Development and cross-validation of a novel multi-omic assay to assess locoregional recurrence risk and adjuvant therapy benefit in early-stage hormone receptor positive invasive breast cancer patients

Abstract Purpose Breast cancer management is shifting towards personalized treatment regimens, particularly for early-stage, hormone receptor positive (HR+) invasive breast cancer (IBC) patients following breast conserving surgery (BCS) where locoregional recurrence (LRR) rates are low. A critical unmet need is the development of tools that can both improve prognostic risk assessment and identify which patients are likely to benefit or not benefit from adjuvant radiation therapy (RT). Herein we developed and cross validated a novel multi-omic assay to assess LRR risk and expected RT benefit for early-stage HR+/HER2-negative IBCs. Methods A retrospective multi-institutional cohort of 922 patients (T1-2, N0-1, HR+, HER2-) treated with definitive breast conserving surgery (BCS) with or without adjuvant treatment was used to develop and cross-validate a test to predict IBC LRR after BCS ± RT. Treatment assignment was not randomized. The test integrated NGS and proteomic assay data using two biosignatures to generate results: a Decision Score (DS) to predict 10-year LRR prognosis and a radiation resistance index (RRI) to predict differential RT effect on LRR. Associations between DS and RRI with LRR risk and RT interaction were tested using multivariable Cox models. Results Increasing continuous DS was associated with increasing LRR risk (HR 3.4 per 5 units; p<.001, n = 922) after adjusting for clinicopathologic risk factors, while RT was associated with reduced LRR risk (HR 0.2; p < .001). Increasing continuous RRI was associated with increasing LRR risk for patients treated with RT (HR = 3.1 per 5 units; RT: RRI pinteraction = 0.002). Biosignature utility was demonstrated for categorical risk groups, which were also associated with differential RT benefit. DS Elevated Risk patients (DS > 5) had higher LRR risk without RT (HR = 4.8; p = .0014) with corresponding 10-year risks of 24% in DS Elevated Risk versus 7% in DS Low Risk (DS ≤ 5). In DS Low Risk patients, a statistically significant reduction in LRR risk (HR = 1.0, p = .96) was not observed. However, in the DS Elevated Risk group, RT was associated with decreased LRR risk (HR = 0.4, p = .0097) for patients with a lower radio resistance index (DS > 5, RRI ≤ 5), whereas RT was not associated with a statistically significant reduction in LRR risk (HR = 0.8, p = .51) for patients with a higher radio resistance index (DS > 5, RRI > 5). Conclusions In this large retrospective, non- randomized multi-institutional cohort, the novel multi-omic biosignature was associated with lower and higher LRR risk after BCS and associated with statistically significant differential RT effect for LRR risk reduction. Clinically meaningful risk groups that were associated with differential RT benefit were identified using the biosignature, supporting its potential utility in the assessment of the LRR risk in early-stage HR+/HER2-negative IBC. Further clinical validation studies are in process, and prospective validation studies are being planned to establish the role of the test in clinical practice.