Nomogram to predict ipsilateral breast relapse based on pathology review from the EORTC 22881-10882 boost versus no boost trial.

ABSTRACT

BACKGROUND AND PURPOSE: The EORTC 22881-10882 trial showed that for patients treated with breast conserving therapy (BCT), a 16Gy boost dose significantly improved local control, but increased the risk of breast fibrosis. A model to estimate the risk of ipsilateral breast relapse (IBR) already exists, but now a model has been developed which takes boost treatment into account and is based on centrally reviewed pathology. MATERIALS AND METHODS: A Cox model was developed based on central pathology review data and clinical data of 1603 patients from the EORTC 22881-10882 trial with a median follow-up of 11.5years. From a predefined set of variables, predictors with a maximal effect on 10-year IBR rate >4% were retained in the model. Bootstrap re-sampling was used to assess model calibration and discrimination. The results are presented in the form of a nomogram. RESULTS: Apart from young age and no boost, presence of DCIS adjacent to the invasive tumor was associated with increased risk of IBR (HR 1.96, p=0.001). Patients with high grade invasive tumors were younger than patients with low/intermediate grade (p<0.0001). The nomogram includes histologic grade, DCIS, tumor diameter, age, tamoxifen, chemotherapy, and boost with a concordance probability estimate of 0.68. CONCLUSIONS: The nomogram for predicting IBR 10years after BCT includes seven factors, with young age, presence of DCIS and boost treatment as the most dominant factors. The nomogram estimates IBR and confirms the importance of a boost dose. Combined with a model to predict fibrosis published previously, the nomogram presented here may assist in decision making for individual patients.