Does vascular imaging with MRI predict response to neoadjuvant chemotherapy in primary breast cancer?

582 Background: The ability to identify early during treatment those women with primary breast cancer (PBC) who will fail to respond to neoadjuvant chemotherapy will enable the use of alternative therapies that may be more effective. Here, we assess the role of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) to identify the non-responders. METHODS 28 patients with PBC (median age 46 years; range 29-70) were imaged prior to & following 2 cycles of 5-fluorouracil, epirubicin & cyclophosphamide (FEC) NAC. DCE-MRI scans were obtained following Gd-DTPA contrast medium using T1 & T2* weighted techniques with derivation of parametric images that reflected tissue microvessel permeability, leakage space & perfusion (Ktrans, ve, kep, MaxGd, rBV, rBF & MTT). Median & 5-95th centile values for each parameter were derived from whole tumour regions of interest. Pre-treatment parameter values & treatment changes were correlated with clinical & pathological response following 6 cycles of NAC using the Mann-Whitney U-test. A cohort of 9 patients was imaged twice prior to therapy to calculate the repeatability statistic for each parameter & hence determine the ability of DCE-MRI to predict pathological non-response on a patient-by-patient basis. RESULTS There were 19 clinical responders & 9 clinical non-responders; 11 pathological responders & 17 pathological non-responders. Pre-treatment parameter values & change in tumour size did not predict for response. Group analysis showed that changes in median Ktrans, kep, rBV & rBF correlated with both final clinical & pathological response to NAC (p<0.005 & p<0.01 respectively) as did changes in 5-95th centile range for these parameters (p<0.05). Application of the repeatability statistic revealed change in median Ktrans to be the best predictor of pathological non-response in individuals (repeatability range -49 to 97%) correctly predicting pathological non-response in all 17 patients (100%) & response in 6/11 patients (55%). CONCLUSION Patients who are destined to fail to respond to NAC can be identified from early changes in DCE-MRI parameters that reflect on microvessel perfusion & permeability. No significant financial relationships to disclose.