Notes: Abstract: Tumor response to preoperative chemotherapy varies from complete to partial to none. To evaluate pathologic predictors of tumor response to preoperative chemotherapy, we reviewed 287 cases of locally advanced breast carcinoma treated with chemotherapy prior to definitive surgery. The patients ranged in age from 18 to 79 years (mean, 48 years). There were 77 (26.8%) patients with stage II disease, 194 (67.6%) with stage III disease, and 16 (5.6%) with stage IV disease. Following the initial diagnosis of invasive carcinoma (by fine-needle aspirate or cutting needle biopsy), the patients received three to four cycles of both doxorubicin-based and cyclophosphamide-based regimens followed by mastectomy or lumpectomy with axillary dissection. The pathologic parameters that were evaluated included stage, clinical tumor size, and tumor nuclear grade (NG). The latter was performed on fine-needle aspirates using Black's nuclear grading system wherein NG1 was considered well differentiated; NG2, moderately differentiated; and NG3, poorly differentiated. Based on pathologic examination of the resected specimens, tumor responses were categorized into complete response, partial response, no response, or progressive disease. The overall response rate was 71% (12% complete and 59% partial responses). In univariate analyses, tumor size and nuclear grade were significantly related to pathologic tumor response to chemotherapy (p = 0.04 and p = 0.0003, respectively), while disease stage was not (p = 0.17). In multivariate analyses, size remained significant even when NG was present in the equation (p = 0.013). Similarly, when size was included, NG remained significant (p = 0.002). NG3 tumors showed better response than NG2 or NG1 tumors did. While 19.3% of NG3 tumors showed complete response, none of the NG1 tumors completely responded to chemotherapy. Initial tumor size was inversely proportional to degree of tumor response. Our findings indicate that tumor clinical size and nuclear grade are important independent predictors of response to preoperative chemotherapy and that poorly differentiated tumors and small tumors showed the most response.