Prospective Evaluation of the 21-Gene Recurrence Score Assay for Breast Cancer Decision-Making in Ontario

PURPOSE: To evaluate the 21-gene recurrence score (RS) on decision-making in a population-based cohort. PATIENTS AND METHODS: Patients with axillary node-negative or nodal micrometastases, estrogen receptor-positive, and human epidermal growth factor receptor 2-negative breast cancer being considered for chemotherapy were eligible. All cancer treatment centers in Ontario, Canada, participated. Oncologists made a preliminary recommendation for endocrine therapy with or without chemotherapy on the basis of Adjuvant! Online (AOL) risk estimation. Patients were asked for their preference regarding chemotherapy. After RSs were available, patients returned for final decision-making. Patient satisfaction was measured by using the decisional conflict scale. RESULTS: Between January 2012 and July 2013, 1,000 patients were recruited. RSs were available for 979 patients. In 58% of patients, risk was categorized as low (RS, 0 to 18); in 33%, intermediate (RS, 19 to 30); and in 9%, high (RS, ≥ 31). Oncologists' recommendations pretest and post-test remained the same in 464 patients (48%), changed from unsure or chemotherapy to no chemotherapy in 365 (38%), and changed from unsure or no chemotherapy to chemotherapy in 143 (15%). After the test, oncologists recommended chemotherapy for 236 patients, 81% of whom received chemotherapy. Of 151 patients in whom risk was classified as intermediate by means of AOL, 41% were a low risk and 44% intermediate risk with RS. Of 298 patients at high risk with AOL, 16% had a high risk RS. None of 236 patients with grade I tumors had a high-risk RS. Mean total decisional conflict scale score significantly improved from pretest to post-test from 34 to 19 (P < .001). CONCLUSION: The RS substantially influenced both oncologists' recommendations and patients' preferences for chemotherapy. The major effect was avoidance of chemotherapy when AOL indicated high or intermediate risk.