ObjectiveTo explore the feasibility and the practical value of conserving upper limb lymph nodes in axillary lymph node dissection (ALND) for early breast cancer. MethodsFrom August 2007 to January 2010, 124 patients with early breast cancer were studied and divided into two phases: phase one, from August 2007 to July 2008; phase two, from August 2008 to January 2010. Five milliliter of methylene blue was injected subcutaneously in ipsilateral forearm in all the patients before operation to locate the upper limb lymph nodes. Routine ALND was performed in 22 patients of phase one. The level Ⅱ lymph nodes and the upper limb lymph nodes were separated from the axillary lymph nodes, respectively. The lymph nodes of level Ⅱ were investigated by combining touch cytology with frozen section during operation. The lymph nodes of level Ⅰ, Ⅱ, Ⅲ, and the upper limb lymph nodes were investigated postoperatively by routine pathological examination to evaluate the feasibility of conserving the upper limb lymph nodes. One hundred and two patients in phase two were divided randomly by lottery into control group (30 cases), and conserving group (72 cases) in which upper limb lymph nodes were selectively conserved. The surgical procedure for control group was same as the phase one blue stained upper limb lymph nodes, in the conserving group were conserved selectively when the lymph nodes metastasis of level Ⅱ were not detected by combining touch cytology with frozen section during operation. The data were collected and analysed on pathological results of all patients and arm circumference was compared between control group and conserving group. Results Total 119 of 124 patients (96.0%) were found with blue stained upper limb lymph nodes. The concordance rate was 99.2% (123/124) between the intraoperative combining pathological method and the postoperative routine pathological examination. No upper limb lymph node metastasis was found in the phase one and the control group of phase two with level Ⅱ group negative. The incidence of arm lymphedema in the control group and the conserving group with level Ⅰ and Ⅱ lymph nodes dissection was 18.2% （4/22） and 20% （1/51）, respectively on 6 months after operation. The difference was statistically significant （χ 2=6,34， Plt;0.05）. ConclusionsMethylene blue being injected subcutaneously in ipsilateral upper limb can be used to show validly lymph nodes of upper limb in the axillary region. ALND with selectively conserving upper limb lymph nodes when level Ⅱ lymph nodes negative in metastasis, can prevent postoperative arm lymphedema.