Pathological complete response (pCR) to
neoadjuvant treatment correlates with outcome in
breast cancer. We determined whether characteristics of
neoadjuvant therapy are associated with pCR. We used multi-level models, which accounted for heterogeneity in pCR across trials and trial arms, to analyze individual patient data from 3332 women included in 7 German neoadjuvant trials with uniform protocols. PCR was associated with an increase in number of
chemotherapy cycles (odds ratio [OR] 1.2 for every two additional cycles; P = 0.009), with higher cumulative
anthracycline doses (OR 1.6; P = 0.002), higher cumulative
taxane doses (OR 1.6; P = 0.009), and with
capecitabine containing regimens (OR 1.62; P = 0.022). Association of pCR with increase in number of cycles appeared more pronounced in
hormone receptor (HR)-positive
tumors (OR 1.35) than in HR-negative
tumors (OR 1.04; P for interaction = 0.046). Effect of
anthracycline dose was particularly pronounced in HER2-negative
tumors (OR 1.61), compared to HER2-positive
tumors (OR 0.83; P for interaction = 0.14). Simultaneous
trastuzumab treatment in HER2-positive
tumors increased odds of pCR 3.2-fold (P < 0.001). No association of pCR and number of
trastuzumab cycles was found (OR 1.20, P = 0.39). Dosing characteristics appear important for successful treatment of
breast cancer. Longer treatment, higher cumulative doses of
anthracyclines and
taxanes, and the addition of
capecitabine and
trastuzumab are associated with better response. Tailoring according to
breast cancer phenotype might be possible: longer treatment in HR-positive
tumors, higher cumulative
anthracycline doses for HER2-negative
tumors, shorter treatment at higher cumulative doses for triple-negative
tumors, and limited number of preoperative
trastuzumab cycles in HER2-positive
tumors.