Predictive value of MR multi-sign analysis for pathological complete response of triple-negative breast cancer after neoadjuvant chemotherapy

XU Xiaoxi; SONG Qiong

doi:10.3969/j.issn.1009-0460.2019.01.013

Abstract：Objective To investigate the predictive value of magnetic resonance (MR) multi-sign analysis in triple-negative breast cancer (TNBC) neoadjuvant chemotherapy with pathological complete response (p CR). Methods 127 patients with TNBC who received neoadjuvant chemotherapy were divided into p CR group (19 cases) and non-pCR group (108 cases) according to their postoperative pathology. The differences of magnetic resonance imaging signs between the two groups were compared. The diagnostic efficacy of the model was evaluated by receiver operating characteristic curve (ROC), and the sensitivity, specificity and area under the curve (AUC) were calculated. Logistic regression model was established to analyze the independent factors predicting p CR in neoadjuvant chemotherapy of TNBC with MR multi-sign. Results After neoadjuvant chemotherapy for TNBC, single factor analysis was performed between p CR group and non-pCRgroup. In non-pCR group, irregular mass, hyperintense intratumoral T2 WI and peritumoral edema were the main MRI signs before chemotherapy. The change rate of ADC value (△ADC％) before and after chemotherapy in nonp CR group was (28. 4 ± 42. 5) ％, lower than (57. 1 ± 30. 2) ％ in p CR group (P ＜ 0. 05). The ROC curve showed that the sensitivity, specificity and the AUC were 78. 9％, 75. 0％ and 0. 762, respectively. Multivariate Logistic regression analysis showed that only△ADC％ was the predictor of p CR in neoadjuvant chemotherapy of TNBC (OR =3. 48, 95％ CI: 1. 169-10. 364, P =0. 025). Conclusion MR multi-sign analysis has a certain correlation with the effect of neoadjuvant chemotherapy for TNBC, which is helpful to predict p CR after chemotherapy.