Introduction: A 21-gene prediction assay (Oncotype DX) is helpful to estimate benefit from adjuvant chemotherapy in patients with hormone receptor-positive, lymph node-negative early breast cancer. This study was conducted to develop a model to estimate high recurrence score (RS) using easily available clinicopathologic parameters in limited-resource countries.
Patients and methods: Hormone receptor-positive, lymph node-negative early breast cancer patients who underwent Oncotype DX were enrolled onto the training set (n = 192). The risk category range of the RS was the same as in the TAILORx study. The multivariable logistic regression model was used to identify significant variables associated with high RS. The independent validation set (n = 264) was established from patients of a different time period.
Results: The median age in the training set was 47 years, and 78.0% were premenopausal. The number of patients with low RS (< 11), intermediate RS (11-25), and high RS (> 25) were 42 (22.0%), 122 (63.9%), and 27 (14.1%), respectively. High nuclear grade, no progesterone receptor expression, and high Ki-67 were associated with high RS, and these variables were used to construct the nomogram. It had significant discriminatory power in internal validation (area under the curve = 0.856) and in the validation set (area under the curve = 0.828). The calibration plot showed optimal agreement between predicted and actual probabilities in both sets.
Conclusion: A nomogram was successfully developed with 3 simple parameters. The probability of high RS can be easily and conveniently estimated using our nomogram. It might be useful to determine whether or not Oncotype DX is conducted in the TAILORx era. Future large-scale prospective studies are warranted.
Keywords: Chemotherapy; Ki-67; Oncotype Dx; Prediction; Progesterone receptor.
Copyright © 2019 Elsevier Inc. All rights reserved.