Retrospective evaluation of the Adjuvant! for breast cancer after 5 years of adjuvant tamoxifen (AAT) and other factors for prediction of recurrences (OFPR) following 5 years of adjuvant tamoxifen (AT).

e11068 Background: AAT is a computer model that predicts 5 year (yr) relapse rate (PRR) for patients (pts) with estrogen receptor positive breast cancer after AT. We used the Cleveland Clinic surgery database to validate the PRR of AAT and investigate OFPR. Methods: 744 Pts surgery 1/1995–12/1999 were reviewed (Table), and 292 were included. Demographic, pathology, staging, and treatment data were entered into AAT and PRR was calculated. Observed outcome (OO) in each pt (no R, local R, systemic R, death due to breast cancer (BC)/non-BC and unknown causes) was recorded for each pt. The association between AAT, PRR and OO were assessed with Cox proportional hazards analysis (CA), concordance index (CI), and calibration plots (CP). CI was calculated as a measure of discrimination for ATT PRR. CP was visually assessed. Deaths from other causes were treated as competing risks for recurrence. Results: Pts characteristics: median age 62, ductal histology 83%, median tumor size 1.5 cm, and median number of nodes ...