The management of patients with NSCLC without signs of mediastinal affectation on CT is still open for discussion. EUS can detect mediastinal nodules of up to 2–3mm.18 Therefore, EUS is able to detect adenopathies smaller than 1cm considered irrelevant by CT and even obtain a sample from them to either confirm or rule out the affectation of the lymph nodes. Four prospective studies have evaluated EUS-FNA in patients with NSCLC without pathological lymph nodes on CT, showing that EUS-FNA is able to detect advanced disease (mediastinal metastasis, suprarenal metastasis, tumor invasion of the mediastinum) in approximately 25% of the patients whose CT showed only non-advanced disease19–22 (Table 2). Wallace et al.19 published a study in which EUS-FNA identified N2–N3 lymph node metastasis in 14 out of 67 patients (21%). There were also 9 false negatives due to either a sampling error (n=5) or to the presence of metastasis in pre- and paratracheal lymph nodes that were not visualized by EUS (n=4). In another study, LeBlanc et al.20 identified N3 metastasis in 5 out of 67 patients (7%) by means of EUS-FNA. EUS avoided surgery in 9 patients (13%), not only due to the N3 affectation (n=5) but also by detecting celiac metastatic lymphadenopathies (n=2), tumor invasion (n=1) and synchronic esophageal cancer (n=1). In another 9 patients, EUS had an impact in management by detecting suprarenal benign lesions (n=8) and a hepatic lesion (n=1). In total, the findings of EUS changed the patient management in 25% of cases. In the surgical staging of the remaining 62 patients in whom mediastinal disease was not diagnosed, lymph node metastasis was found in 17 more patients.

The prevalence of mediastinal metastasis in these two studies (35% and 36%, respectively) is higher than that of other surgical series, which suggests a bias in favor of the EUS either due to the inclusion of patients with lymphadenopathies that are “borderline” in size or because more samples were taken than normal (normally 4 or 5 aspirations per lesion in the study by Wallace et al.). In the study by Fernández-Esparrach et al.,21 EUS-FNA demonstrated metastasis in N2–3 lymphadenopathies in 5 out of 47 patients (11%). There were 5 false negatives that were due to micrometastasis (n=1), the lack of need to perform FNA due to benign morphological appearance (n=3) or to the presence of metastasis in a lymphadenopathy that was not visible by EUS (n=1). In this study, the EUS-FNA was done when one or more of the classic ultrasound criteria of malignancy by EUS were present (hypoechogenicity, round shape, well-defined edges, size>5mm).23,24 Consequently, three metastatic lymphadenopathies (in two patients) visualized by EUS were not considered tributaries of EUS-FNA.

Therefore, in the mediastinal staging of NSCLC, the use of classic criteria for malignancy should be reconsidered.25 The reasons for the low sensitivity of these criteria have not been researched specifically. However, given that it is common to find inflammatory mediastinal lymphadenopathies in healthy patients, one might think that the infiltration of the lymphadenopathies by metastatic cells would not necessarily change the morphologic characteristics of the preexisting lymphadenopathies. In this way, any mediastinal lymphadenopathy would be a candidate for being aspirated in patients with NSCLC. Finally, it is well known that mediastinal lymphadenopathies can contain foci of micrometastasis in patients with NSCLC. Therefore, a more exhaustive analysis of all the visible mediastinal lymph nodes and the addition of NA guided by EBUS for exploring the regions that are not accessible by EUS would improve the diagnostic performance of EUS-FNA in these patients.22

Mediastinal re-staging in patients with NSCLC after induction of chemotherapy is a controversial point. CT and MRI have limited value due to the intrinsic inability of these techniques to differentiate between tumors and necrosis, and therefore their precision is low.26 On the other hand, PET is not a good predictor for the response to chemotherapy, either by the primary tumor or by the lymphadenopathies.27–29 The results of a recent systematic review advise against the use of PET for mediastinal lymph node re-staging as its effectiveness is not very good (sensitivity 63.8% [95% CI: 53.3%–73.7%] and specificity 85.3% [95% CI: 80.4%–89.4%]). As for the response of the primary tumor, the results are very variable, and the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were 80%–100%, 0%–100%; 42.9%–100% and 66.7%–100%, respectively.29

Recent studies have demonstrated that NSCLC patients with metastases in ipsilateral lymphadenopathies that were understaged with induction of chemotherapy had significantly higher survivals after surgical resection compared with those patients with persistent N2 disease.30–33 The three-year survival was 59% in the patients of the first group and 0% among the patients with persistent N2 disease. Therefore, a precise evaluation of the response to neoadjuvant therapy is necessary to decide the therapeutic strategy depending on prognosis.

Mediastinoscopy is still a valid tool in the re-staging of locally advanced NSCLC after neoadjuvant treatment, with a diagnostic precision of 85%.26,34–37 Mediastinoscopy provides histological evidence of the response to treatment and selects those patients who will benefit from a later thoracotomy. However, the frequency of incomplete procedures is high (40%) due to the technical difficulties secondary to the presence of mediastinal fibrosis.38

Annema et al.39 published the first case study with EUS-FNA for mediastinal re-staging in 19 patients with N2 disease who had been treated with induction of chemotherapy and obtained a sensitivity, NPV and diagnostic precision of 75%, 67%, and 83%, respectively. There were three false negatives (sample errors in two cases and a paraesophageal tumor not detected by EUS in one case). These results are lower than those obtained with EUS-FNA in the patients who have not been treated with chemotherapy, but they are very similar to those published by other authors39–42 (Table 3). In the study by Ginès et al.,40 the sensitivity and NPV values are the lowest published to date and could be explained by the location of some lymphadenopathies in stations that are more difficult to visualize by EUS, as is the case of station 4R.

Other authors have compared CT-PET with EUS-FNA. In the study by Stigt et al.42 in 28 patients, EUS-FNA had an NPV and a diagnostic precision of 92%. The agreement between the findings of the re-staging by EUS-FNA and the metabolic response of the metastatic lymphadenopathies occurred in 17 out of 27 patients. However, and due to the fact that the histologic confirmation is higher and therefore necessary, EUS-FNA is the first-choice procedure for mediastinal re-staging.

The lymphadenopathies of the right paratracheal region cannot be correctly visualized by EUS. This part of the mediastinum could be explored by EBUS.43 A recent study of 124 patients treated with induction of chemotherapy for stage IIIA44 reached a sensitivity of 76% for NA guided by EBUS. However, due to the low NPV (20%), the authors suggest that the negative findings should be confirmed by means of surgical staging prior to thoracotomy.

Some of the reasons of this low NPV of NA guided by EBUS are the following:

• 1.

  After chemotherapy, the lymphadenopathies that initially contain tumors present necrosis and fibrosis and are more difficult to biopsy, while the quantity of cell material for analysis is less.

• 2.

  The malignant cells can be found focally inside the lymphadenopathy in the interior of dense areas of extracellular matrix.

• 3.

  The presence of necrosis within the aspirated sample often makes the pathological evaluation more difficult.

• 4.

  There was no in situ evaluation by the cytologist, which possibly would have been able to increase the sensitivity for the detection of malignancy.

It is important to remember the different territories that are usually accessible by EUS and EBUS. EBUS enables access to the paratracheal (stations 2, 3p, and 4), subcarinal (station 7), hilar (station 10) and interlobar (station 11) regions. By means of EUS, we can visualize the regions of the posterior-inferior mediastinum (stations 4L, 5, 7, 8, 9, and occasionally 2 and 4R).11,12 Given the fact that these techniques are complementary, the combination of both could improve their precision. Furthermore, it has been demonstrated that the sequential performance of EBUS and EUS is well tolerated by the patients when done under conscious sedation.45

In conclusion, EUS-FNA should be considered in the mediastinal staging of patients with NSCLC with and without pathologic adenopathies by CT and in the mediastinal re-staging after neoadjuvant treatment.

The authors declare having no conflict of interests.