Assessment of Breast Cancer Response to Neoadjuvant Chemotherapy: Dynamic Contrast-enhanced Magnetic Resonance Imaging (DCE-MRI) vs FDG Positron Emission Tomography (FDG -PET)

Abstract

PURPOSE To retrospectively evaluate the value of the assessment of breast cancer response to neoadjuvant chemotherapy in DCE-MRI and FDG-PET.

METHOD AND MATERIALS Of 155 breast cancer patients who received neoadjuvant chemotherapy between 2006 and 2007, 23 women (mean age: 46.4 years old) underwent concurrent DCE-MRI and FDG-PET before and after chemotherapy. The response evaluation criteria in solid tumors (RECIST) were used for the response assessment in DCE-MRI with the measurement of the maximum diameters of the tumor. The response criteria of FDG-PET were defined as follows; complete response (CR) as no hypermetabolic activity, partial response (PR) as >50% reduction of standard uptake values (SUV), stable disease (SD) as <50% reduction or <15% increase of SUV, and progressive disease as >15% increase of SUV. The reference value of response assessment was microscopic pathologic tumor response, based on National Surgical Adjuvant Breast and Bowel Project. The interpretation of DCE-MRI and FDG-PET was done independently before surgery by the two specialists and the accuracy of response assessment was evaluated. The agreement of response assessment among DCE-MRI, FDG-PET and pathologic response was calculated by weighted kappa statistics (k). The size of residual tumor in DCE-MRI was correlated with that of pathologic residual lesion by Pearson correlation.

RESULTS After chemotherapy, pathologic CR, PR and SD rate was 26.1% (6/23), 30.4% (7/23), and 43.5% (10/23), respectively. The accuracy of response assessment was 65.3% (15/23) in DCE-MRI, and 39.2% (9/23) in FDG-PET. The strength of agreement was moderate (k = 0.572) between pathologic and DCE-MRI assessment, fair (k = 0.322) between pathologic and FDG-PET assessment, and fair (k=0.231) between DCE-MRI and FDG-PET assessment. Among 12 cases of FDG-PET-CR, false positive rate was 50% (6/12). The size of residual tumor in DCE-MRI was correlated with pathologic size (rho = 0.602, p = 0.0047).

CONCLUSION For the assessment of response to neoadjuvant chemotherapy in breast cancer, DCE-MRI showed higher accuracy compared to FDG-PET. The residual lesion in DCE-MRI was correlated well with pathology.

CLINICAL RELEVANCE/APPLICATION No residual uptake in FDG-PET after neoadjuvant chemotherapy should be carefully considered to predict pathologic CR. DCE-MRI would be useful modality to evaluate residual disease.