Introduction

Lung cancer is the leading cause of cancer mortality worldwide among men and women. Accurate staging after the diagnosis is essential to select an appropriate therapy. The current guidelines for the treatment of lung cancer are determined by tumor morphology, the degree of mediastinal lymph node involvement and the presence of distant disease.

In this sense, mediastinal lymph node involvement is a significant prognostic factor. The aim of mediastinal nodal staging is to exclude, with the highest certainty and the lowest morbidity possible, patients with mediastinal lymph node disease as they will not benefit from surgery, at least initially. For this purpose, cervical mediastinoscopy still remains the gold standard technique, with a sensitivity of 0.78 (0.38-0.92) and a negative predictive value of 0.91 (0.80-0.97)¹.

However, this does not occur with supraclavicular lymph node resection, which is no longer routinely performed despite metastases have been described in 23.7% of patients with non-small cell lung carcinoma and non-palpable lymph nodes, representing a non-negligible percentage².

Supraclavicular lymph nodes correspond to station 1 of the lymph node map proposed by the International Association for the Study of Lung Cancer (IASLC). When these lymph nodes are involved, lung cancer is considered as stage IIIB or IIIC of the American Joint Committee on Cancer (AJCC) tumor - node - metastasis (TNM) staging system, and nowadays the international guidelines recommend concurrent chemoradiation as a standard treatment³.

The aim of the present study is to describe the characteristics of patients with non-small cell lung carcinoma in whom supraclavicular lymph node resection would detect non-palpable lymph node metastases (N3 supraclavicular disease).

Material and methods

Data from patients who underwent mediastinal surgical staging for non-small cell lung cancer were collected between December 2016 and December 2019. Staging was performed according to the preoperative images obtained by computed tomography (CT) scan or positron emission computed tomography (PET-CT), using cervical mediastinoscopy and resection of the right or left supraclavicular nodes, depending on tumor location, either through the same incision or extending it (Fig. 1). Patients with a definitive diagnosis of non-small cell lung cancer without distant disease or palpable supraclavicular lymph node were included in the study. Patients with a diagnosis of small cell carcinoma or other carcinomas or cases with incomplete data in the medical records were excluded.

The variables considered for the analysis were histology of the main tumor, tumor site and central location, tumor size, number of lymph node stations evaluated by mediastinoscopy, definitive pN, survival of pN2 patients treated with chemoradiation alone and chemotherapy and surgery, and survival of pN3 patients after one year of follow-up. Central tumors were defined as those located in the inner onethird adopted by drawing concentric lines from the mediastinal midline in axial and coronal images¹³.

Results

A total of 60 patients were included; 46 (76.7%) were men. Tumor size ranged from 1.2 cm to 12 cm (mean 4.7 cm). The histologic variants found are shown in Figure 2 and the distribution of main tumor location is shown in Figure 3.

Mean lymph node stations evaluated by mediastinoscopy was 2.37 ± 1.44 (SD). Most patients had 3 stations evaluated (38.3%), and 20% had 4, while 1 station was evaluated in 8.3% and 2 stations in 11.7%. In 13 cases (21.3%), biopsy of bulky right or left paratracheal lymph nodes was performed, as appropriate. In 50 patients (83%) frozen section and definitive biopsy of the stations reached by mediastinoscopy yielded positive results. All the patients underwent supraclavicular lymph node resection, and the result was positive for epithelial carcinoma in 21 cases (35%). The final pN was N0 (13.3%) in 8 patients, N2 in 29 patients (48.3%), N3 supraclavicular disease in 21 patients (35%) and contralateral N3 disease in 2 patients (3.3%). Of the 21 cases with supraclavicular N3 disease, 2 patients were recorded as skip metastases with negative mediastinoscopy and positive supraclavicular lymph node. The remaining 19 cases were associated with mediastinal N2 disease (p = 0.0424): 11 patients had single nodal station and 8 patients had multistation/ bulky disease (Table 1).

Patients with supraclavicular lymph node metastasis had a mean tumor size of 5.37 cm, with no significant association compared with those without supraclavicular lymph node disease (Table 2).

Most of the tumors with N3 supraclavicular disease were in the upper lobes, with a trend towards right upper lobe involvement in 61% (p = 0.46), but without significant association. On the other hand, there was a significant association between central tumor and N3 supraclavicular disease (p = 0.0148); approximately 60% of patients with central tumor had supraclavicular N3 involvement (Fig. 4), and the likelihood of supraclavicular disease was 5 times greater in central tumors (OR 4.67, 95% CI [1.34 -16.31]).

Adenocarcinoma was the histological type most associated with supraclavicular N3 involvement (p = 0.264). Mean annual survival since diagnosis expressed in months was 9.85 ± 2.53 months for N2 patients and 9.54 ± 2.62 for those with N3 (p=0.746). Only one pN2 patient underwent surgery after chemotherapy and negative re-staging, is still alive nearly 3 years after the procedure. None of the patients died, and only one patient developed a supraclavicular seroma, a complication that required aspiration.

Discussion

Mediastinal lymph node involvement is one of the major factors determining the outcome and defining the therapeutic approach in patients with non-small cell lung carcinoma without extrathoracic extension. Involvement of supraclavicular lymph nodes also affects the outcome, although at present they are not routinely investigated when they are not palpable¹. Therefore, it is not clear when to perform surgical evaluation of non-palpable supraclavicular lymph nodes, although their involvement in lung carcinoma has been reported to be between 1% and 51%, with an average incidence of 20.1%².

In 1949, Daniels was the first to describe that non-palpable prescalene lymph nodes can be pathologically involved with carcinoma or other intrathoracic diseases and that biopsy can frequently provide the same diagnostic and prognostic information as do palpable nodes. Later, the first reports demonstrating the presence of metastatic nonpalpable supraclavicular nodes in lung cancer started to be published in the mid-20th century⁴. Harken et al. reported 31 (39.8%) positive supraclavicular biopsies in 78 patients with known or suspected lung carcinoma⁵. In the study by B. N. Josephs, 16 (27%) of 59 nonpalpable lymph nodes in patients with lung cancer showed metastatic disease⁶.

Moving towards the 21st century, in 1996 Lee and Ginsber combined cervical mediastinoscopy with supraclavicular lymph node biopsy through the same incision using the mediastinoscope in patients with lung carcinoma when contralateral N2 or N3 disease was strongly suspected or identified. Of 81 patients evaluated, 58 had mediastinal lymph node involvement, and of these, 19 (32.8%) also had occult supraclavicular lymph node disease. Tumors were all centrally located, of non-squamous origin, and most were right lung tumors⁷. More recently, Ohno et al. detected involvement in only 5% of non-palpable supraclavicular lymph nodes in patients with non-small cell carcinoma without distant disease, suggesting that biopsies at this level should be limited to appropriately selected cases⁸.

As we have already mentioned, cervical mediastinoscopy is still the surgical method most used for nodal staging due to its high sensitivity and specificity when performed by experienced surgeons. The rate of lymph node involvement is high, between 20 and 25%, in patients with no evidence of mediastinal disease on CT images but with central tumors or suspected N1 disease¹⁰. However, resection of the supraclavicular nodes has fallen into disuse. In this study we performed both procedures within the same operative time and found positive N3 supraclavicular disease in 35% (21 of the 60 cases evaluated), which is within the range described in the literature. Of these patients, 90.4% (19 of 21) also had mediastinal lymph node involvement, with a statistically significant association between the presence of N2 metastases and supraclavicular involvement. In these patients, 58% (11 patients) had N2 single-station and 42% (8 patients) had N2 multi-station/bulky disease. This means that mediastinal lymph node involvement is strongly related with and, in general, precedes nonpalpable supraclavicular disease, without significant differences between N2 single-station and N2 multistation/ bulky. When compared with the literature, the publication by Lee and Ginsberg and by Ohno et al. also demonstrate that mediastinal disease often predicts occult supraclavicular involvement. However, the former study demonstrated that contralateral mediastinal lymph node involvement was more strongly associated with N3 supraclavicular disease while the latter found that the incidence of N3 supraclavicular disease was more strongly associated with N2 multistation involvement^7,8. When compared with our results, patients with N2-multistation/bulky disease evaluated by our team showed N3 supraclavicular disease in 47%, which indicates a non-significant trend.

We have also found 2 cases of supraclavicular N3 skip metastasis, which represents 20% of total negative mediastinoscopies. In both patients the tumor was in the right upper lobe. Due to the multiple connections that exist between the lymphatic channels, mediastinal metastases may occur in any mediastinal lymph node independently of the anatomical origin of the tumor, although they are more common in tumors located in the upper lobes¹⁵. There are no strong data to show the incidence and reveal the outcome of this type of skip metastasis (supraclavicular N3 without N2), but we may say that, although the percentage was not low, one limitation is that the study was performed in a single center and the number of patients is small.

Mediastinoscopy and supraclavicular lymph node resection have been performed on the same side as the major lung tumor in all the cases, except for those located in the left lower lobe, as this lobe, more than any other lobe, is the lobe with the highest propensity to metastasize to the contralateral mediastinal nodes⁹. In fact, of the patients with N3 supraclavicular disease, the only patient with a tumor in the left lower lobe had metastases in the right supraclavicular lymph node.

Primary tumor site, central location and size were other characteristics evaluated in non-palpable N3 supraclavicular disease. Although almost all tumors occurred in the right upper lobes and mean tumor size was > 3 cm, we only found a significant association between central location and occult supraclavicular disease. Shatzlein et al. detected N3 supraclavicular disease in 29% of patients with central carcinoma measuring 3 cm or more in diameter, suggesting that these may be the indications¹⁰, but Ohno et al. found that N3 supraclavicular disease may exist even in peripheral tumors8. There is consensus that lung masses should be considered when resection of a supraclavicular lymph node is evaluated, but tumor central location is the main characteristic contributing to supraclavicular lymph node disease.

Based on our findings, we agree with different authors^7,8 that non-squamous tumors, mainly adenocarcinomas, are those more commonly associated with N3 supraclavicular disease. Survival associated with pN2 and pN3 has not shown significant differences. A study by Oh et al.¹¹ indicates that supraclavicular lymph node metastases do not compromise postoperative outcomes compared to other N3 mediastinal metastases in the setting of definitive chemoradiation. In another recent study, Park et al.¹² noted that N3 supraclavicular involvement could negatively affect the oncologic outcome in stage IIIC non-small cell lung carcinoma. Further clinical research is needed to investigate the prognostic impact of supraclavicular node involvement in stage III patients. The limitations of the present study are its retrospective design, the small sample size, and the fact that it reflects the experience of a single center.

As a conclusion, resection of supraclavicular lymph nodes along with cervical mediastinoscopy through the same incision may provide additional findings to those of the standard mediastinal approach and increase the accuracy of surgical lymph node assessment by identifying patients with stage IIIB-C disease. This approach may be considered it in patients with suspected or confirmed N2 lymph node disease and central tumors, before considering a combined therapeutic approach including surgery.