中文摘要：

目的比较早期乳腺癌病人接受保留乳房（保乳）治疗与根治性手术疗效。方法采用配对的回顾性队列研究的方法。对1995—2002年间在复旦大学肿瘤医院接受手术的乳腺癌病人，根据发病年龄、腋窝淋巴结状态、性激素受体状态、肿块大小4个变量时保乳病人和改良根治的病人进行匹配，比例为1：2。配对结果为保乳病人127例，改良根治病人254例。保乳组中位随访时间为49个月，改良根治组为58个月。比较两组的局部区域复发率以及总存活率和无病存活率。结果保乳组的局部区域复发率为3．39％，改良根治组为1．84％，两组之间差异无显著性（P=0．55）。保乳组总存活率为96．73％，改良根治组97．57％，差异无显著性（P=0．66）。保乳组的无病存活率为86．04％，改良根治组为91．57％两者差异无显著性（P=0．37）。结论对于合理选择的病人采用肿块广泛切除腋窝中下群淋巴结清扫以及手术后辅助性放射治疗可以很好地控制局部复发以及取得理想的存活率。保乳治疗可以取得与改良根治同样的治疗效果，但可拥有较好的外形。

英文摘要：

Objective To compare the effects of breast conservative therapy （BCT） with modified radical mastectomy （MRM） in women with early stage breast cancer. Methods A matched retrospective cohort study using data on patients derived from a prospectively collected breast cancer database was conducted. The database included patients who received MRM or BCT from 1995 to 2002 in Cancer Hospital of Fudan University. The match was conducted according with four variables：age at diagnosis, axillary lymph node status, sexual hormone receptor status and the dimension of tumor. The match ratio was 1：2. Controls were patients who received MRM （ n = 254）. Cases were patients who received BCT （ n = 127）. Median follow-up for the controls and cases were 58 months and 49 months respectively. The differences of incidence of loco-regional recurrence, disease free survival and overall survival at 5 years were compared. There were no significant differences in incidence of loco-regional recurrence, DFS and OS at 5 years between the two groups of patients. Results The incidence of loco-regional recurrence was 1.84% in MRM group and 3.39% in BCT group （ P = 0.55 ）. The OS in MRM and BCT patient were 97.57% and 96. 73% （ P =0.66）. The DFS in MRM and BCT patients were 91.57% and 86. 04% （ P = 0. 37）. Conclusion For appropriate breast cancer patients, classic lumpectomy plus axillary lymph node dissection and post-operative radiotherapy lead to excellent local control and good survival rate. The BCT can result in the same effects as MRM in breast cancer patients with better cosmetic appearances.