The effects of oncotype DX recurrence scores on chemotherapy utilization in a multi-institutional breast cancer cohort.

The use of clinicopathologic features in decision-making in early stage estrogen receptor (ER)-positive breast cancer (BC) may lead to over or under treatment. We investigated the effect of the Oncotype Dx (ODX) on chemotherapy (CTX) utilization in two cancer centers. 276 cases of node-negative ER-positive BC had ODX between 2005 and 2009. Age at diagnosis, tumor size, grade, and progesterone receptor (PR) status were abstracted from records and provided to two medical oncologists blinded to the ODX score. A recommendation for or against CTX was made based on clinicopathologic characteristics. Median age was 55 years. Mean tumor size was 1.6 cm. The median 10-year Adjuvant! Online (AO) mortality risk was 8. The median Nottingham Prognostic Index (NPI) was 3.3. The median ODX recurrence score was 17. Without knowledge of the ODX, oncologists were more likely to recommend CTX to younger women (P < 0.0001), women with negative PR status (P < 0.0001), higher NPI (P < 0.012), and tumors > 1 cm (P = 0.033). On average, CTX recommended patients had larger tumors (2.0 vs. 1.2 cm) and higher AO 10-year mortality (11.4 vs. 4.4%). ODX resulted in a change in management for 38% of women. Of 188 total patients who did not receive CTX, 71 had a recommendation favoring CTX by an oncologist blinded to the ODX score. In our multi-institutional cohort, the ODX score had a significant impact on the receipt of adjuvant CTX and altered management for 38% of women.