We reviewed the data of 513 women, who had stage II or III breast cancer and received NCT and surgery from a single institution. The ability of various clinicopathologic factors to predict the achievement of BCS and tumor size reduction to < or = 3 cm was assessed. Nomograms were built and validated in an independent cohort.
RESULTS:
BCS was performed in 50.1% of patients, with 42.2% of tumors reduced to < or = 3 cm after NCT. A multivariate logistic regressionanalysis showed that smaller initial tumor size, longer distance between the lesion and the nipple, absence of suspicious calcifications on mammography, and a single tumor were associated with BCS rather than mastectomy (p < 0.05). Negative estrogen receptor, smaller initial tumor size, higher Ki-67 level, and absence of in situ component were associated with residual tumor size < or = 3 cm (p < 0.05). Two nomograms were developed using these factors. The areas under the receiver operating characteristic curves for nomograms predicting BCS and residual tumor < or = 3 cm were 0.800 and 0.777, respectively. The calibration plots showed good agreement between the predicted and actual probabilities.
CONCLUSION:
We have established a model with novel factors that predicts BCS and residual tumor size after NCT. This model can help in making treatment decisions for patientswho are candidates for NCT.