Development And Validation Of A Nomogram Which Can Predict 21-Gene Recurrence Score

Background:Breast cancer is a kind of malignant tumor which is harmful to women's health all over the world.The incidence rate of breast cancer is increasing year by year.The application of adjuvant chemotherapy has greatly reduced the metastasis rate and mortality of patients.But breast cancer is a highly heterogeneous disease,not all patients can benefit from adjuvant chemotherapy.How to select high-risk patients who need adjuvant chemotherapy is a challenge for clinicians.Oncotype DX is a kind of gene detection method to assist clinicians to draw up individualized treatment plan,which is widely used worldwide.According to the 21-gene recurrence score,breast cancer patients can be divided into three groups: low risk,medium risk and high risk.The patients of high risk group is most likely to benefit from chemotherapy.At present,the majority of patients in the clinical trials to verify the predictive effect of Oncotype DX are western people.There is still a lack of large-scale clinical research which include different kinds of patient groups.Oncotype DX may not completely cover the prognosis information carried by routine clinicopathological factors and the correlation between 21-gene recurrence score and clinicopathological factors remains to be studied.The long waiting time and the high cost of Oncotype DX limit the clinical use of Oncotype DX so that some eligible patients can not afford this test.The prediction models which can predict 21-gene recurrence score are mostly based on western patients.As for different patient groups,the prediction ability of these prediction models still needs to be verified.Therefore,a prediction model which can predict 21-gene recurrence score of Asian patients can help clinicians to develop individualized treatment programs.Objective:Our purpose is to develop a user-friendly nomogram to predict the 21-gene recurrence score of patients with oestrogen receptor-positive,HER-2 negative invasive breast cancer.Methods:A total of 209 eligible breast cancer patients form the second hospital of Jilin University,Asan Medical Center and Seoul National University Hospital were included in the multicenter study.Univariate analysis was performed to assess the associations between the recurrence score,clinicopathological factors and ultrasound features.Multiple logistic regression analysis was used to identify independent predictors of high RS and the nomogram was bulit by these independent predictors.Discrimination and clinical usefulness of the predicting model were assessed using the C-index and decision curve analysis.Validation of the nomogram was assessed using the bootstrapping validation.Results:Predictors contained in the nomogram included histologic grade,lymphovascular invasion,PR and orientation.The prediction model displayed good discrimination with a AUC of 0.898(95% CI 0.830-0.964)and good calibration.Decision curve analysis showed that our nomogram was clinically useful.High C-index value of 0.878 could be reached using the bootstrapping validation.Conclusion:The novel nomogram in our study showed confidently that clinicopathologic variables and ultrasound features can be used for prediction of high RS.Histologic grade,lymphovascular invasion,PR and orientation are the independent predictors of high RS.Our nomogram shows favorable predictive accuracy for high RS in oestrogen receptor-positive,HER-2 negative invasive breast cancer patients.