The Accuracy of MRI in Predicting Pathologic Complete Response in Invasive Breast Cancer Patients Receiving Neoadjuvant Systemic Treatment.

Abstract

Abstract Background: Magnetic resonance imaging (MRI) has increased sensitivity to detect abnormalities in breast tissue as compared with mammography or ultrasound. Published reports of MRIs ability to predict pathologic response to neoadjuvant chemotherapy have shown conflicting results that vary depending on baseline molecular characteristics and chemotherapeutic regimens, with some studies suggesting higher predictive accuracy in Her2 positive patients receiving trastuzumab. This study examines both the ability of MRI to predict pathologic response and how tumor molecular profiles and treatment regimens influence MRI sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). Sensitivity, specificity, PPV, and NPV for the combination of post-treatment mammogram, ultrasound, and MRI studies in predicting pathologic complete response (pCR) was also explored.Methods: Eighty-one patients with invasive breast cancer treated with neoadjuvant systemic therapy between 2002 to 2009 and imaged pre and post-treatment with breast MRI were reviewed. Patient, tumor, and treatment characteristics including tumor grade, histologic type, receptor status (ER, PR, and Her2), breast cancer subtype [luminal A, luminal B, Her2 positive, triple negative (TN)], correlative pre- and post-treatment mammographic and ultrasound imaging, drug regimen, and pathologic findings were recorded. Complete pathologic response was defined as no residual invasive or pre-invasive disease in the breast.Results: Patients had the following subtypes of breast cancer: 21/81 (26%) luminal A, 13/81 (16%) luminal B, 23/81 (28%) Her2 positive, and 24/81(30%) TN. Of the Her2 positive, only 12/23 (52%) treated after May 2005 received adjuvant trastuzumab. Twenty patients (25%) had a complete radiographic response (rCR) on post-treatment MRI and 23/81(28%) had a pCR. The sensitivity, specificity, PPV, and NPV of MRI for predicting pathologic response was 57%, 88%, 65%, and 84%, respectively. Analysis of breast cancer subtype did not demonstrate a predilection in any particular subtype for correlation of radiographic and pathologic response, although MRI sensitivity (100%) and NPV (100%) were highest in luminal A disease, and specificity and PPV were highest for patients with ER negative (96% and 90%) and triple negative (100% and 100%) disease, respectively. Mammographic and ultrasound post-treatment findings of residual disease did not significantly correlate with pathologic findings in the setting of a rCR on MRI. Multivariate analysis of factors potentially influencing MRIs sensitivity and specificity failed to show that tumor characteristics (ER status, PR status, HER2 status) or neoadjuvant treatment (ACT vs other or trastuzumab) had any effect on these parameters.Conclusions: MRI has the highest specificity and PPV in the setting of ER negative and TN breast cancer. Although treatment results and breast cancer subtype did not influence the sensitivity, specificity, PPV, or NPV of MRI in predicting pathologic response, only half of the Her2 positive subset received trastuzumab. Results from an ongoing large multi-institutional study will further clarify these results. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 4022.