Radiology: MRI predicts breast tumor response to chemo
The study, American College of Radiology Imaging Network 6657, was conducted as the imaging component of the Investigation of Serial Studies to Predict Your Therapeutic Response with Imaging And moLecular Analysis (I-SPY TRIAL) breast cancer trial. It was designed to prospectively assess MRI for its ability to predict response to treatment in patients receiving NACT for stage II or III breast cancer, according to authors Nola M. Hylton, PhD, of the University of California, San Francisco, and colleagues.
The authors noted that changes in tumor size in response to chemotherapy in the neoadjuvant setting have been associated with disease-free and overall survival. “Thus, primary tumor response monitoring has predictive value and has led to greater use of chemotherapy in the neoadjuvant setting for women with breast cancer.”
To investigate the benefits MRI brings in monitoring breast cancer tumors, Hylton and colleagues looked at data from 216 women with invasive breast cancer of 3 cm or greater undergoing NACT between May 2002 and March 2006. Treatment regimens were anthracycline-based, with or without a taxane. MRIs were performed before NACT, after one cycle of anthracycline-based treatment, between the anthracycline-based regiment at taxane, and prior to surgery at the completion of chemotherapy.
Results showed that prediction of pathologic complete response (pCR) following NACT using MRI measurements of tumor size was superior to clinical assessment, with areas under the receiver operating characteristic curve (AUCs) for MRI and clinical size of 0.75 and 0.68, respectively. Volume estimates were shown to be superior to diameter estimates, with AUCs for MRI volume and longest diameter of 0.70 and 0.64, respectively, for the prediction of pCR.
The difference in AUC between MRI volume and clinical size predictors varied over the course of treatment. Differences in AUC between the two methods at early, mid-, and posttreatment time points was 0.14, 0.09 and 0.02, respectively, for prediction of pCR. “Early assessment of treatment response on the basis of MR imaging findings may warrant changes in treatment planning, leading to better clinical outcomes,” wrote the authors.
Furthermore, the highest predictive value of 0.84, when adjusted for age and race, was obtained using a multivariate model that included MRI and clinical measurements, according to Hylton et al.
“With excellent patient compliance in terms of the imaging examinations over the course of this investigation, our study suggests that serial MR imaging examinations can be used effectively to monitor response to treatment.”
For more information on the I-SPY TRIAL, read “Preventive Molecular Imaging Shows Potential, Slow Progress” from the May issue of Health Imaging.