BACKGROUNDRandomized trials comparing neoadjuvant versus adjuvant chemotherapy show that primary chemotherapy allows more frequent breast‐preserving surgery even though no survival advantage has been demonstrated. The aim of the current study was to determine the predicting factors and the survival impact of breast conservation in patients with large breast tumors treated with neoadjuvant chemotherapy.METHODSBetween January 1987 and December 2001, 594 patients with invasive T2–3 breast carcinoma who were ineligible for breast‐conserving surgery (the mean initial tumor diameter was 49 mm) were treated with 3 or 4 courses of an anthracycline‐based primary chemotherapy, surgery, and radiotherapy. Various clinicopathologic factors were tested as possible predicting factors of breast‐preserving surgery. Survival analyses were performed to determine the implications of breast‐conserving surgery on outcome.RESULTSAfter primary chemotherapy, 287 (48%) patients were eligible for breast‐conserving surgery and 307 patients underwent a mastectomy. Initial tumor diameter > 5 cm, low histologic grade, lobular histology, and multicentricity were independent predicting factors of breast conservation ineligibility in the multivariate analysis (logistic regression). In the univariate survival analysis, a failure of breast‐preserving surgery was associated with a poor outcome. Local disease recurrence‐free survival rates were similar in patients treated with lumpectomy and mastectomy.CONCLUSIONSThe results reported in the current study suggested that initial diameter, histologic type and grade, and multicentricity are potential prechemotherapy predicting factors of breast conservation. When carefully selected, patients treated with breast conservation had a risk of local disease recurrence similar to the risk of chest wall disease recurrence after mastectomy. Cancer 2004. © 2004 American Cancer Society.