The value of radical systematic lymphadenectomy for treatment of early-stage bronchial carcinoma is controversial. We performed a prospective randomized study to address this question. Altogether 115 patients with peripheral non-small-cell lung cancers smaller than 2 cm in diameter were enrolled in this study. They were randomly assigned into a lobectomy with lymph node sampling group (sampling group, n = 56) or a lobectomy with radical systematic lymph node dissection group (dissection group, n = 59). Inclusion criteria were based only on preoperative clinical studies. Four tumors were larger than 2 cm postoperatively. One patient had disseminated disease, and two had intrapulmonary metastases discovered at surgery. Two patients had small-cell carcinoma. There were four with pathologic N1 disease and seven with N2 disease in the dissection group and three with N1 and eight with N2 disease in the sampling group. The numbers of local and distant recurrences were two and six, respectively, in the dissection group and two and five in the sampling group. The overall 5-year survival was 81% in the dissection group and 84% in the sampling group. No significant differences in the recurrence rate or survival was seen between the groups. Our results demonstrate that clinically evaluated peripheral non-small-cell carcinomas smaller than 2 cm in diameter do not require radical systematic mediastinal and hilar lymph node dissection.