Early warning model of recurrence site for breast cancer rehabilitation patients based on adaptive mesh optimization XGBoost

Abstract

Objective: Breast cancer is a malignant disease with a high mortality rate. Using postoperative rehabilitation data of breast cancer patients, this study explored the effects of immune, tumor, microenvironmental, psychological, nutritional, aerobic exercise and advanced work indexes on the rehabilitation of breast cancer patients. To determine the weight of the impact of different indications on cancer recovery. By combining the adaptive grid optimization algorithm with the XGBoost (Extreme Gradient Boosting) algorithm, an intelligent prediction model for breast cancer rehabilitation was constructed using patients indexes as input and the recurrence location as the output. Our results showed that the model constructed in this study could effectively predict cancer cell metastasis during breast cancer recurrence in recovered patients. Compared with artificial intelligence algorithm models such as neural network algorithm, support vector machine algorithm, gradient boosting tree algorithm and Adaboost, the model demonstrated a forecast accuracy rate of >93%. The model established in this study could effectively predict the recurrence position of breast cancer and provide an auxiliary reference for doctors to treat breast cancer patients more effectively.