Abstract P5-06-09: Comparative analysis of proposed strategies for incorporating biologic factors into breast cancer staging

Background: Tumor biology is an important prognostic factor in breast cancer. The 8th edition AJCC pathologic prognostic stage incorporates biologic factors (grade, ER, PR, and HER2 status) along with anatomic staging to improve risk stratification. Additional staging models have been proposed including Bioscore and Risk Score. The Bioscore assigns a score of 0-7 based on anatomic stage, ER, HER2, and grade. The Risk Score assigns a score of 0-3 to each anatomic stage group, with 1 point assigned for ER-negative, HER2-negative, or grade 3 status. The objective of this study was to compare the performance of these models in a large, population-based cohort of breast cancer patients. Methods: The Surveillance, Epidemiology and End Results Program (SEER) database was used to identify primary stage I-IV breast cancer patients diagnosed in 2010. Patients were excluded for inflammatory carcinoma, neoadjuvant therapy, and missing data on receptors or tumor grade. 5-year disease specific survival (DSS) estimates were calculated using Cox regression for each staging model: AJCC anatomic stage, AJCC pathologic prognostic stage, Bioscore and Risk Score. Harrell concordance index (C-index) and Akaike Information Criterion (AIC) were used to compare each model in predicting DSS, with higher C-index and lower AIC indicating better predictive value and model fit. Results: 21,901 patients were included with a median age was 60 years. Median follow up was 52 months. All four staging models stratified DSS survival, with stepwise increases in hazard ratios and decrease in DSS for each increase in risk category (Table). The C-index of each model with biologic factors was higher than for anatomic stage, with the Risk Score having the highest C-index (0.832 for the anatomic stage; 0.856 for pathologic prognostic stage; 0.855 for Bioscore; and 0.864 for Risk Score). AIC was lowest for Risk Score (27315 for anatomic stage; 26874 for prognostic stage; 27120 for Bioscore; and 26753 for Risk Score), suggesting the best model fit for staging. Discussion: Staging models incorporating biologic factors perform better than anatomic staging alone. The Risk Score has the lowest AIC and may be the easiest to use as it simply assigns 1 point each for ER-negative, HER2-negative, and grade 3 tumors to each anatomic stage. In addition, the Risk Score has the added benefit of being the only staging system to stratify stage IV disease. We advocate for consideration of the Risk Score as an alternative to the pathologic prognostic stage for ease of use, optimal communication with patients and providers, and to standardize future clinical trials. Table. 5-yr estimates of DSS in 21,464 breast cancer patients in 4 models of breast cancer stagingAnatomic Stage5-yr DSSPrognostic Stage5-yr DSSBioscore5-yr DSSRisk Score5-yr DSSIA98.6IA98.7098.9IA098.9IB96.3IB93.7199.1199.2IIA94.5IIA90.1297.5297.1IIB92IIB84.2393.2392.8IIIA83.3IIIA80.5488.3IIA096.7IIIB73.2IIIB72.3570.2197.7IIIC68.6IIIC53.3656.2291.0IV38.8IV38.8732.2386.8IIB096.6194.7290.4380.4IIIA092.8190.0274.9365.2IIIB096.0180.3269.6353.9IIIC063.3178.0268.2339.7IV061.7147.5230.3310.8 Citation Format: Olga Kantor, Jiangong Niu, Hui Zhao, Sharon H Giordano, Kelly K Hunt, Tari A King, Elizabeth A Mittendorf, Mariana Chavez-MacGregor. Comparative analysis of proposed strategies for incorporating biologic factors into breast cancer staging [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P5-06-09.