Pre-operative breast magnetic resonance imaging (MRI) influences patient selection for partial breast irradiation.

604 Background: Many patients with localized breast cancer undergo breast conservation therapy. Partial breast irradiation (PBI) is a technique that reduces irradiated tissue volume and decreases treatment time. Appropriate patient selection for this technique is critical. We studied the influence of pre-operative breast MRI on patient selection for PBI. METHODS Between 05/04 and 05/05, 231 women with newly diagnosed breast cancer underwent pre-operative bilateral breast MRI (median age = 64, range 34 to 89). Following initial clinico-pathologic review, patients were deemed to be eligible for PBI based on the NSABP B39 protocol if they satisfied its inclusion criteria (tumor size ≤3 cm, ≤3 positive nodes, negative surgical margins, lack of multicentric disease), derived through physical examination, mammography and/or ultrasonography. Breast MRI scans of all patients deemed eligible for PBI by these criteria were then reviewed to determine the impact of the MRI findings on their eligibility for PBI. Only patients who had pathological assessment to correlate with the MRI findings were included in the analysis. RESULTS Based on the NSABP B39 inclusion criteria, 186 patients (81%) were considered to be appropriate candidates for PBI without influence from MRI findings. Of these 186 women, MRI findings led to pathologically proven additional disease that altered the recommendations for PBI in 26 patients (14%). Mammographically occult, biopsy proven multicentic disease was detected in the ipsilateral breast in 13 patients (7%), whereas 5 women (2.7%) had mammographically occult, biopsy proven multicentric disease in the contralateral breast. In 8 patients (4.3%), pathological assessment confirmed more extensive local disease than was detected by breast MRI. CONCLUSIONS Up to 14% of patients initially deemed eligible for PBI were later found to be ineligible because of breast MRI findings. Ipsilateral multicentric disease was identified in 7%, contralateral disease was discovered in 2.7%, and more extensive local disease was noted in 4.3%. Bilateral breast MRI complements other imaging modalities in determining eligibility for PBI by identifying mammographically occult disease in patients for whom PBI may be inappropriate. No significant financial relationships to disclose.