Accurate staging is crucial for determining treatment and prognosis in patients with non–small cell lung cancer (NSCLC). In the absence of distant metastases, lymph node (LN) status is the most important factor in therapeutic decision-making. In disease limited to ipsilateral hilar LNs (N0/N1), surgical resection is the preferred treatment.1,2 However, presence of mediastinal LN involvement (N2/N3) indicates stage III disease,3 where combined chemoradiotherapy is generally the standard of care.2,4
Current mediastinal staging guidelines recommend initial noninvasive imaging with computed tomography (CT) alone or integrated with positron emission tomography (PET-CT) to determine clinical stage.5 In contrast to patients with either a central tumor or N1 LN enlargement, additional preoperative staging is not mandated for small peripheral tumors without radiologic evidence of mediastinal involvement [clinical N0 (cN0)].5,6 However, postsurgical upstaging to pathologic N2 (pN2) disease occurs in 14% to 37% of cN0/N1 NSCLC.7,8 This underscores the potential importance of preoperative pathologic staging of the mediastinum in cN0/N1 cases.
Thoracic surgical guidelines define mediastinal sampling for pathologic staging as selective (suspicious LN only) or systematic (predetermined set of LN stations).9 For cN0/N1 NSCLC, clinical recommendations advocate systematic sampling of at least 3 mediastinal stations including the subcarina.6
Minimally invasive techniques such as endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) have recently become accepted as the tests of first choice to evaluate the mediastinum in lung cancer.5 Previously published meta-analyses report sensitivities of up to 93% in patients with selective sampling of enlarged and/or PET-CT positive LN (cN2/N3).10–12
Recent studies examining the utility of systematic sampling by EBUS-TBNA in cN0/N1 lung cancer have reported inconsistent findings, suggesting that performance of EBUS-TBNA in this population with a lower prevalence of mediastinal metastases may be inferior to that in patients with cN2/N3 NSCLC where prevalence approaches 80%.12 In addition, recent reports have demonstrated the feasibility of endoscopic ultrasound-guided fine needle aspiration (FNA) of mediastinal LN using a gastrointestinal (EUS) or endobronchial (EUS-B) scope to improve diagnostic accuracy of bronchoscopic staging of NSCLC.13
Synthesis of such information may be valuable to assess the utility of EBUS-TBNA for preoperative systematic mediastinal staging in cN0/N1 NSCLC. We hypothesize that the need for surgical staging may be reduced if EBUS-TBNA is performed using the same guidelines for systematic sampling as required by mediastinoscopy. We sought to explore this by systematic review and meta-analysis.
METHODS
Study protocol was registered with PROSPERO (CRD42017057020).
Literature Search
Comprehensive searches of PubMed (MEDLINE), EMBASE, and Cochrane databases were performed from inception to October 2016, to identify studies evaluating EBUS-TBNA for systematic staging of cN0/N1 NSCLC. Search strategies were developed with medical librarian assistance using subject headings and text words based on 3 terms: target condition (lung neoplasm), index test (endobronchial ultrasound), and participants (neoplasm staging) (Appendix 1, Supplemental Digital Content 1, https://links.lww.com/LBR/A170). Further articles were retrieved by manually searching references of included original studies and review papers.
Study Selection
Identified studies were independently assessed by 2 investigators (T.L.L., D.P.S.). Discordance was resolved by consensus. Titles and abstracts were reviewed according to predefined selection criteria. Full-text articles were examined before inclusion in the review.
Eligible studies were identified in accordance with PICOS criteria outlined by the PRISMA guidelines14: Participants (pretreatment patients with cN0/N1 NSCLC for systematic mediastinal staging where we included studies with all or a portion of participants having cN0/N1 disease); interventions (EBUS-TBNA); comparisons (histopathologic analysis and/or clinical follow-up for at least 6 months were used as the reference standard); outcomes (for per-patient analysis, sufficient data to calculate true positive, false positive, false negative, and true negative values); and study design (prospective or retrospective).
We excluded studies not examining EBUS-TBNA; those assessing restaging after induction therapy; review articles; case reports; editorial material; and non-English articles. When multiple articles were published from a single institution, we included papers with no overlapping study periods to prevent double counting of patients. Studies where per-patient outcomes were unclear regarding individual N-status were included in the qualitative review, but excluded from meta-analysis.
Data Extraction
Two investigators independently extracted patient-level data from included studies (T.L.L., P.M.L.). Discordance was resolved by discussion. Additional data were requested from study investigators if needed. Extracted data included: study population (age, sex, prevalence of N2/N3 metastases); study design; radiologic staging (primary tumor size and location, clinical nodal stage); EBUS-TBNA staging (number, size, and station of LNs sampled, EBUS nodal stage); addition of EUS-(B)-FNA; reference standard staging (histopathology alone or combined with clinical follow-up, reference standard nodal stage).
For each study, 2×2 tables were created, categorizing patients as true positive, false positive, false negative, or true negative for detection of occult N2/N3 disease by EBUS-TBNA in patients with a radiologically normal mediastinum. Positive EBUS-TBNA results were considered true positives as chances of contamination are rare.15 From these tables, sensitivity, specificity, positive, and negative predictive values (NPV) were calculated.
All studies were examined using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) tool that evaluates methodological quality of studies validating diagnostic tests.16
Data Analysis
Cohen’s κ coefficient was calculated using GraphPad QuickCalcs (GraphPad Software, La Jolla, CA) to determine interobserver agreement for study selection.
Meta-analysis was performed using Meta-DiSc 1.4 with a random-effects model that allows for study heterogeneity.17 When false negatives or positives were 0, 0.5 was added to all cells of 2×2 tables to allow calculation of sensitivity and specificity.17 Extracted data were pooled with weighted averages applied according to study sample size. As no diagnostic threshold exists for histologic diagnoses, a symmetrical summary receiver operating characteristic curve with area under the curve was constructed to illustrate results quantitatively.18
The pattern of heterogeneity was determined using the χ2 test where P<0.05 was considered apparent heterogeneity. The magnitude of heterogeneity was reflected by the I2 index which describes the percentage of variation across studies that is due to heterogeneity rather than chance19 and where values >50% indicate significant heterogeneity.20
If heterogeneity was demonstrated, prespecified subgroup analyses were performed to explore possible sources of variation. Subgroups included: prevalence of N2/N3 disease; use of PET-CT; use of general anesthetic or Rapid On Site Cytological Evaluation (ROSE) during EBUS-TBNA; number and size of LNs sampled; reference standard used; study design; QUADAS score; and whether all or a portion of patients had cN0 disease.
Linear regression and the Welch t test were used to analyze the difference of sensitivity between subgroups of continuous and categorical variables respectively using GraphPad Prism 7.0a for Mac OSX (GraphPad Software, La Jolla, CA).
Publication bias was examined by funnel plots of sample size against log effect size of individual studies.
P<0.05 was statistically significant.
RESULTS
Literature Search and Study Selection
Bibliographic and manual searching identified 1164 and 9 records, respectively. In total, 28 articles were selected for full-text review. A total of 15 were excluded as they did not meet inclusion criteria. Therefore, 13 formed the basis of our qualitative review21–33 and 9 presented data sufficient for quantitative review (meta-analysis)21–23,26,27,30–33 (Fig. 1). Interobserver agreement was very good with κ=0.808 [95% confidence interval (CI), 0.668-0.947].
;)
FIGURE 1: Study selection. cN0/N1 indicates clinical N0/N1; EBUS, endobronchial ultrasound.
Study Characteristics and Methodological Quality
The mean number of participants with cN0/N1 lung cancer per study was 111 (range, 55 to 220) with a total of 1441 subjects. The prevalence of N2/N3 lung cancer was reported in 10 studies, with a mean of 15.24% (range, 6.19 to 24.00) (Table 1).
TABLE 1: Main Characteristics of 13 Studies Meeting Inclusion Criteria for Qualitative Review
Assessment of study methodological quality revealed generally high scores with a mean of 11 of 14 (range, 9 to 14) (Appendix 2, Supplemental Digital Content 2, https://links.lww.com/LBR/A171). Most studies enrolled patient’s representative of those undergoing EBUS-TBNA in practice, with clear descriptions of selection criteria and diagnostic tests. Because of positive EBUS-TBNA results being regarded as a true positive, many participants did not receive additional surgical staging, and it was variable as to whether the whole cohort received histopathologic verification.
Test Performance
Pooled Data
Nine studies (1146 participants) presented data sufficient for meta-analysis (Table 2).21–23,26,27,30–33
TABLE 2: True Positive, False Positive, False Negative, True Negative, Sensitivity and Specificity Values for Detection of Occult N2/N3 Disease by EBUS-TBNA in Patients Staged cN0/N1 in 9 Studies Meeting Inclusion Criteria for Meta-Analysis
Pooled sensitivity was 49% [95% CI, 41%-57%; I2, 40.6%; χ2, 13.47 (P=0.097)], ranging from 35% to 100% (Fig. 2). When analysis was performed across 8 studies that reported data for EBUS-TBNA accessible LNs, sensitivity increased to 55% [95% CI, 46%-63%; I2, 40.0; χ2, 11.67 (P=0.112)], with the difference being statistically significant (P=0.0005).
FIGURE 2: Forest plot of estimates of sensitivity and specificity for analysis of endobronchial ultrasound-guided transbronchial needle aspiration for systematic mediastinal lymph node staging. CI indicates confidence interval; df, degree of freedom.
Pooled specificity was 100% [95% CI, 99%-100%; I2, 0.0%; χ2, 4.72 (P=0.787)]. As EBUS-TBNA is a test providing tissue diagnosis, similar to a reference standard, specificity for included studies was 99% to 100% as false positive results are highly unlikely.11
These findings corresponded to a pooled positive likelihood ratio of 88.92 (95% CI, 36.66-215.65), pooled negative likelihood ratio of 0.55 (95% CI, 0.48-0.63), and pooled diagnostic odds ratio of 187.51 (95% CI, 72.45-485.33). Mean positive predictive value was 99% and mean NPV was 91% (range, 82% to 100%). The area under the summary receiver operating characteristic curve was 0.9944 (SE, 0.0135), indicating a high level of overall accuracy.
EBUS-TBNA detected radiologically occult N2/N3 disease in 88 of 1146 patients. Therefore, the number needed to test (NNT) with EBUS-TBNA to upstage cN0/N1 patients was 14 (95% CI, 10.8-16.3).
Heterogeneity and Subgroup Analysis
Moderate heterogeneity between the sensitivities of individual studies was observed [I2, 40.6%; χ2, 13.47 (P=0.097)]. To explore sources of heterogeneity, subgroup analyses were performed (Table 3).
TABLE 3: Results of Subgroup Analysis
For each variable, while one subgroup demonstrated no heterogeneity, the other showed significant heterogeneity. Linear regression revealed a significant relationship between QUADAS score and study sensitivity (P=0.0078) (Fig. 3). Therefore, study methodological quality was identified as a possible source of heterogeneity.
FIGURE 3: Results of linear regression analysis of the relationship between methodological quality and sensitivity of individual studies. QUADAS indicates Quality Assessment of Diagnostic Accuracy Studies.
Publication Bias
Funnel plot inspection demonstrated symmetrical scatter, indicating unlikely publication bias (Appendix 4, Supplemental Digital Content 4, https://links.lww.com/LBR/A173).
Qualitative Review
Primary Tumor Characteristics
Primary tumors were poorly described in the 13 studies included in the qualitative review. Two studies reported tumor size, with a mean of 2.85 cm.25,33 Three studies categorized tumor location, with 112 of 395 patients (28.35%) having central lesions.25,31,33 Although subgroup analysis of central versus peripheral tumors was of interest, insufficient data were presented to permit calculations.
LN Characteristics
In total, 11 studies reported the number of LNs sampled per patient, with a mean of 2.47 (range, 1.32 to 4.16).21–25,27,28,30–33 Of these, 8 presented sufficient data for subgroup analysis (Table 3). Seven studies described LN size, with a mean of 7.74 mm (range, 6.90 to 8.70).21–23,25,30,32,33 Of these, 6 presented sufficient data for subgroup analysis (Table 3).
Complete Endoscopic Staging
Three studies examined addition of EUS-(B)-FNA to sample LNs inaccessible to EBUS-TBNA.23,30,32 Of these, 2 presented sufficient data for calculation of summary statistics.23,30
Szlubowski et al23 described EUS-FNA of LNs measuring 3 to 10 mm, followed by EBUS-TBNA of LNs measuring 5 to 10 mm, in 120 patients clinically staged IA to IIB. Occult N2/N3 disease was diagnosed by EUS-FNA alone, EBUS alone, and both techniques in 6 (5.00%), 5 (4.17%), and 8 (6.67%) patients, respectively. This corresponded to EUS-FNA and EBUS-TBNA sensitivities of 50% and 46%, respectively, which significantly increased to 68% with combined procedures.
Oki et al30 performed EBUS-TBNA first, followed by EUS-B-FNA, in 146 patients with cN0/N1 disease. EBUS-TBNA identified N2/N3 metastases in 17 patients (11.64%) and EUS-B-FNA in 15 patients (10.27%), with respective sensitivities of 52% and 45%. Combined EBUS-TBNA and EUS-B-FNA yielded sensitivity of 73%.
On the basis of these studies, the sensitivity of combined endosonography to detect unsuspected mediastinal disease was 71% [95% CI, 57%-82%; I2, 0.0%; χ2, 0.17 (P=0.678)]. Among 266 patients, EBUS-TBNA identified 30 patients (11.28%) and EUS-(B)-FNA identified an additional 13 cases with occult N2/N3 disease (4.89%). This corresponded to NNT of 7 (95% CI, 4.9-8.5). However, the NPV of the combined procedure was 91.77%, comparable to that of EBUS-TBNA alone.
Studies Excluded From Meta-Analysis
In 4 studies,24,25,28,29 although criteria were met for inclusion in a qualitative review, incomplete data were presented to adequately distinguish cN0/N1 cases from the overall cohort that included cN2/N3 patients, and inclusion in meta-analysis was not possible (Appendix 3, Supplemental Digital Content 3, https://links.lww.com/LBR/A172).
Complications
Adverse events occurred in 2 studies with severe cough in 2 (1.37%),30 left mainstem bronchus laceration in 1 (0.06%),29 and massive hemoptysis in 1 (0.06%) patient.29 No deaths were reported.
DISCUSSION
The purpose of this study was to analyze articles addressing the utility of EBUS-TBNA for preoperative systematic mediastinal staging of cN0/N1 NSCLC.
Important findings were that systematic staging by EBUS-TBNA has a sensitivity of 49% (95% CI, 41-57%) and NPV of 91% (range, 82% to 100%) for detection of occult N2/N3 disease in the radiologically normal mediastinum. EBUS-TBNA systematic sampling must be performed in 14 patients to detect 1 additional case with mediastinal metastases. Addition of EUS-(B)-FNA improves sensitivity to 71% (95% CI, 57%-82%) and NNT to 7. Our results confirm that postoperative upstaging can be reduced by preoperative systematic endosonographic staging.
The rate of postoperative upstaging following detection of unrecognized N2/N3 metastases ranges from 9% to 30%.34–36 N2/N3 disease may be radiologically occult due to micrometastases. Consistent with this, our analysis indicated a prevalence of 15% of unsuspected N2/N3 disease with LNs measuring 6.90 to 8.70 mm. This validates the role of preoperative invasive staging of cN0/N1 disease.
Systematic endosonographic staging may contribute where adherence to preoperative and intraoperative surgical staging is incomplete. A survey of 11,000 patients treated surgically for NSCLC revealed that LN sampling was not performed in over 50% who underwent preoperative mediastinoscopy.37 Furthermore, a study of the extent of intraoperative mediastinal LN dissection showed that only 4% of patients received complete dissection in accordance with surgical guidelines.38 Consequently, N2/N3 disease may go undetected even following curative intent resection. This may contribute to high recurrence rates postoperatively, where 20% of stage I and 40% of stage II NSCLC will relapse with metastatic disease that was occult at the time of presentation.39
In recent years, EBUS-TBNA has superseded mediastinoscopy as the recommended first test for preoperative mediastinal staging in patients with high likelihood of mediastinal metastases based on CT or PET.5,6 The potential for minimally invasive staging to reduce postoperative staging was originally demonstrated by Herth et al40 who showed that in patients without enlarged LNs on CT, metastases were diagnosed by EBUS-TBNA in 20%. This corresponds to our findings that the prevalence of mediastinal disease in the CT-negative cohort is 20.79%, compared with 13.86% in the PET-CT-negative group.
For multiple reasons, EBUS-TBNA [±EUS-(B)-FNA] may be better suited to confirm absence of N2/N3 disease, compared with mediastinoscopy. As a minimally invasive procedure with lack of requirement for general anesthetic, it has reduced morbidity and mortality.24 In addition, its reach extends to hilar LNs with ability to diagnose N1 disease.41 This is important due to improved survival of single versus multiple N1 LN involvement42 and the increasing use of stereotactic ablative body radiation (SABR) in stage I (N0) disease.43
SABR is increasingly utilized in high operative risk patients. However, lack of definitive pathologic staging by systematic nodal dissection entails a risk that nodal metastases will go undetected. This is exemplified by a review of 89 patients with stage I surgically treated NSCLC, who would also have been SABR candidates, that revealed occult nodal disease in 8.9%.44 The authors concluded that routine sampling of all LNs accessible to EBUS-TBNA may identify more unsuspected nodal metastases, thereby altering treatment pathways. Indeed, one study has demonstrated EBUS-TBNA detection of mediastinal metastases in 16% of patients under SABR consideration.45
Similarly, there is controversy regarding optimal treatment of N2/IIIA patients due to heterogeneity of nodal involvement where some studies suggest that neoadjuvant chemotherapy before resection has benefit.46 Therefore, careful preoperative mediastinal staging to identify occult N2 disease is essential for therapeutic planning.7,47
In general, mediastinoscopy guidelines have recommended systematic nodal sampling.6 While preoperative systematic staging by EBUS±EUS-(B)-FNA is recommended for specific subgroups,6 it is not universal practice. However, systematic endosonographic staging may have value in specific scenarios. These include invasive staging before SABR45 and systematic mediastinal sampling in stage III NSCLC, where radiation fields based on PET-determined disease extent may result in undertreatment or overtreatment in 33%.48 Therefore, incorporation of recommendations for systematic nodal sampling into endosonographic guidelines may be of value.
Sensitivity in theory should be independent of disease prevalence; however, low prevalence of malignancy has been reported to affect sensitivity of both linear5 and radial probe49 EBUS. We suggest that the reduced sensitivity of linear EBUS in patients staged cN0/N1 likely reflects the different pattern of disease involvement in these studies. Potential explanations include missing micrometastatic deposits, involvement of LN stations not amenable to linear EBUS (station 5,6,8,9),50 and technical challenges of sampling LNs<7 mm.
EUS-(B)-FNA is recognized to potentially address these issues by increasing the number of samples collected, and the number and range of LNs sampled.51 Our findings regarding improved diagnostic performance of combined endoscopic staging in cN0/N1, compared with EBUS-TBNA alone, are consistent with those reported for cN2/N3 NSCLC.13 In fact, the value of combined systematic staging seems greater in this population with low prevalence for mediastinal disease, with NNT for EBUS-TBNA and EUS-(B)-FNA (NNT=7) being half that of systematic EBUS-TBNA alone (NNT=14). Nevertheless, combined endoscopic staging with confirmatory surgical staging in selected high-risk cases continues to have the highest accuracy.52
In summary, minimally invasive preoperative staging has the potential to reduce rates of postoperative upstaging. Equally importantly, diagnostic accuracy is lower than reported in previous meta-analyses in cN2/N3 disease and therefore, minimally invasive staging is not a replacement for definitive surgical staging and mediastinal LN dissection should still be undertaken as per international thoracic surgical guidelines.6
Limitations
The main limitation is that patient population, diagnostic test, and outcome definitions are not identical as evidenced by moderate heterogeneity in the pooled sensitivity statistic. Although study methodological quality was identified as a potential contributor to heterogeneity, it is likely that other unmeasured factors play a role. Of note, QUADAS analysis does not score all potential issues, where in 2 of 9 papers, imaging was not based on PET-CT alone, and 4 of 9 papers were retrospective.
While funnel plot analysis did not identify publication bias, this cannot be excluded with certainty. Funnel plots are widely used in meta-analyses to assess publication bias53; however, they require subjective visual interpretation. True measure of publication bias requires prospective registries of published and unpublished data.54 Such information is not readily available, therefore we cannot be fully confident regarding absence of publication bias.
Lastly, there is risk of verification bias where not all EBUS-TBNA results, particularly positive cases, received surgical confirmation. However, cancer diagnosis by EBUS-TBNA is generally regarded as a true finding not requiring further investigation.
CONCLUSIONS
Preoperative systematic mediastinal staging by EBUS in cN0/N1 NSCLC can reduce postoperative upstaging by identifying radiologically occult mediastinal disease in 1 in 14 patients. Sensitivity in this setting is 49% (95% CI, 41%-57%), which is significantly lower than observed for EBUS-TBNA staging of cN2/N3 NSCLC. However, the low prevalence of disease in this group (15%) means the NPV of 91% is consistent with previous studies. Addition of EUS-(B)-FNA identifies unsuspected disease in an additional 4.89%, halving the NNT. Therefore, a combined approach with systematic sampling by EBUS-TBNA and EUS-(B)-FNA has value before curative intent treatment of early stage NSCLC, but surgical staging is not negated in this population.
REFERENCES
1. Howington JA, Blum MG, Chang AC, et al. Treatment of stage I and II non-small cell lung cancer: diagnosis and management of lung cancer, 3rd ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest. 2013;143(suppl):e278S–e313S.
2. Vansteenkiste J, De Ruysscher D, Eberhardt WE, et al. Early and locally advanced non-small cell lung cancer (NSCLC): ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2013;24(suppl 6):vi89–vi98.
3. Detterbeck FC, Postmus PE, Tanoue LT. The stage classification of lung cancer: diagnosis and management of lung cancer, 3rd ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest. 2013;143(suppl):e191S–e210S.
4. Ramnath N, Dilling TJ, Harris LJ, et al. Treatment of stage III non-small cell lung cancer: diagnosis and management of lung cancer, 3rd ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest. 2013;143(suppl):e314S–e340S.
5. Silvestri GA, Gonzalez AV, Jantz MA, et al. Methods for staging non-small cell lung cancer: diagnosis and management of lung cancer, 3rd ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest. 2013;143(suppl):e211S–e250S.
6. De Leyn P, Dooms C, Kuzdzal J, et al. Revised ESTS guidelines for pre-operative mediastinal lymph node staging for non-small cell lung cancer. Eur J Cardiothorac Surg. 2014;45:787–798.
7. Gómez-Caro A, Garcia S, Reguart N, et al. Incidence of occult mediastinal node involvement in cN0 non-small-cell lung cancer patients after negative uptake of positron emission tomography/computer tomography scan. Eur J Cardiothorac Surg. 2010;37:1168–1174.
8. Watanabe S, Asamura H, Suzuki K, et al. Problems in diagnosis and surgical management of clinical N1 non-small cell lung cancer. Ann Thorac Surg. 2005;79:1682–1685.
9. Lardinois D, De Leyn P, Van Schil P, et al. ESTS guidelines for intraoperative lymph node staging in non-small cell lung cancer. Eur J Cardiothorac Surg. 2006;30:787–792.
10. Dong X, Qiu X, Liu Q, et al.
Endobronchial ultrasound-guided transbronchial needle aspiration in the mediastinal staging of non-small cell lung cancer: a meta-analysis. Ann Thorac Surg. 2013;96:1502–1507.
11. Adams K, Shah PL, Edmonds L, et al. Test performance of endobronchial ultrasound and transbronchial needle aspiration biopsy for mediastinal staging in patients with lung cancer: systematic review and meta-analysis. Thorax. 2009;64:757–762.
12. Gu P, Zhao YZ, Jiang LY, et al.
Endobronchial ultrasound-guided transbronchial needle aspiration for staging of lung cancer: a systematic review and meta-analysis. Eur J Cancer. 2009;45:1389–1396.
13. Korevaar DA, Crombag LM, Cohen JF, et al. Added value of combined endobronchial and oesophageal endosonography for mediastinal nodal staging in lung cancer: a systematic review and meta-analysis. Lancet Respir Med. 2016;4:960–968.
14. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med. 2009;6:1–28.
15. Yasufuku K, Nakajima T, Motoori K, et al. Comparison of endobronchial ultrasound, positron emission tomography, and CT for lymph node staging of lung cancer. Chest. 2006;130:710–718.
16. Whiting P, Rutjes AW, Reitsma JB, et al. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol. 2003;3:25.
17. Zamora J, Abraira V, Muriel A, et al. Meta-DiSc: a software for meta-analysis of test accuracy data. BMC Med Res Methodol. 2006;6:31.
18. Moses LE, Shapiro D, Littenberg B. Combining independent studies of a diagnostic test into a summary ROC curve: data-analytic approaches and some additional considerations. Stat Med. 1993;12:1293–1316.
19. Higgins JP, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses. BMJ. 2003;327:557–560.
20. Dinnes J, Deeks J, Kirby J, et al. A methodological review of how heterogeneity has been examined in systematic reviews of diagnostic test accuracy. Health Technol Assess. 2005;9:1–113.
21. Herth FJF, Eberhardt R, Krasnik M, et al.
Endobronchial ultrasound-guided transbronchial needle aspiration of lymph nodes in the radiologically and positron emission tomography-normal mediastinum in patients with lung cancer. Chest. 2008;133:887–891.
22. Hwangbo B, Kim SK, Lee HS, et al. Application of
endobronchial ultrasound-guided transbronchial needle aspiration following integrated PET/CT in mediastinal staging of potentially operable non-small cell lung cancer. Chest. 2009;135:1280–1287.
23. Szlubowski A, Zielinksi M, Soja J, et al. A combined approach of endobronchial and endoscopic ultrasound-guided needle aspiration in the radiologically normal mediastinum in non-small cell lung cancer staging—a prospective trial. Eur J Cardiothorac Surg. 2010;37:1175–1179.
24. Yasufuku K, Pierre A, Darling G, et al. A prospective controlled trial of
endobronchial ultrasound-guided transbronchial needle aspiration compared with mediastinoscopy for mediastinal lymph node staging of lung cancer. J Thorac Cardiovasc Surg. 2011;142:1393–1400.
25. Cornwell LD, Bakaeen FG, Lan CK, et al. Endobronchial ultrasonography-guided transbronchial needle aspiration biopsy for pre-operative nodal staging of lung cancer in a veteran population. JAMA Surg. 2013;148:1024–1029.
26. Sakairi Y, Hoshino H, Fujiwara T, et al. Validation of EBUS-TBNA-integrated nodal staging in potentially node-positive non-small cell lung cancer. Gen Thorac Cardiovasc Surg. 2013;61:522–527.
27. Yasufuku K, Nakajima T, Waddell T, et al.
Endobronchial ultrasound-guided transbronchial needle aspiration for differentiating N0 versus N1 lung cancer. Ann Thorac Surg. 2013;96:1756–1760.
28. Clementsen PF, Skov BG, Vilmann P, et al. Endobronchial ultrasound-guided biopsy performed under optimal conditions in patients with known or suspected lung cancer may render mediastinoscopy unnecessary. J Bronchol Intervent Pulmonol. 2014;21:21–25.
29. Liberman M, Sampalis J, Duranceau A, et al. Endosonographic mediastinal lymph node staging of lung cancer. Chest. 2014;146:389–397.
30. Oki M, Saka H, Ando M, et al. Endoscopic ultrasound-guided fine needle aspiration and
endobronchial ultrasound-guided transbronchial needle aspiration: are two better than one in mediastinal staging of non-small cell lung cancer? J Thorac Cardiovasc Surg. 2014;148:1169–1177.
31. Shingyogi M, Nakajima T, Yoshino M, et al. Endobronchial ultrasonography for positron emission tomography and computed tomography-negative lymph node staging in non-small cell lung cancer. Ann Thorac Surg. 2014;98:1762–1768.
32. Dooms C, Tournoy KG, Schuurbiers O, et al. Endosonography for mediastinal nodal staging of clinical N1 non-small cell lung cancer. Chest. 2015;147:209–215.
33. Ong P, Grosu H, Eapen GA, et al.
Endobronchial ultrasound-guided transbronchial needle aspiration for systematic nodal staging of lung cancer in patients with N0 disease by computed tomography and integrated positron emission tomography-computed tomography. Ann Am Thorac Soc. 2015;12:415–419.
34. Cerfolio RJ, Bryant AS, Minnich DJ. Complete thoracic mediastinal lymphadenectomy leads to a higher rate of pathologically proven N2 disease in patients with non-small cell lung cancer. Ann Thorac Surg. 2012;94:902–906.
35. Darling GE, Allen MS, Decker PA, et al. Randomized trial of mediastinal lymph node sampling versus complete lymphadenectomy during pulmonary resection in the patient with N0 or N1 (less than hilar) non-small cell carcinoma: results of the American College of Surgery Oncology Group Z0030 trial. J Thorac Cardiovasc Surg. 2011;141:662–670.
36. Licht PB, Jørgensen OD, Ladegaard L, et al. A national study of nodal upstaging after thoracoscopic versus open lobectomy for clinical stage I lung cancer. Ann Thorac Surg. 2013;96:943–949.
37. Little AG, Rusch VW, Bonner JA, et al. Patterns of surgical care in lung cancer patients. Ann Thorac Surg. 2005;80:2051–2056.
38. Verhagen AF, Schoenmakers MC, Barendregt W, et al. Completeness of lung cancer surgery: is mediastinal dissection common practice? Eur J Cardiothorac Surg. 2012;41:834–838.
39. Kelsey CR, Marks LB, Hollis D, et al. Local recurrence after surgery for early stage lung cancer: an 11-year experience with 975 patients. Cancer. 2009;115:5218–5227.
40. Herth FJ, Ernst A, Eberhardt R, et al.
Endobronchial ultrasound-guided transbronchial needle aspiration of lymph nodes in the radiologically normal mediastinum. Eur Respir J. 2006;28:910–914.
41. Ernst A, Eberhardt R, Krasnik M, et al. Efficacy of
endobronchial ultrasound-guided transbronchial needle aspiration of hilar lymph nodes for diagnosing and staging cancer. J Thorac Oncol. 2009;4:947–950.
42. Rusch VW, Crowley J, Giroux DJ, et al. The IASLC Lung Cancer Staging Project: proposals for the revision of the N descriptors in the forthcoming seventh edition of the TNM classification for lung cancer. J Thorac Oncol. 2007;2:603–612.
43. Chang JY, Senan S, Paul MA, et al. Stereotactic ablative radiotherapy versus lobectomy for operable stage I non-small-cell lung cancer: a pooled analysis of two randomised trials. Lancet Oncol. 2015;16:630–637.
44. Robson JM, Vaidyanathan S, Cheyne L, et al. Occult nodal disease in patients with non-small-cell lung cancer who are suitable for stereotactic ablative body radiation. Clin Lung Cancer. 2014;15:466–469.
45. Sarwate D, Sarkar S, Krimsky WS, et al. Optimization of mediastinal staging in potential candidates for stereotactic radiosurgery of the chest. J Thorac Cardiovasc Surg. 2012;144:81–86.
46. Bueno R, Richards WG, Swanson SJ, et al. Nodal stage after induction therapy for stage IIIA lung cancer determines patient survival. Ann Thorac Surg. 2000;70:1826–1831.
47. Uy KL, Darling G, Xu W, et al. Improved results of induction chemoradiation before surgical intervention for selected patients with stage IIIA-N2 non-small cell lung cancer. J Thorac Cardiovasc Surg. 2007;134:188–193.
48. Steinfort DP, Siva S, Leong TL, et al. Systematic endobronchial ultrasound-guided mediastinal staging versus positron emission tomography for comprehensive mediastinal staging in NSCLC before radical radiotherapy of non-small cell lung cancer: a pilot study. Medicine. 2016;95:e2488.
49. Steinfort DP, Khor YH, Manser RL, et al. Radial probe endobronchial ultrasound for the diagnosis of peripheral lung cancer: systematic review and meta-analysis. Eur Respir J. 2011;37:902–910.
50. Taverner J, Cheang MY, Antippa P, et al. Negative EBUS-TBNA predicts very low prevalence of mediastinal disease in staging non-small cell lung cancer. J Bronchology Interv Pulmonol. 2016;23:177–180.
51. Wimaleswaran H, Farmer MW, Irving LB, et al. Pulmonologist-performed transoesophageal sampling for lung cancer staging using an endobronchial ultrasound video-bronchoscope: an Australian experience. Intern Med J. 2017;47:205–210.
52. Annema JT, van Meerbeeck JP, Rintoul RC, et al. Mediastinoscopy vs endosonography for mediastinal nodal staging of lung cancer: a randomized trial. JAMA. 2010;304:2245–2252.
53. Macaskill P, Walter SD, Irwig L. A comparison of methods to detect publication bias in meta-analysis. Stat Med. 2001;20:641–654.
54. Lau J, Ioannidis JPA, Terrin N, et al. The case of the misleading funnel plot. BMJ. 2006;333:597–600.
