The clinicopathological features associated with skip N2 metastases in patients with clinical stage IA non-small-cell lung cancer

OBJECTIVES: Understanding the clinicopathological features of patients with skip N2 metastases (SN2) in clinical early stage lung cancer is important for surgical planning and other treatment considerations; however, the factors associated with SN2 are unclear. This study aimed to investigate the clinicopathological features associated with SN2 in patients with clinical stage IA (cIA) non-small-cell lung cancer (NSCLC). METHODS: We retrospectively studied patients with cIA NSCLC who underwent pulmonary resection (at least lobectomy) and extensive lymphadenectomy (more than ND2a-1) at our institution between January 2004 and December 2010. We investigated the following factors for their association with SN2: age; sex; tumour marker (carcinoembryonic antigen); tumour size on computed tomography (CT), evaluated with a lung-window (LW) and a mediastinal-window (MW) setting; pathology, with or without adenocarcinoma; differentiation; visceral pleural invasion (VPI) and vascular/lymphatic invasion. RESULTS: In total, 422 patients were enrolled, with the following pathological node (pN) statuses: 331 pN0 (78.4%), 39 pN1 (9.3%) and 52 pN2 (12.3%). There were 21 (23.1%) SN2 cases among the patients with nodal metastases. When the cut-off level was defined as a receiver operating characteristic curve with MW (11.5 mm), the sensitivity and specificity of SN2 was 95.2% and 42.9%, respectively. VPI was a statistically independent relevant factor for SN2 in both the patients with cIA and in those with nodal involvement. The VPI classification comprised 59 PL-0 (64.8%), 12 PL-1 (13.2%) and 20 PL-2 (22.0%) with nodal metastases, and there was a significant difference between the three groups (P = 0.03) according to SN2 frequency. There was no difference between VPI 1 and 2 (P = 0.27). CONCLUSIONS: In conclusion, our study suggests that the incidence of SN2 is significantly associated with VPI in patients with cIA NSCLC. Although MW (> 11.5 mm) had a low specificity in the assessment of SN2, it had a high sensitivity, suggesting the possibility of a superior benefit compared with LW. Standard hilar and mediastinal lymph node dissection should be required in patients with suspicious VPI and MW (> 11.5 mm) on preoperative CT.