335P - Real-world data: Correlation of radiological and pathologic complete response after neoadjuvant treatment of breast cancer

Background

Neoadjuvant chemotherapy (NAC) of newly diagnosed breast cancer is used to improve resectability, reduce the extent of breast and axillary surgery and improve overall survival. Accurate assessment of locoregional response to NAC is important for surgical planning. Radiological method of choice is post-treatment breast magnetic resonance imaging (MRI), which determines the rate of response according to RECIST criteria. The primary indicator of outcome of NAC is pathologic complete response (pCR). This study aimed to compare radiologic and pathologic complete response rates after neoadjuvant treatment of early breast cancer.

Methods

A retrospective study of 259 breast cancer cases who received NAC between 01.01.2020 and 31.12.2021 at University Hospital Centre Zagreb was conducted with prior Ethics Committee approval. Radiological and pathohistological characteristics of the tumor at the time of diagnosis and after neoadjuvant treatment were analyzed using the Hospital information system (BIS). Furthermore, post-neoadjuvant therapy MRI findings were correlated with the postoperative pathohistological findings. Sensitivity, specificity, accuracy, and positive predictive value of radiologic complete response (rCR) as a predictor of pCR were calculated.

Results

Complete data was available for 94.98% (246/259) of patients. In 62 (25.20%) patients, rCR was recorded, radiologic partial response (rPR) in 150 (60.98%), stable disease in 22 (8.94%), and progressive disease in 9 patients. In 78 (31.71%) patients pCR was achieved, 46 (58.97%) were also classified as rCR, while 32 (41.03%) of those showed a radiological stable disease or rPR. Out of the 62 patients with rCR, 16 (25.81%) did not achieve pCR. The sensitivity and specificity of rCR were 90.48% and 58.97%, respectively, while positive predictive value was 82.61%. The overall accuracy of breast MRI in predicting pCR was 80.49%.

Conclusions

Although the sensitivity of breast MRI in predicting pCR was high, the specificity was low with only moderate overall accuracy. Low specificity can lead to an incorrect choice of breast surgery, therefore a better method of evaluation is necessary.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

M. Krizic, N. Dedic Plavetic, T. Silovski: Financial Interests, Personal, Research Funding: Roche. All other authors have declared no conflicts of interest.