Feasibility of predicting pathological evaluation by ultrasonic evaluation in initial stage of neoadjuvant chemotherapy for primary breast cancer

Abstract

To investigate the correlation between change of tumor size after 2 cycles of neoadjuvant chemotherapy and pathological evaluation after 4 cycles of neoadjuvant chemotherapy. And to evaluate the feasibility of predicting pathological evaluation by ultrasonic evaluation in the initial stage of neoadjuvant chemotherapy for primary breast cancer. Retrospective analysis was performed in women with primary breast cancer, including 138 patients receiving 4 cycles of anthracycline-based neoadjuvant chemotherapy (CTX500 mg/m(2), D1, D8 Q28D; THP35 mg/m(2), D1, D8 Q28D; 5-Fu200 mg/m(2)/day.ci D1-D28), and 84 patients receiving 4 cycles of taxane-based neoadjuvant chemotherapy (PTX60-80 mg/m(2), D1, D8, D15 Q21D). The ROC (receiver operating characteristic) curve was employed to evaluate whether the product change of 2 largest perpendicular diameters of tumor as observed by ultrasonography after 2 cycles of neoadjuvant chemotherapy could exactly predict the pathologic evaluation by the Miller & Payne grading system criteria after 4 cycles of neoadjuvant chemotherapy. When no response, excellent response or pathologic complete remission to neoadjuvant chemotherapy were predicted by ultrasonic evaluation. And the areas under the curve ROC were 0.689, 0.655 and 0.647 respectively (all P values < 0.05). It was predicted as no response by using the traditional standard of ultrasonic evaluation of < 50% or excellent response at > or = 50% (kappa < 0.40). Pathological evaluation after 4 cycles of anthracycline- or taxane-based primary chemotherapy in breast cancer can't be predicted reliably only by the product change of 2 largest perpendicular diameters of tumor as observed by ultrasound after 2 cycles of neoadjuvant chemotherapy.