CLINICAL STUDY

Risk Factors for Pneumothorax Complicating Radiofrequency Ablation for Lung Malignancy: A Systematic Review and Meta-Analysis Sean A. Kennedy, BASc, Lazar Milovanovic, BASc, Dyda Dao, BS, Forough Farrokhyar, MPhil, PhD, and Mehran Midia, MD, FRCPC

ABSTRACT Purpose: To assess the potential risk factors for pneumothorax secondary to pulmonary radiofrequency (RF) ablation. Materials and Methods: Six electronic databases were searched from inception to February 2014 for studies assessing potential patient-related, tumor-related, or treatment-related risk factors for pneumothorax during pulmonary RF ablation. Study selection, data collection, and quality assessment were done by three independent reviewers. Results: Among 771 studies identified in the search, 10 retrospective cohort studies met inclusion criteria. There were 981 patients (61.5% male) with a mean age of 64.2 years included (259 primary lung tumors, 722 metastatic tumors). The prevalence of pneumothorax was 37% (95% confidence interval [CI], 29%–46%) in 1,916 RF ablation sessions. The potential patient-related and tumor-related risk factors for pneumothorax were increased age (mean difference [MD], 2.09; 95% CI [0.11–4.06]; I2 ¼ 0%), male gender (unadjusted odds ratio [OR], 2.20; 95% CI [1.49-3.27]; I2 ¼ 0%), no history of lung surgery (unadjusted OR, 0.29; 95% CI [0.19–0.44]; I2 ¼ 0%), and a greater number of tumors ablated (MD, 0.50; 95% CI [0.27–0.73]; I2 ¼ 0%). Conclusion: Based on available observational studies, the results suggest risk factors for pneumothorax secondary to pulmonary RF ablation may include increased age, male gender, no history of lung surgery, number of tumors ablated, and increased length of the aerated lung traversed by the electrode. The findings from this systematic review should be interpreted with caution because of the inherent limitations of the retrospective observational design.

ABBREVIATIONS CI = confidence interval, MD = mean difference, NOS = Newcastle-Ottawa scale, OR = odds ratio, PPAP = post–pulmonary ablation pneumothorax

First described by Dupuy et al in 2000 (1), image-guided percutaneous radiofrequency (RF) ablation has become an effective modality to treat both primary and metastatic lung tumors in patients who are not surgical candidates because of advanced stage of the disease or associated comorbidities (2,3). Compared with traditional

From the Michael G. DeGroote School of Medicine (S.A.K., L.M.) and Departments of Surgery (D.D., F.F.), Clinical Epidemiology and Biostatistics (D.D., F.F.), and Diagnostic Imaging (M.M.), McMaster University, 1200 Main Street West, Hamilton, Ontario L7P 4V9, Canada. Received May 28, 2014; final revision received July 13, 2014; accepted July 23, 2014. Address correspondence to M.M.; E-mail: [email protected] None of the authors have identified a conflict of interest. Figure E1 is available online at www.jvir.org. & SIR, 2014 J Vasc Interv Radiol 2014; 25:1671–1681 http://dx.doi.org/10.1016/j.jvir.2014.07.025

surgical options for the treatment of lung tumors, benefits of RF ablation include decreased mortality and morbidity, reduced hospitalization time, preserved pulmonary functional reserve, and reduced cost (1–3). Pulmonary RF ablation has a reported complication rate of 15.2%–55.6% and a mortality rate ranging from 0%–5.6% (2). Potential complications of lung RF ablation include pneumothorax, pleural effusion, pneumonia, pulmonary abscess, hemothorax, pulmonary hemorrhage, and hemoptysis (2–4). Pneumothorax after pulmonary ablation (post–pulmonary ablation pneumothorax [PPAP]) is the most common complication; it is estimated to occur in 4.5%–61.1% of cases (2). Large, symptomatic pneumothoraces requiring chest tube insertion have been reported in 3.3%–38.9% of patients (2). PPAP is the most common cause of morbidity after lung RF ablation (3). Various conflicting risk factors for pneumothorax have been described in the literature,

1672



Postpulmonary Ablation Pneumothorax

including emphysema, number of tumors ablated, and length of lung traversed to ablate the target tumor (2,3). The present systematic review and meta-analysis aims to assess risk factors for PPAP. It is hoped that this review will help interventionalists develop appropriate prevention and monitoring strategies in patients at high risk for developing pneumothorax.

MATERIALS AND METHODS This systematic review and meta-analysis was conducted in accordance with a protocol developed a priori and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (5).

Eligibility Criteria The eligibility criteria were developed based on input of all authors. The study inclusion criteria were as follows: (a) studies of any type examining potential patient-related, tumor-related, and treatment-related risk factors for PPAP following RF ablation; (b) studies including patients Z 18 years old with any type of primary or metastatic lung malignancies; and(c) studies published in the English language at any date. The exclusion criteria were as follows: (a) case reports and case series (o 10 patients), (b) commentaries or editorials discussing pulmonary RF ablation without reporting methodologies and extractable results, (c) conference and meeting abstracts, (d) animal or basic science studies, and (e) review articles. Studies that combined other therapies with RF ablation (ie, microwave ablation, chemoembolization) were also excluded.

Kennedy et al



JVIR

Study Selection Three reviewers (S.K., L.M., D.D.) independently assessed all study abstracts based on criteria defined beforehand. Subsequently, eligibility of full-text articles of the studies that passed the abstract review was assessed by all reviewers. Disagreements on study inclusion were resolved through discussion and unanimous agreement among the reviewers. During discussion, the entire article was read by all reviewers to ensure all eligibility criteria were met. The corresponding author (M.M.) reviewed all articles meeting eligibility criteria as well as articles excluded from the study before data extraction. The authors of the articles included were contacted if pneumothorax risk factors were mentioned in the text but statistical data or information regarding methodologies or both were insufficient in their published series.

Data Extraction Using a data extraction form developed a priori, two reviewers (S.K., L.M., or D.D.) independently extracted the following information from included studies: study design and time frame, study location, number of patients with primary and metastatic lung cancer, patient demographics, number of RF ablation sessions performed, modalities for and timing of pneumothorax diagnosis, definitions, severity and prevalence of pneumothorax, and data on pneumothorax risk factors assessed. For risk factors assessed, we recorded odds ratio (OR), relative risks, corresponding 95% confidence interval (CI), mean, standard deviation (SD), and statistical methods used, where appropriate.

Definition of Pneumothorax

Quality Assessment

In the included studies, pneumothorax was defined as the accumulation of air in the pleural space that was documented by either computed tomography (CT) scan or chest radiograph at any time during or after RF ablation.

Three reviewers (S.K., L.M., and D.D.) assessed the methodologic quality of the included studies using the Newcastle-Ottawa scale (NOS) (6). The NOS is a validated tool that uses a star system to assess observational studies based on selection of the study cohorts, comparability of the cohorts, and assessment of outcome. A maximum of nine stars can be given to a study. Consensus on the number of stars (quality score) given per study was reached through discussion between the three reviewers (S.K., L.M., D.D.) to ensure unanimous agreement. During discussion, the entire study publication was read by all reviewers to assign stars accurately.

Search Strategy An electronic search strategy was developed a priori in conjunction with an experienced medical librarian. Electronic searches were conducted in the following databases from database inception date to February 2014: MEDLINE (Ovid and PubMed), Embase, Cumulative Index to Nursing and Allied Health Literature, and Cochrane Central Register of Controlled Trials. The searches included the following key words or medical subject heading or both: “pneumothorax,” “pneumothoraces,” “pneumatothorax,” “aerothorax,” “collapsed lung,” “catheter ablation,” “radiofrequency ablation,” “RF ablation,” “thermal ablation,” “ablative therapy,” “ablation,” “lung neoplasm,” “lung tumor,” “lung,” and “pulmonary.” Electronic searches were limited to the English language and human study populations (Figure E1 [available online at www.jvir.org]).

Statistical Analysis With respect to the study selection and quality assessment, interreviewer agreement was measured with Fleiss κ statistic and intraclass correlation coefficient using IBM SPSS Statistics for Windows, Version 22.0 software (IBM Corp, Armonk, New York) (7). Weighted means (SD) were also calculated. A random-effects model was used for all meta-analyses based on a priori decision.

Volume 25



Number 11



November



2014

The DerSimonian and Laird random-effects model was used to calculate the pooled weighted proportions of pneumothorax and corresponding 95% CIs using StatsDirect software (www.statsdirect.com). Data on the risk factors were pooled with inverse variance weighting using the Cochrane Collaboration software RevMan Version 6.0 (University of Oxford, Oxford, United Kingdom). For risk factors that were continuous variables, the mean difference (MD) and corresponding 95% CI between patients with and without pneumothorax were calculated. For risk factors that were dichotomized, unadjusted ORs and corresponding 95% CIs were calculated from frequency tables of individual studies and pooled where appropriate. In addition, data were combined via weighted mean (SD) calculations for studies that provided data for different severities of pneumothorax where applicable. Between-study heterogeneity was assessed using the I2 statistic. I2 thresholds of o 25%, 25%–50%, and 4 50% indicate low, moderate, and large heterogeneity.

RESULTS

1673

patient populations, and the more recent of these results was included where overlap occurred (Fig 1) (3,8–10). There were 10 study series from 12 articles included.

Study Characteristics The baseline characteristics of the 10 included studies are presented in Table 1. There were 981 patients (mean age [SD], 64.2 [11.4] y; 61.5% male) included (259 with primary lung tumors and 722 with metastatic tumors). There were 1,916 RF ablation sessions performed, of which 803 resulted in pneumothorax. The definition and characterization of severity of pneumothorax varied among studies (Table 2). The prevalence of PPAP among RF ablation sessions was 37% (95% CI [29%– 46%], I2 ¼ 96.6%). In three studies, symptomatic pneumothoraces were managed with aspiration and later chest tube insertion (8,9,15). Seven studies managed large, symptomatic pneumothorax with chest tube insertion only (3,10,11,13,14,16,17). Subgroup analysis of the studies that discussed intervention for symptomatic pneumothorax showed 29% (95% CI [19%–41%], I2 = 76.6%) of pneumothorax cases required aspiration or chest tube insertion or both for management.

Literature Search and Study Selection Using our search criteria, 771 abstracts were identified. After removal of duplicates, 469 abstracts were screened based on eligibility criteria, and 116 were selected for full-text review (κ ¼ 0.89). From the full-text review, 12 articles were found to fit the eligibility criteria (3,4,8–17). Of the 12 articles, two pairs (one by Yamagami et al (9) and Yoshimatsu et al (8) and the other by Yan et al (10) and Zhu et al (3)) were identified to have overlapping

Methodologic Quality of Included Studies NOS quality scores are listed in Table 1. The intraclass correlation coefficient for consistency in the initial scores given by the reviewers was 75% (95% CI [34%–95%]). All 10 of the included studies were retrospective, singlegroup observational studies. The quality of the selection criteria for each study was assessed based on described patient eligibility criteria and completeness of patient

Figure 1. Systematic study selection process according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement. CINAHL ¼ Cumulative Index to Nursing and Allied Health Literature.

1674

Table 1 . Individual Study Characteristics

Mean (SD) Age (y)

Cancer Type

55

56.8%

62.0 (8.2)

PC: 3

142

224

64.8%

64.0 (10.0)

MC: 34 PC: 30

2002–2010

420

1,000

60.2%

63.0 (14.6)

South Korea

2000–2002

30

30

83.3%

65.2 (8.5)

Germany

2005–2008

82

124

56.1%

64.0 (10.2)

Okuma et al., 2008 (16)

Japan

2000–2007

57

112

78.9%

70.4 (10.9)

MC: 112 PC: 137 Series CC-1-100 (Radionics/ Valleylab) MC: 283 PC: 26 Series CC-1 (Radionics/ Valleylab) MC: 4 PC: 10 POWER System (Celon AG Medical Instruments, Teltow, Germany) MC: 72 PC: 29 RF 2000 (Boston Scientific)

Steinke et al, 2004 (17)

Australia

2000–2003

23

23

60.9%

63.0 (8.3)

MC: 28 PC: 0

22

25

54.5%

62.3 (NR)

MC: 23 PC: 4

68

194

57.4%

68.0 (9.2)

MC: 18 PC: 14

100

129

56.0%

65.0 (8.0)

MC: 54 PC: 6

Country

Gillams and Lees, 2007 (12)

United Kingdom

2002–2006

37

Hiraki et al, 2006 (13)

Japan

2001–2005

Kashima et al, 2011 (4)

Japan

Lee et al, 2004 (14)

Nour-Eldin et al, 2009 (15)

Tajiri et al, 2008 (11)

Japan

Yamagami et al, 2006 (9); Yoshimatsu et al, 2009 (8)

Japan

Zhu et al, 2009 (3); Yan et al, 2006 (10)

Australia

Year

NR

2003–2007

2000–NR

RF Ablation Generator (Manufacturer)

RF Ablation Electrode Type (Manufacturer)

NOS Score

RITA 1500 (Angiodynamics, Expandable (Angiodynamics) Latham, New York) and and internally cooled Cool-tip system (Radionics/ (Radionics/Valleylab) Valleylab, Mansfield, Massachusetts)

9

Series CC-1 (Radionics/ Valleylab) and RF 2000 (Boston Scientific, Natick, Massachusetts)

Internally cooled (Radionics/ Valleylab) and expandable (Boston Scientific)

5

Internally cooled (Cool-tip; Radionics/Valleylab)

9

Internally cooled (Radionics/ Valleylab)

7

Internally cooled (CELON PROSURGE; Celon AG Medical Instruments)

7

Expandable (LeVeen; Boston Scientific)

9

RITA 1500 (Angiodynamics)

Expandable (Starburst XL; Angiodynamics)

5

RF 2000, RF 3000 (Boston Scientific) or Series CC-1 (Radionics/Valleylab)

Internally cooled (Cool-tip; Radionics/Valleylab) or expandable (LeVeen; Boston Scientific)

5

Cool-tip RF ablation system (Radionics/Valleylab)

Internally cooled (Radionics/ Valleylab)

9

RITA 1500 (Angiodynamics)

RITA StarBurst XL (Angiodynamics)

9

JVIR

MC ¼ metastatic lung cancer; NOS ¼ Newcastle-Ottawa scale; NR ¼ not reported; PC ¼ primary lung cancer; RF ¼ radiofrequency.



MC: 94

Kennedy et al

Male (%)

Study

Postpulmonary Ablation Pneumothorax

No. RF Ablation Sessions



No. Patients

Volume 25



Number 11



November



2014

1675

Table 2 . Pneumothorax Definition and Diagnoses in Individual Studies Sessions Resulting in Pneumothorax Study Gillams and Lees, 2007 (12)

Definition Quantitatively using proportion of

Pneumothorax Diagnosis CXR immediately after

Sessions Resulting in Pneumothorax

Pneumothorax Requiring Chest Tube or Aspiration

(%)

(%)

38

15

52

21

RF ablation

hemithorax affected by pneumothorax on either CT or CXR Hiraki et al, 2006 (13)

Kashima et al, 2011 (4)

Quantitatively using proportion of hemithorax affected

up CXR (time not

by pneumothorax on either CT or CXR

specified)

Graded clinical severity

Lee et al, 2004 (14)

CT immediately after RF ablation; follow-

NR

CXR 3–5 h after RF

46.1

NR

30

NR

ablation, next day, 3–5 d later CT immediately after RF ablation, CT next day

Nour-Eldin et al, 2009 (15) Okuma et al, 2008 (16) Steinke et al, 2004 (17)

Lung surface retraction distance on CT scan

CT immediately after

11.3

42.9

6

RF ablation

NR

CT immediately after

31.8

NR

RF ablation CXR within 24 h after

43

60

RF ablation Tajiri et al, 2008 (11)

NR

CXR immediately after RF ablation and next

56

7.1

morning Yamagami et al, 2006 (9); Yoshimatsu et al, 2009 (8)

Quantitatively using proportion of

CT immediately after RF ablation; CXR 3 h

hemithorax affected

after RF ablation if

by pneumothorax on either CT or CXR

symptomatic, CXR 24 h after RF

42.3

36.5

32

48.8

ablation in all Zhu et al, 2009 (3); Yan et al, 2006 (10)

NR

CXR 1 h after RF ablation, 24 h after, CT at 1 wk

CT ¼ computed tomography; CXR ¼ chest radiograph; NR ¼ not reported; RF ¼ radiofrequency.

selection methodology. Two studies did not describe eligibility criteria for patient selection for RF ablation and were rated in selection as two stars out of four; the remaining studies included were rated four stars out of four (11,13). Five studies used multivariate regression analysis to adjust for covariates and control for confounding; these studies were given two stars out of two for comparability (3,4,8,12,16). All studies had sufficient follow-up time of at least 24 hours for pneumothorax occurrence and were rated three stars (Table 1). Patient-Related Factors. Six studies assessed the effect of age on pneumothorax (3,4,8,13,15,16). Only two studies found older age to be more significant in

sessions resulting in pneumothorax in univariate analyses (8,15). Only three studies reported the mean (SD) age between sessions with and without pneumothorax (8,13,16). The pooled MD of these three studies showed age to be significantly greater for sessions resulting in pneumothorax with a MD of 2.09 years (95% CI [0.11–4.06]; P = .04; I2 = 0%) (Fig 2). Six studies examined gender as a potential risk factor (3,4,9,13,15,16). In one study, male patients were found to have approximately 1.84 times higher odds of developing pneumothorax than female patients (unadjusted OR, 1.84, reference: female; P o .03) (13). In multivariable logistic regression analyses, four of the studies did not identify gender to be associated with pneumothorax

1676



Postpulmonary Ablation Pneumothorax

Kennedy et al



JVIR

Figure 2. Meta-analysis of age (y) as a risk factor for pneumothorax. (Available in color online at www.jvir.org.)

Figure 3. Meta-analysis of gender as a risk factor for pneumothorax. (Available in color online at www.jvir.org.)

Figure 4. Meta-analysis of emphysema as a risk factor for pneumothorax. (Available in color online at www.jvir.org.)

occurrence (3,4,8,16). However, meta-analysis of four of the six studies that provided data for unadjusted OR calculations showed that the occurrence of pneumothorax is significantly greater in male patients (OR, 2.20; 95% CI [1.49–3.27]; P o .0001; I2 = 0%) (Fig 3) (9,10,13,16). Eight studies assessed pulmonary emphysema as a potential risk factor for pneumothorax (3,4,8,9,12–16). In univariate analyses, the presence of emphysema was found to be a risk factor for pneumothorax in three of the studies (9,15,16). The pooled results for four of the studies, which used number of RF ablation sessions as the sampling unit and provided sufficient data for unadjusted OR calculations, showed that emphysema did not have a statistically significant effect on pneumothorax (OR, 2.29; 95% CI [0.93–5.66]; P = .07; I2 = 66%) (Fig 4) (8,12,13,16). For two studies, which used number of patients as the sampling unit, emphysema also was found not to have a statistically significant effect (OR, 2.49; 95% CI [0.86–7.17]; P = .09; I2 = 14%) (14,15).

Three individual studies showed through multivariate analysis that emphysema was a risk factor for pneumothorax (4,8,16). These studies were not included in the meta-analysis because they did not provide sufficient data to be pooled and looked at different severities of pneumothorax. Six studies looked at whether previous lung surgery was associated with pneumothorax occurrence (3,4,8,12,13,16). Based on four studies that provided sufficient data for unadjusted OR calculations, no history of previous lung surgery also was shown to be significantly associated with pneumothorax (OR, 0.29; 95% CI [0.19–0.44]; P o .00001; I2 = 0%) (Fig 5) (8,12,13,16). Tumor-Related Factors. Seven studies compared tumor sizes between cases with and without pneumothorax (3,4,8,12,13,15,16). Only one study showed that tumor diameter 4 1.5 cm was significantly associated with pneumothorax (15). Conversely, the remaining individual

Volume 25



Number 11



November



2014

1677

Figure 5. Meta-analysis of history of lung surgery as a risk factor for pneumothorax. (Available in color online at www.jvir.org.)

Figure 6. Meta-analysis of tumor size as a risk factor for pneumothorax. (Available in color online at www.jvir.org.)

Figure 7. Meta-analysis of lobe location as a risk factor for pneumothorax. (Available in color online at www.jvir.org.)

studies found no association between tumor size and pneumothorax in both univariate and multivariate analyses. Five of the seven studies provided sufficient data for MD calculations (8,12,13,15,16). Based on these studies, the mean tumor size was nonsignificantly smaller in sessions that resulted in pneumothorax (MD, 0.24; 95% CI [0.60–0.12]; P = .2; I2 = 52%) (Fig 6). Four studies assessed tumor lobe location as a risk factor for pneumothorax (12,13,15,16). Although two studies, which conducted univariate analysis, showed tumor lobe location was associated with pneumothorax (13,15), pooled data showed no significant association (OR, 0.55; 95% CI [0.25–1.18]; P = .13; I2 = 57%) (Fig 7). Three studies examined tumor depth, defined as the distance from the tumor to the pleura, as a potential risk factor for pneumothorax (8,13,15). Two of these studies individually found significantly higher rates of

pneumothorax with increased mean tumor depth (8,15). However, this trend did not hold on meta-analysis with no significant association between mean tumor depth and pneumothorax risk (MD, 0.97 cm; 95% CI [0.01– 1.93]; P = .05; I2 = 90%) (Fig 8). Six studies assessed whether the number of tumors treated was associated with pneumothorax (3,4,8,12, 13,17). In univariate analyses, two studies found that the mean number of tumors treated was significantly greater in sessions complicated by pneumothorax (12,13). In a multivariable analysis, only Zhu et al (3) showed the number of tumors treated was an independent risk factor for increased occurrence of pneumothorax (adjusted OR, 31.61; 95% CI [6.30–158.62]; P o .001). Likewise, in a meta-analysis of three studies that provided sufficient data, a greater mean number of tumors treated was significantly associated with

1678



Postpulmonary Ablation Pneumothorax

Kennedy et al



JVIR

Figure 8. Meta-analysis of tumor depth (cm) as a risk factor for pneumothorax. (Available in color online at www.jvir.org.)

Figure 9. Meta-analysis of number of tumors treated as a risk factor for pneumothorax. (Available in color online at www.jvir.org.)

Figure 10. Meta-analysis of length of aerated lung traversed (cm) as a risk factor for pneumothorax. (Available in color online at www. jvir.org.)

pneumothorax occurrence (MD, 0.50; 95% CI [0.27– 0.73]; P o .0001; I2 ¼ 0%) (Fig 9) (12,13,17).

cooled). However, none of these factors were significant (Fig 11).

Treatment-Related Factors. Four studies conducted univariate analysis and found the length of aerated lung traversed by the electrode to be significantly associated with pneumothorax (3,12,13,15). Nour-Eldin et al (15) and Zhu et al (3) showed that pneumothorax was significantly associated with lengths 4 2.5 cm (15) and 4 3 cm (3). Zhu et al (3) and Gillams and Lees (12) identified length of aerated lung traversed to be an independent risk factor for pneumothorax on multivariate logistic regression analyses. Meta-analysis of MDs from three studies showed that the mean length of aerated lung traversed by the electrode was 1.7 cm (95% CI [0.37–3.10]) greater in cases with pneumothorax (P ¼ .003; I2 ¼ 83%) (Figure 10) (12,13,15). Metaanalyses of unadjusted ORs or MDs were also done for the following treatment-related factors: patient position during treatment (prone vs supine), ablation time, maximum power output, and electrode type (expandable vs

DISCUSSION Among the included studies, the prevalence of PPAP was 37% (95% CI [29%–46%]; I2 ¼ 96.6%). Although this result is comparable to the current literature, it could be underestimated or overestimated because the pooled value is based only on publications investigating PPAP (3). In a subgroup analysis, 29% (95% CI [19%–41%]; I2 ¼ 76.6%) had pneumothoraces that required chest tube drainage or aspiration or both. The high heterogeneity among included studies is likely due to differences in pneumothorax definition and severity grading among studies (Table 2). In addition, perhaps differences in treatment thresholds and operator experience across institutions led to the marked variation in requirement for pneumothorax aspiration or chest tube drainage (18). A greater number of tumors ablated corresponds to increased number of pleural punctures, which likely

Volume 25



Number 11



November



2014

1679

Figure 11. Meta-analysis of patient position (prone vs supine) as a risk factor for pneumothorax. (Available in color online at www.jvir. org.)

explains the increased risk of pneumothorax (13). Because multiple punctures are a risk factor for pneumothorax, minimizing the number of punctures (ie, ablation of multiple lesions through one percutaneous tract when possible) may help reduce the incidence of pneumothorax after RF ablation. Alternatively, staged, multisession treatment rather than a single-session treatment could be considered in patients with more than one tumor in the same lung and may minimize risk of pneumothorax. A history of lung surgery was found to decrease the risk of pneumothorax significantly (Fig 5). A reduced risk of pneumothorax after lung surgery has also been noted in similar studies looking at percutaneous lung biopsy complications (19). A likely explanation is that adhesions after lung surgery serve as a protective mechanism against pneumothorax. Adhesions between the visceral and parietal pleura reduce the ability of air to accumulate as a pneumothorax (13). Male gender was found to be significant risk factor for pneumothorax on meta-analysis. Previous evidence from the literature on percutaneous lung biopsy suggests greater lung vital capacity, as seen in men, is an independent risk factor for pneumothorax (20). Increased vital capacity translates to greater lung movement during the RF ablation procedure, potentially explaining the increased risk of pneumothorax. Tumor size, location, and depth were found not to be significantly associated with pneumothorax (Figs 6–8). There was high heterogeneity for these three factors. The lack of significance for some or all of these factors may be due to sample size error. Meta-analyses of CT-guided lung biopsies have identified these factors to be significant risk factors for pneumothorax (18). Although emphysema has been identified as a risk factor in individual studies through both univariate analysis and multivariate analysis, it was not a significant risk factor for pneumothorax on meta-analysis (Fig 4). Some studies, such as Kashima et al (4), noted severe emphysema as a risk factor for PPAP and pneumothorax after lung biopsy (4,13). It is possible that heterogeneity of definitions used in documenting the presence and severity of PPAP is skewing the results of this systematic review. Further studies evaluating the

severity of emphysema and the presence of local emphysematous change around the tumor could help further understanding of emphysema as a risk factor for pneumothorax after RF ablation. The only treatment-related risk factor identified as significant was the length of aerated lung traversed (Fig 10). Increased length of lung traversed is technically more challenging. For these procedures, greater requirements for multiple electrodes and traversal of major pulmonary fissures may increase the risk of pneumothorax (13). In regard to individual studies, patient-related, tumorrelated, and treatment-related risk factors for pneumothorax identified to be significant include treatment of multiple lobes, treatment of tumors bilaterally, number of electrode points, traversal of major pulmonary fissure, and operator experience. In particular, lack of operator experience has been identified in similar procedures, such as ultrasound-guided thoracocentesis and CT-guided lung biopsy, to be a significant risk factor for the development of pneumothorax (18). Additionally, a learning curve has been identified in pulmonary RF ablation. Interventionalists with less experience have significantly higher risks of complications, including pneumothorax (10). This fact highlights the importance of having highly trained interventionalists to perform pulmonary RF ablation. Multiple studies compared electrode types, including single tip cooled, clustered, and multitined expandable electrodes, and all studies demonstrated no significant difference in PPAP rate based on type of electrode used (12–14). These studies used different brands, lengths, and gauges of electrodes and could not be compared. Individual study data suggest that electrode type is not a risk factor for pneumothorax, but larger studies comparing electrode types are required. Challenges include operator preference and experience as well as procedure requirements including tumor size and location. Based on our review, limitations in the current literature shed light on the importance of further research to delineate pneumothorax risk factors better. Only a few prospective studies have been carried out to date assessing RF ablation for pulmonary malignancies (21). To develop a better understanding of risk factors for PPAP

1680



Postpulmonary Ablation Pneumothorax

after RF ablation, large prospective studies reporting based on standardized research reporting guidelines are needed. A critical appraisal of existing literature on pneumothorax induced by image-guided lung biopsy could be helpful in developing strategies to address PPAP (22,23). Moving forward, it is critical that studies assessing pulmonary RF ablation standardize the reporting of their results in accordance with the Society of Interventional Radiology (SIR) Research Reporting Standards for Percutaneous Thermal Ablation of Lung Neoplasms (24). This standardization would allow for clear reporting of grades or severity of PPAP among all studies (ie, major vs minor). In the future, these data could allow us to appreciate better the mechanism (direct injury from probe insertion vs secondary injury from ablative therapy or perhaps other extrinsic factors, such as atmospheric pressure) to classify (based on timing and severity) and manage pneumothorax after lung RF ablation (25). Patients at high risk for PPAP likely require closer monitoring and more frequent follow-up imaging. The utility of various techniques that are reported to be effective in preventing pneumothorax after lung biopsy, such as breath hold after forced expiration during needle removal, blood patch, manual aspiration of pneumothorax during guided needle removal, and perhaps periprocedural pleurodesis, may be considered for pneumothorax after pulmonary RF ablation, especially in patients with multiple risk factors for the development of severe pneumothorax (26–28). It is hoped that further research will allow the development of PPAP risk stratification schemas to optimize patient management. In addition, PPAP (high frequency and additional cost) is becoming an important consideration in allocating care path in patients with lung cancer (4,29,30) because the efficacy of pulmonary ablation in treating many lung cancers is close to other therapies, such as stereotactic body radiation therapy and sublobar lung resection (31,32). Also relevant to this work, PPAP has been tied more recently to treatment failure, a possibility that calls for future investigations (33). A limitation to this systematic review is the potential of confounding bias because some of the results are based on ORs that were unadjusted for covariates. Only five studies used multivariate logistic regression to identify potential risk factors, and there were insufficient data from these studies to be pooled in a meta-analysis. Another limitation is the substantial heterogeneity, which is inherent to the retrospective nature of the included studies. A limited number of studies assessing certain variables, small sample sizes, significant differences in tumor size and depth between studies, and lack of standardized definitions of pneumothorax severity (ie, requirements for chest tube) within studies likely contributed to the heterogeneity.

Kennedy et al



JVIR

In conclusion, this systematic review and metaanalysis provides a thorough summary of the current literature assessing risk factors for PPAP following RF ablation of lung cancer. The results suggest that risk factors for PPAP include increased age, male gender, no history of lung surgery, number of tumors ablated, and increased length of the aerated lung traversed by the electrode. The presence of emphysema, tumor size, tumor lobe location, tumor depth, maximum RF ablation power output, ablation time, patient positioning during RF ablation, and electrode type do not appear to be risk factors for pneumothorax. However, the results of this meta-analysis must be interpreted with caution in light of the high heterogeneity and limitations present in the available literature.

ACKNOWLEDGMENT We thank Andrea McLellan at the Health Sciences Library of McMaster University for her help in developing search criteria.

REFERENCES 1. Dupuy DE, Zagoria RJ, Akerley W, Mayo-Smith WW, Kavangh PV, Safran H. Percutaneous radiofrequency ablation of malignancies in the lung. AJR Am J Roentgenol 2000; 174:57–59. 2. Zhu JC, Yan TD, Morris DL. A systematic review of radiofrequency ablation for lung tumors. Ann Surg Oncol 2008; 15:1765–1774. 3. Zhu JC, Yan TD, Glenn D, Morris DL. Radiofrequency ablation of lung tumors: feasibility and safety. Ann Thorac Surg 2009; 87:1023–1028. 4. Kashima M, Yamakado K, Takaki H, et al. Complications after 1000 lung radiofrequency ablation sessions in 420 patients: a single center’s experiences. AJR Am J Roentgenol 2011; 197:W576–W580. 5. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med 2009; 151:264–269. 6. Wells GA, Shea B, O’Connell D, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of non-randomised studies in meta-analyses. Available at: http://www.ohri.ca/programs/clinical_epidemiology/nosgen. pdf. Accessed May 5, 2014. 7. Geertzen J. Inter-rater agreement with multiple raters and variables. 2012. Available at: https://mlnl.net/jg/software/ira/. Accessed May 4, 2014. 8. Yoshimatsu R, Yamagami T, Terayama K, Matsumoto T, Miura H, Nishimura T. Delayed and recurrent pneumothorax after radiofrequency ablation of lung tumors. Chest 2009; 135:1002–1009. 9. Yamagami T, Kato T, Hirota T, Yoshimatsu R, Matsumoto T, Nishimura T. Pneumothorax as a complication of percutaneous radiofrequency ablation for lung neoplasms. J Vasc Interv Radiol 2006; 17:1625–1629. 10. Yan TD, King J, Sjarif A, Glenn D, Steinke K, Morris DL. Learning curve for percutaneous radiofrequency ablation of pulmonary metastases from colorectal carcinoma: a prospective study of 70 consecutive cases. Ann Surg Oncol 2006; 13:1588–1595. 11. Tajiri N, Hiraki T, Mimura H, et al. Measurement of pleural temperature during radiofrequency ablation of lung tumors to investigate its relationship to occurrence of pneumothorax or pleural effusion. Cardiovasc Intervent Radiol 2008; 31:581–586. 12. Gillams AR, Lees WR. Analysis of the factors associated with radiofrequency ablation-induced pneumothorax. Clin Radiol 2007; 62:639–644. 13. Hiraki T, Tajiri N, Mimura H, et al. Pneumothorax, pleural effusion, and chest tube placement after radiofrequency ablation of lung tumors: incidence and risk factors. Radiology 2006; 241:275–283. 14. Lee JM, Jin GY, Goldberg SN, et al. Percutaneous radiofrequency ablation for inoperable non-small cell lung cancer and metastases: preliminary report. Radiology 2004; 230:125–134.

Volume 25



Number 11



November



2014

15. Nour-Eldin NE, Naguib NN, Saeed AS, et al. Risk factors involved in the development of pneumothorax during radiofrequency ablation of lung neoplasms. AJR Am J Roentgenol 2009; 193:W43–W48. 16. Okuma T, Matsuoka T, Yamamoto A, et al. Frequency and risk factors of various complications after computed tomography-guided radiofrequency ablation of lung tumors. Cardiovasc Intervent Radiol 2008; 31:122–130. 17. Steinke K, Glenn D, King J, et al. Percutaneous imaging-guided radiofrequency ablation in patients with colorectal pulmonary metastases: 1year follow-up. Ann Surg Oncol 2004; 11:207–212. 18. Gordon CE, Feller-Kopman D, Balk EM, Smetana GW. Pneumothorax following thoracentesis: a systematic review and meta-analysis. Arch Intern Med 2010; 170:332–339. 19. Hiraki T, Mimura H, Gobara H, et al. Incidence of and risk factors for pneumothorax and chest tube placement after CT fluoroscopy-guided percutaneous lung biopsy: retrospective analysis of the procedures conducted over a 9-year period. AJR Am J Roentgenol 2010; 194: 809–814. 20. Saji H, Nakamura H, Tsuchida T, et al. The incidence and the risk of pneumothorax and chest tube placement after percutaneous CT-guided lung biopsy: the angle of the needle trajectory is a novel predictor. Chest 2002; 121:1521–1526. 21. Lencioni R, Crocetti L, Cioni R, et al. Response to radiofrequency ablation of pulmonary tumors: a prospective, intention-to-treat, multicentre clinical trial (the RAPTURE study). Lancet Oncol 2008; 9:621–628. 22. Lim CS, Tan LE, Wang JY, et al. Risk factors of pneumothorax after computed tomography-guided coaxial cutting needle lung biopsy through aerated versus nonaerated lung. J Vasc Interv Radiol 2014; 25: 1209–1217. 23. Boskovic T, Stanic J, Pena-Karan S, et al. Pneumothorax after transthoracic needle biopsy of lung lesions under CT guidance. J Thorac Dis 2014; 6(Suppl 1):S99–S107. 24. Rose SC, Dupuy DE, Gervais DA, et al. Technology Assessment Committee of the Society of Interventional Radiology. Research reporting

1681

25.

26.

27.

28.

29.

30.

31.

32.

33.

standards for percutaneous thermal ablation of lung neoplasms. J Vasc Interv Radiol 2009; 20:S474–S485. Haga T, Kurihara M, Kataoka H, Ebana H. Influence of weather conditions on the onset of primary spontaneous pneumothorax: positive association with decreased atmospheric pressure. Ann Thorac Cardiovasc Surg 2013; 19:212–215. Min L, Xu X, Song Y, et al. Breath-hold after forced expiration before removal of the biopsy needle decreased the rate of pneumothorax in CTguided transthoracic lung biopsy. Eur J Radiol 2013; 82:187–190. Malone LJ, Stanfill RM, Wang H, Fahey KM, Bertino RE. Effect of intraparenchymal blood patch on rates of pneumothorax and pneumothorax requiring chest tube placement after percutaneous lung biopsy. AJR Am J Roentgenol 2013; 200:1238–1243. Yamagami T, Terayama K, Yoshimatsu R, Matsumoto T, Miura H, Nishimura T. Role of manual aspiration in treating pneumothorax after computed tomography-guided lung biopsy. Acta Radiol 2009; 50: 1126–1133. Zheng A, Wang X, Yang X, et al. Major complications after lung microwave ablation: a single-center experience on 204 sessions. Ann Thorac Surg 2014; 98:243–248. Gurley MB, Richli WR, Waugh KA. Outpatient management of pneumothorax after fine-needle aspiration: economic advantages for the hospital and patient. Radiology 1998; 209:717–722. Renaud S, Falcoz PE, Olland A, Massard G. Is radiofrequency ablation or stereotactic ablative radiotherapy the best treatment for radically treatable primary lung cancer unfit for surgery? Interact Cardiovasc Thorac Surg 2013; 16:68–73. Bott MJ, Crabtree T. Treatment of stage I lung cancer in high-risk and inoperable patients: SBRT vs. RFA vs. sublobar resection. Ann Cardiothorac Surg 2014; 3:167–169. Legras A, Mordant P, Cazes A, Riquet M. Radiofrequency of lung metastases: should initial pneumothorax predict treatment failure? Rev Pneumol Clin 2013; 69:336–339.

1681.e1



Postpulmonary Ablation Pneumothorax

Figure E1. Data collection strategy used.

Kennedy et al



JVIR

Risk factors for pneumothorax complicating radiofrequency ablation for lung malignancy: a systematic review and meta-analysis.

To assess the potential risk factors for pneumothorax secondary to pulmonary radiofrequency (RF) ablation...
2MB Sizes 0 Downloads 5 Views