Abstract
BACKGROUND: The use of the Oncotype DX test reduces the rate of adjuvant chemotherapy recommendations. Few in-depth analyses have been performed on this decision-making process. METHODS: We retrospectively analyzed patient data based on available Oncotype DX test results (RS) irrespective of nodal status at a single center. We collected recommendations from six oncologists, first without RS (pre-RS) and then with RS results (post-RS). We investigated changes in recommendations, agreement between oncologist decisions, and the effect of different National Comprehensive Cancer Network (NCCN) recommendation categories (for, against, and considering chemotherapy). RESULTS: Data from 201 patients were included in the analysis. Recommendation of chemotherapy decreased by an average of 39.5%. Agreement improved substantially with RS, with a kappa value pre-RS of 0.37 (fair agreement) and post-RS of 0.75 (substantial agreement). Discordance remained substantial in cases where the NCCN recommendations considered chemotherapy only (32%). Pre-RS consensus against chemotherapy predicted low RS results (50 out of 51 patients). Post-RS consensus was highest in the NCCN chemotherapy recommendation group. CONCLUSIONS: The Oncotype DX test substantially improves decision accuracy in recommending adjuvant chemotherapy. It may be further improved with a consensus decision. In the case of pre-RS consensus against chemotherapy, the test can be spared.