Discovery and validation of a genomic signature to identify women with early-stage invasive breast cancer who may safely omit adjuvant radiotherapy after breast-conserving surgery.

Abstract

512 Background: Adjuvant radiotherapy (RT) is currently the standard of care for women with early-stage invasive breast cancer (BC) treated with breast conserving surgery (BCS). However, some women may have very low risk of recurrence and could safely be spared RT. This study aimed to identify these women using a molecularly-based approach. Methods: We performed an analysis of the SweBCG91-RT cohort, a trial randomizing women with node-negative stage I-II invasive BC +/- RT following breast conserving surgery, with sparse use of adjuvant systemic therapy. Only patients with ER+, HER2- tumors, and not treated with adjuvant systemic therapy, were included in this analysis. Transcriptome-wide profiling of tumors was performed using the Affymetrix Human Exon 1.0 ST microarray. The SweBCG91-RT cohort was divided into a training cohort of 243 patients and a validation cohort of 354 patients. Biological gene sets and individual genes related to locoregional recurrence in patients not receiving RT of the training set were identified, and a 16-gene signature was trained using elastic net regression. The signature, named Profile for the Omission of Local Adjuvant Radiation (POLAR), was locked prior to validation. Results: In the validation cohort, POLAR was prognostic for locoregional recurrence (LRR) in patients not treated with RT (multivariable Cox model adjusting for age, grade, tumor size, and luminal A vs luminal B: HR = 1.7 [1.2,2.3], p < 0.001). Patients categorized as POLAR low-risk had a 10-year locoregional recurrence rate of 7% in the absence of RT. Notably, there was no significant benefit from RT for these POLAR low-risk patients (HR = 1.1 [0.38,3.3], p = 0.83), whereas patients categorized as POLAR high-risk had a significant decreased risk of locoregional recurrence when treated with RT (recurrence rate without RT at 10-years 19%, HR = 0.43 [0.24,0.78], p = 0.0053). Conclusions: These data suggest that the novel POLAR genomic signature based on LRR biology can not only identify patients who have a low risk of LRR without adjuvant RT after BCS but who also would not benefit from RT, thus being prime candidates for RT omission.