c-T2N1M0 or lower stage disease. These data correspond with the frequency of sub-carinal nodes metastasis previously reported in LSD-NSCLC (9.8-13.9%) [6, [12] [13] [14] . We previously analysed the lymph node involvement patterns from each lobe or division including LSD-(in 32 p-N-positive patients) and LLD-NSCLC (in 4 p-N-positive patients) limited to c-T2N1M0 or less extensive disease [7] . In our previous study, we showed that the most frequent metastatic site from LSD-NSCLC was stations #5

more »

... (n = 13), #12u (n = 10) and #11 (n = 9). In addition, there were no sub-carinal or inferior mediastinal node metastases from LSD-NSCLC with c-T2N1M0 or less disease. On the other hand, LSD-NSCLC showed metastasis to #4 or #6 without #11 or #12u metastasis, and therefore we determined the regional lymph nodes of LSD-NSCLC as #4, #5 and #6. In regard to LLD-NSCLC, we had determined the regional lymph node stations as #5, #7 and #11 in the previous study. However, as we described in the previous report, the number of LLD-NSCLC was too small to obtain definitive conclusions. Therefore, in this study, we focused on the LLD-NSCLC and analysed the lymph node metastasis patterns from LLD-NSCLC thoroughly. As a result, we revised the regional lymph node stations as #5, #11 and #12u, using a larger series of LLD-NSCLC patients. If these revised regional node stations had been pathologically examined intraoperatively for the 160 LLD-NSCLC patients with c-T2N1M0 or lower stage disease, selective MLND would have been performed on 125 p-N0 and 5 p-N1 patients diagnosed with no metastasis, and complete MLND would have been carried out on 14 p-N1 and 16 p-N2 patients diagnosed with metastasis. Since those 5 p-N1 patients, initially diagnosed as having no metastasis, developed metastasis only to the lymph node stations within the area of a selective MLND, a selective MLND in association with a lingulectomy would have completely removed the local tumour. Thus, based on our revised node station examination, all patients would have been accurately assigned either a selective or complete MLND, resulting in no false negatives. Limitations of this study include its retrospective nature and a relatively small sample number enrolled from each institution. With regard to prognosis, adjuvant chemotherapy regimens, which might influence the prognosis to some degree, differed between institutions. In conclusion, an intraoperative pathological examination, using our proposed LLD-specific regional lymph node stations for LLD-NSCLC (#5, #11 and #12u), may accurately diagnose the status of node metastasis, and appropriately lead to a selective or complete MLND in patients with c-T2N1M0 or lower stage disease. Conflict of interest: none declared. eComment. Radical vs selective lymphadenectomy for non-small-cell lung cancer: which is the optimal procedure?