Tumor histology predicts mediastinal nodal status and may be used to guide limited lymphadenectomy in patients with clinical stage I non-small cell lung cancer.

Abstract

Methods to minimize surgical trauma from mediastinal lymphadenectomy in patients with early-stage lung cancer are still immature. This study aimed to identify predictors of negative pathologic N2, which may be used to select patients for limited mediastinal lymphadenectomy. Clinicopathologic features of 1430 patients with resected clinical stage I non-small cell lung cancer and complete mediastinal lymphadenectomy were retrospectively analyzed for variables associated with negative N2 nodal metastasis (2008-2015). Overall and recurrence-free survival in patients after complete or limited mediastinal lymphadenectomy were assessed via Kaplan-Meier survival analysis and log-rank testing. The accuracy of frozen section diagnosis for predicting final pathology was retrospectively assessed in 126 randomly selected patients after the surgery. Multivariable analysis revealed that tumor size ≤2 cm, negative pN1, lymphovascular invasion, and lepidic adenocarcinoma were associated with negative mediastinal nodal metastasis. Notably, none of the patients with histology of adenocarcinoma in situ, minimally invasive adenocarcinoma, or lepidic pattern-predominant adenocarcinoma on final pathology had pN2 disease, and the 5-year overall and recurrence free-survival of these patients (99.3% and 99.3%, respectively) was not different from those after limited mediastinal lymphadenectomy (98.7% and 100%, P = .582 and .511, respectively). If these subtypes were classified together as the low-risk group, the concordance rate between frozen section and final pathology diagnosis was 88.9% in the retrospective test cohort. Tumor histology may predict negative mediastinal metastasis in patients with early-stage lung cancer. Future prospective studies are merited to validate the feasibility of using frozen section to select patients for limited mediastinal lymphadenectomy.