e12655 Background: Pathological complete response (pCR) after neoadjuvant chemotherapy (NACT) for breast cancer predicts the risk of recurrence and increasingly may indicate the need for additional therapy postoperatively. Methods: A retrospective analysis was performed in two cohorts of patients (pts) treated with docetaxel, trastuzumab and carboplatin (TCH) or with docetaxel, carboplatin and dual blockade (TCH-P) in the neoadjuvant setting in patients with early breast cancer (tumor size < 50 mm and > 10 mm and cN0 or cN1) in our Clinic, and who had definitive surgery was conducted. Demographic data, size, grade, tumor type, receptor status prior to neoadjuvant treatment, pathological complete response (pCR) rates, and adverse effects were analyzed. The pCR was defined as ypT0 ypN0. Results: Patient in cohort A (n = 58) received TCH x 6 cycles and in cohort B (n = 25) TCH-P x 6 cycles. Median age was 51 (range 23 to 76 years) in cohort A and 46 (range: 30-68) in cohort B. In cohort A 37 (64%) of pts was HR-positive, in cohort B only 9 (36%) pts . The most common adverse events in both groups were neutropenia, diarrhea, chemotherapy induced polyneuropathy and febrile neutropenia. There are no significant differences in the frequency of adverse events in two cohorts. There was no symptomatic heart failure, but 6 pts (10%) in cohort A and 5 pts (16%) in cohort B had > 10% asymptomatic decrease in LVEF. All patients were evaluable for pCR. Higher rates of pCR were achieved in the HER2pos/HRneg pts: 66% (n = 14) in cohort A, and 87% (n = 14) in cohort B. In group HER2pos/HRpos pts, the pCR rate was 48% (n = 18) vs 55% (n = 5) respectively. Conclusions: In HER2positive early breast cancer, a dual blockade (trastuzumab and pertuzumab) together with carboplatin and docetaxel neoadjuvant chemotherapy achieved higher rates of pCR ( 76%) compared with pts treated with trastuzumab, carboplatin and docetaxel (56%). However, a much higher percentage of pCR was observed in the group of patients with non-luminal cancers, who received a double blockade (87% vs 66%).
Prediction of pathological complete response to neoadjuvant chemotherapy in breast cancer using a deep learning (DL) method
