Prediction of pathological response to neoadjuvant chemotherapy in breast cancer patients by imaging.

BACKGROUND

Diagnostic imaging is important for predicting the pathological response to chemotherapy during neoadjuvant chemotherapy (NAC) and for considering the surgical management with appropriate resection after NAC. This study was performed to examine the accuracy of the present radiological imaging for predicting the pathological complete response (pCR).

METHODS

From 188 patients in our previous JONIE1 Study, a randomized controlled trial comparing chemotherapy with and without zoledronic acid for patients with human epidermal growth factor receptor 2-negative breast cancer, we evaluated 122 patients whose tumor size was examined by magnetic resonance imaging or ultrasound at three points: before NAC; after administering fluorouracil, epirubicin, and cyclophosphamide; and after NAC. The maximum tumor diameter was evaluated by magnetic resonance imaging or ultrasound. Tumor reduction ratios were calculated at the same three points. The association between the radiological clinical response and the pCR was examined.

RESULTS

Among the 122 patients evaluated, there were 98 and 24 patients with luminal (Lum) and triple-negative (TN) subtypes, respectively. There were no patients who showed tumor progression after treatment. The radiological size of the tumors was finally reduced by an average of 58.4%. Clinical complete response and pCR were achieved in 22 (18.0%) and 15 (12.3%) patients, respectively. In the overall population (n = 122), the accuracy, sensitivity, and specificity for predicting pCR were 86.1%, 88.8%, and 66.7%, respectively. The negative predictive value and false-negative rate were 45.5% and 11.2%, respectively. According to subtypes, the accuracies were 83.7% and 95.8% in Lum and TN, respectively. Negative predictive value and false-negative rate were markedly different between the Lum (29.4% and 13.5%) and TN subtypes (100% and 0%), respectively.

CONCLUSIONS

This randomized clinical trial demonstrated that NAC was safe for operable breast cancer patients with appropriate radiological monitoring. Radiological evaluation after NAC may be a reliable method for predicting pathological response in the TN subtype, but not in the Lum subtype.