Predicting two-year quality of life after breast cancer surgery using artificial neural network and linear regression models

摘要: The purpose of this study was to validate the use artificial neural network (ANN) models for predicting quality life (QOL) after breast cancer surgery and compare predictive capability ANNs with that linear regression (LR) models. European Organization Research Treatment Cancer Quality Life Questionnaire its supplementary measure were completed by 402 patients at baseline 2 years postoperatively. accuracy system evaluated in terms mean square error (MSE) absolute percentage (MAPE). A global sensitivity analysis also performed assess relative significance input parameters model rank variables order importance. Compared LR model, ANN generally had smaller MSE MAPE values both training testing datasets. Most ranging from 4.70 19.96 %, most high prediction accuracy. outperformed According analysis, pre-operative functional status best predictor QOL surgery. conventional more accurate patient-reported higher overall performance indices. Further refinements are expected obtain sufficient improvements routine clinical practice as an adjunctive decision-making tool.