Abstract PD7-04: A novel nomogram model can predict oncotype DX results thus reducing healthcare expenditures

Abstract

Abstract Oncotype DX (ODX) is a commercially available 21-gene recurrence score (RS) assay for breast cancer (BC), which has both prognostic and predictive recurrence value for estrogen receptor-positive (ER+)/HER2-negative (HER2-)/lymph node-negative (LN-) BC. ODX is currently endorsed for us by ASCO, the NCCN, and others for routine guideline application. The RS predicts benefit of adding adjuvant chemotherapy to hormonal manipulation. ODX is costly, a factor which contributes to the test being performed in only ∼1/3 of ER+BC in the USA. Disparities of its use in the US and other developed countries were recently published by us and others. The need for finding an accurate widely applicable, readily available surrogate for ODX is apparent. This study serves to develop and validate nomograms which can be used in place of ODX. National Cancer Data Base (NCDB) analysis from 2010-2012 of ODX tested ER+/HER2-/LN- patients (pts) with 6-50mm tumor size was used to create nomograms for predicting a high or low ODX RS test results. Analysis of NCDB ODX tested pts in 2013 was used for external validation. Age, tumor size, grade, progesterone receptor (PR) status, lymph-vascular invasion (LVI), and the four most frequent BC histologic types were chosen as variables for creating nomograms based on the published methods (JCO, 26(8):1364-70, 2008). Logistic regression was used to generate the scores and predicted probabilities. The predictive accuracy of the regression model was yielded using a Receiver Operator Characteristic (ROC) analysis and model fit was analyzed by plotting the predicted probabilities against the actual probabilities. Nomograms predicting a high- or low-risk ODX RS test results were created based on results from 27,719 ODX tested pts (2012-2012) and were validated on 12,763 ODX tested pts (2013) (table1). Table 1.Points assigned for original cohort nomogram (2010-2012; N=27,719)Points assigned for external validation cohort nomogram (2013; N=12,763)Clinicopathologic CharacteristicsHigh-risk ODX RSLow-risk ODX RSHigh-risk ODX RSLow-risk ODX RSAge0.1/yr0.1/yrTumor size0.8/1mm0.9/1mmGrade 101000100Grade 240403737Grade 310001000PR Positive ≥1%076075PR Negative <1%760750LVI Yes0.5020LVI No00.502Invasive ductal carcinoma036032Invasive lobular carcinoma360320Invasive ductal and lobular317302Invasive ductal mixed with other types731230Maximum points214210C-index.89.89p value<.001<.00195%CI.88-.89.88-.90Note: actual nomogram to be published in manuscript Grade and PR status were shown to be the highest predictors for a low- or high-risk ODX RS. The ROC curves for the probabilities of a low- or high-risk RS showed excellent agreement between the nomogram prediction and actual observation with high, acceptable C-indexes-.89 for both internal and external validation cohorts. We created and validated nomograms that accurately predict for a high- and low-risk ODX score based on the large, National Cancer Data Base which is comprised of >1500 Commission on Cancer accredited facilities. These nomograms may serve to select pts for which further ODX testing is not necessary and may be an excellent surrogate for BC pts for which ODX testing is not available. Citation Format: Orucevic A, Bell JL, Heidel RE. A novel nomogram model can predict oncotype DX results thus reducing healthcare expenditures [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr PD7-04.