E1527. MRI Accurately Predicts pCR in ER/PR- HER2+ Breast Cancer Treated With Neoadjuvant Immunotherapy Followed by Neoadjuvant Chemotherapy
  1. R Weinfurtner; Moffitt Cancer Center
  2. Brian Czerniecki; Moffitt Cancer Center
  3. Hatem Soliman; Moffitt Cancer Center
Patients with ER/PR-HER2+ breast cancer demonstrate higher pathologic Complete Response (pCR) rates after neoadjuvant chemotherapy (NAC) than patients with ER/PR+ breast cancer. The addition of neoadjuvant immunotherapy (NAI) has the potential to increase pCR rates further. However, this adds additional challenges in response monitoring with MRI due to the potential difficulty in discerning inflammatory change from residual invasive malignancy. The purpose of this study was to evaluate patients with HER2+ breast cancer in an NAI/NAC pilot study to determine if MRI response could predict pCR at surgery.

Materials and Methods:
An IRB-approved retrospective analysis was performed on patients in a single institution pilot study of NAI using HER2 pulsed dendritic cell vaccines (DC1) injected intratumorally and intranodally followed by NAC for the treatment of ER/PR-HER2+ breast cancer. Baseline and post neoadjuvant therapy MRIs were reviewed to determine MRI complete vs MRI partial response to therapy (mCR vs mPR). These were compared to surgical pathology results to evaluate for pathologic complete vs pathologic partial response (pCR vs pPR). The definition of pCR was no residual invasive carcinoma in the breast or axilla.

Twenty female patients were included in the trial with a median age of 59 years (range 31 - 73). Most pretreatment MRIs demonstrated a dominate mass (17/20) with 3 patients presenting with NME only. Patients presenting with masses also had satellite lesions in 9/17 and associated suspicious NME in 3/17. Average longest diameter of the dominate finding was 45mm (range 14 - 140 mm). In addition, 16/20 (80%) patients had biopsy proven ipsilateral axillary metastasis prior to treatment. Posttreatment MRIs demonstrated mCR in 17/20 (85%). Surgical pathology yielded pCR in 14/20 (70%). As a diagnostic test with mCR as a true positive predictor of pCR, accuracy was 85% (62 - 97% confidence interval) with sensitivity 100%, specificity 50%, positive predictive value 82% and negative predictive value 100%.

Post-NAI/NAC presurgical MRI demonstrated high accuracy in predicting pCR in patients with HER2+ breast cancer, despite the potential for confounding inflammatory changes that might arise with the addition immunotherapy to the neoadjuvant treatment regimen.