT RCT RCT RCT

Faggiano (41)

2012

Italy

RCT

LA LA LA LA LA LA LA LA RFA RFA RFA RFA

16 15 15 15 10 13 14 21 15 15 15 20

47.0a 46.0a 46.0 45.0 46.5a 68.0 58.0 44.9 40.9 37.5 37.7 58.3

13:3 15:0 14:1 15:0 9:1 11:2 11:3 18:3 12:3 13:2 15:0 16:4

Abbreviation: OS, observational study. Data indicate the number of nodules. The percentage mean change 关absolute mean change兴 in nodule volume was preferentially selected at the 6-month follow-up and was chosen as the next nearest follow-up if there had not been a 6-month follow-up. a

The values are median age.

of the percentage mean change [absolute mean change] in nodule volume from baseline to the 6-month follow-up as well as its 95% CI for each treatment group. The estimates between the two treatments, assuming that each treatment group was well-balanced and homogeneous from the perspective of its clinical characteristics, were directly compared. Relevant study and clinical characteristics are given in Tables 1 and 2. The heterogeneity of the pooled data was statistically assessed using the Cochran Q method. Also, we assessed publication bias in traditional pooling using funnel plots and Begg’s test among both the LA and RFA groups. Bayesian network meta-analysis to allow indirect comparisons among treatment interventions was performed using both fixed- and random-effects models in WinBUGS (version 1.4.3; MRC Biostatistics Unit). The pooled estimates and posterior distributions were obtained using the Markov Chains Monte Carlo simulation. In our Bayesian analysis, we used noninformative (vague), prior distributions that allow data to drive the posterior distributions. For each model, we generated 300 000 Markov Chains Monte Carlo simulations to estimate the pos-

Table 2.

terior probabilities and relative efficacy and discarded the first 30 000 simulations as a burn-in period. The achievement of convergence was evaluated using the Brooks-Gelman-Rubin statistics, which was provided in WinBUGS. The median of the posterior distribution based on 300 000 simulations was reported as the point estimate, and we obtained the relative efficacy between treatments measured with differences in percentage mean change [absolute mean change (baseline minus follow-up)] and its 95% credible intervals (CrIs) using the 2.5th and 97.5th percentiles of the posterior distribution. We also estimated the probability of being best by ranking the treatments on a relative scale. The model made the conventional assumption that the data are based on the normal likelihood and according to a generalized linear modeling theory. The model specification becomes a normal likelihood using the identity link function. The observed data, yi, arise from a normal distribution with a mean ␪i, and variance sei ⫺ yi ⫽ (␪i,k,sei,k). The model fit was compared between random- and fixed-effect models using the deviance information criterion (DIC) to determine the statistically more appropriate model (Supplemental Data).

Nodule Characteristics in the Included Studies Initial

Døssing (16) Døssing (35) Døssing (28) Døssing (36) Gambelunghe (37) Døssing (38) Papini (39) Baek (40) Huh (24) Faggiano (41)

Outcomes

Volume, mL

Solidity

Treatment Session

10.0 9.0 10.1 10.7 10.6 10.4 10.6 11.7 7.5 13.3 13.0 13.3

Solid Solid Solid Solid Solid Solid Solid Solid Solid Solid Solid Solid

1 1 1 3 1 1 1 1 1 1 2 1

Follow-Up, mo

Volume, mL

Complication

Pain

6 6 6 6 6 7.5 6 12 6 6 6 6

5.4 5.3 5.7 4.6 6.5 5.7 4.6 6.2 1.3 3.8 3.0 3.2

⫺ ⫺ ⫺ ⫺ ⫺ ⫹a ⫺ ⫺ ⫺ ⫺ ⫺ ⫺

⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫺

The percentage mean change 关absolute mean change兴 in nodule volume was preferentially selected at the 6-month follow-up and was chosen as the next nearest follow-up if there had not been a 6-month follow-up. a

Three of the 13 patients reported fever during the first 12 hours and eight of the 13 patients reported transient hyperthyroidism.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 May 2015. at 04:06 For personal use only. No other uses without permission. . All rights reserved.

1906

Ha et al

Comparative Efficacy of RFA and LA

J Clin Endocrinol Metab, May 2015, 100(5):1903–1911

Figure 1. Flow diagram of study selection process.

Results Literature search The selection process is described in Figure 1. The literature search generated 128 initial article candidates, of which 85 were screened for eligibility after removal of duplicates. Fifty-two papers were excluded after review of the titles and the abstracts, ie, 13 review articles; four case reports; 14 letters, editorials, or conference abstracts; and 21 articles that were not in the field of interest of this study. The full text of each of the remaining 33 articles was retrieved. Search of the bibliographies of these articles yielded no additional eligible studies. Of the 33 articles, 23 were subsequently excluded after reviewing the full text: 17 were retrospective, one was not in the field of interest, two studies combined ethanol and RFA therapy, two studies dealt with cystic thyroid nodules (11, 33), and one included thyroid nodules larger than 20 mL (34). Thus, 10 studies fulfilled the eligibility criteria and were included in the meta-analyses (16, 24, 28, 35– 41). Characteristics of the included studies Characteristics of the 10 included studies are summarized in Tables 1 and 2. Eight studies were RCTs, and two studies were designed as prospective, observational studies. Of these 10 studies, three compared LA to an observational group (35, 37, 39), one compared a single vs three LA treatment sessions (28), one compared LA vs radioiodine treatment (38), two trials compared RFA vs observation (40, 41), and one compared a single vs two RFA treatment sessions (24). There was no patient overlap among the studies by Døssing et al (16, 28, 35, 36, 38). All included studies were well-balanced and homogeneous regarding nodule characteristics. The mean initial nodule volume was approximately 10 mL (range, 7.5–13.3 mL), and the proportion of the solid component was greater than 50%. All patients were followed for more than 6 months and we include a prospective, single-treatment

Figure 2. Risks of bias in the trials included in the meta-analysis. ⫹, Low risk of bias; ⫺, high risk of bias; ?, unclear risk of bias.

session study. Thus, all data could be used in the traditional, frequentist meta-analysis. The risks of bias in the trials are detailed in Figure 2. Among the eight RCTs, seven were submitted to Bayesian network meta-analysis. An RCT by Døssing et al (38) was excluded because it compared a single LA treatment session with radioiodine treatment. Figure 3 shows the two types of network diagrams used to graphically present the comparison of treatment interventions. Solid lines present the direct comparisons, and dotted lines present the indirect comparisons conducted in the network analysis. The quality of the selected studies and the risk of bias were measured using the Cochrane Review Checklist. Figure 2 shows the overall quality of the studies. Most of the studies did not report on “allocation concealment,” and “blinding of participants and personnel” was not satisfied in all studies (Figure 2).

Figure 3. Evidence network among one treatment session (A) and different treatment sessions (B) studies in the RFA and LA groups included in the meta-analysis. Direct and indirect comparisons for all lines (dotted line, indirect; solid, direct).

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 May 2015. at 04:06 For personal use only. No other uses without permission. . All rights reserved.

doi: 10.1210/jc.2014-4077

Table 3.

jcem.endojournals.org

1907

Results of the Pooled Percentage of Mean Changes in the LA and RFA Groups

Abbreviation: NA, not applicable. The difference in the pooled percentage mean change in nodule volume between the LA and RFA groups is 26.2 (95% CI, 13.6 –38.8) (P ⬍ .001). * Ref. 28. † Ref. 24.

Pooled estimates of the percentage of mean change for each intervention group The pooled estimates for the percentage mean change in nodule volume from baseline to the 6-month follow-up after RFA or LA therapy are summarized in Table 3 with corresponding forest plots. Using both the fixed- and random-effect models, both RFA and LA showed a statistically significant reduction in nodule volume at the 6-month follow-up. However, RFA achieved a higher pooled percentage mean change than LA (76.1%; 95% CI, 70.1– 82.1%; vs 49.9%; 95% CI, 41.4 –58.5%, respectively). The difference in the pooled percentage mean change in nodule volume between the RFA and LA groups is 26.2 (95% CI, 13.6 –38.8; P ⬍ .001). No statistical heterogeneity was found when analyzing the pooled data (P value by Q test in the LA group ⫽ .931; P value using the Q test in the RFA group ⫽ .980). The pooled estimates for the absolute mean change in nodule volume from baseline to the 6-month follow-up after RFA or LA therapy are summarized in Supplemental Table 1. The difference in the pooled absolute mean change in nodule volume between the RFA and LA groups is 3.7 (95% CI, 1.7–5.7; P ⬍ .001).

We did not find any publication bias in the traditional pooling above. The funnel plot was quite symmetrical, suggesting that there is no publication bias. Also, Begg’s test showed there was no statistical significance when investigating publication bias among both the RFA group (P ⬎ .99) and LA group (P ⫽ .497). Bayesian network meta-analysis: indirect and direct comparisons Tables 4 and 5 and Supplemental Table 2 show the results of the fixed- and random-effect models of the Bayesian network meta-analyses. Based on the values of the DIC (Supplemental Table 3), there were no substantial differences in model fits between the random-effect and the fixed-effect models because there was no numerical difference greater than 5 in the DIC between the two models. Therefore, we have selected the fixed-effect model to present the outcome, whereas the results from the random-effect model are presented in the Supplemental Data. Tables 4 and 5 show the distribution of probabilities for each treatment being best ranked. Both RFA (pooled mean change, 77.8%; 95% CrI, 67.7– 88.0) and LA (49.5%; 95% CrI, 26.7–72.4) significantly reduced the nodule vol-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 May 2015. at 04:06 For personal use only. No other uses without permission. . All rights reserved.

1908

Ha et al

Comparative Efficacy of RFA and LA

J Clin Endocrinol Metab, May 2015, 100(5):1903–1911

Table 4. Results of the Percentage of Mean Changes in the RFA, LA, and Control Groups Using the Bayesian Network Meta-Analysis (Fixed-Effect Model): Network 1—LA ⫽ 1, RFA ⫽ 2, Control ⫽ 3 Tx

LA

LA RFA Control

RFA

Control

Prob Best

Rank

⫺28.31 (⫺53.26, ⫺3.39)

49.53 (26.66, 72.36) 77.84 (67.7, 87.97)

.01316 .9868 .0

2 1 3

Abbreviations: Tx, treatment; Prob Best, probability of being the most efficacious treatment. Data indicate the percentage of mean change (95% CrI).

ume at the 6-month follow-up compared to the control group. RFA thus achieved a larger pooled percentage mean change at the 6-month follow-up compared to LA (28.3%; 95% CrI, 3.4 –53.3). The RFA showed the highest probability (98.7%) of being the most efficacious treatment. We also compared treatment strategies classified according to different treatment sessions, as presented in the second network diagram in Figure 3. Two treatment sessions of RFA showed the highest probability (53.0%) of being the most efficacious treatment, followed by one session of RFA (27.5%), three sessions of LA (19.2%), one session of LA (0.2%), and the control group (0.0%). The corresponding values using the absolute mean changes in the RFA and LA are summarized in Supplemental Table 4. Complications Complication data were not subjected to formal metaanalysis. The systematic review revealed that overall complications were minor for both RFA and LA (Tables 1 and 2). Major and minor complications were those defined according to that suggested by the Society of Interventional Radiology (42, 43). There were no major complications such as voice change or hypothyroidism after either RFA or LA. One study reported minor complications, including fever and transient thyrotoxicosis after LA (37). In this study, self-limiting thyrotoxicosis occurred in eight of 13 patients at 3 days after LA. Importantly, thyroid function was normalized in all at 6 weeks of follow-up, and none had hyperthyroid symptoms. Almost all patients complained of various degrees of neck pain during the procedure, although this was usually self-limiting. No pa-

tient, despite some requiring mild pain-killers, had to be hospitalized due to pain, and none of these patients had inadequate ablation. Because no study evaluated the degree of pain using a validated questionnaire, direct comparison of the therapy options could not be carried out.

Discussion This systematic review and the meta-analyses demonstrated that both RFA and LA show a significant reduction in nodule volume at the 6-month follow-up; however, RFA reveals superior efficacy to LA for volume reduction of benign solid thyroid nodules and shows that both intervention modalities are devoid of major complications. The difference in volume reduction between the RFA and LA was 29.8% using traditional, frequentist meta-analysis and 28.3% using Bayesian network meta-analysis. Based on the above, RFA showed the highest probability (98.7%) of being the most efficacious treatment. In Bayesian network meta-analysis, some controversy exists related to the difficulty in evaluating the underlying assumptions of exchangeability, consistency, and homogeneity. As presented in Tables 1 and 2, the study and patient characteristics were relatively homogenous, as also shown statistically by the I2 value in pooled estimates of the percentage mean change in each group. Importantly, in most studies, treatment interventions were randomly assigned. However, the assumption of consistency, which refers to the agreement between direct and indirect evidence, could not be evaluated because there is currently no closed loop with both direct and indirect evidence.

Table 5. Results of the Percentage of Mean Changes in the RFA, LA, and Control Groups Using the Bayesian Network Meta-Analysis (Fixed-Effect Model): Network 2—LA1 ⫽ 1, RFA1 ⫽ 2, Control ⫽ 3, LA3 ⫽ 4, RFA2 ⫽ 5 Tx LA1 RFA1 Control LA3 RFA2

LA1

RFA1

Control

LA3

RFA2

⫺27.67 (⫺52.42, ⫺2.85)

50.08 (27.41, 72.85) 77.75 (67.65, 87.86)

⫺13.29 (⫺63.54, 37.23) 14.38 (⫺41.61, 70.57) ⫺63.37 (⫺118.6, ⫺8.08)

⫺34.99 (⫺94.87, 24.77) ⫺7.32 (⫺62.52, 47.91) ⫺85.07 (⫺141.1, ⫺29.1) ⫺21.7 (⫺100.1, 56.52)

Prob Best

Rank

.002037 .2752 .0 .1924 .5304

4 2 5 3 1

Abbreviations: Tx, treatment; Prob Best, probability of being the most efficacious treatment. Data indicate the percentage of mean change (95% CrI).

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 May 2015. at 04:06 For personal use only. No other uses without permission. . All rights reserved.

doi: 10.1210/jc.2014-4077

Results related to other studies and reviews When comparing the outcomes of different treatments, it is desirable to match the treatment groups in order to control for confounding that may affect the treatment outcomes (44). Previous studies of nonsurgical treatment for benign thyroid nodules have shown that four factors are related to nodule-volume reduction after thermal ablation. Lim et al (23) reported that efficacy is related to the initial nodule volume and solidity, as seen in their 4-year follow-up of 111 patients. Huh et al (24) and Døssing et al (28) demonstrated that multiple-session ablation was superior to single-session ablation (24, 28). Long-term follow-up studies after treatment by Lim et al (23) and Valcavi et al (30) have visualized the importance of the length of the follow-up period. Therefore, we matched for pertinent initial phenotype features and length of follow-up for the treatment groups. This led to the exclusion of three studies: two that included patients with cystic nodules (11, 33), and one that enrolled patients with nodules greater than 20 mL (34). Following this matching, our meta-analyses demonstrated a significantly higher volume reduction after RFA compared to LA, despite the smaller number of treatment sessions. This is in line with previous publications indicating that the efficacy of RFA may be superior to that of LA (45– 47). Although there have been no direct comparison studies between the two modalities in terms of long-term follow-up, the studies with more than 2 years of follow-up suggest that the volume reduction after RFA (up to 90 –92%) may be superior to that following LA (around 50%) (23, 30). One explanation for the seemingly superior efficacy of RFA may be the use of the moving-shot technique and straight, internally cooled electrodes (23, 24). This technique is based on the concept of moving the electrode during the procedure (5) and thereby maximizing ablation of the entire tumor margins, which is essential for preventing marginal regrowth and to effectively reduce the nodule volume (23). In addition, it is also effective in the treatment of hypervascular nodules, using the friction heat and thus minimizing the heat-sink effect. Also, moving the electrode tip unit-by-unit allows the entire tumor to be treated safely. Initial LA studies placed the needle/fiber in the center of the nodule during the ablation (5). Such a placement would not achieve treatment of the nodule periphery and would therefore lead to regrowth of the nodule margins, as seen with long-term follow-up (30). In our study, one RFA session, using the moving-shot technique and straight internally cooled electrodes, achieves the largest volume reduction and seems to have the highest probability of being the most efficacious treatment even when compared to three sessions of LA. However, because recent studies employing LA use a similar technique with

jcem.endojournals.org

1909

moving the fiber during therapy (2, 48), the difference in efficacy is difficult to explain. It may also rely on differences in type of energy and the energy delivered per milliliter of thyroid tissue. It is, therefore, unclarified whether the two modalities differ in efficacy and overall superiority if factors such as energy delivered per milliliter of thyroid tissue, treatment time, and treatment technique were congruent. A fair comparison would need to be based on a prospective randomized study additionally taking overall cost and quality of life issues into consideration in longterm follow-up studies. Regarding the devices, the modified straight type of internally cooled electrodes has been widely used for thyroid RFA (5). This electrode is short (7 cm) and thin (18 gauge), for easy control when using the moving-shot technique. Internally cooled electrodes allow more pronounced ablation volume by minimizing the carbonization. Variable-sized active tips (0.5, 1.0, 1.5, or 2.0 cm) can be used depending on the nodule size. In comparison, LA uses a 300- to 400-␮m diameter, plane-cut optic fiber through the sheath of a 21-gauge needle, and with 5–10 mm of the bare fiber being in direct contact with thyroid tissue. Although multiple fibers can achieve a larger ablation volume, controlling several fibers may be relatively difficult when compared with the single electrode used in RFA. However, from a safety standpoint, both seem safe with very limited side effects (2, 48). Limitations of this study Our conclusion from this meta-analysis may be influenced by various factors. First, there are relatively few studies included in this review, and many have been excluded to obtain a more homogeneous group of studies allowing comparisons. This can lead to false-positive or false-negative conclusions, ie, the risk of random errors, and a potential publication bias might also exist. We used Begg’s method to determine the presence of publication bias for traditional pooling of mean change among the LA group and the RFA group. The fact that we did not find any statistical significance in the assessment of publication bias, with symmetrical distribution as shown in the funnel plot, argues against such bias. Second, none of the trials were double-blinded, which, given the nature of these techniques, is not possible. However, those who assessed the treatment outcome were blinded in most of the studies, suggesting that outcome is unlikely to be affected by the lack of blinding. The randomization procedure was unclear or inadequate in some of the trials. Third, we compared neither the cost effectiveness nor the influence on quality of life in this study (49). Further studies regarding cost-effectiveness and quality of life, preferably using a thyroid-specific questionnaire, are required in the future.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 May 2015. at 04:06 For personal use only. No other uses without permission. . All rights reserved.

1910

Ha et al

Comparative Efficacy of RFA and LA

Despite all of the above limitations, this meta-analysis constitutes the best available evidence for RFA and LA treatment efficacy of benign solid thyroid nodules. Conclusions Based on this review, both RFA and LA treatment offer a significant reduction in nodule volume; however, RFA appears to be superior to LA in reducing benign solid thyroid nodule volume despite the smaller number of treatment sessions. Both techniques are devoid of major complications. Further investigations including randomized prospective studies focusing on efficacy, side effects, costs, and quality of life, preferably utilizing disease-specific questionnaires, are warranted.

Acknowledgments Address all correspondence and requests for reprints to: Jung Hwan Baek, MD, Department of Radiology and Research Institute of Radiology, University of Ulsan College of Medicine, Asan Medical Center, 86 Asanbyeongwon-Gil, Songpa-gu, Seoul 138-736, Korea. E-mail: [email protected]. Disclosure Summary: The authors have nothing to disclose.

References 1. Gharib H, Papini E. Thyroid nodules: clinical importance, assessment, and treatment. Endocrinol Metab Clin North Am. 2007;36: 707–735, vi. 2. Gharib H, Hegedu¨s L, Pacella CM, Baek JH, Papini E. Clinical review: nonsurgical, image-guided, minimally invasive therapy for thyroid nodules. J Clin Endocrinol Metab. 2013;98:3949 –3957. 3. Na DG, Lee JH, Jung SL, et al. Radiofrequency ablation of benign thyroid nodules and recurrent thyroid cancers: consensus statement and recommendations. Korean J Radiol. 2012;13:117–125. 4. Hegedu¨s L. Clinical practice. The thyroid nodule. N Engl J Med. 2004;351:1764 –1771. 5. Baek JH, Lee JH, Valcavi R, Pacella CM, Rhim H, Na DG. Thermal ablation for benign thyroid nodules: radiofrequency and laser. Korean J Radiol. 2011;12:525–540. 6. Monzani F, Lippi F, Goletti O, et al. Percutaneous aspiration and ethanol sclerotherapy for thyroid cysts. J Clin Endocrinol Metab. 1994;78:800 – 802. 7. Zingrillo M, Torlontano M, Chiarella R, et al. Percutaneous ethanol injection may be a definitive treatment for symptomatic thyroid cystic nodules not treatable by surgery: five-year follow-up study. Thyroid. 1999;9:763–767. 8. Zingrillo M, Torlontano M, Ghiggi MR, et al. Percutaneous ethanol injection of large thyroid cystic nodules. Thyroid. 1996;6:403– 408. 9. Valcavi R, Frasoldati A. Ultrasound-guided percutaneous ethanol injection therapy in thyroid cystic nodules. Endocr Pract. 2004;10: 269 –275. 10. Sung JY, Kim YS, Choi H, Lee JH, Baek JH. Optimum first-line treatment technique for benign cystic thyroid nodules: ethanol ablation or radiofrequency ablation? AJR Am J Roentgenol. 2011; 196:W210 –W214. 11. Sung JY, Baek JH, Kim KS, et al. Single-session treatment of benign cystic thyroid nodules with ethanol versus radiofrequency ablation: a prospective randomized study. Radiology. 2013;269:293–300.

J Clin Endocrinol Metab, May 2015, 100(5):1903–1911

12. Bennedbaek FN, Hegedu¨s L. Treatment of recurrent thyroid cysts with ethanol: a randomized double-blind controlled trial. J Clin Endocrinol Metab. 2003;88:5773–5777. 13. Paschke R, Hegedu¨s L, Alexander E, Valcavi R, Papini E, Gharib H. Thyroid nodule guidelines: agreement, disagreement and need for future research. Nat Rev Endocrinol. 2011;7:354 –361. 14. Kim YJ, Baek JH, Ha EJ, et al. Cystic versus predominantly cystic thyroid nodules: efficacy of ethanol ablation and analysis of related factors. Eur Radiol. 2012;22:1573–1578. 15. Spiezia S, Vitale G, Di Somma C, et al. Ultrasound-guided laser thermal ablation in the treatment of autonomous hyperfunctioning thyroid nodules and compressive nontoxic nodular goiter. Thyroid. 2003;13:941–947. 16. Døssing H, Bennedbaek FN, Karstrup S, Hegedu¨s L. Benign solitary solid cold thyroid nodules: US-guided interstitial laser photocoagulation–initial experience. Radiology. 2002;225:53–57. 17. Pacella CM, Bizzarri G, Guglielmi R, et al. Thyroid tissue: USguided percutaneous interstitial laser ablation–a feasibility study. Radiology. 2000;217:673– 677. 18. Papini E, Guglielmi R, Bizzarri G, Pacella CM. Ultrasound-guided laser thermal ablation for treatment of benign thyroid nodules. Endocr Pract. 2004;10:276 –283. 19. Deandrea M, Limone P, Basso E, et al. US-guided percutaneous radiofrequency thermal ablation for the treatment of solid benign hyperfunctioning or compressive thyroid nodules. Ultrasound Med Biol. 2008;34:784 –791. 20. Spiezia S, Garberoglio R, Di Somma C, et al. Efficacy and safety of radiofrequency thermal ablation in the treatment of thyroid nodules with pressure symptoms in elderly patients. J Am Geriatr Soc. 2007; 55:1478 –1479. 21. Jansen JP, Fleurence R, Devine B, et al. Interpreting indirect treatment comparisons and network meta-analysis for health-care decision making: report of the ISPOR Task Force on Indirect Treatment Comparisons Good Research Practices: part 1. Value Health. 2011; 14:417– 428. 22. Hoaglin DC, Hawkins N, Jansen JP, et al. Conducting indirecttreatment-comparison and network-meta-analysis studies: report of the ISPOR Task Force on Indirect Treatment Comparisons Good Research Practices: part 2. Value Health. 2011;14:429 – 437. 23. Lim HK, Lee JH, Ha EJ, Sung JY, Kim JK, Baek JH. Radiofrequency ablation of benign non-functioning thyroid nodules: 4-year follow-up results for 111 patients. Eur Radiol. 2013;23:1044 –1049. 24. Huh JY, Baek JH, Choi H, Kim JK, Lee JH. Symptomatic benign thyroid nodules: efficacy of additional radiofrequency ablation treatment session–prospective randomized study. Radiology. 2012; 263:909 –916. 25. Jeong WK, Baek JH, Rhim H, et al. Radiofrequency ablation of benign thyroid nodules: safety and imaging follow-up in 236 patients. Eur Radiol. 2008;18:1244 –1250. 26. Jang SW, Baek JH, Kim JK, et al. How to manage the patients with unsatisfactory results after ethanol ablation for thyroid nodules: role of radiofrequency ablation. Eur J Radiol. 2012;81:905–910. 27. Lee JH, Kim YS, Lee D, Choi H, Yoo H, Baek JH. Radiofrequency ablation (RFA) of benign thyroid nodules in patients with incompletely resolved clinical problems after ethanol ablation (EA). World J Surg. 2010;34:1488 –1493. 28. Døssing H, Bennedbaek FN, Hegedu¨s L. Effect of ultrasoundguided interstitial laser photocoagulation on benign solitary solid cold thyroid nodules: one versus three treatments. Thyroid. 2006; 16:763–768. 29. Døssing H, Bennedbæk FN, Hegedu¨s L. Long-term outcome following interstitial laser photocoagulation of benign cold thyroid nodules. Eur J Endocrinol. 2011;165:123–128. 30. Valcavi R, Riganti F, Bertani A, Formisano D, Pacella CM. Percutaneous laser ablation of cold benign thyroid nodules: a 3-year follow-up study in 122 patients. Thyroid. 2010;20:1253–1261. 31. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7:177–188.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 May 2015. at 04:06 For personal use only. No other uses without permission. . All rights reserved.

doi: 10.1210/jc.2014-4077

32. Demets DL. Methods for combining randomized clinical trials: strengths and limitations. Stat Med. 1987;6:341–350. 33. Døssing H, Bennedbæk FN, Hegedu¨s L. Interstitial laser photocoagulation (ILP) of benign cystic thyroid nodules–a prospective randomized trial. J Clin Endocrinol Metab. 2013;98:E1213–E1217. 34. Spiezia S, Garberoglio R, Milone F, et al. Thyroid nodules and related symptoms are stably controlled two years after radiofrequency thermal ablation. Thyroid. 2009;19:219 –225. 35. Døssing H, Bennedbaek FN, Hegedu¨s L. Effect of ultrasoundguided interstitial laser photocoagulation on benign solitary solid cold thyroid nodules - a randomised study. Eur J Endocrinol. 2005; 152:341–345. 36. Døssing H, Bennedbaek FN, Hegedu¨s L. Beneficial effect of combined aspiration and interstitial laser therapy in patients with benign cystic thyroid nodules: a pilot study. Br J Radiol. 2006;79:943–947. 37. Gambelunghe G, Fatone C, Ranchelli A, et al. A randomized controlled trial to evaluate the efficacy of ultrasound-guided laser photocoagulation for treatment of benign thyroid nodules. J Endocrinol Invest. 2006;29:RC23–RC26. 38. Døssing H, Bennedbaek FN, Bonnema SJ, Grupe P, Hegedu¨s L. Randomized prospective study comparing a single radioiodine dose and a single laser therapy session in autonomously functioning thyroid nodules. Eur J Endocrinol. 2007;157:95–100. 39. Papini E, Guglielmi R, Bizzarri G, et al. Treatment of benign cold thyroid nodules: a randomized clinical trial of percutaneous laser ablation versus levothyroxine therapy or follow-up. Thyroid. 2007; 17:229 –235. 40. Baek JH, Kim YS, Lee D, Huh JY, Lee JH. Benign predominantly solid thyroid nodules: prospective study of efficacy of sonographically guided radiofrequency ablation versus control condition. AJR Am J Roentgenol. 2010;194:1137–1142. 41. Faggiano A, Ramundo V, Assanti AP, et al. Thyroid nodules treated

jcem.endojournals.org

42.

43.

44.

45.

46.

47.

48.

49.

50.

1911

with percutaneous radiofrequency thermal ablation: a comparative study. J Clin Endocrinol Metab. 2012;97:4439 – 4445. Burke DR, Lewis CA, Cardella JF, et al. Quality improvement guidelines for percutaneous transhepatic cholangiography and biliary drainage. J Vasc Interv Radiol. 2003;14:S243–S246. Lewis CA, Allen TE, Burke DR, et al. Quality improvement guidelines for central venous access. The Standards of Practice Committee of the Society of Cardiovascular & Interventional Radiology. J Vasc Interv Radiol. 1997;8:475– 479. McDonald RJ, McDonald JS, Kallmes DF, Carter RE. Behind the numbers: propensity score analysis-a primer for the diagnostic radiologist. Radiology. 2013;269:640 – 645. Ha EJ, Baek JH, Lee JH. The efficacy and complications of radiofrequency ablation of thyroid nodules. Curr Opin Endocrinol Diabetes Obes. 2011;18:310 –314. Shin JE, Baek JH, Lee JH. Radiofrequency and ethanol ablation for the treatment of recurrent thyroid cancers: current status and challenges. Curr Opin Oncol. 2013;25(1):14 –19. Wong KP, Lang BH. Use of radiofrequency ablation in benign thyroid nodules: a literature review and updates. Int J Endocrinol. 2013;2013:428363. Papini E, Pacella CM, Hegedus L. Diagnosis of endocrine disease: thyroid ultrasound (US) and US-assisted procedures: from the shadows into an array of applications. Eur J Endocrinol. 2014;170: R133–R146. Watt T, Hegedu¨s L, Groenvold M, et al. Validity and reliability of the novel thyroid-specific quality of life questionnaire, ThyPRO. Eur J Endocrinol. 2010;162:161–167. Døssing H, Bennedbaek FN, Hegedu¨s L. Effect of ultrasoundguided interstitial laser photocoagulation on benign solitary solid cold thyroid nodules: one versus three treatments. Thyroid. 2006; 16:763–768.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 May 2015. at 04:06 For personal use only. No other uses without permission. . All rights reserved.

Comparative efficacy of radiofrequency and laser ablation for the treatment of benign thyroid nodules: systematic review including traditional pooling and bayesian network meta-analysis.

To compare the efficacy of radiofrequency ablation (RFA) and laser ablation (LA) for treatment of benign solid thyroid nodules, using a systematic rev...
1MB Sizes 0 Downloads 5 Views