Development and Evaluation of Prediction Models for Breast Cancer Prognosis

点击翻译按钮获取中文摘要

PROJECT SUMMARY One in eight US women will be diagnosed with breast cancer during her lifetime, making it the second leading cause of cancer deaths among US women. Currently, there are about 3.1 million women living with breast cancer in the US. Despite the overall favorable prognosis for breast cancer, the risk of recurrence persists for the remainder of the patient's lifetime, with a 15-year recurrence rate greater than 40%. Accurate prediction of breast cancer outcomes, therefore, is essential for developing personalized treatment, which would maximize treatment efficacy, spare patients unnecessary treatment, and identify women at high risk of recurrence for preventive intervention. However, existing prediction tools and models for breast cancer prognosis, including Adjuvant! online, the most widely used prediction tool in the US, include only clinicopathological prognostic factors (e.g., age, tumor grade and size, and lymph node, hormone receptor , and estrogen receptor status ), without considering well-recognized lifestyle predictors such as obesity, weight gain, and post-diagnosis physical activity. In addition, no prediction tool or models have been developed and validated among Asian women with breast cancer. In the proposed study, we will utilize the resources of the Shanghai Breast Cancer Survival Study, a well characterized, population-based prospective cohort study of 5,042 breast cancer survivors with detailed information on lifestyle factors, to build prediction models that incorporate both clinicopathological and lifestyle factors. Separate prediction models will be built first for 5-and 10-year overall survival, breast cancer-specific survival, and disease-free survival based on the clinicopathological factors that are included in Adjuvant! online, and then expanded to incorporate lifestyle factors. Models will be validated and evaluated for predictive ability, then compared with existing breast cancer prediction tools/models. The proposed study is cost-efficient, feasible, and will fill gaps in breast cancer outcomes prediction research. While the prevalence of lifestyle factors may differ between the US and China, we have specifically demonstrated this in our investigations of lifestyle factors and breast cancer outcomes in several large-scale studies that physical activity, pre-diagnosis BMI, and soy food intake, are associated with breast cancer outcomes, regardless of country or race/ethnicity. Therefore, the knowledge gained from the proposed study should be generalizable for building prediction models for breast cancer patients of other racial and ethnic groups.