Oncotype DX is the only multigene assay supported by the National Comprehensive Cancer Network (USA) with Level 1 evidence for use on node-negative and postmenopausal node-positive patients with estrogen receptor (ER)+/human epidermal growth factor receptor 2 (HER2)-breast cancer to predict the prognosis and to estimate chemotherapy add-on effects. However, the test's high cost prevents its use in most cases. Therefore, we aimed to obtain an alternative recurrence score (RS) prediction formula using the optimal clinicopathological factors. We retrospectively reviewed data of 81 patients with ER+/HER2- primary breast cancer in our hospital where Oncotype DX RS was measured. Stepwise multivariate linear regression analysis was conducted with several selected clinicopathological factors of 60 consecutive cases in the training group. The obtained RS-predicted values were validated against Oncotype DX RS using 21 additional consecutive cases. The RS prediction formula derived from the combination of ER, tumor-infiltrating lymphocytes (TILs), progesterone receptor (PgR), and Ki-67-labeling index produced a favorable model with a correlation coefficient (r) of 0.731103 for the Oncotype DX RS (p = 0.0002) and an adjusted R2 coefficient of 0.510013. The RS-predicted values and the actual Oncotype DX RS were classified into four 2 × 2 groups, by using an RS of 26 as a threshold for adding chemotherapy with a concordance rate of 95.2 % (20/21) and a kappa coefficient of 0.829. RS-predicted values of combined ER, TILs, PgR, and Ki-67 may be an appropriate substitute for Oncotype DX RS in certain situations.
Keywords: Alternative RS prediction formula; ER+/HER2− breast cancer; Multiple regression analysis; Oncotype DX; Progesterone receptor (PgR); RS-predicted value; Tumor-infiltrating lymphocytes.
Copyright © 2024 The Authors. Published by Elsevier Inc. All rights reserved.