Tournoy, Kurt G. MD*; Ryck, Frédéric De MD†; Vanwalleghem, Lieve MD‡; Praet, Marleen MD‡; Vermassen, Frank MD†; Maele, Georges Van PhD§; van Meerbeeck, Jan P. MD*
In non-small cell lung cancer, accurate staging of the mediastinal lymph nodes (MLN) allows the selection of those patients who are candidates for radical resection.1,2 Although imaging techniques such as computed tomography (CT) and fluorodeoxyglucose positron emission tomography (FDG-PET) with or without integrated CT scan contribute to mediastinal staging, their diagnostic performance is insufficient to make invasive staging redundant.1,3–5 By consequence, accurate mediastinal staging requires tissue biopsy in case of either enlarged or FDG-avid MLN but also in some specific conditions such as centrally located tumors abutting the mediastinum, broncho-alveolar cell carcinoma and non–FDG-avid primary lung tumors.2 Surgical staging procedures such as cervical mediastinoscopy or video-assisted thoracoscopy are widely used to obtain MLN biopsies for staging. In addition to surgical procedures, minimally invasive out-patient endoscopy techniques have in recent years become available, allowing the investigation of MLN; in particular, linear transoesophageal endoscopic ultrasound with fine-needle aspiration (EUS-FNA) and endobronchial ultrasound with transbronchial-needle aspiration.6,7 These minimally invasive techniques are now also accepted as mediastinal staging tools.2 This recommendation is based on the substantial amount of nonrandomized data suggesting that in selected patients with enlarged MLN, EUS-FNA has a sufficiently high yield (>90%) to confirm MLN involvement.6,8–10 Series addressing the accuracy of EUS-FNA in lung cancer patients presenting with small but otherwise suspicious MLN are scarce.
This report studies the diagnostic performance and the clinical impact of EUS-FNA in a homogenous cohort of lung cancer patients, according to the size of the MLN.
PATIENTS AND METHODS
Patients and Study Design
We did a retrospective study of a prospectively gathered cohort and analyzed with EUS-FNA 100 lung cancer patients, all with suspicion of malignant MLN invasion. This suspicion was based on the fact that enlarged MLN were found on the CT-scan (short axis ≥10 mm) or if the patient had MLN that were FDG-avid on the PET scan regardless the MLN size, or if one or more of the following presentations were present: a primary tumor without FDG uptake, presence of FDG uptake in hilar lymph nodes, or when the primary tumor was located in the inner third of the lung (in the vicinity of the mediastinum—regardless the FDG-PET data). The FDG-PET data available in this study are heterogenous because the patients in this study were recruited from over 20 referring hospitals.
The null-hypothesis was that the test characteristics and the impact of EUS-FNA would not be influenced by the size of the MLN. The study was registered (B67020072199) and approved by the local ethics committee of the Ghent University Hospital (Trial Nr EC UZG 193).
If no enlarged MLN were seen on CT-scan, the patient was categorized in the group with small-sized MLN (group A). If the short axis of at least one MLN on CT was at least 10 mm, the patient was categorized as having enlarged MLN (group B).
EUS-FNA was performed by a trained operator (K.G.T.) in an outpatient setting, using the curved linear scanning ultrasound endoscope (GF-UCT160-OL5, Olympus, Aartselaar, Belgium) that was connected to the ultrasound unit (ALOKA, Mechelen, Belgium).10 There was monitoring of heart rate and oxygen saturation.11 Punctures were performed with a 22-gauge fine needle (EUS-needle, Olympus, Aartselaar, Belgium). Smears of the aspirates obtained by EUS-FNA were processed for rapid on site evaluation (ROSE) to evaluate the cellular contents of the air-dried specimens with a quick staining method (Diff-Quick). If necessary, several lymph nodes were sampled. Specimens were categorized positive (tumor cells) or negative (lymphoid but no tumor cells). Punctures were continued until the cytopathologist (L.V.W.) was able to make a formal conclusion. As such, representative MLN sampling was available for all patients.
All 100 patients would routinely have been scheduled for further invasive diagnostic or staging surgical procedures targeting the MLN. Only those patients in whom EUS-FNA did not confirm malignant MLN invasion were referred to the thoracic surgeon (primarily for cervical mediastinoscopy, with a systematic sampling of 2R/L-4R/L and 7). If a patient was operable and when the tumor was resectable, a thoracotomy with MLN dissection was performed.12 Hence, a formal pathology of the MLN investigated with EUS-FNA was available for all patients.
All categorical variables are reported as proportions. Continuous variables are reported as medians with minimum–maximum range and are compared with the Mann-Whitney U test. Diagnostic performance of EUS-FNA and FDG-PET, calculated at the patient level and defined in terms of sensitivity and specificity, are estimated accordingly in the group of 100 patients in whom a formal pathologic diagnosis was available. The test-characteristics are given with 95% confidence intervals (CIs) and are analyzed with the Fishers' exact test. The clinical impact was calculated as the proportion of patients in whom EUS-FNA demonstrated malignant MLN disease, making a surgical approach of the MLN redundant. The significance level was set at α = 0.05. The analysis was done using SPSS 15.0 (SPSS Inc., Chicago, IL).
Table 1 shows the characteristics of the 100 lung cancer patients, all with suspicion of malignant MLN invasion and in whom a formal pathologic diagnosis of the MLN was available. Twenty-five patients (25%) had only small-sized MLN on CT and were categorized in group A, whereas 75 patients (75%) had at least one enlarged MLN and were categorized in group B. In 41%, a diagnosis of the primary cancer was available before the EUS-FNA procedure. As such, EUS-FNA was performed for staging only in 56% versus 36% of group A and group B patients, respectively (p = 0.10). The overall histology subtype distribution of the primary lung tumor was comparable between group A and group B. The median duration of the EUS-FNA procedure, the dose of sedation, and the number of EUS-FNA passes were comparable between two groups.
FDG-PET data were available for 69 patients (69%) and suspicion of MLN invasion was noted in 56 patients (81%). Of group A, 17 patients (85%) had positive or indeterminate MLN by FDG-PET versus 46 of the patients (94%) of group B (p = 0.31; Table 1). The remaining eight patients of group A with a distinct negative FDG-PET or in whom no FDG-PET was performed were suspected to have MLN invasion mainly because the primary lung tumor was in the vicinity of the mediastinum. All patients of group B underwent EUS-FNA because there was at least one enlarged MLN (regardless of the PET data).
On the basis of the unequivocal presence of FDG uptake in discrete MLN, the sensitivity of FDG-PET to predict malignant MLN invasion was 78% (95% CI, 39–97) for group A versus 95% (95% CI, 83–99) for group B (p = 0.14). The accuracy of FDG-PET was 53 and 87% for the respective groups (p = 0.001).
EUS-FNA data are presented in Figure 1. The overall sensitivity of EUS-FNA was 95% (95% CI, 88–98) with a corresponding negative predictive value (NPV) of 81% (95% CI, 58–94). In the 25 patients of group A, the median size of largest MLN was 7 mm; whereas this was 18 mm in the patients of group B. The prevalence of malignant MLN invasion was 56 versus 92% (p < 0.001) in the respective groups. The sensitivity of EUS-FNA to detect malignancy is 93% (95% CI, 66–99) in group A patients versus 96% (95% CI, 87–99) for group B (p = 0.53). This reflects in NPVs of 92% (95% CI, 61–99) and 67% (95% CI, 29–92), respectively, for both groups (p = 0.27). The clinical impact in terms of avoided surgical procedures is 52% for the patients of group A versus 88% for group B (p < 0.001).
In Table 2, the final pathology data of the MLN are shown. There was no difference in the distribution of histology subtypes of the malignant MLN between both studied groups. All 21 patients in whom EUS-FNA did not demonstrate malignant MLN invasion were investigated by ensuing surgical procedures targeting the MLN. Seventeen of these patients had N0-1 by surgical staging. In four patients, surgical staging showed malignant MLN invasion: one by mediastinoscopy (station 7, enlarged MLN, FDG-avid), one by video-assisted thoracoscopy (station 5, enlarged MLN, FDG-avid), and two by thoracotomy (station 7, enlarged MLN but no FDG uptake and station 4L-5, small MLN but intense FDG uptake). In two of the false negative findings, the surgical specimen showed a discrete tumor infiltration of MLN in the subcapsular sinus (Figure 2), whereas in the other, the malignant infiltration was massive (including the patient from group A).
This study shows that the sensitivity of ROSE-assisted EUS-FNA to detect malignant MLN in patients with lung cancer is high, not only for enlarged MLN but also for small MLN. The impact of EUS-FNA to avoid surgical mediastinal investigation is however greater when enlarged MLN are present.
In the absence of distant metastasis, the accurate mediastinal staging in patients with lung cancer determines both prognosis and treatment options.13 It was previously shown that the likelihood of metastases increases with MLN size measured on CT-scan.3,14 Although at a cut-off of 1 cm transversal diameter, the best accuracy was found, it was recognized that up to 33% of the MLN measuring over 3 cm did not have metastasis, and inversely, that 13% of the subcentrimetric MLN were metastatic.3 In addition, Seely et al.15 showed that in patients with T1 lesions and nodes with a short axis <1cm at CT-scan, 21% had malignant nodal invasion. The accuracy of FDG-PET/CT scan to predict on the presence or absence of malignancy in the MLN is significantly higher when compared with the CT-scan.4 Nevertheless, FDG-PET/CT is more sensitive but less specific with increasing lymph node size.4,16 The sensitivity of FDG-PET/CT to detect invasion in small MLN averages only 82%.4 These data challenge the statement that based on FDG uptake, one can confidently predict on the presence or absence of malignant invasion, even in small MLN.
The ability of EUS-FNA to contribute to the diagnosis and staging of lung cancer is now well established based on several observational series.6,9,10,17–19 Its sensitivity ranges between 77 and 100% and depends on the prevalence of invasion in enlarged MLN.6 EUS-FNA impacts on the diagnostic decision in a substantial number of patients since surgical procedures can be cancelled in 49 to 70%.6,8,10,20,21 These data have recently been strengthened by a randomized controlled trial that directly compared surgical staging with EUS-FNA.22
It remains however unknown whether the accuracy reported for EUS-FNA in lung cancer patients with enlarged MLN can be inferred to patients without enlarged MLN.8 EUS-FNA could be prone to false-negative findings because it is technically easier to puncture enlarged MLN. Two reports have analyzed the yield of EUS-FNA in patients in whom the CT-scan showed no enlarged MLN at all.23,24 The sensitivity of EUS-FNA ranges in these series from 29 to 61%.23,24 The data also indicate that EUS-FNA detects unexpected locally advanced disease in 23% of the patients24 and impacts on the clinical management in 36%.23
The current series is the first to compare directly the yield and impact of EUS-FNA in lung cancer patients with either small or enlarged MLN. The results indicate that the sensitivity to detect malignancy in small MLN is not different as compared with enlarged MLN. The impact in terms of avoided surgical staging procedures is higher in patients with enlarged MLN as compared with the patients in whom only small MLN were found. The clinical impact in the latter group is however higher than what would be expected based on the available data.23,24 Several factors can explain this: first, the patients in this study were selected based on the a priori clinical suspicion for MLN invasion and by consequence on the high prevalence of MLN invasion. Second, the data are biased by that all patients had MLN within reach of EUS-FNA. This series includes no patients with suspicious MLN located solely in the pretracheal (4R) or on the aortic arch (region 6). Third, we10 and others8,25 believe that ROSE during EUS-FNA enhances the yield because the chance of having inconclusive samples is likely to be much higher in small MLN.
An important issue is also the value of a negative finding and the inherent decision whether additional confirmatory procedures are necessary. EUS-FNA punctures can be false negative despite the representativity assured by ROSE. The first reason is of course that the wrong MLN is sampled (i.e., anatomic miss). We document here that discrete tumor infiltration in the subcapsular sinuses is another reason. In those cases, it is obvious that a fine-needle microbiopsy has a great chance to miss those tumor islets. We found that the NPV of EUS-FNA was 81%, with a range for enlarged and small MLN between 67 and 92%. This confirms that, even in the presence of ROSE, a negative result should always be confirmed by a surgical procedure.
Although it was shown that invasive staging techniques have a better test performance as compared with imaging techniques,5 we analyzed in addition the performance of FDG-PET as function of the size of the MLN. The overall sensitivity to detect malignant invasion MLN was 92% in our series, which largely compares with the values available in a meta-analysis on FDG-PET-scan accuracy.4 The sensitivity also tended to be higher for enlarged MLN. The specificity and accuracy was however low, which can be attributed to the small patient numbers, the low threshold for suspicion of MLN invasion and to the exclusion of patients with an indeterminate FDG-PET result. Technical factors also could have contributed since the FDG-PET data were obtained from different hospitals. Nevertheless, the data show that mediastinal staging based solely on FDG-PET data remains unreliable.5
We conclude that the sensitivity to detect malignant MLN invasion is comparably high for both enlarged and small MLN. Hence, the clinician should consider EUS-FNA, even if otherwise suspicious MLN are not enlarged on CT.
1. Silvestri GA, Gould MK, Margolis ML, et al. Noninvasive staging of non-small cell lung cancer: ACCP evidenced-based clinical practice guidelines (2nd Ed.). Chest 2007;132:178S–201S.
2. Detterbeck FC, Jantz MA, Wallace MB, et al. Invasive mediastinal staging of lung cancer. ACCP evidence based clinical practice guidelines (2nd Ed.). Chest 2007;132:202S–220S.
3. McLoud TC, Bourgouin PM, Greenberg RW, et al. Bronchogenic carcinoma: analysis of staging in the mediastinum with CT by correlative lymph node mapping and sampling. Radiology 1992;182:319–323.
4. Gould MK, Kuschner WG, Rydzak CE, et al. Test performance of positron emission tomography and computed tomography for mediastinal staging in patients with non-small-cell lung cancer: a meta-analysis. Ann Intern Med 2003;139:879–892.
5. Tournoy KG, Maddens S, Gosselin R, et al. Integrated FDG-PET/CT does not make invasive staging of the intrathoracic lymph nodes in non-small cell lung cancer redundant: a prospective study. Thorax 2007;62:696–701.
6. Herth FJ, Rabe KF, Gasparini S, et al. Transbronchial and transoesophageal (ultrasound-guided) needle aspirations for the analysis of mediastinal lesions. Eur Respir J 2006;28:1264–1275.
7. Vilmann P, Krasnik M, Larsen SS, et al. Transesophageal endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) and endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) biopsy: a combined approach in the evaluation of mediastinal lesions. Endoscopy 2005;37:833–839.
8. Annema JT, Versteegh MI, Veselic M, et al. Endoscopic ultrasound-guided fine-needle aspiration in the diagnosis and staging of lung cancer and its impact on surgical staging. J Clin Oncol 2005;23:8357–8361.
9. Eloubeidi MA, Cerfolio RJ, Chen VK, et al. Endoscopic ultrasound-guided fine needle aspiration of mediastinal lymph node in patients with suspected lung cancer after positron emission tomography and computed tomography scans. Ann Thorac Surg 2005;79:263–268.
10. Tournoy KG, Praet MM, Van Maele G, et al. Esophageal endoscopic ultrasound with fine-needle aspiration with an on-site cytopathologist: high accuracy for the diagnosis of mediastinal lymphadenopathy. Chest 2005;128:3004–3009.
11. Kramer H, van Putten JW, Douma WR, et al. Technical description of endoscopic ultrasonography with fine-needle aspiration for the staging of lung cancer. Respir Med 2005;99:179–185.
12. Semik M, Netz B, Schmidt C, et al. Surgical exploration of the mediastinum: mediastinoscopy and intraoperative staging. Lung Cancer 2004;45(Suppl 2):S55–S61.
13. 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.
14. Libshitz HI, McKenna RJ Jr. Mediastinal lymph node size in lung cancer. AJR Am J Roentgenol 1984;143:715–718.
15. Seely JM, Mayo JR, Miller RR, et al. T1 Lung cancer: prevalence of mediastinal nodal metastases and diagnostic accuracy of CT. Radiology 1993;186:129–132.
16. de Langen AJ, Raijmakers P, Riphagen I, et al. The size of mediastinal lymph nodes and its relation with metastatic involvement: a meta-analysis. Eur J Cardiothorac Surg 2006;29:26–29.
17. Ramos CD, Erdi YE, Gonen M, et al. FDG-PET Standardized uptake values in normal anatomical structures using iterative reconstruction segmented attenuation correction and filtered back-projection. Eur J Nucl Med 2001;28:155–164.
18. Caddy G, Conron M, Wright G, et al. The accuracy of EUS-FNA in assessing mediastinal lymphadenopathy and staging patients with NSCLC. Eur Respir J 2005;25:410–415.
19. Kramer H, van Putten JW, Post WJ, et al. Oesophageal endoscopic ultrasound with fine needle aspiration improves and simplifies the staging of lung cancer. Thorax 2004;59:596–601.
20. Larsen SS, Krasnik M, Vilmann P, et al. Endoscopic ultrasound guided biopsy of mediastinal lesions has a major impact on patient management. Thorax 2002;57:98–103.
21. Larsen SS, Vilmann P, Krasnik M, et al. Endoscopic ultrasound guided biopsy performed routinely in lung cancer staging spares futile thoracotomies: preliminary results from a randomised clinical trial. Lung Cancer 2005;49:377–385.
22. Tournoy KG, De Rijck F, Vanwalleghem LR, et al. Endoscopic ultrasound reduces surgical mediastinal staging in lung cancer: a randomized trial. Am J Respir Crit Care Med 2008. DOI: 1164/rccm200708-1241oc
23. LeBlanc JK, Devereaux BM, Imperiale TF, et al. Endoscopic ultrasound in non-small cell lung cancer and negative mediastinum on computed tomography. Am J Respir Crit Care Med 2005;171:177–182.
24. Wallace MB, Ravenel J, Block MI, et al. Endoscopic ultrasound in lung cancer patients with a normal mediastinum on computed tomography. Ann Thorac Surg 2004;77:1763–1768.
25. Klapman JB, Logrono R, Dye CE, et al. Clinical impact of on-site cytopathology interpretation on endoscopic ultrasound-guided fine needle aspiration. Am J Gastroenterol 2003;98:1289–1294.