Impact on Radiation Therapy Recommendation and Treatment Modality for Patients with Ductal Carcinoma In Situ Using the 7Gene Biosignature: Analysis of the PREDICT Study

Abstract

Breast conserving surgery (BCS) followed by adjuvant radiotherapy (RT) has been a mainstay in the treatment of DCIS based on multiple randomized trials demonstrating a local recurrence benefit with RT. However, these studies have failed to identify subsets of patients who did or did not benefit from adjuvant RT after BCS, raising concerns regarding both over and undertreatment. Thus, better prognostic and predictive tools are needed to appropriately risk stratify patients and understand their benefit of RT. The 7-gene predictive DCIS biosignature provides a validated score (DS) for women undergoing BCS that assesses their 10-year risk of in-breast and invasive recurrence with and without adjuvant RT. This trail was designed to evaluate the decision impact of the 7-gene predictive biosignature score on DCIS treatment recommendations. The PREDICT study is a prospective, multi-institutional trial for patients who received DCISionRT testing as part of their routine care. The registry includes females 26 and older who are diagnosed with DCIS, are candidates for BCS, and eligible for RT. Treating physicians completed treatment recommendation forms before and after receiving test reports to capture surgical, radiation and hormonal treatment (HT) recommendations and patient preferences. Analysis was performed in 2,012 patients treated at 63 clinical sites. Median age was 62 years old with 32% grade 3 and 10% size 2.5 cm or greater. Post-test, RT recommendation changed for 38% of patients (p<0.001), with a net reduction of 20% in patients recommended to receive RT(p<0.001). The DCISionRT test results had the greatest impact (OR 26.2, 95% CI 19.1-36.4, when analyzed categorically using DS>3 cut-off; 2.3 per DS, 95% CI 2.1-2.6, when evaluated continuously) on post-test RT recommendation in multivariable analysis when compared to all other factors including patient preference, patient clinical and tumor pathological factors, patient race/ethnicity, treatment facility, physician specialty. The post-test RT recommendation rate increased with increasing DS (0-2, 2-4, 4-10) on a categorical basis, with odds ratios of 6.8 DS (2-4 vs 0-2), and 35.0 for DS (4-10 vs 0-2). After DCISionRT test result, patient preference was the second most important factor in post-testing RT recommendation. There was also a significant change in the modality of RT recommended to 34% of those patients recommended RT pre-test and post-test by radiation oncologists (n = 937), with intensified RT modality for higher DS (p<0.001) and de-escalation for lower DS (p<0.001). This analysis of over 2,000 patients demonstrates significant changes in recommendations to add or omit RT based on the 7-gene predictive. The integration of DCISionRT into clinical decision processes has substantial impact on recommendations aimed at optimal management to prevent over- or under-treatment.