The role of adjuvant radiotherapy (RT) following breast-conserving surgery (BCS) for women with ductal carcinoma in situ (DCIS) remains controversial. Although definitive trials provide level 1 evidence supporting the use of RT in reducing the risk of local recurrence, the same randomized trials demonstrate that approximately 70%-80% of DCIS patients do not have a local recurrence at 10 years after BCS alone. The DCISionRT® test is a 7-gene predictive biosignature that uses tumor biology in conjunction with clinicopathologic factors. The test provides a validated score (DS) for women receiving BCS that assesses the 10-year risk of DCIS recurrence and development of invasive breast cancer with and without adjuvant RT. We established a registry to evaluate the decision impact of the 7-gene predictive biosignature on DCIS treatment recommendations.
The PREDICT study is a prospective, multiinstitutional registry for patients who received DCISionRT testing as part of their routine care. The registry includes females 26 and older who are diagnosed with DCIS and are candidates for BCS and eligible for RT or systemic therapy. Treating physicians completed treatment recommendation forms before and after receiving test reports to capture surgical, radiation, and hormonal treatment (HT) recommendations and patient preferences. The primary end point is to identify the proportion of patients where testing led to a change in RT recommendation. Additional analyses include changes in recommendations in patient subgroups based on clinicopathologic factors or clinician specialty.
Analysis was performed in 2304 patients treated at 63 clinical sites. The median age of patients was 62 years old (18% < 50 years old), the nuclear grade was high in 33%, and tumor size was 2.5 cm or greater in 11%. Test results were DS low risk (DS ≤ 3) for 63% of women and 37% were DS elevated risk (DS > 3). Overall, the RT recommendation (yes/no) was changed for 38% of women after the 7-gene biosignature testing and the HT recommendation was changed for 11%. There was a net decrease in RT recommendation from 71% pre-assay to 53% post-assay (P <.001), where RT recommendations decreased by 53% in DS low-risk patients but increased by 25% in DS elevated risk patients.
Surgeons were more likely to change their RT recommendation (47%) than radiation oncologists (35%). When test results indicated DS elevated risk, both surgeons (79%) and radiation oncologists (88%) were likely to recommend RT, but when the results were DS Low risk, surgeons were more likely than radiation oncologists to recommend omitting RT (82% vs. 60%, respectively). Compared to traditional clinicopathologic features, the factor most strongly associated with RT recommendation was the biosignature result with other factors of importance being patient preference, tumor size, and grade.
This analysis demonstrates significant changes in recommendations to add or omit RT based on the 7-gene predictive biosignature in 2304 patients. The integration of DCISionRT into clinical decision processes has a substantial impact on recommendations aimed at optimal management to prevent over- or under-treatment.
AFFILIATIONS:
Pat W. Whitworth,1 Steven C. Shivers,2 Chirag Shah,3 Rakesh Patel,4 Karuna Mittal,2 Troy Bremer,2 Charles E. Cox5
1University of Tennessee, Knoxville, TN.
2PreludeDx, Laguna Hills, CA.
3Cleveland Clinic, Cleveland, OH.
4Good Samaritan Hospital, Los Gatos, CA.
5University of South Florida Morsani College of Medicine, Tampa, FL.