[Application of Decision Curve on Evaluation of MRI Predictive Model for Early Assessing Pathological Complete Response to Neoadjuvant Therapy in Breast Cancer].

Y J He,X T Li,Z Q Fan,Y L Li,K Cao,Y S Sun,T Ouyang

DOI: https://doi.org/10.3760/cma.j.issn.0376-2491.2018.04.004

2018-01-01

Abstract:Objective: To construct a dynamic enhanced MR based predictive model for early assessing pathological complete response (pCR) to neoadjuvant therapy in breast cancer, and to evaluate the clinical benefit of the model by using decision curve. Methods: From December 2005 to December 2007, 170 patients with breast cancer treated with neoadjuvant therapy were identified and their MR images before neoadjuvant therapy and at the end of the first cycle of neoadjuvant therapy were collected. Logistic regression model was used to detect independent factors for predicting pCR and construct the predictive model accordingly, then receiver operating characteristic (ROC) curve and decision curve were used to evaluate the predictive model. Results: ΔArea(max) and Δslope(max) were independent predictive factors for pCR, OR=0.942 (95%CI: 0.918-0.967) and 0.961 (95%CI: 0.940-0.987), respectively. The area under ROC curve (AUC) for the constructed model was 0.886 (95%CI: 0.820-0.951). Decision curve showed that in the range of the threshold probability above 0.4, the predictive model presented increased net benefit as the threshold probability increased. Conclusions: The constructed predictive model for pCR is of potential clinical value, with an AUC>0.85. Meanwhile, decision curve analysis indicates the constructed predictive model has net benefit from 3 to 8 percent in the likely range of probability threshold from 80% to 90%.