In the evolution of solid cancer, there are four steps: noninvasive tumor, local invasive cancer without metastasis, local invasive cancer with lymph node metastasis, and eventually systemic disease. For the first three phases, local treatment, including lymph node dissection, may cure the disease. The choice of local treatment depends on the tumor characteristics, but surgery remains important in many of these cancers. Gastric cancer is one of the typical tumors which remain locally invasive, with or without nodal metastasis, but without systemic metastasis for a rather long period. Metastasis to lymph nodes occurs, frequently even in T1 tumors, but seldom to other sites until the late stage. Thus, the target of local control is the regional lymph nodes. The Intergroup study IT-0116 proved the effect of chemoradiotherapy (CRT) for curable gastric cancer, and thus proved the insufficiency of limited surgery (D0/1). The conventional method of local control for gastric cancer is surgery, including regional lymph node dissection (D2). However, the superiority of D2 has not been proven by randomized controlled trials (RCTs). But all RCTs so far have a crucial problem in the quality of treatment given in the D2 arm. D2 is not a dangerous procedure if done by specialists in large-volume hospitals. D0/1 plus CRT is better than D0/1 alone, but it may be worse than D2 alone. The survival benefit of CRT after D2 is an open question. Establishing standard adjuvant chemotherapy after D2 is a more urgent clinical issue, and there is no reason to abandon D2 gastrectomy for curable gastric cancer in Japan.